Anda di halaman 1dari 400

UPGRADING EXISTING OR DESIGNING NEW

DRINKING WATER TREATMENT FACILITIES


UPGRADING EXISTING
OR DESIGNING NEW
DRINKING WATER
TREATMENT FACILITIES

James E. Smith, Jr.


U.S. Environmental Protection Agency
Center for Environmental Research Information
Cincinnati, Ohio

Robert C. Renner and Bob A. Hegg


Process Applications. Inc.
Ft. Collins, Colorado

Jon H. Bender
U.S. Environmental Protection Agency
Technical Support Division
Cincinnati, Ohio

NOYES DATA CORPORATION


Park Ridge, New Jersey. U.S.A.
Copyright © 1991 by Noyes Data Corporation
Library of Congress Catalog Card Number: 90-23205
ISBN: 0·8155·1262·7
ISSN: 0090·516X
Printed in the Un ited States

Published in the United States of America by


Noyes Data Corporation
Mill Road, Part< Ridge, New Jersey 07656

10987 65432 1

Library of Congress Cataloging-in·Publication Data

Upgrading existing or designing new drinking water treatment


facilities / James E. Smith, Jr .... [et aU.
p. em... (Pollution technology review, ISSN 0090-516X ; no.
198)
Includes bibliographical references and index.
ISBN 0·8155-1262-7 :
1. Water--Purification -- Equipment and suppl ies. I. Smith, James
E.,1941- . II. Series
TD434.U63 1991
628.1'62 --dc20 90·23205
CIP
Foreword

This book describes technologies for upgrading existing, or designing new, drinking water treat-
ment facilities. Technologies which address prefiltration, filtration, disinfection, and organic and
inorganic contaminants are covered. Particular solutions for small community water treatment
plants (2,500-100,000 gallons per day) are described; and a composite correction program for
evaluating and optimizing plant performance is included, along with 13 "case studies" using this
method.

The 1986 statutory provisions of the Safe Drinking Water Act (SDWA) amendments bring a large
number of previously unregulated or minimally regulated water systems under significant regula-
tory control. This book covers both establ ished and emerging technologies needed to comply with
these new regulations. Descriptions of the various technologies include an overview of the process,
performance, design considerations, operating and maintenance aspects, costs, and experiences.
This information is meant to assist public water system engineers, operators, and decision-makers
faced with the many new regulatory requirements in the selection of methods for compliance with
the SDWA.

The information in the book is from the following documents:

Technologies for Upgrading Existing or Designing New Drinking Water Treat-


ment Facilities, prepared by James E. Smith, Jr. of the U.S. Environmental Pro-
tection Agency, Center for Environment Research Information, March 1990.
Environmental Pollution Control Alternatives: Drinking Water Treatment for
Small Communities, prepared by James E. Smith, Jr. of the U.S. Environmental
Protection Agency, Center for Environmental Research Information, April 1990.
Summary Report-Optimizing Water Treatment Plant Performance with the
Composite Correction Program, prepared by Robert C. Renner and Bob A. Hegg
of Process Applications, Inc. and Jon H. Bender of the U.S. Environmental Pro-
tection Agency, Technical Support Division for the U.S. Environmental Pro-
tection Agency, March 1990.

The table of contents is organized in such a way as to serve as a subject index and provides easy
access to the information contained in the book.

Advanced composition and production methods developed by Noyes Data


Corporation are employed to bring this durably bound book to you in a min-
imum of time. Special techniques are used to close the gap between "manu-
script" and "completed book." In order to keep the price of the book to a
reasonable level, it has been partially reproduced by photo-offset directly
from the original reports and the cost saving passed on to the reader. Due to
this method of publishing, certain portions of the book may be less legible
than desired.

v
NOTICE

The materials in this book were prepared as accounts of


work sponsored by the U.S. Environmental Protection
Agency, the American Water Works Association, and the
Association of State Drinking Water Administrators.
This information has been sUbject to the EPA's peer and
administrative review and has been approved for publica-
tion. On this basis the Publisher assumes no respOnsibil-
ity nor liability for errors or any consequences arising
from the use of the information contained herein. Men-
tion of trade names or commercial products does not
constitute endorsement or recommendation for use by
the Agency or the Publ isher.

The book is intended for information purposes only.


The reader is cautioned to obtain expert advice before
implementation of any procedures described for up-
grading existing or designing new drinking water treat-
ment facil ities, and to exercise caution when handl ing
any materials which could be potentially hazardous.
Final determination of the suitability of any informa-
tion or procedure for use by any user, and the manner
of that use, is the sole responsibility of the user.

All information pertaining to law and regulations is pro-


vided for background onl y. The reader must contact the
appropriate legal sources and regulatory authorities for
up-to-date regulatory requirements, and their interpre-
tation and implementation.

vi
Contents and Subject Index

PART I
TECHNOLOGIES FOR UPGRADING EXISTING, OR DESIGNING NEW,
DRINKING WATER TREATMENT FACILITIES

1. INTRODUCTION 3

2. SELECTING AND EVALUATING TREATMENT PROCESSES 4


2.1 Overview of Federal Drinking Water Regulations 4
2.2 Selecting Treatment Technologies 6
2.3 Overview of Available Treatment Alternatives 8
2.3.1 Filtration ~ 8
2.3.2 Disinfection 10
2.3.3 Organic Contaminant Removal. 12
2.3.4 Inorganic Contaminant Removal and Control 13
2.3.4.1 Corrosion Controls 13
2.3.4.2 Inorganic Contaminant Removal 13
2.4 Final Process Selection and Design 15

3. PREFIL TRATION TREATMENT ELEMENTS 18


3.1 Modifying Chemical Feed 19
3.1.1 Chemical Type 19
3.1.2 Chemical Dosage Management 20
3.1.3 Chemical Application Methods and Considerations 22
3.2 Modifying or Adding Rapid Coagulant Mixing 23
3.3 Improving Flocculation 26
3.3.1 Improving Mixing 26
3.3.2 Improving Flocculator Inlet and Outlet Conditions 27
3.3.3 Improving Basin Circulation with Baffles 28
3.4 Improving Sedimentation 28
3.4.1 Horizontal Flow Sedimentation Basins 30
3.4.2 Upflow Solids Contact Clarifiers 31

4. FI LTRATION TECHNOLOGIES 34
4.1 Modifying Filtration Systems 34
4.1.1 General Effectiveness of Filtration Systems 35
4.1.2 Filtration System Improvements 35

vii
viii Contents and Subject Index

4.1.3 System Design Checklist 38


4.2 Direct Filtration 38
4.2.1 Process Description 38
4.2.2 System Performance 39
4.3 Slow Sand Filtration 39
4.3.1 System Design Considerations 40
4.3.2 Operation and Maintenance 40
4.3.3 System Performance 42
4.3.4 System Costs 42
4.4 Package Plant Filtration 42
4.4.1 Selecting a Package Plant System 43
4.4.2 System Description and Design Considerations 43
4.4.3 Operation and Maintenance 44
4.4.4 System Performance 45
4.4.5 System Costs 45
4.5 Diatomaceous Earth Filtration 45
4.5.1 System Design 47
4.5.2 Operation and Maintenance 47
4.5.3 System Performance 48
4.5.4 System Costs 48
4.6 Other Filtration Systems 49
4.6.1 Membrane Fi Itration 49
4.6.2 Cartridge Filtration 52
4.7 Selecting the Appropriate Filtration Treatment System 53
4.7.1 Steps in an Evaluation 53
4.7.2 Need for Pilot Studies 54
4.7.3 Flocculation and Sedimentation Studies 54
4.7.4 Filtration Studies 54

5. DISINFECTION AND DISINFECTION BY-PRODUCTS 57


5.1 The Objectives of Disinfection 57
5.1.1 CT Values 58
5.2 Disinfection By.Products 59
5.2.1 The Chemistry of Oxidation 60
5.2.2 The Presence of Disinfection By-Products in Drinking Water 61
5.2.3 Strategies for Controll ing Disinfection By-Products 61
5.3 Comparing Disinfectants 63
5.3.1 Chlorine 63
5.3.2 Chlorine Dioxide 63
5.3.3 Monochloramine 63
5.3.4 Ozone 64
5.3.5 Ultraviolet Radiation 65
5.3.6 Advanced Oxidation Processes 65
5.4 Primary Disinfection Technologies 65
5.4.1 Chlorine 66
5.4.1.1 Process Description , '" " , .66
5.4.1.2 Disinfection with Chlorine Gas 69
5.4.1.3 Disinfection with Sodium Hypochlorite Solution 71
5.4.1.4 Disinfection with Solid Calcium Hypochlorite 72
5.4.2 Ozone 72
5.4.2.1 Process Description 72
5.4.2.2 System Design Considerations 73
5.4.2.3 Costs of Ozonation Systems 79
5.4.3 Chlorine Dioxide 81
Contents and Subject Index ix

5.4.3.1 Process Descri ption 82


5.4.3.2 Establishing a Chlorine Dioxide Residual. 83
5.4.3.3 Chlorine Dioxide System Design Considerations 85
5.4.3.4 Costs of Chlorine Dioxide Disinfection 86
5.4.4 Ultraviolet Radiation 87
5.4.4.1 Process Description 88
5.4.4.2 UV Disinfection System Design Considerations 90
5.4.4.3 UV System Operating and Maintenance Considerations 91
5.4.4.4 Costs of UV Radiation Systems 92
5.5 Secondary Disinfectants 92
5.5.1 Chloramination 92
5.5.1.1 Process Description , , , , .. , 93
5.5.1.2 Establishing a Chloramine Residual 93
5.5.1.3 Chloramination System Design Considerations 93
5.5.1.4 Costs for Chloramination 93

6. TREATMENT OF ORGANIC CONTAMINANTS 96


6.1 Pretreatment for Natural Organic Contaminant Removal. 96
6.1.1 Coagulant Pretreatment 98
6.1.2 Oxidation Pretreatment 99
6.2 Granular Activated Carbon 99
6.2.1 Process Design Considerations 100
6.2.2 Tests for Deriving Carbon Usage and Other Design Criteria 103
6.2.3 Least Cost Design Criteria 103
6.2.4 Facility Design Criteria 104
6.2.5 Operation and Maintenance 104
6.2.6 System Performance 105
6.2.7 System Costs 105
6.3 Packed Column Aeration 108
6.3.1 System Design Considerations 111
6.3.2 Pilot Testing PCA 111
6.3.3 VOC Emission Control 112
6.3.4 Operation and Maintenance 114
6.3.5 System Performance 114
6.3.6 System Costs 114
6.4 Powdered Activated Carbon Plus Conventional Treatment 117
6.4.1 PAC Application Techniques 117
6.4.2 System Design Considerations 118
6.4.3 System Performance , 119
6.5 Diffused Aeration 119
6.5.1 System Design Considerations 120
6.5.2 System Performance 120
6.6 Multiple Tray Aeration 120
6.6.1 System Design 120
6.6.2 System Performance 120
6.7 Emerging Applications of Treatment Technologies for Organic Contaminants 120
6.7.1 Oxidation Including Ozone 121
6.7.2 Reverse Osmosis .. , 123
6.7.3 Mechanical Aeration 123
6.7.4 Catenary Grid 123
6.7.5 Higee Aeration 123
6.7.6 Resins 124

7. TREATMENTS FOR INORGANIC CONTAMINANTS 129


7.1 Techniques for Controlling Corrosion 129
x Contents and Subject Index

7.1.1 The Problem of Corrosion 130


7.1.2 Diagnosing and Evaluating the Problem 131
7.1.2.1 Consumer Complaints , " 131
7.1.2.2 Corrosion Indices 131
7.1.2.3 Sampling and Chemical Analysis 132
7.1.2.4 Scale or Pipe Surface Examination 132
7.1.2.5 Rate Measurements 133
7.1 .3 Corrosion Controls 133
7.1.3.1 Distribution and Plumbing System Design Considerations 134
7.1.3.2 Water Quality Modifications 134
7.1.3.3 Corrosion Inhibitors 136
7.1.3.4 Cathodic Protection 137
7.1.3.5 Coatings and Linings 137
7.2 Treatment Technologies for Controlling Inorganic Contaminants, Including
Radionuclides 137
7.2.1 Removing Radionuclides in Drinking Water 139
7.2.1.1 Costs , 139
7.2.2 Conventional Treatment: Coagulation and Lime Softening 139
7.2.2.1 Coagulation 140
7.2.2.2 Lime Softening 142
7.2.3 Reverse Osmosis 142
7.2.3.1 Design Considerations 143
7.2.3.2 System Performance 146
7.2.3.3 System Costs 146
7.2.4 Ion Exchange 146
7.2.4.1 System Performance 146
7.2.5 Activated Alumina 147
7.2.5.1 System Performance 149

8. CURRENT AND EMERGING RESEARCH 150


8.1 Current Research on Disinfection By·Products 150
8.1.1 Identifying and Controll ing Chlorination By-Products 151
8.1.2 Identifying Ozone By·Products 151
8.2 Treatment of Organic and Inorganic Contaminants 151
8.2.1 Granular Activated Carbon Systems 152
8.2.1.1 Establishing Carbon Usage Rates 152
8.2.1.2 Field Tests of Granular Activated Carbon Systems 153
8.2.2 Ozone Oxidation Systems 154
8.2.3 Ultraviolet Treatment 154
8.2.4 Reverse Osmosis 154
8.2.5 Ultrafiltration 154
8.2.6 Packed Tower Aeration 154
8.2.7 Conventional Treatment 155
8.2.8 Ion Exchange 155
8.2.9 Technologies for Removing Radionuclides 155
8.2.10 Secondary Sources of Pollution , 156
8.2.11 Small Systems Technologies 156
8.3 Mandatory Disinfection 156
8.3.1 Treatment/Distribution Microbiology 156
8.3.2 Bacterial Detection/Monitoring 157
8.4 Prohibition of Lead Materials 157
8.5 Systems and Cost Modeling Studies 157
8.6 Future Directions 158
Contents and Subject Index xi

9. REFERENCES , 159

APPENDIX A-EXPERIENCE MODIFYING EXISTING FILTRATION SYSTEMS 165


A.1 Upgrading Existing Treatment Facilities 165
A.1.1 Horizontal Flow Basin Example 165
A.1.2 Upflow Solids Contact Clarifier Example 166
A.1.3 Sacramento, California (Sequeira et ai, 1983) 166
A.1.4 Erie County, New York (Westerhoff, 1971) ., 166
A.1.5 Corvallis, Oregon (Collins and Shieh, 1971) 167
A.1.6 Novato, California (Culp, 1976) 168
A.2 Slow Sand Filter Systems 169
A.2.1 Idaho State (Tanner, 1988) 169
A.2.2 New York State (Letterman and Cullen, 1985) 169
A.2.3 Mcindoe Falls, Vermont (Pyper, 1985) 170
A.2.4 Village of 100 Mile House, British Columbia, Canada (Bryck et ai, 1987) 171
A.3 Package Plants 171
A.3.1 Conventional Package Plants (Morand and Matthew, 1983) 171
A.3.2 Adsorption Clarifier Package Plants 172
A.4 Diatomaceous Earth Filters 176
AA.1 Colorado State University Study (Lange et ai, 1984) 176
AA.2 Mcindoe Falls, Vermont (Pyper, 1985) 176
A.5 Selecting a Filtration System 176
A.5.1 Lake County, California (Conley and Hansen, 1987) 176

APPENDIX 8-CASE HISTORIES OF EMERGING DISINFECTION TECHNOLOGIES 180


8.1 Ozone Case Histories 180
B.1.1 Primary Disinfection with Ozone: North Andover, Massachusetts 180
B.1.2 Preozonation for THM Control: Kennewick, Washington (Cryer, 1986) 181
8.2 UV Radiation Case Histories 183
B.2.1 Ultraviolet Radiation for Primary Disinfection: Fort Benton, Montana 183
8.3 Chlorine Dioxide Case Histories 183
B.3.1 Predisinfection for THM Control: Evansville, Indiana (Lykins and Griese,
1986) 183
B.3.2 Primary and Secondary Disinfection with Chlorine Dioxide: Hamilton, Ohio
(Augenstein, 1974; Miller et ai, 1978; U.S. EPA, 1983) 184
8.3.3 Preoxidation with Chlorine Dioxide, Postchlorination with Chlorine Dioxide
and Chloramine: Galveston, Texas (Myers et ai, 1986) 186
8.4 Chloramine Case Histories 187
BA.1 Prechlorination, Postchloramination: Bloomington, Indiana (Singer, 1986) 187
BA.2 Prechlorine Dioxide, Prechlorination, and Postchloramination: Phi ladel ph ia,
Pennsylvania (McKeon et ai, 1986) 187

APPENDIX C-EXPERIENCE WITH TREATMENT TECHNOLOGIES FOR ORGANIC


CONTAMINANTS 190
C.1 Experience with Granular Activated Carbon 190
C.1.1 GAC for VOC Removal: Washington, New Jersey (Chrobak et ai, 1985) 190
C.1.2 GAC for Contaminant Control: Cincinnati, Ohio (DeMarco, 1983)
(Westerhoff and Miller, 1985) 190
C.1.3 EPA Health Advisory Example 191
C.2 Experience with PTA: Scottsdale, Arizona (Cline et ai, 1985) 191
C.3 Experience with PAC 192

APPENDIX D-EXPERIENCE WITH TREATMENT TECHNOLOGIES FOR


INORGANIC CONTAMINANTS 197
D.1 Corrosion Control 197
xii Contents and Subject Index

0.1.1 Controlling Lead: Seattle, Washington 197


0.1.2 Control Iing Lead with pH Adjustment: Boston, Massachusetts 197
D.2 Coagulation to Control Barium: Illinois 198
D.3 Reverse Osmosis: Sarasota, Florida 198
D.4 Ion Exchange: McFarland, California 199
0.5 Activated Alumina: Gila Bend, Arizona 200

APPENDIX E-SUMMARY OF CORROSION INDICES 203

PART II
WATER TREATMENT TECHNOLOGIES FOR SMALL COMMUNITIES

INTRODUCTION 207

1. DRINKING WATER TREATMENT: AN OVERVIEW 208


Why Do We Need Drinking Water Treatment? 208
How Is Drinking Water Treated? 210
Fi Itration 211
Disinfection 211
Treatment of Organic Contaminants 211
Treatment of Inorganic Contaminants 211

2. NEW AND PROPOSED DRINKING WATER TREATMENT REGULATIONS: AN


OVERVIEW 213
Compliance Schedules 214
Maximum Contaminant Levels 214
MCLs for Volatile Organic Compounds 215
MCLs for Inorganic and Synthetic Organic Compounds 215
MCLs for Microbiological Contaminants 215
MCLs for Radionuclide Contaminants 215
MCLs for Disinfectants and Disinfection By-Products 215
Monitoring , , 217
Volatile Organic Chemicals 217
Fluoride 217
Radionuclides 217
Microbiological Contaminants 217
Laboratory Analysis and Sampling Requirements 220
Surface Water Treatment Requirements 221

3. SOLUTIONS TO DRINKING WATER TREATMENT PROBLEMS: AN OVERVIEW 223


Questions to Consider in Choosing Treatment Technologies 223
What Are the Requirements for Drinking Water Suppl ied by the System? 225
Are Nontreatment Alternatives Available? 225
What Are the Characteristics of the Raw Water? 225
What Is the Configuration of the Existing System? 226
What Are the Costs of the Treatment Options? 227
What Are the Treatment Technology's Operating Requirements? 227
How Compatible Are the Processes Used? 227
What Waste Management Issues Are Involved? 227
What Are the Future Needs of the Service Area? 227
Special Issues for Small Systems 227
Financial/Capital Improvement 227
Multicommunity Cooperative Arrangements (Regionalizationl 228
Operator Capabilities 229
Contents and Subject Index xiii

Selecting a Consulting Engineer/Equipment Vendor 232


Using a Point-of-Use/Point-of·Entry (POU/POE) System 232

4. FI LTRATION TECHNOLOGIES FOR SMALL SYSTEMS 233


Processes Preceding Filtration 233
Choosing a Filtration Technology 234
Slow Sand Filtration 234
Diatomaceous Earth Filtration 237
Package Plants 238
Membrane Filtration (Ultrafiltration) 238
Cartridge Filtration 239
Innovative Filtration Technologies 240

5. DISINFECTION 241
Chlorination 241
Disinfection Terminology 241
Factors Affecting Chlorination Efficiency 241
Chlorination Chemicals 242
Disinfection with Sodium Hypochlorite Solution 242
Disinfection with Solid Calcium Hypochlorite 243
Hypochlorination Equipment 244
Disinfection with Chlorine Gas 244
Chlorination Monitoring 245
Ozonation 245
Monitoring the Ozonation System Operation 247
Ultraviolet Radiation (UV) 247
Obtaining Effective Disinfection: CT Values 248
Disinfection By-Products and Strategies for Their Control 248

6. TREATING ORGANIC CONTAMINANTS IN DRINKING WATER 250


Granular Activated Carbon 250
Aeration 252
Packed Column Aeration 253
Diffused Aeration 253
Multiple Tray Aeration 254
Emerging Technologies for Organics Removal 254
Mechanical Aeration 254
Catenary Grid. . . 254
Higee Aeration 254

7. CONTROL AND REMOVAL OF INORGANIC CONTAMINANTS 256


Corrosion 256
Controlling Lead Levels in Drinking Water 256
Techniques for Controlling Corrosion 257
Treatment Technologies for Removing Inorganic Contaminants 258
Coagulation/Filtration 260
Reverse Osmosis and Electrodialysis 261
Ion Exchange 261
Activated Alumina 262
Technologies for Radon Removal: Aeration and Granular Activated
Carbon 263

8. RESOURCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . 264
Safe Drinking Water Hotline ... . 264
xiv Contents and Subject Index

U.S. Environmental Protection Agency Regional Offices 264


State Drinking Water Agencies 264
Organizations Assisting Small Systems 264
American Water Works Association (AWWAl Small Systems Program 264
National Rural Water Association (NRWA) 264
Rural Community Assistance Program (RCAP) 264
Farmers Home Administration (FmHAl 264
Publications 264
General 264
Sampling 270
Filtration 270
Disinfection 270
Corrosion Control 270
Radionuclide Removal 270
Well head Protection 271
Costs/Financial Management 271
Consultants 271
The Federal Register 271

APPENDIX A-HOW TO TAKE BACTERIOLOGICAL SAMPLES 272

APPENDIX B-CHECKLlST: SOME FACTORS AFFECTING WATER TREATMENT


SYSTEM PERFORMANCE 273
Administration 273
Maintenance 273
Design 273
Operation 274

APPENDIX C-SELECTING A CONSULTING ENGINEER 275

APPENDIX D-CHLORINE RESIDUAL MONITORING 278


Demonstration of Maintaining a Residual. 278
Maintaining a Residual Entering the System 278
Maintaining a Residual Within the System 279

APPENDIX E-CT VALUES 280

APPENDIX F-SAMPLE OF CT CALCULATION FOR ACHIEVING 1-LOG


GIARDIA, 2-LOG VIRUS INACTIVATION WITH CHLORINE DISINFECTION 282

PART III
THE COMPOSITE CORRECTION PROGRAM FOR OPTIMIZING WATER TREATMENT
PLANT PERFORMANCE

1. INTRODUCTION 287
Purpose 287
Background 287
Content. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 287

2. THE COMPOSITE CORRECTION PROGRAM APPROACH 288


The Comprehensive Performance Evaluation 288
Design Components 288
Operation and Maintenance 288
Administration 289
Contents and Subject Index xv

Evaluating the Factors that Limit Performance 289


Reporting 289
The Composite Correction Program 291
Implementing the Composite Correction Program 291
Maintaining Long- Term Involvement 291

3. RESULTS OF CASE STUDIES 293


CPE Findings 293
CCP Findings 296
Overall Factors Limiting Performance 296

4. CASE STUDIES 298


Plant 1 298
Plant 2 304
Plant 3 312
Plant 4 317
Plant 5 323
Plant 6 331
Plant 7 336
Plant 8 342
Plant 9 350
Plant 10 356
Plant 11 364
Plant 12 370
Plant 13 : 378
Part I

Technologies for Upgrading Existing,


or Designing New,
Drinking Water Treatment Facilities

The information in Part I is from Technologies for Up-


grading Existing or Designing New Drinking Water
Treatment Facilities, prepared by James E. Smith, Jr.
of the U.S. Environmental Protection Agency, Center
for Environmental Research Information, March 1990.
Acknowledgments

This document is a compilation ofthe material presented in a series of workshops on "Emerging


Technologies for Upgrading Existing or Designing New Drinking Water Treatment Facilities." Some
of these workshops were sponsored by local sections of the American Water Works Association
(AWWA) and others by the Association of State Drinking Water Administrators (ASDWA). Other
sponsors included the United States Environmental Protection Agency's (EPA) Offices of Drinking
Water and Research and Development and Regions.

Overall management of the workshop series and preparation of this document was provided by Dr.
James E. Smith, Jr., of EPA's Center for Environmental Research Information. He was assisted in
providing technical direction by representatives of ASDWA, the local AWWA Sections, EPA Regional
Personnel, John Trax and Ken Hay of EPA's Office of Drinking Water, and Walter Feige and other
members of EPA's Drinking Water Research Division.

Un~ess otherwise noted, the technical material in this document was taken from handout materials
and presentations provided by the following speakers in the workshops: Dr. Robert M. Clark, Director
of EPA's Drinking Water Research Division, Cincinnati, Ohio; John F. Dyksen, Director, Department
of Water Supply, Village of Ridgewood, New Jersey; Sigurd Hansen, HDR, Inc., Cameron Park,
California; Dr. Gary S. Logsdon, Chief of EPA's Microbiological Treatment Branch, Cincinnati, Ohio;
Dr. Rip G. Rice, Rice International Consulting Enterprises, Ashton, Maryland; and Dr. Edward
Singley, James Montgomery Consulting Engineers, Gainesville, Florida. Appreciation is also
expressed to those individuals around the country who participated in panels and presented case
histories of their treatment experiences.

Organization and presentation of the material, technical writing, and editorial work was provided by
Lynn Knight, John Reinhardt, and Susan Richmond of Eastern Research Group, Inc., Arlington,
Massachusetts. Appreciation is expressed to those individuals in EPA's Offices of Drinking Water and
Research and Development who assisted in reviewing drafts of this publication; especially Peter Cook,
Ronnie Levin, Michael Schock, and Tom Sorg. The workshop presenters are acknowledged for their
assistance in reviewing materials, especially John Dyksen who provided overall technical review.

In addition, EPA gratefully acknowledges the assistance of:

Kalyan Ramen, Environmental Engineer, Malcolm Pirnie,Inc.


Ross Hymen and Paul Anderson, Massachusetts Dept. of Environmental
Quality Engineering.
Tom Boshar, Lally Associates
Dr. Carl Nebel, PCI Ozone Corporation
Fred Zinnbauer, Aquionics, Inc.
Gary Swanson, Robert Peccia Associates

2
1. Introduction

This document discusses drinking water treatment Chapter 3 covers prefiltration elements of a water
technologies that address contaminants and treatment system, including rapid mixing, chemical
contaminant categories regulated under the Safe dosage, coagulation, flocculation, and sedimentation.
Drinking Water Act (SDWA - 42 U.S.C. 300f, et seq.) While these elements usually precede filtration, they
and its 1986 amendments. The information was may be found with other treatment technologies as
distilled from materials used in a series of workshops well. These elements impact the performance of
conducted from 1987 through 1989 in locations subsequent components in the treatment train,
throughout the United States. The workshops were which are described in Chapters 4 through 7.
sponsored by the Offices of Drinking Water (ODW)
and Research and Development (ORO) of the United Chapter 4 describes filtration technologies that
States Environmental Protection Agency (EPA), and address removal of turbidity and microbial
the Association of State Drinking Water contamination. Technologies covered include
Administrators (ASDW A). conventional, direct, slow sand, diatomaceous earth,
membrane, and cartridge filtration systems.
The 1986 statutory provisions of the SOWA
amendments bring a large number of previously Chapter 5 reports on the five major disinfection
unregulated or minimally regulated water systems technologies: chlorine, ozone, ultraviolet (UV)
under significant regulatory control. This document radiation, chlorine dioxide, and chloramines. The
covers both established and emerging technologies problem of disinfection by-products and strategies for
needed to comply with these new regulations. their control are also addressed.
Descriptions of each technology include an overview
of the process, performance, design considerations, Chapters 6 and 7 describe technologies that address
operating and maintenance aspects, costs, and organic and inorganic contamination, respectively.
experiences. This information is meant to assist Treatment technologies for organics removal include
public water system engineers, operators, and granular activated carbon, packed column aeration,
decision-makers faced with the many new regulatory powdered activated carbon, diffused aeration,
requirements in selecting methods of compliance. multiple tray aeration, oxidation, mechanical
aeration, catenary grid aeration, Higee aeration, and
Chapter 2 is an overview of the selection process and membrane filtration. Treatment technologies for
the potential technological solutions for each inorganics removal include corrosion control, reverse
contaminant or contaminant category. This chapter osmosis, ion exchange, activated alumina, aeration,
serves as a guide to subsequent chapters that discuss and powdered activated carbon.
each treatment technology in more detail. It includes
many tables that compare relevant information Chapter 8 reviews the recent research activities of
between technologies. Case histories illustrating EPA's Drinking Water Research Division.
experience with each technology are provided in
Finally, Chapter 9 lists the references used in the
appendices.
entire document.

3
2. Selecting and Evaluating Treatment Processes

This chapter serves as a guide to the entire document. contaminants d.elivered to a user of public water
It begins in Section 2.1 with a brief summary of the supplies.
statutory and regulatory framework that applies to
public drinking water systems. The new statutory • Maximum Contaminant Leuel Goals (MCLG):
and regulatory provisions drive the treatment These are nonenforceable limits that indicate the
objectives for drinking water. Identifying these level of the contaminant that does not cause any
objectives and selecting treatment alternatives are known or anticipated adverse effect on human
discussed in Section 2.2. Section 2.3 summarizes the health. They were formerly termed Recom-
various technological alternatives available to water mended Maximum Contaminant levels (RMCLs).
utilities for complying with the new regulatory
provisions. Section 2.4 discusses the final selection • Secondary Maximum Contaminant Levels
process. (SMCLs): These are nonenforceable goals for
preserving the aesthetic qualities of drinking
water.
2.1 Overview of Federal Drinking Water
Regulations The amendments authorize EPA to set treatment
This overview provides a general context for technique requirements in lieu ofMCLs when it is
discussing the available treatment technologies not economically or technically feasible for water
presented in this document. It is not intended as a suppliers to determine the level of contaminants. In
substitute for the Code of Federal Regulations that addition, the amendments require EPA to set best
describes each provision's many regulatory available technology (BAT) for purposes of complying
specifications, variances, and exceptions. with National Primary Drinking Water Regulations.
For instance, the statute lists granular activated
The 1986 amendments significantly strengthen and carbon (GAC) as BAT for synthetic organic chemical
expand the 1974 Safe Drinking Water Act (SDW A). (SOCl removal. For purposes of complying with the
A major focus of the amendments is establishing a regulations, any other technology, treatment
strict schedule for promulgating drinking water technique, or control strategy selected to remove
regulations that use numerical standards, referred to SOCs must be as effective as GAC. BAT is also set for
as Maximum Contaminant Levels (MCLs), or purposes of issuing variances under Section 1415.
treatment technique requirements for 83
contaminants (see Table 2-ll. Prior to the 1986 In addition to the quantitative provisions noted
amendments, only 22 MCLs had been set. In addition, above, there are several far-reaching provisions on
EPA must regulate 25 additional contaminants every monitoring, filtration, disinfection, and use of lead
3 years beginning in 1991. materials included in the 1986 Amendments to the
SDWA, as follows:
The regulatory provisions derived from the 1986
amendments use the following terms to describe • Monitoring: The amendments instruct EPA to
controls for water contaminants: promulgate regulations requiring monitoring of
certain unregulated contaminants. The
• Maximum Contaminant Levels (MCLs): These monitoring schedule is varied based on the
are the maximum permissible levels of number of persons served by the system, the

4
Selecting and Evaluating Treatment Processes 5

Teble 2·1. ContamInants Required 10 be Regulated Under water sources and procedures for States to
the SOWA Amendmen1. of 1986 determine which systems must filter.
VOl. tile Organic Chemicals
Tnchloroethylene Benzene • Disinfection: All public water supplies will be
Tetrachloroethylene Chiorobenzone required to disinfect their water.
Carbon letrachlonde DichlorObenzene
I, I, 1. Tnchloroelhane TnchlorObenzene • Lead: The amendments prohibit the use of
1,2-0,chloroethane 1,1 -D,chloroethylene solders and fluxes containing more than 0.2
Vinyl chloride trans·1,2 Dlchloroethylene percent lead and pipes and pipe fittings
Methylene chloride CIS'1,2-DichIOfoethyiene containing more than 8 percent lead.
Microbiology and Turbidity
Total coliforms· Viruses
Turbidity. Slandard plate count The three primary entities involved with the
Giardia lamblis Legionel/a regulatory effort for the 1986 amendments are:
Inorg.nlcs
Arsenic· Molybdenum
• U.S. EPA, with the primary roles of national
Asbestos
primary and secondary drinking water
Banum·
regulations, designating BATs, and overseeing
Cadmium' Sulfate
Copper
State programs and enforcement.
Chromium'
Lead' Vanadium
Mercury· Sodium
• States, with the primary responsibility of
Nitrate" Nickel
implementation, program administration, and
Zinc
enforcement.
Selenium'
Silver' Thallium
Fluonde" Beryllium
• Utilities, which will have to increase monitoring,
Cyanide
install new treatment processes, and increase
Aluminum
Anumony
public awareness of contamination problems.
Synttletlc Organics
Endnn' 1,1 ,2,TnchlOfoethane
These are very simplified descriptions of each role.
Lindane- Vydate The statute and regulations require significant
Methoxychlor· Slmazlne
interplay between the three groups.
Toxaphene" PAHs
2,4·0" PCBs
The regulatory effort is divided into the following five
2,4,5·TP' Atrazine
phases:
Ald,carb Phthalates
Chlordane Acrytamlde
• Phase I: Volatile organic compounds
Dalapon Dibromochloropropane (DBCP)
OlQuat 1,2-Dlchloropropane
• Phase II: Synthetic organic compounds,
EndothaJi Pentaehlorophenol inorganic compounds, and unreg-
Glyphosate PlChlOfam ulated contaminant monitoring
Carbofuran Dinoseb
Alachor Ethylene dlbrom,de IEDB)
• Phase III: Radionuclide contaminants
EptChlOfOl1ydnn D,bl'Omomethane
Toluene Xylene
• Phase IV: Disinfectant and oxidant by-products
Adlpates Hexachlorocyclopentadiene
2,3.7,8·Chlonnated dibenzo Tota! tnhaJomethanes· • Phase V: Inorganic compounds and synthetic
lurans (DIOXin)
organic compounds
R.d/onuel/des
Radium 226 and 228' Gross alpha par1Jcle acUVlty· On July 8,1987, EPA promulgated final regulations
Beta parucle and photon Radon for eight VOCs and fluoride. The final MCLGs and
radioactJvlty-
Uranium
MCLs for the VOCs are shown in Table 2·2. The
regulations published July 8, 1987, also include
'Reoulaled poor 10 1986 amendments to SDWA. monitoring requirements for 51 unregulated
Source: U.S. EPA Fact Sheet; February 1989.
contaminants, which are shown in Table 2-3.

source of the supply, and the contaminants likely In May 1989, EPA issued proposed regulations for 30
to be found. SOCs and 8 inorganic chemicals (IOCs), which are
listed in Table 2-4. For two contaminants, epichloro-
• Filtration: EPA was required to specify criteria hydrin and acrylamide, EPA proposed a treatment
under which filtration is required for surface technique in lieu of an MCL and monitoring
6 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Table 2-2. Final MClGa and MCla lor VOCa Table 2-3. Monitoring lor Unregulated VOea
Final MCLOo Final MCl
(mgIL) (mgIL) Required for AU Systems-
Tnch/()(oethylene lero 0.005 ChlOr01orm l,2-Dtchforopropane
Carbon tetrachloride zero 0.005 Bramadichloromethane , ,I,2,2-Tetrachloroethane
Vinyl chlonde zero 0.002 Chlarooibramomethane Ethylbenzene
, .2·D,chloroethane zero 0.005
Bromoform 1,3-Dichloropropane
Benzene zero 0.005
trans-' .2-DlChloroethylene Styrene
para-Dichlorobenzene 0.075 0.075
Chlarobenzene Chloromethane
, . 1 ·Dichloroethylene 0.007 0.007
'.'.' -Trichloroethane 0.2 0.2 m-DichlorObenzene Bromomethane
Dichlararnelhane 1,2,3-Trichioropropane
• Final MClO~ were published Nov. 13. 1985. The MClG and
MCl lor p-dlchlorallenzene were reproposed at zero and 0.005 cis-' .2-DlChloroethyfene 1,1,1,2-Tetrachloroethane
mgIL on Apnl '7. , 987; comment was rt!QueSled on levels of
0.075 mgIL and 0.075 mgIL. respecbvely. O-DlChlorobenzene Chloroelhane
Source: U.S. EPA Fact Sheet February 1989. Dibtomomethane l,l,2·Trichloraethane
1,l-DichlorOpropane 2,2-Dichloropropane
requirements. The proposal also established nine
new secondary MCLs, which are listed in Table 2-5. Tetrachlaroethylene o-Chlorotoluene
Toluene p-Chlorotoluene
In June 1988, EPA issued proposed regulations to p-Xylene Bromobenzene
define MCLs and MCLGs for lead and copper. The
o-Xylene 1,3-Dichloropropane
MCLG proposed for lead is zero, and 1.3 mglL for
copper. The proposed MCLs applicable to water m-Xylene Ethylene dibrom,de
entering the distribution system were 0.005 mglL for , ., -Dichloroethane , .2-Dibromo-3-chloropropane
lead and 1.3 mglL for copper. EPA also was
Required for Vulnerable
considering MCLs and/or action levels for lead and Systems Only:
copper and other related water quality parameters at
the consumers' tap. These proposed regulations are in 1,2-Dibtarno-3-Chloropropane
(DBCP)
various stages oflinalization.
Ethy\erledibromide (EDB)
On June 29, 1989, the Surface Water Treatment Rule At Each State's Discretion:
(SWTR) and the Coliform Rule were promulgated. , .2,4-Tnmelhylbenzene 1,3,5-Tnmethylbenzene
According to the SWTR, all public water systems , .2.4-Trichlorobenzene p-lsopropylloluene
using surface water or ground water under direct
, .2.3-Tnchlorobenzene lsopropyfbenzene
influence of surface water, must disinfect and may be
required to filter if certain source water quality n-Propylbenzene ler1-buty1benzene
requirements and site-specific conditions are not met. n-Buty1benzene sec-butylbenzene
The MCLGs established in the rule are: NaphlhaJene Fluarotrichloromelhane
• Giardia lamblia - 0 Hexac!llarOOuladlene Dichlorooifluoromethane
• Viruses - 0 Bromochloromethane
• Legion.ella - 0 Source: U.S. EPA Fact Sheet Feb<uary 1989.
No MCLGs were set for turbidity and heterotrophic
plate count. Treatment requirements also were
established in lieu ofMCLs for Giardia, viruses,
HPC, Legionella, and turbidity. Treatment must 2.2 Selecting Treatment Technologies
reliably achieve at least 99.9 percent Defining treatment objectives and selecting control
removalJinactivation of Giardia lamblia cysts and technologies involve eight basic considerations:
99.99 percent removalJinactivation of viruses.
• Emuent requirements
The Coliform Rule requires all public water systems • Influent characteristics
to meet the coliform MCL and monitor total coliform • Existing system configuration
with frequencies depending on population served, • Cost
and requires small systems to conduct a sanitary • Operating requirements
survey. To comply with the coliform MCL, no more • Pretreatment and posttreatment components
than 50 percent of all total coliform samples per • Waste management
month can be total coliform-positive. • Future needs of the service area
Selecting and Evaluating Treatment Processes 7

Table 2·4. Proposed MCLG••nd MCLa lor SOC•• nd T.bIe 2·5. Proponc:t SMCl."
IOC. Conlaminant level (mgIL)
EXisting Propo$lld Proposed Aluminum 0.05
NPOWR" MClG MCl
o·Dichlorobenzene 0.0 1
COnlalTlInant (mgIL) (mgIL) lmgi1..l
AcrylamiOe _0 IT" p-DocI1lorobOnzene 0.005
Alachlor Z8fO 0.002 Ethylbenzene 0.03
AldlCarb 0.01 am Pentachlorophenol 0.03
Aldicarb sulfoxide 0.01 0.01
Silver 0.09
Aldicarb sulfone 0.04 0.04
Styrene 0.0 I
Alrazme 0.003 0.003
Carlloturan 0.04 0.04 Totuene 0.04
Chlordane Z8fO 0.002 Xylene 0.02
CIS-l.2· 0.07 0.07 " Other secondary regulabons may be
D,chloroethylene proposed In !he future. as appropnate.
D,bromochloropropane zero 0.0002 Source: U.S. EPA (1989).
(DBCP)
1.2·Dichloropropane zero 0.005 effiuent criteria from regulatory recommendations
o·Dichlorobenzene 0.6 0.6 (e.g., U.S. EPA Health Advisories), professional
2.4-0 0.1 0.07 0.07 organizations, State regulations, and consumer
Ethylenodibromlde Z8fO 0.00005 preferences are also considerations. Comparisons of
(EDB) existing influent characteristics with effiuent
Eplchlorphydrin zero ITb requirements provide the basis for identifying
Ethylbenzene 0.7 0.7 treatment needs. The influent is properly
Heptachlor zero 0.0004 characterized by an historical profile of the chemical
Heptachlor epoxide z8t'O 0.0002 and biological constituents of the water source. Non-
lindane 0.004 0.0002 0.0002 contaminant parameters, such as pH levels, are also
Methoxychlor 0.1 0.4 0.4 useful in characterizing influent because of their
Monochlorobenzene 0.1 0.1 impact on treatment process efficiencies.
PCBs (as zero 0.005
decachloroblphenyl)
Pentachlorophenol 0.2 0.2 Existing System Configuration
Styrene" _010.1 0.00510.1 The adaptability of an existing system configuration
Telracnloroethylene _0 0.005 to larger or different processes is an important
TOluene 2.0 2.0 consideration in assessing treatment options. Also, a
2.4.5· TP2.4.5· TP am 0.05 0.05 system's ability to blend treated water with raw
Toxaphene 0.005 lero 0.005 water can affect its achievement of regulatory
lrans·l.2· 0.1 0.1 standards, which are generally based on contaminant
DIChioroe1l1ylene
concentrations.
Xylenes (total) 10.0 10.0

IOC Cost
Asbestos 7 Fll.d 7 Fll.d Total treatment costs are divided into one-time
Sanum 1.0 50 5.0 capital costs and annual operating and maintenance
Cadmium 0.010 0.005 0.005 costs. Each treatment option has a different
Chromium 0.05 O. I 0.1 combination ofcapital and operating and
Mercury 0.002 0.002 0.002 maintenance costs, with higher capital costs often
Nitrate (as nitrogen) 10.0 10.0 10.0 being associated with lower operating and
Nitnte (as nllrogen) 1.0 1.0 maintenance costs. Smaller systems with limited
Selenium 0.01 0.05 0.05 capital budgets are generally saddled with higher
• NPDWR • NabonaJ Primary Dnnkmll Water RegulatJons. operating and maintenance costs because they cannot
bIT· .Treatment TechnIQue. afford the appropriate capital equipment.
c EPA proposes MCls of O. I mgIl. based on a group C carcmogen
clsssificabon and 0.005 mgIl. based on a B2 classrticatJon.
d 7 million fibersl1iter (only fibers longer than 10 l'IJ1). Operating Requirements
Source: U.S. EPA (1969)
Consistency in the nature and volume of the influent
is a crucial operating parameter. As the consistency
Effluent Requirements and Influent of the influent decreases, the necessity of monitoring
Characteristics and the operating complexity of most systems
Federal drinking water regulations, which include increases. Consequently, the level of operator
effiuent requirements, are the primary factors training and attention required, as well as the
determining water treatment goals. Supplementary amount of instrumentation , controls, and
8 Upgrading Existing or Designing New Drinking Water Treatment Facilities

automation, also increases as influent consistency • Degree of acceptance


decreases. Other important operating considerations • Conditions required for effective operation
include: • Operating and maintenance requirements
• Compatibility with other treatment processes
• Energy requirements
• Chemical availability and consumption rate Table 2-6 lists the treatment options associated with
• Instrumentation and automation each of five categories of treatment objectives:
• Preventative maintenance filtration, disinfection, organic and inorganic
• Noise contaminant removal, and corrosion control. The
• Esthetics table indicates four levels of treatment technology
• Backup/Redundant systems acceptance: experimental, emerging, established,
• Startup phase requirements before full removal and BAT. Experimental technologies have shown
capacity is achieved promise in some applications, but have not been
• Cleaning and backwashing requirements extensively tested. Emerging technologies have
proven themselves in the laboratory, but not in the
field. Established treatments are commonly used in
Pretreatment and Posttreatment Processes the water industry. BAT is a regulatory designation
All water treatment technologies perform differently that indicates the level of contaminant removal
with different pretreatment and posttreatment achievable through specification of a technology or a
processes. The compatibility of all the. processes in technology equivalent, rather than an MCL.
the treatment train is key to achieving individual
treatment goals. For instance, a lower pH is desirable Detailed discussions ofal1 technologies are provided
for maximum efficiency of chlorine disinfection; in Sections 3 through 7, as follows:
however, lower pH accelerates corrosion of the water
distribution system. For successful implementation • Prefiltration and filtration for treatment of
of a treatment technology, all elements of the train primarily turbidity and color - Chapters 3 and 4.
must interact efficiently and effectively.
• Disinfection for treatment of pathogenic
organisms, including Giardia cysts, bacteria, and
Waste Management viruses - Chapter 5.
Waste management is a concern associated with the
removal ofcontaminants in drinking water. Most • Organic contaminant removal, including volatile
treatment processes concentrate contaminants, organic chemicals and other synthetic organic
sometimes hazardous, into residuals that requires chemicals - Chapter 6.
special handling. Such wastes are in the form of
solutions, gases, sludges, and solids. Extensive • Inorganic contaminant removal, including
regulations cover the management of these wastes, products of distribution system corrosion and
particularly those classified as hazardous. radioactive contaminants - Chapter 7.
2.3.1 Rltratlon
Future Service Area Needs
The last major factor in selecting a treatment The three basic regulatory requirements associated
technology is the future of the service and supply with filtration systems are control of turbidity, color,
areas. Population and economic forecasts of the and biological contamination (Giardia cysts, enteric
service area are the basic tools for evaluating future viruses, and coliform bacteria). Other key consid-
demands. Examination and analyses of the erations in selecting a filtration system include:
watershed(s) of present and potential water supplies
are important to determine the vulnerability of • Frequency of the cleaning cycle
supplies to anthropogenic and natural threats. • Chemical requirements
• Operational complexity
2.3 Overview of Available Treatment • Sludge volume and toxicity
Alternatives Each of these factors is site specific. In addition,
This section provides an overview of the available climate will determine whether it is necessary to
treatment technologies associated with all categories house the filter operation, which in some cases is
of treatment objectives. Summary data are presented prohibitively expensive.
for each technology and alternatives are compared on
the basis of: In most conventional filtration systems, mixing,
flocculation, and sedimentation processes typically
• Performance precede actual filtration. While these pretreatment
• Suitability to treatment plant size elements are always found in conventional filtration
Selecting and Evaluating Treatment Processes 9

Table 2-6. Overvl_ 01 Water Treatment Technologies


Technological Options
Trealment Requirements Under the New Stage of Size
to Meet RegulatoIY
Regulations ReqUllements Acceptability SUitability Comments
Filtration of surface water supplies to control
turbidity and micrObial contamination
Conventional filtration Established All Most common: adaptable for
adding other processes
Direct filtration Established All Lower cost alternative to
conventional filtration
Slow sand filtration Established Especially Operationally simple; low cost. but
small, but all requires large land areas
sizes
Package plant tiltratlon Established MosUy small Compact: vanety of process
combinations aVaJlable
Diatomaceous earth Established MosUy small Limited applicability: potentially
filtrabOn expensive for small systems
Membrane filtrabOn Emerging MosVy small Expenmental, expenSive
Cartndge filtration Emerging Small Expenmental, expensive
D,s,nfecbon of all public water supplies
Chlanne Established All Most Widely used method:
concerns about health eltects of
by' products
Chlanne diOXide Established All Relatively new to the United
States: concerns about Inorganic
by· products
Monochloramlne Established All Secondal'f diSinfectant only: some
by-product concerns
Ozone Established All VeIY effective and reqUlles a
secondal'f diSinfectant
UltraViolet radiabOn Established All Simple, no established harmful by-
products and reqUlles secondaIY
diSinfectant
Advanced OXIdation Emerging All Not much Information concerning
(ozone piuS HzO z and diSinfection aspects of this process
ozone plus ultraViolet
radiation)
Orgamc contamination contrOl, Including 50
specific compounds
Granular activated BAT All Highly effective; potential waste
carbon disposal issues
Packed column aeration BAT All Highly effectIVe for volatile
compounds: potential all
emiSSions Issues
Powdered activated Established Large ReqUires conventional treatment
carbon process traJn for application
Diffused aeraMn Established All Variable removal effectiveness
Multiple tray aeration Established All Vanable removal effectiveness
Oxidation Expenmental All By-products concerns
Reverse OSmosiS Emerging Small to Variable removal effectiveness;
medium expensive
Mechanical aeration Expenmental All Mostly for wastewater treatment;
high energy reqUllements, easy to
operate
Catenal'f gnd ExpenmentaJ All Performance data are scarce:
potential cur emiSSions Issues
Higee aeration Experimental Small Compact, high energy
reqUllements: potential aJr
emissions issues
(Continued)
10 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Table 2·8. (Continued)


Technological Options
Slage of SIZe
Treatment Requorements Under the New to Meet Regulato<y
RegulahOns Requirements Acceptability SUItability Comments
ReSins Expenmental Small Data scarce
Ultrafiltration Emerging Small Primarily for turbidity; data for
organics removal are scarce
InorganiC contamination control. Including 36
specific Inorganic contaminants. and 5
radioactive contaminants
Reverse osmosis Established Small to Highly effectJve; expensive:
medium potential waste disposal issues
Ion exchange Established Small to Highly effective; expensive;
medium potential waste disposal Issues
Activated alumina Established Small Highly effectlV9; expensive:
potenlial waste disposal Issues
Granular activated Expenmental Small Expenmental for radionuclide
carbon removal; potential waste disposal
issues
Corrosion centrols pH control Established All Potential to conflict WIth other
Ireatments
CorrOSion Inhibitors Established All Vanable effectiveness depending
on type of Inh,bttor
Source: Denved from tables and text Included in Chapters 3 through 7.

systems, they are also sometimes found in filtration units do not have a coagulation step. The
nonconventional systems. In addition to the three complexity of operating a package filtration plant
prefiltration steps, additives, such as chemicals for varies with the manufacturer and model. Package
pH adjustment, coagulants, coagulant aids, and plants, slow sand, membrane, and cartridge filtration
polymers, are commonly used in conjunction with are most applicable to small systems.
filtration. All of these elements affect the
performance of the filter system. These effects are The removal capacities for Giardia cysts and viruses
discussed in detail in Chapter 3. of the seven filter systems are presented in Table
2-8.
The seven filtration options reviewed in this
document include:
2.3.2 Disinfection
• Conventional filtration
• Direct filtration The recently promulgated SWTR requires water
• Slow sand filtration treatment systems to inactivate 99.9 percent of
• Package plant filtration Giardia cysts and 99.99 percent of enteric viruses.
• Diatomaceous earth filtration Disinfection systems alone, or in conjunction with
• Membrane filtration (reverse osmosis) filtration systems, can meet these requirements. At
• Cartridge filtration the same time, regulations regarding disinfection by-
products in the finished water must be met.
The p<:rformance of each filter type depends on the Currently, total trihalomethanes are the only
quality of the influent and proper design and disinfection by-products regulated by EPA; however,
operation. The range of influent characteristics for new disinfection by-product regulations are
which various filters are effective is provided in anticipated.
Table 2-7.
Primary disinfectants are those used for the
Conventional filtration, with rapid mix, flocculation, inactivation of Giardia cysts, viruses, and bacterial
and sedimentation, is clearly the most versatile in its contaminants, while secondary disinfectants
effectiveness in treating variable influents. suppress biological regeneration in the distribution
Coagulation/filtration systems are more difficult to system. Common primary disinfectants are chlorine,
operate compared to slow sand or diatomaceous earth chlorine dioxide, ozone, and UV radiation. Most are
filters because they involve adjusting water suitable for both ground water and surface water. An
chemistry for proper coagulation. Slow sand, exception is UV radiation, which is only suitable for
diatomaceous earth, membrane, and cartridge ground water because it is not effective against
Selecting and Evaluating Treatment Processes 11

Table 2-7. Typical Influenl CharaeterfallCi and Capacl1lea for Fillralion Technologlea
Coliform Count TYPical Capacity
Filtration Options Turbidity (NTUsl Color (in color units) (per 100 mll (MOD)
Conventional NO rest'lCtJons <75 < 20,000 > All sizes
<14 <40 <500 > All sizes
Slow sand <5 < 10 <800 <15
Package Plant [depends on processes utilized I <6
Diatomaceous earth <5 <5 <50 < 100
Membrane <1 [fouling IndeX ot < 101 <0.5
Cartndge <2 NA NA < 1.0
NA • not available.
1 MOD • 0.044 m3/sec
Source: See tables and text In Chapters 3 and 4.
The preferred application point for the various
Table 2-3. Removal Capacltlea ot Seven Filter Optlona
(percenl removal)
disinfectants are described in Table 2-11. Table 2-12
compares basic operational considerations, pH,
Achievable presence of by-products, operational simplicity, and
Giardia Cyst Achievable Virus maintenance required.
Filtration OplJOns levels levels
Conventional 99.9 99.0
Direct 99.9 99.0 One of the most important considerations in
SlOw sand 99.99 99.9999 assessing disinfectants is balancing inactivation or
Package plant vanes WIth manufac1urer biocidal effectiveness with by-product production.
Diatomaceous earth 99.99 > 99.95
The by-products of greatest current concern are
trihalomethanes and other halogenated organic
Membrane 100 Very low
compounds; chlorine has the greatest potential for
Cartridge > 99 lIttle 'data aV81lable generating harmful by-products. The amount of these
Source: See tallies and text In Chapters 3 and 4. by-products produced by chlorine is affected by:

Giardia cysts. Secondary disinfectants include • Chlorine dosage


chlorine, chlorine dioxide, and chloramines. • Types and concentrations of organic material in
the influent
Disinfection effectiveness is measured in terms of the • Influent temperature
residual concentration and length of contact time • Influent pH
necessary to achieve the desired inacti vations. Four • Contact time for free chlorine
chemical disinfectants listed in descending order of • Nature of residual (free chlorine vs. combined
their effectiveness are ozone, chlorine dioxide, chlorine)
chlorine, and chloramines. The effectiveness of a • Presence of bromide ion
particular disinfectant is also influenced by water
quality, temperature, and pH. Lower disinfectant Ifchlorine produces an unsatisfactory level of by-
dosages may be used when: products, then other disinfectants are potential
alternatives. Chlorine dioxide is effective, but the
• There is filtration or oxidation prior to the total levels of chlorine dioxide and its
disinfection step. oxidation/reduction products may limit its
• The water temperature is high. applicability. Ozone and UV radiation are very
effective primary disinfectants, but require the use of
The dosage of chlorine required for effective secondary disinfectants. Ozone will produce harmful
disinfection is reduced as the pH of the water is halogenated by-products with influent containing
reduced. A qualitative summary of the advantages bromide ions. It will also produce harmful oxidation
and disadvantages of the five disinfectants is products in the presence of certain synthetic organics
provided in Table 2-9. such as heptachlor. Chlorine dioxide and
monochloramines are effective secondary
Some disinfectants are unstable and, therefore, have disinfectants, but require careful dosage and
to be generated on site. Production considerations for application management to avoid producing harmful
the five disinfectants are contained in Table 2-10. by-products in finished water. (Chapter 5 contains a
The table also provides the number and type of more complete discussion of disinfection by-products
alternative production methods. and strategies for their con troLl
12 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Table 2-9. Advantages and Disadvantages of Five DlslnfectanlS

Disinfeclant Advantages Disadvantages


Chloone EffectMl. Widely used. Vaflety of possible awtication Harmful halogenated by·products. Potential conn,ct With
points. InexpensIVe. ApprQpnate as both pnm3IY and corrOSion contrOl pH levels. when used as a second3IY
second3IY dislnfeclant dlSinfeclant.
Ozone Very effectMl. Minimal harmful by'products Identified 10 ReQUt'es secondary disinleclanl Relatively high cost
date. Enhances slow sand and GAC fillers. Provides More complex operattons because It must be generated
oXldabOn and disinfecnon In the same step. on·Slte.
Ultraviolet radianon Very effectMl for VIrUseS and bactena. Readily Inappropnate lor water With Glard;a cysts. high
available. No harmful residuals. Simple operation and suspended solids, high color. high lurbldity, or soluble
mainl8nance. organics. ReqUires a secondary diSinfectant.
Chlonne dioXIde Effecnve. Relanvety low cost. Generally does not Some harmful by·products. Law dosages currently
produce THMS. recommended by U.S. EPA may make It IneffectIVe.
Must be generatad on·Slte.
Chloramlnes Mildly effective 10< baclena. Long-lasting reSidual. Some harmful by·products. ToxIc effects lor kidney
Generally does not produce THMS. dialYSIS pattents. Only recommended as a second3IY
diSinlectant. Ineffecttve against Ytruses and cysts.

Source: See tables and text in Chapter 5.

Table 2-10. Disinfectant Production Considerations Table 2-11. Desired Points of Disinfectant Application>
Chlonne Mono- Ultraviolet ChlOrine Towards the end of the water treatmenl
Chlonne Dioxide chloramine Ozone Radiabon process to minimize THM formatton and
Chemically Yes Yes Yes No NA proVide secondary d,s,nlecnon
stable Ozone Pnor to the rapid mixing step In all Ireat·
On-Slle No Yes Yes Yes Yes ment processes. except GAC and conven·
producnon tlonal treatment processes: pno< to filtra-
reQUt'ed tton for GAC; post·sed,mentation lor con-
ventional treatment. In addition. suffiCient
Number of NA 3> 2b 3< NA ttme lor biodegradation 01 the OXidation
alternanve products of the ozonallon of organic com·
on-sote pounds IS recommended pnor to
generation secondary diSinfection.
techniQues
UltraViolet radiation Towards the end of the water treatment
>Includlng: process to minimize presence of other
,. Treallng sodium chlonte solutIOn wllh chlorine gas contarnlnanlS that Interlere With thiS
2. Treating sodium chlorite soIubOn With sodium diSinfectant
hypochlO<lte and mineral acid
Chlorine dioxide Pnor 10 filtratton; 10 assure low levels of
3. Treattng sodium chlorile solution With minerai acid
blncludlng: CIO" CIO,-, and CIO,. Ireat With GAC after
diSinfection.
1. Adding ammonia to a water and chlOrine solution
2. Adding chlonne 10 a water and ammonia solution Monochiorarnines Best applted towards the end 01 the
<Including: process as a second3IY disonlectant
1. Ambient (lIr
2. Pure oxygen > In general, diSinfectant dosages Will be lessened by placing the
3. Oxygen-ennched air point of appltcatton towards the end 01 the waler treatment process
NA • not applicable. because 01 the lower levels 01 contaminants that would Interlere
Source: See text and tables In Chapter 5. WIth effH:lent diSinfection. However, water plants With short
detention limes In clear wells and With nearby first customers may
be reQUired to move thel' point 01 diSinfection upstream to attain
the approprtale CT value under the $urlace Water Treatment
Rule.
2.3.3 Organic Contaminant Removal Source: See Ie", and tables ,n Chapter 5.

The 1986 Amendments to the SDWA require the


establishment of new MCLs for many organic The removal capabilities of selected treatment
contaminants, including disinfection by-products. technologies for several organic contaminants and
The regulations designate BATs as well as MCLs for contaminant classes are provided in Table 2-13. The
organic contaminants. The July 1987 regulations first listed treatment, coagulation/filtration is widely
specify packed column aeration (PCA) and CAC as used primarily for turbidity and microbial
the BAT for seven of the eight VQCs. PCA is BAT for contaminant control, but can be somewhat effective
vinyl chloride. the eighth VQC. Water utilities with in removing certain organic compounds. CAC and
these contaminants will have to provide removals at PCA are designated as BAT for many of the organic
least equivalent to those achieved by the designated chemicals and are much more effective. Costs for
BATs. either technology vary depending on the contaminant
Selecting and Evaluating Treatment Processes 13

• Water quality characteristics


Table 2·'2. Disinfectant Application Considerations
• Materials subject to corrosion
Mono-
CI 2 CI0 2 chloramine 0, UV
Opbmum 7 6-9 7·8 6 NlA The two most significant water quality
water pH
characteristics that influence corrosion are pH and
By·products Yes Yes Yes Yes No carbonatelbicarbonate alkalinity. Components of the
present
distribution system subject to corrosion include pipes,
OperabonaJ Yes No No NO" Yes valves, meters, plumbing, solder, and flux. The
simplicity
longer the contact time between the water and
MainlllnatlCe Low Low Low High High corrodible materials in the distribution system, the
reqUifed
higher the concentration of dissolved metals in the
OOperabonaJ/y simplified with an automated system. drinking water.
NA s not applicable.
Source: See teXl and tables ,n Chapter 5.
The four general types of corrosion controls are:

removed and whether waste management is a • Adjustments to water pH


potential issue. Powdered activated carbon (PAC), • Addition of corrosion inhibitors to the water to
the fourth treatment listed in the table, is only suited form protective coatings over the potentially
for application to conventional systems with rapid corrodible metal
mix, flocculation, sedimentation, and filtration • Electronic cathodic protection
components. PAC also addresses some taste and odor • Applied coatings and linings
problems. This form of activated carbon is especially
suitable for seasonal organic contaminant problems The most commonly used corrosion control is pH
since it can be added as needed. In some instances, adjustment because it is inexpensive and easily
however, such high dosages of PAC are required to applied. Table 2-15 compares corrosion controls.
achieve organic removal that waste management
becomes a problem. Corrosion controls involving pH adjustment may
conflict with ideal pH conditions for disinfection and
All other listed treatments, including the different control of disinfection by-products. The treatment
aeration configurations, are less established. The methods selected for both treatment objectives should
aeration treatments, developed for specific be carefully coordinated to avoid diminishing the
applications, are generally equal to or less effective effectiveness of either process.
than PCA, but have higher energy requirements. For
some, controlling biological growth in the systems is
a problem. The remaining treatments, besides the
aeration technologies, all tend to be very effective in 2.3.4.2Inorganlc Contaminant Removal
removing contaminants, but their application to There are 10 treatment processes for addressing the
specific organics is still experimental. The many inorganic contaminants, including
efficiencies of all treatments depend on the type and radionuclides. Most treatment processes are effective
concentration of the contaminants. While for only a specific set of contaminants under certain
pretreatment is not al ways required, it can increase circumstances. Fortunately, all of the inorganic
the effectiveness of some of the treatments. Table 2- contaminants do not often occur simultaneously. The
14 illustrates the .variation in operating conditions most appropriate applications for each treatment
for these treatments. process are shown in Table 2-16. This table
distinguishes between ground water and surface
water by the generally higher concentrations of
2.3.4 Inorganic Contaminant Removal and suspended solids in most surface water. Table 2-17
Control provides removal efficiencies for specific
Inorganic contaminant treatments are categorized as contaminants with the same treatments. For radon
prevention strategies or removal technologies. removal, which is not included in the table, aeration
Corrosion controls prevent or minimize the presence is highly effective.
of corrosion products (inorganic contaminants) at the
point of use. Removal technologies treat source water The most important factors affecting inorganic
that is contaminated with metals or radionuclides. contaminant removal are:

• Contaminant type and valence


2.3.4.1 Corrosion Controls • Influent contaminant concentration
Corrosion controls address the two primary aspects of • Influent levels of dissolved solids and pH
corrosion: • Desired effiuent concentration
14 Upgrading Existing or Designing New Drinking Water Treatment Facilities

T8ble 2·'3. Tre81ment Technology Remov8l ENec1tv_ Reported lor Org8nlc Cont8mln8nt8 (percent)
Coagula1iool Diffused Reverse
Contafnlnanl Filtration GAC PCA PAC Aeration OXidation" OSmosiS
Acrylamide 5 NA 0·29 13 NA NA 0·97
Alachlor 0-49 70-100 70-100 36-100 NA 70·100 70-100
Aldica1l NA NA 0·29 NA NA NA 94-99
Benzene 0-29 70-100 70·100 NA NA 70·100 0-29
Carboluran 54-79 70-100 0-29 45-75 11-20 70·100 70-100
Carbon tetrachloride 0-29 70-100 70-100 0-25 NA 0-29 70-100
Chlordane NA 70·100 0-29 NA NA NA NA
Chlorobenzene 0-29 70-100 70-100 NA NA 30-69 70-100
2.4-0 0-29 70-100 70-100 69-100 NA W 0·65
1,2-Dichloroethane 0-29 70-100 70-100 NA 42·77 0-29 15-70
1,2-DichIorQPfopane 0-29 70-100 70-100 NA 12-79 0-29 10-100
Dibromochlompropane 0-29 70-100 30-69 NA NA 0-29 NA
Dichlorobenzene NA 70-100 NA NA NA NA NA
o-Dichlorobenzene 0-29 70-100 70-100 38-95 14-72 30-88 30·69
p-Dichlorobenzene 0-29 70-100 70-100 NA NA 30-69 0-10
1,I-DichlO<OBthyiene 0·29 70-100 70-100 NA 97 70·100 NA
cis- I ,2·DlChloroethytene 0-29 70·100 70·100 NA 32·85 70·100 0·30
trans-l.2- 0-29 70·100 70-100 NA 37-96 70-100 0·30
Dichlo<oethylene
EplChlorohydnn NA NA 0-29 NA NA 0-29 NA
Ethylbenzene 0-29 70-100 70-100 33-99 24·89 70-100 0·30
Ethylene dlbmmlde 0-29 70-100 70-100 NA NA 0-29 37-100
Heptaehlpr 64 70·100 70-100 53-97 NA 70·100 NA
Heptachlor epaxlde NA NA NA NA NA 26 NA
High molecular wlltght NA W NA NA NA NA NA
hydrocartlons (gasoline.
dyes. armnes. humlCS)
Llridane 0-29 70-100 0-29 82-97 NA 0-100 50-75
Methoxychlor NA 70·100 NA NA NA NA >90
Monochlorobenzene NA NA NA 14-99 14·85 86-98 50-100
Natural organiC material P P NA P NA W P
PCBs NA 70·100 70-100 NA NA NA 95
Phenol arid NA W NA NA NA W NA
chloropllenols
Pentaehlorophenol NA 70·100 0 NA NA 70-100 NA
Styrene 0-29 NA NA NA NA 70-100 NA
Telrachlo<oelhylene NA 70·100 NA NA 73·95 W 70·90
Tnchlo<oelhylene 0-29 70-100 70-100 NA 53-95 30-69 0·100
Tnchlo<oethane NA 70·100 NA NA NA NA NA
1. 1.1-Trichloroelhane 0-29 70·100 70-100 40-65 58·90 0-29 15·100
Toluene 0·29 70·100 70·100 0-67 22·89 70-100 NA
2,4,5·TP 63 70-'00 NA 82·99 NA 30-69 NA
Toxaphene 0-29 70·100 70-100 40-99 NA NA NA
Vinyl chloMe 0-29 70·100 70-100 NA NA 70-100 NA
Xylenes 0-29 70·100 70-100 60-99 18-89 70-100 10·85
W well removed.
P poorty removed.
NA ~
nol avaJiable.
"The specifics of the OXidation processes effective in remOVIng each contaminant are provided In Chapter 6.
Note: Little or no specific performance data - . , aV31lable tor:
1. Multiple Tray Aeration
2. Calenary Aeratron
3. Hlgee Aeratron
4. ReSins
5. Ultrafiltration
6. Mechanical Aeration
Source: See text and tables In Chapter 6.
Selecting and Evaluating Treatment Processes 15

Table 2·14. Operational CondltJonl for OrganIC Table 2·16. MOlt Probable Treatment Application tor
Trutmenta Inorganic Treatments

level ot level ot Energy Most Probable Application


Operational Skill Maintenance Require-
Treatment Source Water Contaminants
Technology Required Required ments
CoagulatlOnl High High low Conventional WIth Surface Ag. As. Cd. Cr. or
Filtration coagulation Pb

GAC Medium lime softening Surface (hard) As V. Cd. Cr III, F,


low low
or Pb
PCA low low Varies
Ground water Ba or Ra
PAC low Medium low
Ion exchange: cation Ground water Ba or Ra
Diffused low low Vanes anl()(l Ground water NO),Se
aeration
Reverse osmOSIS Ground water All
Multiple tray low low Low
aeration Powdered ac~vated Surface water Organic Hg from
carbon spills
Ox1da~on High High Vanes
Granular ac~vated Surfac9 waler Organic Hg
Reverse High High Htgh carbon Ground water Organic Hg
osmosis
Activated alumina Ground water As, F. or Se
Mechanical Low Low Low
aeration Source: See tables and text In Chapter 7.
Catenary goo Low Low High
Higee Low Medium High
aeration chemicals (particularly VOCs), radon, lead, and
Res",s Medium Medium low disinfection will have the most impact.
Ultrafiltration Medium High Medium
Balancing the processes selected for microbiological
Source: See text and tables In Chapter 6.
control, disinfection by-product removal, and lead
control will be important for both surface and
ground-water treatment. One water quality
Table 2-15. Con-Olion Control Conllclenltlone parameter that affects all three processes is pH. A
low pH is desirable for disinfection, disinfection by-
Amount ot Completely Optimum
System
product removal, and minimizing by-product
Agreeable Level of
Covered by willl Additive in formation in the distribution system. However,
CorrOSIon ContrOls the Control Consumers the Water minimizing lead corrosion in the distribution system
Water Treatment requires a higher pH. One option to meet all
pH adjustment Total Yes >6.0 pH conditions involves use of a corrosion inhibitor, thus
permitting a lower pH in the finished water for
Inorganic Total Usually Vanes
phosphates disinfection and by-product control.
Silicates Total Yes 2·12 mgIL
For ground-water supplies, process selection should
Cattlodic Partial Yes NA consider treatments capable of controlling a number
ProtectIon
ofcontaminants. For example, PCA can remove both
CoaMgs and P~ Yes NA
Llmngs
VOCs and radon. It also removes carbon dioxide from
the water, which raises the pH to levels desirable for
NA • not applicable. lead corrosion control. As another example, oxidation
Source: See tables and text In Chapter 7. techniques, such as ozone, may oxidize organic
chemicals as well as provide disinfection. Also,
The first three factors are site specific. Table 2-18 reverse osmosis might be considered when both
summarizes the advantages and disadvantages of organic and inorganic chemical contaminants are
each of the treatment mechanisms. present in the raw water. Use of one treatment
technique to meet multiple regulatory requirements
will help reduce operating complexity and costs.
2.4 Final Process Selection and Design
When modifying an existing or designing a new Evaluation of potential treatment technologies
treatment system, consideration must be given to all proceeds from literature searches to bench-scale or
the regulations impacting the system. For surface field tests. Sludge volume and composition are also
water, regulations regarding microbiological, important considerations in the fina I design
disinfection by-product. and lead contaminants are selection. Construction costs for the final alternatives
most likely to impact the treatment system. For are used to select the appropriate option before
ground-water systems. regulations regarding organic developing the final design.
..
-;
0-
I en
iD
~ C
U
:-' to
-.
Q)
::c
lI>
0-
3 :::l
0 to

Table 2·170 Removal Effectiveness lor Nine Processes by Inorganic Contaminant


<
!!!. m
X
!:! <:no
Contaminant iO =0
Hg(l, 2
Treatment Ag As

-
ASIU Asv Sa Cd Cr

-
Crill CrVl F Hg

-
HgIO' N0 3 Pb Ra Rn Se

-
Sel(\l' Se(1l1 U
.
<" to
:::l

..."
ConvenlJOnal H M H L H H H L M M L H L - M L M 0
-.
lI>
treatment
0
CoagulatIOn
aluminum
° H - - H - M - H - - M - - - H - - - - - - Q
z '"
.0
It>
0
;- :::l
CD 50
CoagulatIOn ° M - - H - - - H H - - - - - - - - - - - - ." to
Iron 0
....
n Z

...
It>
lime softening - - M H H H - H L M - L M L H H - - M l H ~

Reverse OsmoSIS H - M H H H H - - H H - - M H H - H - - H 0-
-<
-.0
and electrodialySIs ;- :::l

"50
Cation exchange - l - - H H - H l l - - - l H H - l - - H ."
0
.0
to
n° ~
Anion exchange - - - - M M - M H - - - - H loA loA - H - - H n°
0 ....
Q)

Activated alumina - - H - l L - - - H - - - - - l - H - - - "iO3 -.It>


-i
Powdered
aCllValed carbOn
L - - - l M - l - L - M M l - l - - - - - ..
;-
:!
-.
It>
....
Q)

3
Granular - - - - l M - L - L - H H, L - l H - - - - It>
....:::l
activated carbon
"T1
Q)
H • High • > 80% removal. n
M • Med,um • 20-80% removalo
L • Low • < 20% removal. ....(b0
- • indicale no data were providedo en
Source: See text and tables in Chapter 7°
Selecting and Evaluating Treatment Processes 17

The chapters that follow elaborate on each group of


Table 2-18, Advantage. and Dlaadvantalln 01 Inorganic treatment techniques. Filtration technologies are
Contaminant Removal Procnan
discussed in Chapters 3 and 4, disinfection in
Precipitation and Coprecipltation Used in CoagulaoonlFillrati'on Chapter 5, organic contaminant treatments in
Advan1ages Chapter 6, and technologies that control corrosion
• Low cost for high volume and remove inorganic compounds and radionuclides
• Often Improved by high Ionic strength in Chapter 7,
• Reliable process well SUited to automatIC control
Disadvantages
• Stoichiometric chemical addi60ns reQUl/ed
• High-water-content sludge must be disposed of
• Parts-per-billion effluent contaminant levels may reQuire two-
stage preclplta60n
• Not readily applied to small, intermittent flows
• Copreclpitatlon effiCiency depends on Inlbal contaminant
concentra60n and surface area of pnmary floc
Ion Exchange
Advantages
• Operates on demand
• Rela6vely insensl6ve to flow variations
• Essenbally zero level of effluent contamlna6on poSSible
• Large variety 01 specific resins aV81lable
• Beneficial selectivity reversal commonty occurs upon
ragenerabon
Disadvantages
• Potential for chromatographIC effluent peaJ<ing
• Spent ragenerant must be diSposed of
• Vanable etlluent quality with respect to background ions
• Usually not feasible at high levels of total dissolved solids
Activated Alumina
Advantages
• Operates on demand
• InsenSitive to flow and total dissolved solids background
• Low effluent contaminant tevel posSible
• Highly selectrve for fluollde and arsenic
Disadvantages
• Both acid and base are required for ragenerabon
• Media tend to dissolve. prodUCing fine partICles
• SlOw adsorption kInetICS
• Spent ragenerant must be disposed ot
Membranes (Reverse Osmosis and Bectrodialysis)
Advantages
• All contaminant IonS and most dissolved non·ions are
removed
• Relawely ,nsens,t,ve to flow and total dissolved solids
level
• Low etlluent concentration PQSSlble
• In reverse osmOSIS. bacteria and parbCles are removed
as well
Disadvantages
• High capital and operating costs
• High level of pretreatment required
• Membranes are prone to fouling
• Relect slteam IS 20-90% of feed flow
Source: Clifford (1986),
3. Prefiltration Treatment Elements

Conventional treatment is the most widely used train Table 3-1. Summary Of COnventional Treatment Operating
luu. .
of processes to control microbial and turbidity levels
in surface supplied drinking water. The precise order Treatment
and composition of conventional treatment elements Element Potential Causes of Problems
are not invariably the same, but commonly include ChemICal feed -Choice at chemlCal(s)
chemical feed, rapid mix, flocculation, sedimentation, -Choice of chemical dose and pH
-Control of chemical addit>on; perlormance
filtration, and disinfection. of chamlC3l pumping equipment
-Mamtananca at chemICal fead hoes
The performance ofa filtration technology is greatly -f'lexibility in teed system 10 allow tor
impacted by the processes that precede it. changing the poont of addition. adding
Prefiltration elements, including chemical feed, chemicals at more than one poon~ etc.
rapid mix, flocculation, and sedimentation may need -sequence of adding different chemicals
-Degree of dilution of ctlemlC3ls before injection
upgrading to improve overall system performance or
Rapid mIx -Type 01 rapod mIx; irHioe versus mechanlC3l mix
accommodate system expansions. This chapter -Number of rapid mixers
presents typical problems and recommended -Method of chemICal addition
solutions associated with prefiltration processes. -Mixing spaadldatention time
Flocculation ~ptimum detentIOn time
It is often necessary to modify existing conventional ~ptimum mixing intensity
treatment plants due to changes in population. -Number of stages
financial constraints, raw water quality. and -AclaQuate baffling to approximale plug now
conditIOns
regulatory atmosphere. These changes may upgrade
Sedimentation -Surlaca loading rates
the present system to: -Short Circuiting due to wIOd. temperature.
density differences, Inlet and ouUet deSIgn
• Improve the water quality -Amount and rate of accumulation of sludge
• Increase plant capacity -Sludge removal
• Improve reliability Filtration -f'ilter rate and rate control
• Reduce maintenance -Hydraulics
-ChemlC3l pretreatment of water reaching the
• Reduce costs tilter
-l nadequate backwashlng
Upgrading conventional treatment plants can effect
significant improvements. For example, upgrading
can increase the capacity of most soundly constructed
plants by 100 to 200 percent, provided traditional
design parameters were used in the original plant deficiencies for each treatment element. Techniques
installation. for upgrading fall into five basic categories, according
to treatment element:
Any of the five basic components of conventional
treatment can be upgraded. alone or in combination. • Modifying chemical treatment or dosage
A deficiency, or a set of deficiencies, in one or more • Modifying or adding rapid coagulant mixing
components of conventional treatment will typically • Improving flocculation
prompt an evaluation of the potential for upgrading. • Improving sedimentation
Table 3-1 summarizes the causes of potential • Modifying filtration processes

18
Prefiltration Treatment Elements 19

The remainder of this chapter discusses the first four


Table 3·2. Chemical Costs for Conventional Treatment
categories of modification. The last category, ($1986)
modifying filtration processes, is discussed in Large ( > 1 MGD)
Chapter 4. Small « 1 MGD)
Chemical Systems, $/lon Systems. $/ton
Alum (dry) $ 500 $ 250
3.1 ModIfying Chemical Feed Alum (liquid) 300 125
Proper chemical feed is critical to optimal Lime (Quick) 100 75
performance of flocculation, sedimentation, Lime (hydrated) 150 100
filtration, and disinfection systems. The three basic Ferric chloride 500 275
aSj)f!cts of chemical feed systems are chemical type,
Ferrous sunate 277 250
dosage management, and method of chemical
application. This chapter addresses these topics and Ferric sultate 200 155
the costs associated with modifying chemical feed Soda ash 250 200
systems. Sodium hydroXide 595 316
Chlonne 500 300
3.1.1 Chemical Type Sodium hypochlonte 190 150

The three types of chemicals usually applied to the LKluid carbon diOXide 350 100
raw water at the beginning of the conventional Sodium 1,160 1.100
treatment train are coagulants, coagulant aids, and hexametaphosphate
pH control substances. Some polymer coagulants are Zinc orthophosphate 1.520 1,000
used to coat filters during backwash cycles to AmmonlB. aqua 230 200
increase their effectiveness. However, this Ammonia, anhydrous 410 350
application is discussed in the chapter on filtration. Sulfuric acid 140 100
Hydrochlonc acid 171 166
Coagulants Acbvated carbon 950 800
Coagulants are chemicals used to remove turbidity powdered
and organic substances from raw water by Activated carbon 1,900 1.600
granular
precipitation. Coagulation also removes bacteria,
algae, color, iron oxide, magnesium oxide, calcium Acuvated alumina 1.694 1,156
carbonate, and clay. Coagulation's effectiveness in Potasslum 2.800 2,500
removing humic acids is important because humic permanganate
acids combine with chlorine to form trihalomethanes. Sodium bisulfate 909 673
Coagulants act by overcoming the charges of (anhydrous)
suspended particles thus allowing larger particle Sodium SIlicate 400 200
groupings to form. They are introduced during the Sodium chloride 105 85
initial stage of water treatment just before the rapid Polyelectrolyte 1.500 1,000
mixing step. Coagulants are effective by themsel ves
Diatomaceous earth 680 310
and in conjunction with coagulant aids and pH
modifiers. The two most common coagulant types are Magnesium 650 582
metallic salts and polymers; the most common Sodium chlonte 3.200 2,800
metallic salt coagulants are aluminum sulfate (alum) SodJum hydrOXide 590 316
and ferric chloride. 76%
Sodium bicarbonate 490 380
The selection of a particular coagulant depends on Calcium hypochlonte 2,700 1.540
the required level of effectiveness. A standard jar test
1 MGD ~ 0.044 mJ/sec
is the recommended method to determine the relative
effectiveness of various coagulants for a particular
raw water supply. Jar test procedures are described
later in this chapter. If different coagulants are Ferric chloride, other metallic salts, and polymers
equally effective, then the factors that should be are less widely used. Alum's performance, however,
considered are cost; availability; overall safety; and is greatly affected by the pH of the influent. The
ease of storage, handling, and application. Metallic recommended dosage of alum ranges from 5 to 150
salts, for example, are much less expensive than mg/L, but the problem of sludge disposal increases
polymers, as shown in Table 3-2, which lists a variety with higher alum dosages.
of chemical costs.
Due to health concerns about aluminum, some water
Alum is the most widely used coagulant because of its utilities are considering switching to ferric chloride.
availabilitY,low cost, ease of use, and ease of storage. Although ferric chloride is not always as effective as
20 Upgrading Existing or Designing New Drinking Water Treatment Facilities

alum in reducing trihalomethane formation potential pH and Alkalinity Controls


(THMFP) and total organic carbon (TOC), it is more
effective than alum for water with high dissolved The optimum pH range makes metallic coagulants
color, low turbidity, and a moderate pH. insoluble and improves the strength of the floc, and
stronger floc enhances turbidity removal through
Polymers are effective coagulants, coagulant aid~, sedimentation and filtration. The water must contain
and filter aids. They consist of chains of monomers sufficient alkalinity for aluminum or iron floc
and are classified according to their charge or lack of formation. When pH is too low, adding soda ash,lime,
charge. A polymer with a charge is an ionized sodium hydroxide, or sodium bicarbonate will
polymer, or a polyelectrolyte. Ifa polymer has a increase the water's pH and alkalinity. Without
positive charge, it is cationic; if it has a negative sufficient alkalinity. the coagulation process will not
charge, it is anionic; and if it has no charge, it is non- proceed to completion. Very low pH levels will also
ionic. lead to high aluminum ion residual in the finished
water, which is undesirable.
Cationic polymers are effective for coagulating
. negatively charged clay particles, while anionic When the pH level is too high due to factors such as
polymers are not very effective for this purpose. algae and aquatic plant activity, adding acid is
Anionic polymers are generally more effective for necessary. Typical acids used for this application are
removing certain positively charged colloids and sulfuric acid and hydrochloric acid.
when used in conjunction with alum or other metallic
coagulants. Nonionic polymers are effective filter
aids. Polymers are not effective for coagulating some 3.1.2 Chemical Dosage Management
organic colors. Polymer selection requires frequent Coagulant dosage management is a universal
use of the standard jar test to adjust and compensate problem that is critical for achieving low turbidities
for seasonal changes in the influent characteristics. in any plant, both well equipped and poorly equipped.
Too little coagulant will result in excessive turbidity;
In applications where polymers are effective, dosages too much coagulant could also result in excessive
are generally lower than alum dosages for the same turbidity and wasted coagulant. Coagulant aid and
effect. Typical polymer dosages range from 1.5 to 10 pH control dosages are determined using jar test
mglL. Consequently, polymer coagulants produce procedures.
less residual sludge than alum. To avoid large sludge
volumes, there is a trend towards replacing alum While coagulant type is largely determined by
with either cationic polymers, cationic polymers plus influent characteristics, coagulant dosage is chiefly
alum, or sodium aluminate. The use of cationic determined by effiuent testing. The tests are
polymers instead of alum is most applicable to raw performed either in a batch reactor or on a continuous
water with turbidity of less than 5 NTU and where basis. Batch tests provide indications of proper
direct filtration is feasible. coagulant dosage selection at a single point in the
influent stream, but continuous tests are preferable
because they can be used to constantly monitor the
coagulation process and adjust operations. The three
Coagulant Aids basic tests are the jar test, particle charge
Coagulant aids are added to the influent after or observation technique, and pilot filters. Turbidity
simultaneously with the primary coagulant to monitoring provides an indication of coagulation
improve particle capture during flocculation, poor effectiveness.
flocculation basin performance,clarification
efficiency, and retention in filters of the coagulated
particles that agglomerate (floc) Coagulant aids, Jar Tests
through formation of faster settling floc, will improve The most common technique, the standard jar test,
the sedimentation process and reduce turbidity involves laboratory tests of samples, the results of
loadings on filters, thus potentially extending filter which are extrapolated to full-scale operations.
operating cycles. Nonionic and anionic polymers are Different coagulant dosages are applied to influent
used as coagulant aids. Nonionic polymers samples, with each dosage level in a different jar. For
strengthen floc. prescribed periods oftime, the resultant turbidity
levels are observed to derive the optimal dosage.
Coagulant aid dosages are stated in terms of their
ratio to the amount of alum added to th:l influent. Jar tests are simple. but certain aspects require
Dosages are determined by raw water characteristics attention. Because jar tests are not very accurate for
and operational factors. The ratio of alum to polymer low turbidity water, coagulant dosages for influent
dosages ranges from 100: 1 to 50: 1. Again, standard turbidities less than 5 NTU are difficult to ascertain
jar tests are required to determine precise coagulant with only jar test results based on settled water
aid dosage. turbidity. In such cases, it may be appropriate to
Prefiltration Treatment Elements 21

filter settled water through filter paper before operations, because they require a high level of skill
measuring turbidity. Careful attention is necessary and patience to perform and interpret the results.
to ensure that the jar test water temperature is the
same as that of the plant influent. In addition, jar
tests are inappropriate for establishing expected
performance ofdirect filtration plants because the
liedimentation process simulated in the jar test is not
part ofa direct filtration system.

Zeta Potential Observation Procedures


The zeta potential observation technique is a batch
test method that measures particle charge with
electrodes. However, it can be adapted to a
continuous test method using a streaming current
detector (SCD), which provides a continuous
indication of partic Ie charge.

The particle charge test relates electrode output to


the zeta potential, a measure of the electrical
Synchronous
potential between the water and the counter ions Reclifler
surrounding the colloid particles. Colloids are
usually negatively charged and resist coagulation.
Different colloids require different coagulants to Figure 3·1. Simplified diagram of streaming current
detector,
adjust this zeta potential and thus enhance their
removal from the water stream. Zeta potential after
coagulation ranges from + 5 to -10 minivolts (mv).
Organic colloids require near 0 zeta potentials for Pilot Filters
coagulation, while clay colloids require negative zeta Pilot filters measure the water quality on a
potentials. Particle charge observation tests are continuous basis. They provide an accurate and
calibrated with jar tests or filtrate monitoring. direct measure of the expected plant emuent
turbidity, avoiding the intermediate step of
The results of an SCD, which continuously measures extrapolating lab tests to predict plant-scale
indicators of zeta potential, can be used to continually performance. Pilot filters determine the adequacy of
control the chemical feed pumps. Streaming current the coagulant dosage and are used with manual or
detectors provide information on coagulant dose more automated coagulant adjustment mechanisms. Their
rapidly thanjar tests. l Therefore, they enable oper- accuracy in monitoring emuent turbidity can achieve
ators to optimize chemical doses and cut chemical significant cost savings through improved efficiency
costs, especially in systems with variable influent of coagulant usage. Pilot filters have the advantage of
quality. Use of an SCD improves plant operations, providing indications of floc strength and filter
but is not a substitute for an effective operator. headloss rate. This information is useful for filter
Figure 3-1 is a schematic diagram of an SCD. management and minimizing backwash water.

The matching of the zeta potential with specific plant In general, pilot filters are used to determine whether
output must be carefully observed because, while an applied coagulant dosage is adequate to produce
there is a high correlation between zeta potential and acceptable quality water from the plant filters, A
finished water quality, there is no causal small sample of water, taken immediately after the
relationship. The lack of careful scrutiny of the addition of coagulants, is processed through a small
relationship between particle charge observation test (usually an 11.43-cm [4.5-in] inside diameter
results and plant emuent will lead to false plexiglass column) pilot filter. The turbidity of the
conclusions about the finished water's clarity. Bench- pilot filter is monitored continuously, and the
scale particle charge tests are best suited for water information provides advance warning of improper
with small annual variations in quality and are also coagulant dosages, which may lead to excessively
useful in research. They are not practical for plant high turbidity levels from the plant filter. Improper
coagulant dosages can be detected with pilot filters
within 10 to 15 minutes of coagulant application.

Floc breakthrough in the pilot filter can contribute to


, For a tho<ough discussion of SCDs, see Sleven K. Oertel and
Christine M. Kingery; November 1988; • An Evaluation of
emuent turbidities and interfere with accurate
Sirearnlng Currenl Detectors," American Water Works ASSOCiation; coagulant dosage determination. To prevent floc
Denver. Colorado.
22 Upgrading Existing or Designing New Drinking Water Treatment Facilities

breakthrough, excessive do~es of polymer filter aids use dry bagged chemicals that are mixed with water
are applied to pilot filter influent. to produce desired treatment-strength solutions.

Automated coagulant control systems can be used The point of application of any chemical will have a
with the pilot filter systems. Some units are equipped great impact on its effectiveness. Chemicals are
with plant flow pace meters to account for flow added at seven basic points in the water treatment
variation. Override controls can also be placed process:
throughout the system.
• To raw water before rapid mixing
• During rapid mixing
Turbidity Monitoring • Before flocculation
Turbidity is a measure of suspended particles. • Before sedimentation
Monitoring filtered water turbidity is important • Before filtration
because it reflects the effectiveness of the coagulation • After fil tration
process. In addition, turbidity monitoring is: • Prior to filter backwash

• Required of all surface water systems by the The optimal application point depends on site-specific
recently passed Surface Water Treatment Rule factors. The best situation is to have the capability to
• Related to probable level of potentially add chemicals at several points, so that the plant
pathogenic organisms in the finished water operator can perform tests to determine the ideal
• Reflective of the overall treatment effectiveness application point.
• A primary aesthetic indicator of the finished
water quality Most chemical feed systems commonly include
several basic elements such as mixing tanks, mixers,
The current availability of inexpensive and accurate and metering p.umps. In the chemical feed process,
turbidimeters makes monitoring the turbidity of operators manually add chemicals either wet or dry,
finished water very cost effective. Continuous in specific concentrations to the mixing tank. A
monitoring of each filter unit is especially useful for metering pump ensures accurate feeding of known
optimizing the quality of effiuent blended from concentrations of coagulants, polymers, coagulant
different filters. aids, or pH adjustment chemicals into the water
stream.
Particle counting is another method of measuring
turbidity that provides a profile description of the
particle sizes. Particles as small as 1 micron can be Uquld Alum Feed
measured by this method, but since many particles Specialized liquid alum feed systems become
are less than a micron in size, the method is of limited appropriate for plants larger than about 0.12 m3/sec
applicability. However, particle counting is very (2.5 MGD). Design elements for these facilities
useful for pilot plants and research applications, include:
especially for asbestiform fiber detection.
• Storage facilities for a 15-day supply
(recommended)
• Fiberglass reinforced or polyethylene tanks
3.1.3 Chemical Application Methods and • Indoor uncovered tanks for smaller systems
Considerations • Outdoor, covered, and heated tanks for larger
The following discussion covers the equipment and systems, because crystallization occurs at under
methods used to apply treatment chemicals, 10·C (50·F)
including operation and maintenance considerations.
The following four chemical feed systems are
reviewed: liquid alum, polymer, sodium hydroxide, Polymer Feed
and sulfuric acid. In plants of less than about 0.12 m3/sec (2.5 MGDJ,
polymer application can be performed using a basic
Chemical feed system options change as plants chemical feed system. In small plants, a polymer feed
increase in size. Plants with a capacity of 0.12 m3/sec may include two tanks of equal size, a mixer, and a
(2.5 MGDl or greater generally are able to justify metering pump. One tank is used for mixing and one
bulk purchases and storage of chemical stock and for storage of the polymer feed solution. The polymer
install the requisite storage and handling equipment. mixing tank is elevated above the storage tank to
Special storage facilities for chemicals are also permit gravity feed of the polymer solution.
usually needed for larger plants. For example, liquid Optionally, a single tank suitable for water
alum is available from 15,140-L (4,OOO-gall tank treatment operations can be interrupted to prepare
trucks for larger plants, while smaller plants often the polymer solutions. A preparation of 0.25 percent
Prefiltration Treatment Elements 23

stock polymer solution is generally used for all plant 3.2 Modifying or Adding Rapid
sizes.
CoagUlant Mixing
Plants greater than 0.12 m3/sec (2.5 MGD) generally Rapid mixing is generally the first stage of the
require two tanks of equal size and a mixer, but, in treatment process. This step is essential for quick and
addition, require a dry chemical storage hopper and a complete dispersion and mixing of the coagulant; the
dry chemical feeder. absence of complete mixing could result in excessive
effiuent turbidity. While some plants only have one
rapid mixer to add coagulants, other plants have
multiple rapid mixers to achieve optimum
Sodium Hydroxide Feed performance.
Plants less than about 0.12 m3/sec (2.5 MGDl are able
to use basic chemical feed systems to apply sodium Increasing the plant flow rates often necessitates
hydroxide solution prepared from up to 90.72 kg (200 improvements to existing rapid mixing facilities to
Ib)/day of dry sodium hydroxide. These plants ensure complete and thorough mixing of coagulants
require: with incoming water. Effective rapid mixing also can
reduce coagulant dosage requirements.

• Piping and valves to convey water for mixing Improvements to rapid mixing involve adjusting both
• 98.9 percent pure dry sodium hydroxide the means of mixing the coagulant with the influent
• A volumetric feeder for mixing and the coagulant's point of application. Refinements
• Two mixing tanks because of the slow mixing to the rapid mixing process usually result in
rate necessary to accommodate the high heat of significant improvements to the finished water. The
the process savings in coagulant usage justify the low capital cost
• Application of the sodium hydroxide in a 10 of the improvements.
percent solution
• A dual headed metered pump to deliver the Rapid mixer effectiveness is usually gauged by
sodium hydroxide solution to the point of contact time and the velocity gradient imparted to
application the water stream. Contact time refers to the
detention time of the water in the rapid mixer. The
velocity gradient is the difference in water speeds in
Plants greater than 0.12 m3/sec (2.5 MGD) are the rapid mixing chamber (where water with the
designed to use liquid sodium hydroxide, which highest speed is in proximity to the mixing device
requires equipment over and above the basic feed and the slowest water is near the chamber wall).
system. Plants using liquid sodium hydroxide feed a Velocity gradient is related to the amount of energy
premixed 50 percent solution containing 6.38 pounds imparted to the water during mixing.
of sodium hydroxide per gallon. To store a 15-day
supply of this stock solution, plants need fiberglass Complete coagulant dispersion is accomplished in
reinforced or steel tanks. The sodium hydroxide must very little time and with a wide range of velocity
be kept indoors or stored in heated tanks to avoid gradients. The reaction times for coagulation range
crystallization of the solution that occurs at or below from microseconds up to 30 seconds, and the velocity
12.2°C (54°Fl. gradients of various mixers range from 700 to 5,000
sec-I. Table 3-7 presents contact times and velocity
gradients for four types of rapid mixers.
Sulfuric Acid Feed The contact time required for successful coagulation
Sulfuric acid feed systems are designed to directly depends on which of the two primary coagulation
apply a 93 percent solution to the water through a processes - adsorption or sweep-floc (enmeshment) - is
metering system that includes standby metering necessary for the particular raw. Adsorption takes
pumps. place very quickly; sweep-floc coagulation takes
longer. In adsorption, if polymers are not formed
during the application process, then coagulation can
Most plants wiIl purchase acid in bulk, which must occur in microseconds. Ifpolymers are formed,
be stored in fiberglass reinforced tanks. Only very coagulation will still take less than a second. Sweep
small plants will use sulfuric acid delivered in drums floc, a process that is triggered by large doses of alum,
that will require indoor storage facilities. All plants ferric chloride, and other precipitative coagulants,
should have storage capacity for a 15-day supply. takes from 1 to 7 seconds to complete.

Capital and operating and maintenance costs for The types of devices used for rapid mixing include
alum, polymer, sodium hydroxide, and sulfuric acid mechanical mixers, static mixers, jet injection
feed systems are presented in Tables 3-3 through 3-6. mixers, and coagulant diffusers. Table 3-8 presents
24 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Table 3-3. Estimated Costs for Supplementing Surface Water Treatment by Adding Alum Feed facilltiea (to
"'9'1) ($t978)

Plant Capacity. Average Flow. Capital Cost. OperatJon and Maintenance Costs Total Cost.
Category MGD MGD $1.000 $1.0001yr ¢II ,000 gal ¢ll.ooo gal
I 0.026 0.013 12.2 2.3 47.8 87.1
2 0.068 0.45 15.9 2.6 16.0 26.2
3 0.166 0,13 20.3 5.4 11.5 16.5
4 0.50 0.40 28.0 7.7 5.3 7.5
5 2.50 1.30 40.7 4.1 0.9 1.9
6 5.85 3.25 47.1 7.9 0.7 1.1
7 11.59 6.75 58.1 14.9 0.6 0.9
8 22.88 11.50 70.4 24.4 0.6 0.8
9 39.68 20.00 81.6 41.3 0.6 0.7
10 109.90 55.50 128 112 0.55 0.63
11 404 205 250 402 0.54 0.58
12 1,275 650 575 1,245 0.52 0.55
Source: U.S. EPA (1979).

Tsble 3·4. Esllmated Costs for Supplementing Surface Water Treatment by Adding Polymer Feed Facilities
(0.3 mgtl) ($1978)

Plant Capacity, Average Flow, Capital Cost. Operation and Maintenance Costs Total Cost,
Category MGD MGD $1,000 $1.000Iyr· ¢/l,OOO gal ¢/l.000 gal
1 0.026 0.013 2.8 1.3 28.3 35.3
2 0.068 0.045 3.3 1.4 8.6 11.0
3 0.166 0.13 3.8 3.5 7.3 8.2
4 0.50 0.40 4.6 3.7 2.5 2.9
5 2.50 1.30 66.1 5.8 1.2 2.9
6 5.85 3.25 66.7 6.8 0.6 1.2
7 11.59 6.75 68.8 8.4 0.3 0.7
8 22.86 11.50 72.7 10.7 0.2 0.5
9 39.68 20.00 78.6 14.6 0.2 0.3
10 109.90 55.50 107.2 31.1 0.2 0,2
11 404 205 159.9 100 0.1 0.2
12 1,275 650 227.3 299 0.1 0.1
'Polymer cost lor Categones 1 through 4 is 75¢!1b, and lor Calegones 5 through 12 ,t IS $I,ooolton
Source: U.S. EPA (1979).

Table 3-5. Estimated Costs for Supplementing Surface Wster Treatment by Adding Sodium Hydro.lde
Facilitles ($1978)
Plant Capacity, Average Flow. Capital Cost. Operation and Maintenance Costs Total Cost.
Category MOD MOD $1,000 $I,OOO/yr' ¢II ,000 gal ¢II ,000 gal
1 0.026 0.013 2.4 1.2 24.3 30.2
2 0.068 0.045 2.7 1.4 8.8 10.7
3 0.166 0.13 3.3 3.3 7.6 8.4
4 0.50 0.40 5.4 6.2 4.2 4.7
5 2.50 1.30 33.3 8.8 1.8 2.7
6 5.85 3.25 36.9 18.4 1.6 1.9
7 11.59 6.75 43.0 35.8 1.4 1.6
8 22.86 11.50 58.3 59.4 1.4 1.6
9 39.68 20.00 76.4 101 1.4 1.5
10 109.90 55.50 159 275 1.4 1.4
11 404 205 353 1.015 1.4 1.4
12 1.275 650 697 3.210 1.4 1.4
• Costs Include storage and feed facIlities to add NaOH at a concenlTation of 10 mgtl. Dry sodium hydroxide IS used tor
Categones , through 4. while a liquid soIutJon is used lor bulk defivefy for Categories 5 1Ivough 12.
Source: U.S. EPA (1979).
Prefiltration Treatment Elements 25

Table 3-6. Estimated Costs lor Supplementing Surface Water Treatment by Adding Sulfuric Acid Feed
Facilities ($1978)

Plant Capacity. Average Flow. Capital Cost. Operation and Maintenance Costs Total Cost.
Calegory MGD MGD $1.000 $1.000Iyr· ¢/1.000 gal ¢/1.000 gal
1 0.026 0.013 2.8 0.9 20.0 27.0
2 0.068 0.045 3.2 1.3 7.6 9.9
3 0.166 0.13 40 3.2 6.7 76
4 0.50 0.40 5.7 5.6 3.9 4.3
5 2.50 1.30 21.6 2.4 0.5 1.0
6 5.85 3.25 25.6 3.6 0.3 0.6
7 11.59 6.75 29.7 5.8 0.2 0.4
8 22.86 11.50 379 8.8 0.2 0.3
9 39.68 20.00 49.9 14.1 0.2 0.3
10 109.90 55.50 86.9 36.3 0.2 0.2
11 404 205 210.3 128.3 0.2 0.2
12 1.275 650 431.0 404.7 0.2 0.2
• Costs Include storage (15 days). feed. and metering facilities for delivenng concentrated acid directly from Slorage to
application point. Categones 1 through 6 Include delivery in drums. stored Indoors. Categories 7 through 12 Include bulk
delivery and outdoor storage In FRP tanks. Application rate IS 2.5 mgiL.
Source: U.S. EPA (1979).

Table 3-7. Typical Rapid MIMer Contact Times and In-Line Static Mixers
Velocity Gradients
In-line static mixers consist of a series ofbaffie
TYPical Contact Typical VelOCity elements placed in a pipe section to impart
Time Range Gradients alternating changes in flow direction and intense
Mixer Type (In seconds) (In sec")
mixing action as water flows through the device.
Mechanical mixers 10-30 700-1.000 Headloss ranging from 0.3 to 1.8 m (1 to 6 ft) occurs as
In·line mechanIcal 0.5 3.000-5.000 the flow passes through the static mixer. Static
and StatiC blenders
mixers achieve virtually instantaneous mixing, are
In·hne Jet Injectors 0.5 1.000-2.000 relatively maintenance free, and are less expensive
Hydraulic mixers 2 800 than other rapid mixing processes. The only
Source: Williams and Culp (1986). disadvantage of in-line static mixers is that mixing
intensity depends upon the plant's water flow rate.
As rates decrease, the mixing intensity slows.

the estimated capital and operating costs associated In-Line Mechanical Blenders
with adding rapid mixing.
In-line mechanical blenders provide rapid mixing of
treatment chemicals with water flowing in a pipeline
under pressure. These devices consist of a housed
Mechanical Mixers propeller driven by an electrical motor. They have
The most common mixing devices are mechanical the advantage over static mixers of imparting
mixing tanks, also termed completely mixed or back- considerably less head loss and not being affected by
mixed units. They use turbines or impellers to mix change in flow. In addition, in-line blenders offer the
following advantages:
coagulant with water. The three advantages of
mechanical mixers are that they (1) are effective, (2)
impart little head loss (headloss being loss of pressure • Virtually instantaneous mixing with a minimum
of Short-circuiting
or the reduction of the water velocity within the
plant). and (3) are unaffected by the volume of water • Minimal head losses
and flow variations. • Less expense than more conventional rapid mix
units

A variable speed drive can alter mixer speeds to Jet Injection Blending
achieve different velocity gradients. Lower velocity Jet injection mixers, a type of in-line blender, are
gradients are used following the application of used for attaining nearly instantaneous dispersion of
polyelectrolyte coagulant aids. Labor is required for coagulant with raw water, usual1y at larger
daily jar testing operations and routine inspections, treatment plants. Jet injectors can avoid the
as well as annual inspections, cleaning, tank back mixing inefficiencies of turbine or impel1er
drainage, and oil changes. mixers, and the recommended detention times are
26 Upgrading Existing or Designing New Drinking Water Treatment Facilities

TatlM 3-8. Eltlm81ed ca." for Supplementl~ Surface Water Treatment by Adding Rapid Mix ($1978)
Planl Capacily. Average Flow. Capital Cost, Opera~on and Maintenance Costs Total Cost,
CatllgOly MGD MGO $1.000 $1.000tyr ¢/1.000 gal ¢/I,OOO gal
1 0,026 O,Ot3 13,2 2,8 58,6 91,3
2 0,068 0,045 17,5 2,9 17.6 30.1
3 0.166 0.133 22.5 70 14.7 20.3
4 0,50 0.40 30,9 7,9 5.4 79
5 2,50 1.30 47.7 13.3 2.8 4.0
6 5,85 3.25 63.7 22.4 1.9 2.5
7 11.59 6,75 88,2 38.2 1.6 2.0
8 22.86 11.50 139 69.2 1.6 2.0
9 39,68 20.00 218 116 1.6 1.9
10 109,90 55.50 587 313 1.5 1.9
11 404 205 2,100 1,130 1.5 1.8
12 1,275 650 6,670 3,540 1.5 1.8
Source: U.S. EPA (1979).

shorter than for mechanical mixers. The primary but gentle enough to prevent disintegration of
disadvantages are that the orifices in the injection existing flocculated particles.
pipe tend to become plugged and the mixing intensity
cannot be varied. Figure 3-2 illustrates a jet injection Effective flocculation is important for the successful
mixer. operation of the sedimentation process, Expanding
plant capacity will require flocculation improvement
Coagulant Diffusers to maintain plant performance. For example, a
doubling of plant capacity resulting from a change in
Coagulant diffusers are also used to improve the filter media from rapid sand to mixed- or dual-media
rapid mixing process. Figure 3-3 is a schematic can increase flow capacity from 1.34 to 2.7 Usedm 2 (2
diagram of a coagulant diffuser. Coagulant diffusers to 4 GPM/sq ftl. The increased flow will cut the
are similar to jet-injection mixers in design, except flocculation detention time in half, thereby
that jet injectors are usually used in a pipe setting preventing completion of the flocculation process,
and coagulant diffusers are typically used in basin and reducing sedimentation effectiveness.
settings. Coagulant diffusers are designed to apply
the coagulant at the point of maximum turbulence. Three basic methods of improving the flocculation
In some plants, multiple application points spaced process are:
uniformly across the flow cross section permit rapid
and thorough contact of chemical solution with the • Installing new mixing equipment to improve
entire incoming water flow. This rapid and uniform agitation
dispersion of coagulant prevents its hydrolysis, which • Improving inlet and outlet conditions to
is a common problem in systems with single minimize destabilizing turbulence
application points. Where turbulence in incoming • Adding baffiing to reduce basin short-circuiting
channels or pipelines provides good mixing, the
simple addition of a coagulant diffuser could improve
coagulation and lead to improved filtered water 3.3.1 Improving Mixing
quality. Different mixing equipment can improve turbulence
patterns in the basin to maximize the formation of
flocculated particles and minimize the destruction of
previously formed floc. There are several types of
3.3 Improving Flocculation mixers used for flocculation, with the mechanical
Flocculation usually follows the rapid mixing process mixer being the most common, Mechanical mixers for
in conventional treatment plants. Flocculation is a flocculation basins are differentiated by overall
time-dependent process that directly affects design and type of agitator
clarification efficiency by providing multiple
opportunities for particles suspended in water to The most important goal in flocculator design is the
collide through gentle and prolonged agitation. The efficient removal of floc during sedimentation and
process takes place in a basin equipped with a mixer filtration. Many flocculator units, especially low-
that provides agitation. This agitation must be energy mixers, tend to maximize floc size rather than
thorough enough to encourage interparticle contact, floc density, which affects the speed with which floc
Prefiltration Treatment Elements 27

Vertical Turb.ne Pump ~


r-- liquid Alum Feed POint

• I • t •• ..- •• I 1

Raw Water

Radial Flow DINuser

To Floccu·
latlon Basin

,-
-. .. 72 in Influent Pipe

Figure 3·2. In·llne jet Injection mix....

settles. High-energy mixers create smaller and 3.3.2 Improving Flocculator Inlet and Outlet
denser floc which settles faster and occupies less Conditions
volume in the filter bed than the larger floc created
by low-energy mixers. The desired detention time (i.e., residence time)
determines the size and occasionally the design
configuration of the flocculation basin. The basin's
Paddle, walking-beam, and flat blade turbine mixers internal design then determines detention time
are known as low-energy mechanical mixers. The effectiveness. The design of the basin inlet, outlet,
paddle or reel mixer operates with low tip speeds of2 and internal circulation patterns all affect floc
to 15 revolutions per minute (rpm) to prevent floc formation and destruction.
destruction. The flat blade turbine typically operates
at 10 to 15 rpm. Flat blade turbines can produce Typical detention times range from 15 to 45 minutes.
excessively high velocity gradients of over 45 rpm. These are influenced by the influent water
conditions, type of coagulant used, and requirements
of downstream processes. Cold, low turbidity water
Axial flow propellers or turbines, known as high- could require 30 minutes of detention, while the same
energy mechanical flocculators, typically operate at water undergoing direct filtration at higher
150 to 1,500 rpm with no limit on blade tip speed. temperatures could require only 15 minutes of
Axial flow propellers are favored in some situations detention. In small plants with high efficiency
because they produce uniform turbulence and are volumetric mixing, 10 to 15 minutes of detention
simple to install and maintain. might be acceptable.

[fwater passes through the flocculation basin in


Careful evaluation is needed to determine whether a much less than the volumetric residence time, the
plant should change to high-energy flocculators. influent stream is said to have "short-circuited." To
Sometimes, higher energy gradients can be achieved achieve effective flocculation and minimize short-
through modifying paddle-style units by adding more circuiting, designers must pay particular attention to
paddles or installing a new higher speed mechanical inlets, outlets, and internal baming. Inlet and outlet
drive. turbulence is the major source of destructi ve energy
28 Upgrading EXisting or Designing New Drinking Water Treatment Facilities

.............
.----- -
.... -..
....~...- ··.'

... ~
'
r- Coagulant
Feed LIne

~ Multijet
Slide
Gate
~ •• C~,,~
I~ 14 In

f f'\
Source
1i ~ Mo-f-'l==lI-H-- Water

.." . /~1t+~91-1~- Flow

.' ,-------
,
I I I Coagulant

1 Coagulant
Oiffuser '.
-...·::. . '.
....--===----::.,...J •.
-... ::

Front View SecllOn View

Figure 3-3. Coagulant dlffu.....

in the flocculation basin contributing to short- flocculation. Baffles, commonly made of wood,
circuiting. plastic, concrete. or steel are either "over-under" or
"around-the-end" in design. "Over-under" designs
Effective baffle, inlet, and outlet design reduces direct water flow either over or under the baffle.
short-circuit problems. Improvements that are often while "around-the-end" designs direct the water
overlooked can include: around either end of the baffle. Baffles should be
designed to direct the water flow such that the
• Adding inlet diffusers to improve the uniformity velocity gradients are less than those produced in the
of the distribution of incoming water flocculation process but greater than those in the flow
• Enlarging connecting conduits to reduce floc- moving laterally across the inlets. An example of a
disrupting turbulence basin divided with baffiing is illustrated in Figure
• Adding a secondary entry baffie across the inlet 3-4.
to the flocculation basin to impart increased
headloss for uniform water entry Estimated costs for adding flocculation are presented
in Table 3-9.
Improving inlet and outlet conditions will reduce
basin turbulence and thus lessen floc breakup,
increase detention times, and allow more efficient 3.4 Improving Sedimentation
coagulant usage. Sedimentation is the step in conventional water
treatment systems that follows flocculation and
precedes filtration. Its purpose is to enhance the
3.3.3 Improving Basin Circulation with filtration process by removing particulates.
Baffles Sedimentation requires that the water flow through
Flocculation basins can be baffied to direct the the basin at a slow enough velocity to permit the
movement of the water through the basin. Baffling, particles to settle to the bottom before the water exits
usually near the basin entrances and exits, can the basin. The equipment required for this process
improve basin circulation and achieve more uniform includes a settling basin of either rectangular.
Prefiltration Treatment Elements 29

Up
Wood On
Perlorated

.'
Wall ""
Entrance
Baffle B B B

CRaPld
M,x
Apply 0.25 mglL
Product B (985 NI ...
-j
457m


(15 hI 1.22m
(4 h)
-t 1 I Typical Baffle Would be the

TypIcal Wood Well


With Port ,n
---
J TypIcal Wood Well
WIth Port In
~~
'" :!.
- - 0 Same Elevation as the
Port Extending 122m
(4 ") Into the Compartment

Top of Wall Bouom

FllIurll 3·4. Divided Ilocculation balln.

Table 3-9. Eltlmated Caets lor Supplementlnll Surface Water Treatment by Addlnll Flocculation ($1978)

Plant Capacity, Average Flow, Capttal COS1, Opera\lon and Maintenance Costs Total Cost.
CaJegoIy MGD MGO $1,000 $1,0001yr ¢/l.oo0 gal ¢It,000 gal
I 0.026 0.013 10 1.0 21.7 45.2
2 0.068 0.045 18 1.1 69 20.1
3 0.166 0.133 34 2.3 4.9 13.3
4 0.50 0.40 73 2.7 1.8 7.7
5 2.50 1.30 217 3.8 0.8 6.2
6 5.85 3.25 325 5.6 0.5 3.7
7 11.59 6.75 418 8.7 0.4 2.3
8 22.86 11.50 587 14.5 0.4 2.0
9 39.68 20.00 840 22.9 0.3 17
10 109.90 5550 1,830 53.9 0.3 1.3
11 404 205 6.060 182 0.2 1.2
12 1,275 650 19,200 569 0.2 1.2
Source: U.S. EPA (1979).

square, or circular configuration. The basin includes Improvements in other elements of the treatment
provisions for inlet and outlet structures, and a process often require modification of the settling
sludge collection system. In addition, sedimentation basins to maintain effiuent water quality. Typical
systems are optionally equipped with tube or plate system improvements that necessitate sedimentation
settlers to improve performance. modification involve expanded flow rates, resulting
30 Upgrading Existing or Designing New Drinking Water Treatment Facilities

in decreased detention times and increased There are several dilTerent types of commonly used
clarification rates. tube settling basin designs. The two most common
choices are horizontal flow basins (either rectangular
or circular) and upflow solids contact clarifiers.
Problems with settling basins that might require Either design uses tube settling modules or plates to
modification include: apply the principles of shallow depth sedimentation,
permitting operation at higher clarification rates
• Poor entry and/or exit conditions than conventional clarifiers.
• Destructive basin turbulence
• Excessive clarification rate
• Inadequate sludge collection and removal The major by-product of sedimentation is a volume of
sludge, which must be removed on a continuous
Settling basin inlets are often responsible for basis. Sludge collection systems are either manual or
creating turbulence that results in breaking floc. One automated. Automated mechanical sludge collection
way to a void such destructi ve turbulence is to equipment is suitable for larger plants to attain
eliminate conduits and pipes between flocculation continuous and complete sludge removal.
and settling basins by conducting both flocculation
and sedimentation in one basin separated by
perforated bames. Baffies will minimize the breaking
of floc and evenly distribute the flocculated water to 3.4.1 Horizontal Flow Sedimentation Basins
the settling basin to maximize settling efficiency. As the name implies, the water generally follows a
horizontal flow path in this type of basin. Water is
Improperly designed outlets are often responsible for introduced at one end of the basin and suspended
sedimentation basin short-circuiting. One method of particles settle to the bottom as the water moves
reducing short-circuiting in rectangular basins is to through to the exit. The particular basin
use finger launders extending inward from the exit configuration will alTect the design requirements for
wall in a uniform pattern. ringer launders are small installing tube or plate settlers to improve
troughs with V-notch weir openings that collect performance or increase c.lpacity.
water uniformly over a large area of the basin. Weir
loading rates ranging from 126.2 to 252.3 m 3 /day/m The terms overflow rate, clarification rate, and
(10,000 to 20,000 gpdIftl of weir length are optimal. loading rate all refer to the amount of water that can
ror circular or radial flow basins, clarified water is be adequately processed by sedimentation (i.e., the
collected in a continuous peripheral weir trough. velocity of water through the tubes expressed in
These collection troughs should be designed with the GPM/sq ftl. The clarification rate at which tubes or
same loading rate as weir troughs used in plate settlers can be operated depends on the type of
rectangular basins. clarifier and water characteristics, including
temperature and the desired effiuent quality. For
Tube or plate settlers are often added to settling example, the tube clarification rate must be reduced
basins to improve their efficiency, especially if flows when influent turbidity increases or water
are to be increased beyond original design conditions. temperature decreases. The specific relationships
This established technology is widely used to reduce between tube loading rates and influent and emuent
floc carryover in existing basins. Tube settlers can be turbidity for cold and warm water are presented in
installed in most conventionally designed settling Table 3-10.
basins and, in many cases, result in twice the basin's
original settling capacity. Also, the use of tube and The preferred location for tube modules in horizontal
plate settlers in new plant construction minimizes flow basins is away from areas of possible turbulence
settling basin costs by minimizing the basin size such as basin entrances. For example, in a horizontal
necessary to attain a desir'!d 1""·,,,1 of treatment. Tube basin, often as much as one third of the basin length
settlers can also be used with vacuum sludge removal at the inlet end may be left uncovered by the tubes so
systems. that it may be used as a "stilling" zone, where water
turbulence is diminished in preparation for
Tube settlers typically use 0.6 to 0.9 m (2 to 3 ftl long sedimentation. This is permissible in most basins
tubes or plates placed from 5.1 to 10.2 cm (2 to 4 in) because the required number of tube settlers will
apart and installed over part or all of a basin. They cover only a portion of the basin. Tube modules are
are generally designed to accept flow rates ranging usually placed at least 0.6 m (2 ft) below the surface
from 0.68 to 2.04 Usedm 2 (l to 3 GPM/ft21. The in standard depth (3 to 3.9 m [10 to 13 ftJ) settling
shallow settling depths and the large surface area basins. They can be submerged as much as 1.2 m (4
provided by tube settlers permit elTecti ve ft) for deeper basins (4.8 to 6 m [16 to 20 ftll. In all
clarification at detention times of several minutes cases sufficient clearance must be maintained
versus several hours in conventional sedimentation beneath the tube settling module for movement of
basins. sludge collection equipment.
Prefiltration Treatment Elements 31

open area between the modules and the center well to


Table 3·10. Horizontal Flow Ballna LoadIng Rates
dissipate inlet turbulence.
Overllow Rate. Probable
Overflow Rate. Based Raw Portion of Basin Effluent 3.4.2 Upflow Solids Contact Clarifiers
on Total Clarifier Area Water Covered by Tubes Turbidity.
(GPMIft2) Turbidity (OPM/ft 2 ) NTU The design of upflow solids contact clarifiers is based
For water temper· upon maintenance of a layer or blanket of flocculated
atures 40°F or less: material through which water flows in a vertical
2.0 0-100 2.5 1-5 direction in the clarifier. The purpose ofthe layer,
2.0 0-100 3.0 3-7 known as a sludge blanket, is to entrap slowly
2.0 100·1.000 2.5 3-7
settling small particles and achieve a high level of
clarification. The sludge blanket is maintained at a
2.0 100-1,000 3.0 5-10
certain level and concentration by the controlled
For water tempera- removal of sludge. The precise height is determined
tures 50 ° F or above:
by the clarification rate. When the flow is increased,
2.0 0·100 2.5 1-3 the clarification rate is greater and the level of the
2.0 0·100 3.0 '·5 blanket rises.
2.0 0·100 40 3-7
In solids contact clarifiers, the clarification efficiency
3.0 0-100 3.5 '-5 of tube settlers depends upon both the clarification
2.0 100·1.000 2.5 '-5 rate and the concentration of floc in the sludge
2.0 100·1.000 3.0 3-7 blanket. The allowable loading rate or velocity of flow
through tube settlers is dependent upon;
1 GPM 30.06308 lisee: 1 n2 • 0.0929 m2;
°C • (OF - 32) x .556. • Average settling velocity of the floc layer or
sludge blanket
Figure 3-5 shows a typical tube settler installation in • Ability of the clarifier to concentrate solids
a rectangular basin illustrating the arrangement of • Ability of the sludge removal system to maintain
tube settling modules and the utilization of collection an equilibrium solids concentration; if the sludge
troughs in the covered area to ensure uniform flow is not removed quickly enough, then the solids
collection. In radial flow basins (see Figure 3·6), the layer will pass through the tubes and into the
required quantity of modules can be placed in a ring effiuent
around the basin periphery, leaving an inner-ring • Temperature of the water

! I
/ I
/ I
Influent Zone / Effluent
Collecnon Zone
/
/ I
/ I
I r y
~.
/ I \

V V
New Baffle ___ E.. s~ng Launders New Launders
- -
l
-' -' -' J
(FlOW Panern) -------------~--12----~r:--L----~--/ I '
I /
Tube Supports Tube Mcx1ules

Figure 3-5. Typical tube settler Installation In rectangular basin.


32 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Vanable Speed
RecirculatIOn Pump r Collector Onve

:be Sel1ling Modules - ;

/
~ " M
u rfijgp I I
/
.,~
'.
, 111J .. 11]tt\~-:-:3-
MIxIng and ,~
-.-!Z_-
-=:=
.'

.
.' Rer~!)~ _
----.-----
.,

.. suppon/ / \I; Flocculanon' and


Beam Secondary Reaction Zone ..
:.
- f-
'.

..
; I
- t-
I

.'

~ll Ir~ 4
'. .' ~
'. ..'
'\\1}/ '
.'

:
..... .
Sludge Collector
BI~wdown
" ~" .. " .,
'SIUdg'e
" Sludge Wasnng Sump

SectIOnal Elevanon

Figure 3·8. Radial solids contact clarifier with lube settl....

Recommended loading rates for upflow clarifiers are original design capacities is limited to 50 to 100
provided in Table 3-11. percent.

Tube module placement is governed by the same


Table 3·11. Upflow Clarifier loading Rat" for Cold Water
general rule as that for horizontal flow basins; that
Overflow Rate. Based Overflow Rate, Pon,on Probable EI- is, tube modules are best located away from influent
on Total Clanlier Area of Basin Covered by nuent Turbid- turbulence. In terms of placement depth and general
(GPMtn2) Tubes (GPM/ft2) l!y, NTU
proximity, the recommendations regarding distance
For water temperatures
40'F or less:
from the inlet and depth from the surface are also the
same. Lastly, collection launders should be placed at
1.5 2.0 1-3 3- to 4.5-m (l0- to 15-ft) intervals to ensure uniform
'.5 3.0 '-5 collection of flow over the area covered by the tube
1.5 4.0 3·7 settling modules.
~ ~ 1~

2.0 3.0 3-7 Estimates of the costs of adding tube modules to


conventional water systems are provided in Table 3-
For water temperatures
50' F or above: 12. These costs include modules, supports, anchor
braces, transition baffles, effiuent launders, and
2.0 2.0 1·3
installation.
2.0 3.0 1·5
2.0 4.0 3-7
2.5 25 3-7
2.5 30 5-10
1 GPM : 0.06308 Usee: 1 ft2 : 0.0929 m2:
'C : ('F. 32) x 556.

Expansion of the capacity of this type of clarifier is


limited by its ability to remove solids from the
system. Generally, the maximum expansion from
Prefiltration Treatment Elements 33

Table 3-12. Estimated Costs for Supplementing Surface


Water Treatment by Adding Tube Settling
Modules ($1978)

Plant
Cate- Capacity, Average Capital Cost Total Cost.
gory MGD Flow. MGD $1,000' ¢/l ,000 gal
1 0.026 0.013 1.1 27
2 0.068 0.045 2.3 1.6
3 0.166 0.133 3.9 0.9
4 0.50 0.40 8.2 0.7
5 2.50 1.30 25.2 0.6
6 5.85 3.25 46.8 0.5
7 11.59 6.75 92.0 0.4
8 22.86 11.50 163.0 0.4
9 39.68 20.00 250 0.4
10 109.90 55.50 684 0.4
11 404 205 2.500 0.4
12 1,275 650 7,850 0.4
• Costs Include modules, sUPPDns and anchor brackets, tranSillon
baffle. effluent launders, and Installation.
1 MGD = 0.044 m3 /sec; 1,000 gallon = 3.785 m3
Source: U.S. EPA (1979).
4. Filtration Technologies

This section describes several available filtration The sections describing filtration technologies
technologies ranging from commonly used provide information on each type of system covering
conventional systems to new and emerging the following aspects:
technologies. Filtration is one of the most important
elements in traditional water treatment systems and
also plays an important role in controlling some • Process description
organic contaminants. • System design'
• Operation and maintenance
The filtration process itself may need upgrading due • System performance
to expansions or the need to meet stricter effiuent • System costs
quality limits. Section 4.1 addresses typical concerns
and describes beneficial modifications. Many water The last subsection, Section 4.8, provides some
supply systems will have to add filtration to comply guidance in selecting the appropriate filtration
with the 1986 Safe Drinking Water Act (SDWA) technology for particular applications. Appendix A
Amendments and the resultant Surface Water includes descriptions of experiences with the various
Treatment Rule. Sections 4.2 through 4.6 provide filtration technologies.
overviews of the following filtration technologies:

• Conventional treatment and direct filtration


• Direct filtration (gravity and pressure filters) 4.1 Modifying Filtration Systems
• Slow sand filtration Filtration is usually the final step in conventional
• Package plants treatment trains, although disinfection frequently
• Diatomaceous earth filtration follows filtration. This section addresses
• Membrane filters improvements that can be made to plants that use
• Cartridge fil ters rapid sand, mixed-media, or dual-media filters.

Conventional treatment and direct filtration (Section


4.2) are the most widely used systems. Slow sand Filtration systems are regarded as effective for
filters (Section 4.3), package plants (Section 4.4), and removal of turbidity and microbial contaminants.
diatomaceous earth filtration (Section 4.5) are Microbial contaminants of special concern include
considered newer technologies, but have broad coliform bacteria, Giardia lamblia, enterovirus, and
applicability. While slow sand filters have been used Legionella.
for many decades, they do not have an established
record of performance with a large number of water
systems in this country. Membrane and cartridge The Surface Water Treatment Rule (SWTR) requires
filtration systems (Section 4.6) are considered that filters achieve turbidities of less than 0.5 NTU
emerging technologies because they show promise in 95 percent of the finished water samples, and
but have not generally been used for treatment of never exceed 5 NTU. Turbidity is a measure of
drinking water. Package plants, slow sand, diatoma- suspended particles, which can include organic
ceous earth, membrane, and cartridge filters are solids, viruses, bacteria, and other substances.
considered best suited for small systems (less than Turbidity particles range in size from less than I
0.44 m3/sec (10 MCDIl. micron to 100 microns.

34
Filtration Technologies 35

originally designed to operate at 1.36 Llseclm 2 (2


4.1.1 General E"ectiveness of Filtration GPMlsq ftl, is usually achieved by either changing
Systems the filter media to dual- or mixed-media or replacing
The effectiveness of filtration systems is determined the top layer of sand with anthracite coal (also
by their ability to remove microorganisms, turbidity, termed "capping" the filter). With these
and color. Color is imparted to water supplies by modifications, filter rates often can be increased to
organic material and can be removed by chemical 2.7 to 3.4 Useclm 2 (4 to 5 GPM/sq ftl, doubling plant
coagulation. Once the color is coagulated and capacity and producing higher quality effiuent.
combined with the floc particles, color removal can be
related to turbidity removal. The measurement and
control of microbes and turbidity are distinct A thorough hydraulic study is required to determine
procedures. The effecti veness of fil ter systems in the feasibility of improving a filtration system. For
removing microbial contaminants, however, is example, if a plant is going to be expanded,
heavily related to influent turbidity. The general sedimentation basins and interprocess transfer
rule of thumb is to minimize turbidity to maximize systems have to be evaluated in terms of their ability
microbial removal. to accommodate the increased flow from larger
filters. Assessing the maximum carrying capacities
The filtration systems discussed in this document are for all components of the system is a prerequisite to
appropriate for raw water with varying evaluating the expansion potential of a treatment
characteristics. Table 4-1 contains the recommended plant. Field testing of the entire system is the
upper limits for several influent variables, including preferred method for conducting hydraulic
coliform bacteria, turbidity, and color, for four evaluations.
filtration technologies. Conventional treatment is
clearly the most accommodating of the filtration
systems in the table because it includes flocculation [f a hydraulic evaluation reveals that it is feasible to
and sedimentation, which reduce turbidity before the increase plant capacity, a specific evaluation offilter
water is actually processed by the filter. modification is appropriate. Such an evaluation will
Diatomaceous earth filtration systems, which include include consideration of filter box design, underdrain
little pretreatment, require high-quality influent. type, surface wash system, air scour system, flow-
limiting devices and piping, filtration rate controls,
Giardia LambLia is of particular concern in drinking filter aid application, and backwash additives. Figure
water supplies because it forms a cyst that cannot be 4-2 illustrates the basic components of a sand filter
effectively killed by traditional chlorine disinfection. system that need to be evaluated.
Effective removal can be attained only by filtration.
Conversely, viral and bacterial pathogens are
relatively easy to destroy with disinfection.
Filter Boxes
Figure 4-1 is a graph of the relationship between Filter boxes, the structures that contain the filter
filter effiuent turbidity and cyst removal efficiency media, should be examined first to see if they can
after filtration. The graph clearly shows that filtered accommodate a filter that will operate at higher
water with very low levels of turbidity, ranging flows. Most rapid sand filters are easily converted to
below 0.1 NTU, contained almost no cysts (OeWalle higher volume dual- or mixed-media filters.
et aI., 1984). The four filtration systems shown in Backwash rates, the primary operational
Table 4-2 are very effective in removing Giardia; the consideration, are about the same for these two filter
only exception is direct filtration without types.
coagulation, which does not effectively reduce
turbidity. If water supplies containing Giardia are
not effectively coagulated, the cysts will pass through When flow capacities are increased using either dual-
the entire treatment process, including the filters. or mixed-media filters, wash troughs and certain
filter cleaning technologies will require specific
The viral removal efficiencies offour filter consideration. Filter wash troughs, structures at the
technologies found to be very effecti ve are shown in top of filter boxes that collect waste backwash water
Table 4-3. Of these processes, direct filtration during filter cleaning, may need to be deepened to
provides the poorest removal of viruses, ranging from prevent loss of the lighter anthracite coal media.
90 to 99 percent. Wash troughs that are shallow and wide and spaced
closely together can create overly high rise velocities
that wash out anthracite coal during backwash. If
4,1.2 Filtration System Improvements filter troughs cover more than half the area of the
Improving filtration systems can increase plant filter box, narrower and deeper troughs covering less
capacity and improve effiuent quality. Increasing the surface area may be needed to prevent excessive coal
capacity and effectiveness of rapid sand filters, loss during filter back washing.
36 Upgrading EXisting or Designing New Drinking Water Treatment Facilities

Tallie 4·1. Generalized Capallility 01 Fll1ratlon Systems To Accommodate Raw Water Quality Conditions
Recommended Upper Limits For Influent
Treatment Technology Total Coliforms. 1/100 mL Turbidity. NTU Color. CU
Conven~onaJ treatment < 20.000 No Restncuons 75
- With no preclisinfectlon < 5.000 No Reslnc~ons 75
Direct Fillrat,on <500 <7·14 <40
Slow Sand Filtra~on <800 <10 <5
Diatomaceous Earm Fillratlon <50 <5 <5

Tallie 4-3. Removal Elliciencies 01 Viruses By Water


1.0r-------------------, Treatment Processes

Percent
Unit Process RemovaJ

0.5 Slow Sand Fillrauon 99.9999


Diatomaceous Earth Filtration (with hlter aid) >99.95'
Direct FiltraMn 90-99
ConvenbQnaJ Treatment 99
::>
>- ., No viruses recovered.
Z

2>
~
:; 0.1
>- Underdralns
E
" ~
The condition of the underdrain gravel, the coarse
E layer underneath the filter media, is difficult to
w
0.05 assess in a filter evaluation. Removing filter media
and visually inspecting the condition of the
underdrain is the only reliable mathod of
ascertaining whether it is satisfactory. However,
indirect indications of underdrain failure include
boils incurred during the backwash process or an
uneven mounding of the filter surface. As a general
practice, underdrain gravel is removed and replaced
with new material along with the filter media.
0.01 '--.l-...........--'_'---'--...................._'--.l-.............- - l Frequently. the existing backwash water
40 50 60 70 80 90 100
distribution system, which commonly consists of pipe
Cyst RemovaJ - "to laterals, is also replaced at the same time.
Figure 4-1. Relationship between cyst removal and IIltered
water tUrbidity.
Surface Wash Systems
Surface wash systems are used to scour the upper
layers of the filter during or prior to backwashing.
Tallie 4-2. Removal Eltlclencles 01 Giardia Lamblia By These systems are more important for dual- and
Water Treatment Proc81lses
mixed-media filters than for sand filters. Surface
Percenl washers prevent the formation of "mud-balls," which
Unrt Process Removal form within the filter media, most frequently at the
RaplG FIllration With Coagulation. sed,menianon 96.9·99.9 sand-coal interface. Also, surface washers improve
D,rect FiltralJon
the process of cleaning dual- and mixed-media filters
. With coagula~on 95.9·99.9 when polyelectrolytes are added, a practice used at
Without coagula~on 10-70 filter flow rates of2.7 to 4.1 Usedm 2 (4 to 6 GPMlsq
. With flocculation 95-99 ft) to control floc breakthrough.
Diatomaceous Earth Flllra~on 99-99.99'
Slow Sand Flltra~on 98-100· Surface wash systems may be designed with one or
two rotary arms. The single-rotary-arm systems
• Aided by coagula~on. position the arm at the filter surface. The two·arm
• Fully-ripened Iiller. system has one arm on the surface and one arm
within the filter bed at the sand and coal interface in
Filtration Technologies 37

Filter Tank

Cast·lron Manifold

Figure 4-2. Cutaway view 01 typical rapid sand IIIter.

a dual media filter. The arms deliver a high velocity current system. Backwash water supply and waste
stream of water that violently agitates the filter piping in rapid sand filter plants, however, generally
media as it turns. have adequate capacity for expansions. Design
changes may be necessary to accommodate other
modifications that diminish the already limited
Air Scour Systems available space in filter pipe galleries. For instance,
Air scour systems can be used to clean rapid sand or cast iron piping and gate valves can be replaced by
dual-media filters, eliminating the need for surface more compact fabricated steel piping and less bulky
wash systems. They require specially designed wash butterfly val ves.
water distribution systems and generally do not use a
coarse grade gravel support layer, because the violent
mixing action caused by introducing air beneath a
filter would upset a graded gravel underdrain system Method of Filter Rate Control
and destroy the filter bed. Instead, specially designed There are two methods for controlling filter rates:
nozzles with retaining screens placed on a false floor constant rate filtration and variable declining rate
in the filter box under the filter media are commonly filtration. Constant rate filtration is the most
used. The air introduced beneath either dual- or common method. While constant rate filters have
mixed-media filters destratifies the media. To return provisions to control both the influent and effiuent
a destratified filter to service involves backwashing flow of each filter, variable declining control systems
at 10.2 LJsedm 2 (15 GPMlsq ftl to reclassify the filter do not use filter rate-of-flow controllers and have a
media. common filter effiuent collection flume or pipe. In
variable declining rate filtration, filtration rate
declines as headloss builds up during the filtering
Piping and Controls process. In plants designed with multiple filters,
Filter effiuent piping and filter flow rate controllers filter influent flow frequently is directed
may be too small to accommodate expansion of a automatically to another cleaner filter.
38 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Variable declining rate filtration can be used where • Provide for the housing of filter controls and pipe
the capacity of an existing rapid sand filter is to be galleries.
increased. Its major advantages over traditional • Use color-coded filter piping.
constant rate filtration methods are its relatively low • Provide for the easy removal of pumps and valves
cost and its elimination of rate-of-flow controllers for maintenance.
which require maintenance. Also, the filters are • Equip each tilter with a filter-to-waste cycle for
never operated at negative heads with this type of unsatisfactory water.
control, thus eliminating the potential for air • Provide for filter cleansing with either surface·
binding. The major limitation is that the filtration wash or air scouring.
rate ofindividulll filters is uncontrolled at the • Equip with an automatic filter backwash system.
beginning of a filter cycle. • Select backwash rate based on filter media and
expected wash water temperatures.
• Select backwash supply storage capacity to
Filtration Aid Application accommodate a minimum of two filter
Filtration aids prevent premature turbidity backwashes.
breakthrough by controlling floc penetration into the • Provide backup capacity to the largest single
filter. Chemically, they are usually non ionic pump for backwashing and surface washing.
polyelectrolytes, and are very useful in maximizing • Equip pressure filters with pressure and vacuum
the performance of high-rate filters and systems with release valves.
cold influent water. These aids should be applied in • Equip backwash supply lines with air release
systems that upgrade from sand to dual- or mixed- valves.
media filters. An optimal dosage of filtration aid
minimizes the effiuent turbidity until the filter Appendix A presents brief case histories of plant
achieves its maximum headloss. Typical filtration upgrades, discussing some of those steps in more
aid dosages are from 0.02 to 0.1 mg/L. detail.

Addition of Polymer to Backwash 4.2 Direct Filtration


The addition of polymers to the backwash water can Direct filtration systems are similar to conventional
reduce the initial turbidity peaks during filter systems, but omit sedimentation. This section
ripening following backwash and extend filter describes direct filtration technology.
operation before breakthrough occurs. Polymer
enhances the ability of the filter to retain turbidity
particles. Adding polymer also improves the settling 4.2.1 Process Description
and thickening rate of the backwash solids. In Direct filtration is an established technology that
addition, savings in backwash water can be achieved was developed because dual- and mixed-media filters
by precoating filters with polymer. are able to effectively process higher influent
turbidities without the use of sedimentation. The
Polymer should be added during the tirst 5 to 7 direct filtration process is expected to be more widely
minutes of the 10- to IS-minute backwash period. used on water supplies that, up until now, only
Polymer additions ranging from 0.1 to 0.15 mg/L performed chlorination.
have reduced the initial turbidity breakthrough by
one half. Direct filtration is only applicable for systems with
high-quality and seasonally consistent influent
supplies. The influent generally should have
4.1.3 System Design Checklist turbidity ofless than 5 to 10 NTU and color ofless
The following steps are recommended to maximize than 20 to 30 units.
the efficiency of any tilter modification:
Direct filtration consists of several combinations of
• Use pilot test data to select tilter size and treatment processes. It always includes coagulation
medium; however, if pilot test data are not and filtration, and sometimes includes flocculation or
available, use data from analogous applications. a contact basin after coagulation addition. Typical
• Provide for the addition of disinfectant directly to coagulant dosages range from less than I to 30 mg/L.
tilter influent. Cationic polymers often successfully coagulate water
• Provide for the addition of polyelectro lytes supplies and assist direct filtration. Nonionic
directly to filter influent for dual- or mixed-media polymers sometimes are added to the filtration step to
filters. increase filter efficiency. A flow diagram of a typical
• Provide for the continuous monitoring and direct filtration system is shown in Figure 4-3.
recording of turbidity from each filter.
• Provide for the continuous monitoring and In-line direct filtration, the simplest version of direct
recording of flow and head loss from each fiI ter. filtration, is commonly used in pressure filtration
Filtration Technologies 39

Influent

Figure 4-3. Flow diagram of typical direct IIltratlon Iystem.

systems (see Figure 4-4). In this process, chemical


coagulant application and mixing are followed by 4.3 Slow Sand Filtration
addition ofa nonionic polymer aid and filtration. Slow sand filtration systems have a long history,
There is no separate flocculation step. This treatment having been used without disinfection at least since
is effective and commonly used in the removal of iron the 1850s in London, England. Slow sand filters were
and manganese from ground water when also commonly used in the United States in smaller
accompanied by appropriate coagulant doses. In-line facilities. Rapid sand filters have replaced many of
direct filtration, however, is only applicable for them to accommodate the increase in water demand;
surface waters with seasonal average turbidities of consequently, slow sand filters are now not very
less than 2 to 3 NTU and free of contamination by common. According to the Surface Water Treatment
wastewater sources (when permitted by regulatory Rule under the SDWA, many small systems could
agencies). meet their regulatory filtration obligations with the
simple slow sand filters. With the requirement for
the multiple barrier disinfection approach in water
treatment, slow sand filters are almost always
accompanied by disinfection. Because slow sand
Influent
filters with disinfection have not been used
extensively, they are classified as "new" technology
in the current literature.
Figure 4-4. Flow diagram of In-line direct flltratlon system.
Slow sand filters are similar to single media rapid
rate filters (conventional systems) except that they:
Another version of a direct filtration system is the
"modified" system, which substitutes a contact basin • Are 50 to 100 times slower than normal filtration
for the flocculation basin found in the typical direct rates
filtration process. Figure 4-5 is a flow diagram of • Use biological processes, which may enhance
modified direct filtration. The I-hour detention chemical/physical removal processes
contact basin serves primarily to condition the floc • Require a ripening period before operation
prior to filtration. Contact basins also provide • Use smaller pores between sand (i.e., smaller
pretreatment by decreasing turbidity peaks in the sand particles)
influent, providing silt and sand removal, and • Do not require back washing
allowing for more equal distribution ofcoagulant. • Have longer run times between cleanings

Slow sand filters are most attractive for smaller


4.2.2 System Performance systems with high quality raw water. Specifically,
Direct filtration units can perform effectively within water which comes from a protected surface water
the following general influent parameters: supply, has previously received only chlorination as a
treatment, contains less than 10 NTU, and has no
• Less than 500 total coliforms per 100 mL color problems is the most suitable for slow sand
• Less than 14 NTU of turbidity filtration. While their operational simplicity makes
• Less than 40 color units them very suitable for small plants, slow sand filters
are also applicable for medium to large plants.
While the direct filtration process is able to operate
satisfactorily with influent turbidities as high as 14 The advantages of slow sand filtration include its
:-ITU, optimally, influent turbidity should be less simplicity and reliability, low cost, and ability to
than 5 NTU. Effective direct filtration performance achieve greater than 99.9 percent Giardia cyst
ranges from 90 to 99 percent and from 10 to 99.99 removal. In addition, these systems do not require
percent for virus and Giardia removal, respectively. continuous turbidity monitoring since they are
The wide variation in direct filtration's Giardia applied to water supplies with relatively low
removal efficiencies is due to the wide range of turbidity.
available system configurations. The most effective
direct filtration configurations for Giardia removal Slow sand filters have several limitations, however.
must include coagulation. Due to a low filtration rate, these filters require
40 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Inftuent

Nonionic Polymer
Filtration Aid
Figure 4-5. Flow diagram ot modlfl8Cl direct lIIttlltlon system.

relatively extensive land area. Without pretreatment a cleaning cycle to become a functioning biological
(particularly coagulation), limitations exist on the filter. Although Giardia removal is not usually
quality of raw water that is suitable for treatment affected by the ripening period, the overall poor water
using slow sand filtration. Also because pretreatment quality during this period requires provision of a
is minimal or nonexistent at slow sand filter plants, filter·to-waste cycle.
other contaminants such as synthetic organic
chemicals, disinfection by·product precursors, or These filters require continuous operation under
inorganic chemicals are not readily removed. submerged conditions, ranging in depth from 0.9 to
Systems with raw water containing high color or 1.5 m (3 to 5 ftl. Hydraulic filter outlet controls keep
algae content are probably not appropriate for slow the filters submerged at all times. The difference
sand filtration, because these contaminants are not between the level of water in the collection gallery
removed by slow sand filtration and the raw water and the level of water above the filter media is called
will likely contain precursors for chlorination by· the headloss through the media. The initial headloss
products. is about 0.06 m (0.2 ftl. Maximum headloss should be
less than the submerged depth of 1.2 to 1.5 m (4 to 5
ft) to avoid air binding and the uneven flow of water
4.3.1 System Design Considerations through the filter medium. The buildup of the
Slow sand filters require influent that does not maximum headloss is slow, taking up to 6 months.
exceed the following parameters:
Redundant, or stand·by, systems are required to
e Turbidity of < 10 or up to 20 NTU depending on accommodate the extended cleaning periods
other operating characteristics associated with slow sand systems.
e Color of less than 30 units as prescribed by the
Ten State Standards In cl imates subject to below freezing temperatures,
e Algae of less than 5 mg per cubic meter of slow sand filters usually must be housed. Unhoused
chlorophyll A filters in harsh climates develop an ice layer that
prevents cleaning during the winter months.
These are maximum limits, not typical operational Uncovered slow sand filters will operate effectively,
parameters. Design flow rates range from 0.94 to 9.35 however, if influent turbidities are low enough to
m3 /m 2 (l to 10 million gallons per acre)per day, with permit the filter to operate through the winter
the usual range from 2.8 to 5.6 m3/m 2 (3 to 6 million months without cleaning. An illustration of an
gallons per acre)per day. Slow sand filters require unhoused system is provided in Figure 4-6; Figure 4-
sand with particle sizes ranging in effective 7 shows an example of a housed system.
diameters from 0.25 to 0.35 mm, with a uniformity
coefficient of 2 to 3. A higher uniformity coefficient is Because of filter housing costs, slow sand filters are
acceptable for filters with pilot test verifications of most appropriate for small systems. Due to this
the performance. expense, large systems can usually only utilize slow
sand filters when they are located in moderate
Typical underdrains for slow sand filters are made climates and therefore do not require housing. The
from split tile with laterals placed in coarse stone. slow sand filter in Salem, Oregon, is an example of a
These drains routinely discharge to a tile or concrete large system that is able to use unhoused slow sand
main drain. Recently constructed slow sand systems filters because of climate.
are equipped with perforated polyvinyl chloride pipe
as laterals. Manually adjusted weirs, outlets, and
valves are adequate for these systems. Inlet 4.3.2 Operation and Maintenance
structures may be located at the end or side of the The primary operational consideration for a slow
filter. sand filtration system is maintaining a clean filter.
Cleaning becomes necessary when headloss reaches
Slow sand filters perform poorly during the first 1 to about 1.2 m (4 ftl. The normal length oftime between
2 days of operation, called the "ripening period." The cleanings ranges from 20 to 90 days, but will vary
ripening period is the time required by the filter after depending on raw water quality, sand size, and
Filtration Technologies 41

t:rtluent Flow
Slow Sand Filter Control Structure Clearwell

Raw
Water a::::::I><t:;:f==~~-~:;;:;:;-;~;;-7
Supernatant Water

Supernatant
Water Dram

Figure 4-6. Typical unhoused slow sand Ill1e, Inatallation.

Effluent Flow
Slow Sand Filter Clearwell
Control Structure

Head Space

Raw Water DC><Ir:::~

Supernatant Water

AdJustable
Welf l1_-
Supernatant ..
'-,'
Waler Dram

Filtered
Water for
Backfilling
Sand Fliter Bed
SupPOll Gravel Flow Meier
Dram Tile - - - - - Control Valve

Pump to
Filters for
Back Filling

Figure 4-7. Typical housed slow sand IiIte, inatallatl,,".

filtration rate. Filters should be cleaned every 1 to 2 small systems use manual cleaning techniques, but
months. very large systems may use mechanical scrapers.

A number of precautions must be observed in regard


to cleaning operations. Cleaning filters necessitates
Cleaning involves manually scraping ofT the top 2.5 removing the filter from service, after which a
to 5 cm (l to 2 in) of the filter media. The removed ripening period is required to bring the filter back
sand is either discarded or treated separately. Most into operation. Operators must minimize
42 Upgrading Existing or Designing New Drinking Water Treatment Facilities

intermittent operation by coordinating maintenance The process energy requirements of slow sand filters
tasks. While cleaning, operators should take care to are negligible assuming they have gravity-fed source
avoid disturbing the filter bed during back washing, water. The housing structure energy requirements
shock chlorination (if practiced), and raking the include heating, lighting, and ventilation, and are
surface of the filter bed. Scouring and subsequent directly related to the size of the structure.
erosion of the filter media from refilling after Maintenance supply requirements are also negligible
cleaning must also be avoided. Gradual filling from since the filters are assumed to operate virtually
the bottom of the filter during the startup phase unattended. Replacement sand is not included
partially addresses this problem. Excessive drying of because it is needed so infrequently.
the filter beds also should be avoided during
cleanings. During the postcleaning ripening period,
filtered water should be directed to waste. Operators 4.4 Package Plant Filtration
should periodically remove algae from the filter and, Package plants are categorized as "new" treatment
to accommodate seasonal changes, reduce filtration technology. They are not altogether different
rates during the winter months. processes from other treatments mentioned in this
section, although several models contain treatment
If sand depth drops below 61 em (24 in), new sand elements that are innovative, such as adsorptive
needs to be added. Bed depths of only 30.5 to 50.8 cm clarifiers. The primary distinction, however, between
(12 to 20 in) have had poor performance records. package plants and custom-designed plants is that
Replacing sand, however, is not considered a normal package plants are built in a factory, skid mounted,
operational or maintenance task because with and transported virtually assembled to the operation
careful cleaning, resanding may be necessary only site.
once every 10 years.
Package plants are designed to effectively remove
[n addition to the above maintenance considerations, turbidity and bacteria from surface water with
slow sand filters require daily inspection, control generally consistent low-to-moderate turbidity
valve adjustment, and turbidity monitoring. The levels. While package plants can treat influent
filter housing structure, if present, also requires streams with highly variable characteristics, they
routine maintenance. would require more skilled operators and closer
attention.
4.3.3 System Performance There are about 650 to 700 package plants operating
Slow sand filtration units perform best within the in the United States with capacities ranging from
following general influent parameters: 27.3 m3/day to 0.18 m 3 /sec (7,200 GPO to 4 MGO).
Many are built to conventional design standards.
• Less than 800 total coliforms per 100 mL Others, using tube settlers, have reduced size and
• Less than 10 NTU of turbidity larger capacities. More recently, package plants have
• Less than 5 color units been built with efficient and compact adsorptive
clarifiers. Package plants with adsorptive clarifiers
Effective slow sand filtration performance, with have capacities ranging from 54.5 m 3/day to 0.26
respect to virus and Giardia removal, ranges from 91 m 3 /sec (14,400 GPO to 6 MGO).
to 99.9999 percent and 100 percent, respectively.
The four major advantages of package plants are
their compact size, cost effectiveness, relative ease of
operation, and design for unattended operation.
4.3.4 System Costs Typically, these types of filtration plants are used to
Slow sand filter cost data are presented in Table 4-4. treat small community water supplies and for a
Smaller plants are typically designed with cast-in- variety of special applications, including:
place concrete structures with wood or concrete slab
covers. The piping is either cast iron or steel and • Emergency supplies
below grade. Flow meters are used to monitor the • Recreational areas
output for each filter. A below-grade concrete clear- • State parks
well is another cost element included in Table 4-4. • Construction sites
• Ski areas
Larger plants, on the other hand, use uncovered • Military installations
earthen berm structures. These on-grade plastic • Other areas not served by municipal supplies
membrane-lined structures have all below-grade
piping, and some plants use above-grade steel tanks. Package plants can differ widely with regard to
F low meters are a Iso used for eac h fi 1te r in the design criteria, and operating and maintenance
emuent control structure. considerations.
Filtration Technologies 43

Table 4-4. Estimated Costs For Supplementing Surface Water Treatment By Slow Sand Filtration ($1978)
Operallon and Maintenance Costs
Plant Capacity, Average Flow. Capital Cost. Total Cost,
Category MGO MGO $1,000 $1,OOOlyr' ¢/l,OOO gal ¢/l ,000 gal

0.026 0.013 145 0,8 17,9 377,8


2 0.068 0.045 273 1.6 9.9 205,1
3 0.166 0.133 508 5,1 10,5 133.4
4 0.5 0.40 603 9.0 6.2 54.7
5 2.5 1.30 1,213 20.5 4.3 34.3
6 5.85 3.25 2.573 38.0 3,2 28.7
7 11.59 6.75 4.782 62,3 2,5 25.3

• Processes include slow sand filters and clearwell storage. Sand filters in Categories 1 through 3 are constructed of concrete and are
covered. Sand filters in Categories 4 through 7 are constructed of membrane-lined earthem berms and are uncovered.
1 MGO 8 0.044 mJ/S: 1.000 gal • 3,78 mJ,
Source: U.S. EPA (1979),

4.4.1 Selecting a Package Plant System Conventional Package Plants


Package plant systems are most appropriate for plant Conventional package plants are manufactured by
sizes ranging from 94.6 to 22,710 m3/day (25,000 to several firms to a variety of specifications, As their
6,000,000 GPO). Common treatment elements in a name indicates, they contain the conventional
package plant are chemical coagulation, flocculation, processes of coagulation, flocculation, sedimentation,
settling, and filtration. and filtration, Typical design standards for these
units are:
Some package plant manufacturers have used new
treatment technologies to improve their product's • 20- to 30·minute flocculation detention time
performance, for example, tube settlers, adsorptive • 2-hour sedimentation detention time
clarifiers, and high flow rate mixed- or dual-media • Rapid sand filters rated at 1.34 Useclm 2 (2
filters. Package plants employing innovative GPM/ft 2)
technologies sometimes encounter regulatory
barriers, because many existing State design
standards recognize only conventional treatment Tube- Type Clarifier Package Plants
processes. A flow diagram of a tube-type clarifier package plant
is illustrated in Figure 4-8. This type of plant has two
Influent water quality is the most important versions with different capacity ranges; one version
consideration in determining the suitability of a can treat from 0,63 to 6.3 Usec (l0 to 100 GPM) and
package plant application. Complete influent water the other, equipped with dual units, can treat from
quality records should be examined to establish 12.6 to 88.3 Usec(200 to 1,400 GPM).
turbidity levels, seasonal temperature fluctuations,
and color level expectations. Both high turbidity and In these package systems, the disinfectant, primary
color may require coagulant dosages beyond many coagulant, and polyelectrolyte coagulant aid are
package plant design specifications, [n cases of added before the influent enters the flash mixer.
consistently high levels of turbidity and color, the After the flash mixer, the water enters the
package plant capacity must be down-rated or a flocculation chamber where mechanical mixers
larger model selected. Where turbidity exceeds 100 to gently agitate the water for 10 to 20 minutes
200 NTU, presedimentation may be required as a depending on the flow,
pretreatment. Pilot tests may be necessary to select a
package plant for more innovative designs using high The flocculated water then enters the tube settlers,
flow rates and shorter detention time unit processes, which consist of many 2.5-cm (l·in) deep, 99-cm (39-
in) long split-hexagonal-shaped passageways, The
large surface area of the many 2.5-cm (l-in) deep tube
settlers achieves an effective clarification overflow
4.4.2 System Description and Design rate ofless than 6.1 m 3/day/m 2 (l50 GPD/ft 2 ),
Considerations Adequate clarification is attained with less than 15-
The three basic types of package plant filter systems minute detention times.
are conventional package plants, tube-type c1arifi·
cation package plants, and adsorption clarifier The clarified water then enters a gravity flow mixed-
package plants, media filter, A constant filtration rate is maintained
44 Upgrading Existing or Designing New Drinking Water Treatment Facilities

I
<Jl Polyelectrolyte
Clear Water
ana

~ III
Backwash
Siorage
Alum Mixed
Media
fii u.. Filter I
.!2 -!'
Chlonne Influent
Connecbon
'-------,.

Figura 4-8. Flow diagram 01 a package plant.

by a low-head filter effiuent pump discharging cleaning is initiated more frequently than filter
through a float operated, level control valve. After a backwashing because more solids are removed by the
preset headloss is reached, backwashing of the filter clarifier. The clarifier cleaning process is
is initiated automatically. A manual backwash cycle automatically initiated either by a timer or a
can be initiated any time (if desired). Settled sludge pressure switch that continuously monitors headloss
from the tube settlers is flushed during the across the adsorption media.
backwashing cycle. Combining backwashing and
tube settler flushing simplifies operations and The air introduced under the adsorption media
reduces operator skill requirements. causes a vigorous scrubbing action. The scrubbing
action dislodges solids, which are washed away by
the flow of the incoming water. Flushing is generally
Adsorption Clarifier Package Plant timed to occur between every fourth and eighth hour.
The adsorption clarifier package plant is a new Complete cleaning of the adsorption media is not
concept developed in the early 1980s. It uses an desired because performance is enhanced by some
upflow filter with low density plastic bead media residual solids. Diagrams illustrating the various
(called the adsorption clarifier) followed by a mixed- cycles of an adsorption clarifier package plant are
media filter to complete the water treatment. The contained in Figure 4-10.
adsorption clarifier replaces the flocculation and the
sedimentation basins, combining flocculation and
sedimentation into one step. A typical example is 4.4.3 Operation and Maintenance
contained in Figure 4-9. Package plant operation is simplified by automated
features, and maintenance requirements are well
While passing through the adsorption media, the documented in manuals. However, the operator needs
coagulant and water are mixed, contact flocculated, to be well acquainted with water treatment
and clarified. The mixing intensity, as measured by principles and the plant manual, and should have
the mean temporal velocity, ranges from 150 to 300 attended a comprehensive training session.
per second. Flocculation is accomplished through
turbulence as the water passes through the Common automated devices found in package plants
adsorption media. In addition, flocculation is are effiuent turbidimeters and chemical feed
enhanced by contact between the flocculated controls. The effiuent turbidimeters and fail-safe
materials and the floc-coated media. Turbidity is controls are built into many plants to ensure that the
reduced through adsorption of the coagulated and finished water does not exceed set turbidity levels.
flocculated solids onto the adsorption media and the Automated chemical feed systems are especially
previously adsorbed materials. The adsorption appropriate for plants without full-time operators or
clarifier can achieve 95 percent or greater removal at with highly variable influent characteristics.
6.8 Usedm 2 (10 GPMfft 2 1. This highly efficient
clarification method results in extremely compact Typical plant operator and maintenance manuals
designs. contain operating principles, methods of establishing
proper chemical dosages, operating instructions, and
Adsorption clarifiers are cleaned by a combination of troubleshooting guides.
air scouring followed by water flushing. The air
scouring starts the cleaning process for the plastic Periodic visits by the manufacturer to make
media used in the clarifier. Adsorption clarifier adjustments to the plant and inspect the equipment
Filtration Technologies 45

Figure 4·9. Adsorption clarifier package plant (courtesy of Mlcrotloc products).

operations and performance are recommended. The Diatomaceous earth filters are effective in removing
first visit should be no more than 6 months after cysts, algae, and asbestos. For water supplies with
initial operation: the next should follow in another 6 low amounts of suspended solids, they have lower
months. Subsequently, annual visits should be initial costs than conventional rapid sand filtration
sufficient. systems. Diatomaceous earth filters are especially
effective against Giardia cysts.
4.4.4 System Performance The difficulties in maintaining a perfect film of
According to extensive manufacturer evaluations, diatomaceous earth as the filtering layer have
system performance, in general, is improving discouraged wide usage of diatomaceous earth filters
because of better equipment, more highly skilled for potable water treatment, except in favorable
operators, and greater surveillance by regulators. circumstances (i.e., waters with low turbidity and low
Table 4-5 contains summaries of evaluations at three bacterial counts). Consequently, while this
plants in Oregon. All finished water reported in the technology has been used extensively in specialized
table had less than 0.53 :-<TU turbidity. applications, such as swimming pools, it is a "new"
technology for water supply treatment.
4.4.5 System Costs Diatomaceous earth filter plants have been chosen
for projects with limited initial capital, and for
Table 4-6 presents capital, operating, and emergency or standby capacity to service large
maintenance costs for package plants. Costs range seasonal increases in demand. Since these systems
from $0.14 to $2.50/m 3 ($0.52 to $9.44/1,000 gall. are most suitable for applications where influent is
low in turbidity and bacterial counts, water supplies
4.5 Diatomaceous Earth Filtration presently receiving just chlorination may consider
Diatomaceous earth filtration, also known as precoat using diatomaceous earth to meet the filtration
or diatomite filtration, relies on a layer of requirements of the SWTR.I
diatomaceous earth about 0.3-cm (liS-in) thick placed
on a septum or filter element. The septums may be 1 "Ten Stale Standards" refers 10 the publicallon: Recommended
placed in pressure vesse Is or operated under a Standards for Water WorkS' PolicIes for the ReVIew and
vacuum in open vessels. A schematic diagram ofa Approval of Plans and SpecificatIOns for PublIc Water
typical pressure system is shown in Figure 4-11. Supplies. 1982. by Committee of the Great Lakes/Upper
MISSISSIppI River Board 01 SaOilary Engineers. published by Heallh
Educallon Service. P.O. Box 7283. Albany. New York 12224.
46 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Media
Retainer


Filter Effluen!

FluSh Cycle
Water Level

Media
Retainer

Filler
Adsorpnon Clan"er Flush Cycle Effluent

Wash Cycle
Water Level

Media
Ret<lJner

1'1 Backwash
\} Inlet

Filler
Etlluent
Backwash Cycle

Figure 4·10. Operating cycles of package plant (courtesy of Mlcrofloc products).


Filtration Technologies 47

Table 4·5. Summary Of R. . ulta Of AdlOrptlon Clarifk:atlon Packlllle Plants


Turb'dity, NTU
Rates. GPM/ft 2 (rnth) Adsorpt,on
Adsorption Water Temp.• Clanf,er M,xed-Med'a
Location Clarifier M,xed·Media OF (OC) Color Units Influent ENluent ENluent

CONal,S. OR 5 (12.2) 9 (22.0) 68 (20) a 100 8 0.27


10 (24.4) 10 (24.4) 68 (20) 0 46 10 0.52
8 (19.5) 7 (17.1) 57 (14) 0 103 21 0.30
Raln,er, OR 10 (24.4) 5 (12.2) 41 (5) 30 9.3 1.4 0.24
15 (36.6) 5 (12.2) 41 (5) 30 8.2 1.6 0.22
20 (48.8) 5 (12.2) 41 (5) 30 8.1 1.7 0.20
Newport. OR 10 (24.4) 5 (12.2) 45 (7) 8 19 4.3 0.13
15 (36.6) 7.5 (18.3) 45 (7) 8 15 3.7 0.11
20 (48.8) 10 (24.4) 45 (7) 9 9 38 0.23

1 GPMlft2 • 0.679 Llseclm2.


Source: MlCroflow ProduC1S Group.

Table 4-6. Estimated Costa for Supplementing Surface Water Tn8atment by Complete Treatment Package
Planta (S1978)

Plant Capac,ty, Average Flow. Capital Cost. Operat,on and Maintenance Costs Total Cost,
Category MGD MOD $1.000 $1.000/yr· ¢/l ,000 gal ¢/l.000 gal
1 0.026 0.013 278 12.2 2552 9445
2 0.068 0.045 295 15.9 87.5 277.4
3 0.166 0.13 428 42.4 89.2 195.1
4 0.50 0.40 773 75.1 51.4 113.6
5 2.50 1.30 1.770 137 29.0 72.8
6 5.65 3.25 2.952 274 23.1 52.4
• Processes ,nclude chemical feed (alum. soda ash. and polymer); complete trealmant package plant (f1occulabon, tube
sel1l,ng, and m,xed medoa filtration); backwash storagelclearwell baSinS; and sludge dewatering lagoons. A separate
pump,ng station 's used to transm't unlll,ckened sllKlge to Ille stlKlge dewatenng lagoons ,n Categories 5 and 6. SllKlge
pump,ng 's ,ncluded ,n Ille cost of I!le packalle plant in Categones 1 Illrough 4.
1 MGD • 0.044 m3 /sec; 1,000 gallons • 3.78 m3
Source: U.S. EPA (1979).

4.5.1 System Design 4.5.2 Operation and Maintenance


The majority of the minimum design criteria in the Operating a diatomaceous earth filter requires:
Ten State Standards for diatomaceous earth systems
meet the SWTR. However, two design criteria in • Preparation of filter body feed (diatomaceous
addition to the Ten State Standards are necessary to earth) and precoat
meet current regulations: • Verification of proper dosages
• Periodic back washing
• Disposal of spent filter cake
1. The minimum amount of filter precoat should be • Periodic inspection of filter septum for
1 kg/m 2 (O.21b/ft2 ) to enhance Giardia cyst cleanliness and damage
removal. • Verification of the emuent quality

2. The minimum thickness of the precoat should be A common operating difficulty is maintaining a
increased from 0.3- to 0.5-cm (118- to 1I5-inch) complete and uniform thickness of diatomaceous
(found to be more important than the size earth on the filter septum.
graduation of the diatomaceous earth), also to
enhance Giardia cyst removal. In some cases, alum precoating of the diatomaceous
earth can improve performance. Typical alum doses
range from I to 2 percent by weight (I to 2 kg/lOa kg
An additional recommendation is to use coagulant ofdiatomaceous earth applied). Typical precoats of
(alum or a suitable polymer) to coat the body feed to diatomaceous earth of 0.49 to 0.98 kg/m 2 (0.1 to 0.2
improve removal rates for viruses, bacteria. and Ib/ft 2 ) are applied to prepare the filter. These filters
turbidity. Adding these chemicals to the coating does also require a continuous supplemental body feed of
not improve Giardia removal rates. diatomite because the filter is subject to cracking. If
48 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Filtrate

Clear Liquid Line

Septum
Backwash
Line

Filter

Precoat Body Precoat


Tank Feed Dram
Tank Line
Filter Cake

Body
Raw Feed
To
Dram
Water Pump
Source Backwash
Dr&n LIne

Figure 4·11. Typical pressure diatomac80U8 e.rth IIltratlon system.

the filter has no body feed, there will be rapid levels. Effective removals of viruses and Giardia
increases in head loss due to buildup on the surface of range up to 99.95 percent and from 99 to 99.99
the filter cake. Body feed rates must be adjusted for percent, respectively.
effective turbidity removal. Diatomaceous earth
filters do not need a filter-to-waste cycle, because of
the precoating process. Some researchers have found that diatomaceous
earth filters, with added operational steps, are
Regular cleaning will maintain up to 95 percent of
effective in removing polioviruses. The additional
the filter septum area available for filtration after
steps include coating the filter surface with filter aid
100 filter runs. The filter cake drops off the septum
or mixing the filter influent with a cationic polymer.
during an interruption in flow, such as occurs during
The limited research found:
cleaning. During operating interruptions, clean
diatomaceous earth and filter water should be used to
recoat the filter to reduce the potential for passage of
• No viruses were detected in 11 effiuent samples
pathogens.
from a 12-hour run of a filter coated with 1 mg of
Filter runs typically range from 2 to 4 days. The rate cationic polymer per gram of diatomaceous earth.
of the body feed and the diatomite media size are
critical in determining the filter run length. Shorter • Only 1 of 12 effiuent samples contained viruses
filter runs will minimize filtered water odor and taste during the operation of another fil ter coated with
problems stemming from the decomposition of 1 mg of cationic polymer per gram of
organic matter trapped in the filter. diatomaceous earth.
Vacuum diatomaceous earth filters are a variation of • No viruses were detected in the effiuent in 12
this technology that offer the advantages of visibility samples from another 12-hour run of an uncoated
during backwashing and of not requiring pressure filter, and the influent was mixed with 0.14 mg of
vessels. Their primary disadvantage is that they run cationic polymer per liter of water.
an increased risk of the release of gases in the filter
cake that shorten filter runs.
4.5.4 System Costs
4.5.3 System Performance Costs for diatomaceous earth filters are provided in
Diatomaceous earth filtration units perform well on Table 4-7. Costs range from $0.09 to $1.78/m 3 ($0.35
waters with low influent turbidity and bacteria to $6.73/1,000 gal) depending on plant size.
Filtration Technologies 49

Table 4·7. Estimated Costs lor Supplementing Surface Water Treatment by Direct Filtration Using
Diatomaceous Earth ($1978)

Plant Capacity. Average Flow. Capital Cost. Operation and Maintenance Costs Total Cost.
Category MGD MGD $1.000 $1.000/yr· /t/l.000 gal /tIl .000 gal
1 0.026 0.013 221 6.0 127.0 672.9
0.068 0.045 285 8.0 43.7 2272
0.166 0.130 374 20.0 42.2 134.7
4 0.50 0.40 570 30.4 20.8 66.6
5 2.50 1.30 1.573 128 270 660
6 5.85 3.25 2.538 214 18.0 43.1
7 11.59 6.75 4.433 369 15.0 36.1
8 22.86 11.50 10.713 762 18.1 48.1
9 39.68 20.00 15.982 1.165 16.0 41.7
10 109.90 55.50 37.733 2.730 13.5 35.4
a Processes Include pressure diatomaceous earth IiIlratlon un,ts. diatomaceous earth teed eqUipment. filtered water storage
clearwell: and sludge dewatering lagoons. A separale administratIOn. lab. and mamtenance bUilding IS Included In
Categones 5 through 10. Sludge pumps are Included In the package tacllitles used ,n Categones 1 through 4. but separate
sludge pumping stalJons are Included in Categones 5 through 10. Categones 8 through 10 Include sludge holding tanks.
sludge dewatenng With Iilter presses. and hauling of dewatered solids 10 landfill diSposal.
1 MGD = 0.044 mJ/sec: 1.000 gallons = 3.78 mJ
Source: U.S. EPA (1979).

4.6 Other Filtration Systems designed to pass influent to either the inside or
outside ofthe membrane. The hollow fiber
This section covers the "emerging technologies" of membranes are contained in a pressure vessel or
membrane and cartridge filtration. Since there are cartridge and operate over a pressure range of 10 to
limited operational data concerning these 100 psi. Contaminants collect on the end of the hollow
technologies, pilot and case studies assume more fiber and are discharged to waste by reversing the
importance. water flow. A sample membrane system is shown in
Figure 4-12.
4.6.1 Membrane Filtration Periodic backfiushing and occasional chemical
Membrane filtration, also known as ultrafiltration, is cleaning is necessary to maintain the membrane
extremely compact and does not require coagulation. fibers. Treatment of the flush water containing solids
Membrane filters use hollow fiber membranes to requires separate coagulation and clarification. The
remove undissolved, suspended, and emulsified clarified flush water is either recycled or discharged
solids. The membranes are typically able to exclude after treatment. The sludge collected from this type of
all particles greater than 0.2 microns from the water system is typically dried and disposed of in a landfill.
stream.
Customarily, 90 percent of the feed water is collected
Membrane filters are typically used for specialized as permeate; the other 10 percent is discharged along
applications that require highly purified water, and with the contaminants. These filters can be designed
often serve as: to exclude particles larger than 0.01 11m. Unlike
reverse osmosis, this process does not exclude
• Pretreatment processes for reverse osmosis inorganic salts and electrolytes. Hollow fiber filters
with the finest membranes remove bacteria, Giardia,
• Pretreatment to remove colloidal silica from and some viruses.
boiler feed water
The hollow fiber membranes vary in size and porosity
• Treatment for drinking water supplies with and in their corresponding effectiveness, yet all
influent turbidities of 1 NTU or less, and fouling membrane filter fibers are sensitive to influent
indexes of less than 10 concentrations of suspended and colloidal solids.
Specifically, the flux level <the volume of permeate
Typically, ground water and surface water of high produced per unit area of membrane filter per day)
clarity have fouling indexes ofless than 10. Fouling and the flux stability are affected by:
of the hollow fibers by turbidity is the major problem
preventing widespread application of this technology. • Quality of the influent
• Filter cycle duration
Traditional membrane filters introduce feed water to • Quality of the back washing water
the inside of the hollow fiber membrane, with the
permeate or filtrate emerging on the outside of the In general, influent water with a fouling index ofless
membrane. State-of-the-art membrane filters are than 10 will permit filter cycles of8 hours with a 10
50 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Filtered Water

.... Hollow Fiber Membranes

Raw Water Supply Membrane


Pump Cartndge

Unfiltered
Recycle

Chemical
Coagulant-

Clarified
" ' - - - Air Inlet tor
Recycle - - - - - Backwashing

Backnush
- Wastewater

Backnush
Clarifier

Figure 4-12. Flow diagram of -.nbrane filtration system.


Fi Itration Technologies 51

percent reduction in flux. In between backflushes, a Turbidity removal was also evaluated by Colorado
high degree of the original fiber porosity can be State University. Prior to development of
retained with fast forward flushes oflhe influent, intermittent, air-assisted membrane cleaning,
which can then be routed to waste. After two to three membrane filters would experience rapid loss of flux
fast forward flushes, a full backflush is required to or flow rates at medium or high turbidity levels. The
restore the initial flux level. "radial pulse" system permits cleaning units without
such reductions in flux rates. The test results of this
system found that:
Application Considerations • Turbidity up to 30 NTU created from the
Membrane filtration is an attractive option for small introduction of bentonite clay were reduced to 0.2
systems because of its small size and automated NTU.
operation. Because it is best suited for small systems, • Actual field tests with 190 NTU influent
many typical membrane systems are skid-mounted turbidity resulted in effiuent with 0.6 NTU.
units. The diagram in Figure 4-13 illustrates this • Tests with effiuent ranging from 2.4 to 3.0 :-iTU
type of membrane system, and includes the following resulted in effiuent of 0.25 to 0.57 NTL'.
elements:

• Hollow membranes in cartridges Operation and Maintenance


• Automatic and manual valves for back washing Effective cleaning is essential for the successful
and unit isolation operation ofa membrane filter system. It is achieved
• Flow meters by backflushing and chemically cleaning the hollow
• Pressure gauges fibers to prevent fouling. Chlorine is sometimes
• Integral supply pump added to the backflush to destroy bacteria.
• Control panel Backflushing and chemical cleaning restore the
original porosity and flux rates, permitting the filter
Three other necessary components of membrane to operate indefinitely.
systems are (1) a separate supply pump and
interconnecting piping for plants with multiple filter Chemical cleaning is routinely done only onl:e a
units, (2) storage tanks and chemical feed pumps for week, unless there are unusually high levels of
membrane cleaning solutions, and (3) filtered water suspended solids in the influent. The cleaning
storage with chlorination capacity. solution is a mixture of caustic-based detergent and
hydrogen peroxide disinfectant. Membranes of water
systems with iron in their influent may require
cleaning with hydrochloric acid. Methods for disposal
System Performance of the spent cleaning solutions must comply with all
System performance data for membrane filter applicable regulations.
systems include data on Giardia, coliform, and
turbidity removal. The following results were from One manufacturer has developed the self-cleaning
tests conducted at Colorado State University's "radial pulse" system. This cleaning innovation
Department of Pathology. periodically injects gas under high pressure into the
center of the hollow fibers. The specially designed
membrane expands as the gas passes through, thus
One test evaluated Giardia removal effectiveness for removing materials fouling the membrane. Radial
one manufacturer's hollow membrane unit. The 0.2 pulse cleaning prolongs the effectiveness of the
I-lm membrane filters were found to be 100 percent membrane and thus extends the time between
effective in removing Giardia cysts at influent chemical cleanings.
concentrations of 1,100 cysts per liter. In addition,
researchers found that the unit's "radial pulse" Another manufacturer has developed a self·cleaning
cleaning mechanism was effective in preventing hollow-fiber membrane filter system that uses a flow
membrane fouling and caused no reductions in of water from the outside to the inside of the fibers to
Giardia cyst removal. clean the membranes. A water stream is tangentially
introduced outside of the fiber. The filter system then
Coliform removal was evaluated for a manufacturer's collects filtrate in the interior of the fiber, and the
hollow membrane unit. This evaluation used water rejected stream of concentrate is either recycled or
seeded with Escherichia coli bacteria ranging from 20 discharged. Continuous recycling permits extremely
to 30 miHion organisms per 100 mL. During a 130- high filtered water recovery rates of up to 99 percent.
minute test run, the effiuent contained less than I This system uses a more porous fiber than the more
coliform bacteria per 100 mL. No membrane traditional inside-to-outside design that is only about
breakthrough was experienced during three test 70 percent porous with a 0.21-lm nominal pore size.
runs. The higher porosity of these membranes allow for
52 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Chemical Resistant Membranes


Non·Corroslve Piping System - - - - . . .

Microprocessor Controls for


Radial Pulse. CIP System, and
Membrane Integnty Check

Clean In Place System - - - -

1 - - - - Stainless Steel

.,••••••
Stainless Steel Centrifugal Pumps
Skid Mounted
Frame
----;~.iiiii
PrefiltrabOn and Pretreatment System
When ReqUired

Figure 4-13, Skid-mounted membrane filtration aasembly.

treatment of influent with higher suspended solids effectively remove organic materials contributing
concentrations, These filters will meet turbidity color. Smaller inside-to-outside membranes of about
standards. exclude all Giardia cysts. and remove 0.1 11m will remove smaller particles, but when
coliform and other bacteria. but will not remove influent contains organic materials that contribute to
viruses. color, such as humic or fulvic acids, supplemental
treatments may be necessary.
System Costs Pilot testing of membrane systems is generally
Cost data for membrane filter units for small systems necessary to establish design criteria and operating
are contained in Table 4-8. The total costs range from parameters. The paucity of available data
SO. 371m 3 for the largest systems to $1.21/m 3 for the necessitates pilot testing. The only known membrane
smallest systems ($1.38 to $4.56/1,000 gal). filter used for water supply treatment is in Keystone,
Colorado. Most membrane experience has been in the
medical and electronic fields where tap water is
Potential Problems treated for special high purity applications.
The primary application concerns with membrane
filters are membrane failure and organics removal
effectiveness, Pilot testing is needed to qualify each
application. because of the lack of experience with the 4.6.2 Cartridge Filtration
process. Membrane failure is a major concern because Cartridge filters are considered an emerging
the membrane is the only barrier between potentially technology suitable for removing microbes and
pathogenic microbial contaminants and the finished turbidity in small systems. These filters are
water. Most other treatment systems have multiple mechanically simple but manually operated, so they
barriers to pathogenic breakthrough. To guard could be expensive to operate. They also require low
against this potential problem, some membrane turbidity influent. Cartridge filters use ceramic or
systems include features that trigger an operational polypropylene microporous filter elements that are
shutdown or activate an alarm. One manufacturer packed into pressurized housings. They operate by
includes a manual or automatic device to verify the the physical process of straining the water through
integrity of each membrane. porous membranes and can exclude particles down to
0,2 11m. The pore sizes that are suitable for producing
The second concern is that membrane filters may not potable water range from 0.2 to 1.0 11m. The ease of
be effective in removing certain organic components, operation and maintenance of cartridge filters makes
Larger membranes of about 0.2 11m will not them very attractive for small systems,
Filtration Technologies 53

Table 4-8. Estimated Costs lor Supplementing Surface Water Treatment by Package Membrane Filtration
Plants ($1978)

Planl Capaclly. Average Flow. Capital Cost. Operation and Malnlenance Costs Total Cost.
Category MGO MGO $1.000 $1.000Iyr' el'.OOO gal el'.OOO gal
1 0.026 0.013 142 50 105.2 4556
2 0.068 0.045 269 9.8 53.7 2268
3 0.166 0.130 503 26.0 54.7 179.2
4 0.50 0.40 1.144 67.7 46.4 1384
• Processes Include a complete package membrane hltratlOn uOll. clearwell storage. and sludge dewalering lagoons With
liquid sludge hauling.
1 MGD 2 0.044 m3tsec: 1.000 gallons ~ 3.78 m3
Source: U.S. EPA (1979).

One manufacturer uses single microporous ceramic turbidities of2 NTU or less, depending on the pore
filter elements packaged together in a cartridge size of the filter. Another manufacturer incorporates
housing with flow capacities of up to 1.5 Usec (24 particles of silver into the ceramic filters to prevent
GPM). This filter has pore sizes as small as 0.2 11m bacterial growth.
and is suitable for influent with moderate levels of
turbidity, algae, and microbial contaminants. The
4.7 Selecting the Appropriate Filtration
clean filter pressure drop is about 3.2 kglcm 2 (45 psi),
while the pressure drop during cleaning reaches Treatment System
about 6.2 kglcm 2 (88 psil. This section discusses considerations in selecting an
appropriate filtration technology. First, it discusses
the steps involved in selecting any filtration system.
Application Conslderat/ons Next, it discusses the role and objectives of pilot
Roughing filters, for pretreatment prior to cartridge studies, flocculation and sedimentation studies, and
filtration, are sometimes necessary to remove larger filtration studies in selecting the specific operating
suspended solids and prevent the rapid fouling of the characteristics for selected filtration technologies.
cartridges. Roughing filters can be rapid sand filters,
multimedia filters, or fine mesh screens.
4.7.1 Steps in an Evaluation
Prechlorination is recommended to prevent surface- The first step in selecting a new water treatment
fouling microbial growth on the cartridge filters and technology is to review all raw water quality data to
reduce microbial pass-through. Except for chlorine, establish the requirements for the potential
no chemical additions are necessary. There is a lack alternatives. Next, a list of alternative technologies
of data concerning the efTecti veness of cartridge should be compiled. The considered alternatives must
filters for viral removal. be able, in theory or as proven under similar
conditions, to solve the problems identified with the
current filtration system. Once the potential
Operation and Maintenance alternatives are selected, one must determine the
These systems are operationally simple, apart from necessity of pilot or bench-scale tests. If the desired
cleaning and membrane replacement. There is no performance of one or more of the alternatives is in
need for skitted personnel; personnel are necessary doubt, testing is appropriate. (Testing is always
only for general maintenance. Ceramic membranes useful if time and budget allow.) Otherwise,
may be cleaned and used for repeated filter cycles. literature surveys, bench-scale studies, or pilot test
Polypropylene cartridges become fouled relatively results can be used to derive performance
quickly and must be replaced with new units. characteristics and design considerations for each
alternative. For small systems, the alternatives for a
In one manufacturer's unit, cleaning the ceramic particular application may include slow sand filters,
cartridge filters entails cleaning each vertical filter package plants, diatomaceous earth ftltration, or
element with a hand-operated hydraulically driven membrane filters.
brush that fits over the element. Some
manufacturers use disposable polypropylene filter Following this initial selection, the basic process
elements in multi-cartridge stainless steel housing to concerns for the various alternatives should be
avoid the cleaning procedures. This type of unit is identified, including the following:
available with capacities ranging from 0.13 to 45.4
Usee (2 to 720 GPM). The primary disadvantage of • Turbidity removal performance
the disposable polypropylene membrane is that it can • Color removal performance
only be cleaned once before disposal. Manufacturers' • Giardia removal performance
guidelines state that these filters have service • Cleaning cycle frequency
periods ranging from 5 to 20 days with influent • Necessary chemicals
54 Upgrading Existing or Designing New Drinking Water Treatment Facilities

• Chemical dosages also particularly necessary for plants that use new
• Requisite operational skill short-detention and high-rate processes or where
• Applicable regulatory standards direct filtration is being considered.
• Necessary sludge management
4.7.3 Flocculation and Sedimentation
In the next stage, conceptual designs and Studies
preliminary layouts for selected alternatives are
prepared. Also, construction costs for the alternatives Full-scale flocculation and sedimentation studies are
should be developed. One should compare all necessary because it is difficult to extrapolate the
alternatives for reliability, simplicity, flexibility, and performance of pilot-scale flocculation and
ease of implementation, to select the appropriate sedimentation tanks to full-scale plants. Since
application. Finally, one can proceed with the final flocculator efficiency is directly related to volume,
design. smaller flocculators perform more efficiently than
their full-scale counterparts. Consequently, pilot
flocculation studies do not provide results applicable
4.7.2 Need tor Pilot Studies to full-scale facilities.
A pilot study is a broad term used for small-scale
testing of either complete water treatment processes Traditional sedimentation basins present even
or merely individual processes. Pilot studies are used greater extrapolation problems. Since small-scale
to evaluate alternate treatment options and versions cannot duplicate the process of traditional
operating techniques. Pilot tests can be relatively basins, which are generally 2.7 to 4.5 m (9 to 15 ft)
short or very long in duration. For example, time deep, full-scale sedimentation studies are also
requirements for pilot tests of rapid filtration are necessary. However, tube settlers can be evaluated
about 2 weeks; for slow sand filtration, 2 to 4 months; on a pilot scale. Figure 4-14 shows a sample test
and for corrosion inhibitors, up to 6 months. These module for a settling tube. These pilot-scale units can
tests may be continuous or intermittent. The longer, be suspended from existing basin walls or collection
more involved tests are more expensive. To avoid launders and operated at various flow rates.
unnecessary costs, pilot tests should have clearly
delineated objectives to preventincluding extraneous
evaluations. Yet they should be performed long 4.7.4 Filtration Studies
enough to encounter the full range of raw water Filtration studies can successfully employ pilot tests.
conditions and process design parameters. They are necessary for plants considering direct
filtration and very useful for plants with unusual raw
Simple jar tests are usually sufficient to evaluate water characteristics. One of the problems that can
procedures for traditional coagulation, flocculation, be identified and evaluated with pilot tests is the
sedimentation, and filtration processes. However, presence of diatoms (filter clogging algae) that do not
new or innovative technologies usually require the show up as high turbidity, yet can clog filters.
more extensive pilot tests. Another such problem involves the presence of paper
fiber, a common situation for water intakes below
Pilot tests have been endorsed by field experts for paper plant effiuent outfalls. These fibers also may
many years to assess precise design specifications, not show up as high turbidity but present filter
operational recommendations, and chemical clogging problems. Filtration pilot tests establish
applications procedures. In addition, pilot tests are whether the presence of paper fiber or diatoms will
required for some technologies to adjust individual cause problems.
processes to specific local water conditions. Pilot tests
are also used to identify unforeseen design and Side-by-side pilot filters will provide useful
operating problems, demonstrate operation to comparisons of different filter media designs being
regulatory authorities, and develop better considered for a particular application. A schematic
information concerning capital and operating costs. ofa pilot filter is shown in Figure 4-15. Each of the
Pilot studies may also be a prerequisite to obtaining three filters can be operated at the same flow rate and
conditional regulatory approval. They are especially is designed to maintain a constant flow as headloss
needed for new and emerging technologies and for increases. Pilot filters also include provisions to
accepted technologies with innovative elements, such measure headloss. Measurements of the filtered
as tube settlers or mixed-media filters. water turbidity and filter head loss are two of the
most useful results of side- by-side pilot tests to
Pilot studies are important to ensure the suitability predict filtration efficiency and filter run length.
of a small-scale prototypical plant for a particular
application. This is especially true for raw water with Pilot tests are especially recommended when high-
difficult treatment"aspects or poor quality, such as rate clarification/filtration processes are being
highly polluted water with high concentrations of considered. When such systems have short contact
organics, iron, manganese, and algae. Pilot tests are times, a significant risk of poor performance is
Filtration Technologies 55

Discharge
Flow Control Valve _ _ _ _ _

sample Pump~ ~
Discharge Pipe _ "'

Suppot1 Bracket

Perlorated CollectIOn Pipe

Corrugated Fiberglass
Side Panels

2" 6 In

FlQure 4-14. Tube ..ttllng tnt module.

present if a maladj ustment of the system occurs. Due


to the small permissible margin of error, pilot tests of
actual site conditions are necessary to supplement
system designs with extensive operational and pilot
test histories.
56 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Mixer Coagulant

Flash Mix. Overflow to Drain


Constant
Level Tank

DifferennaJ
( ' Filter AId Pressure Gage

Surface
Wash

Drain
Drain Drain

Backwash~
Pressure
Regulator

F1llU" 4-15. Pilot finer achematlc.


5. Disinfection and Disinfection By-Products

Disinfection is a process that kills or inacti vates of disinfection by-products and strategies for their
pathogenic microorganisms that occur in drinking control. The advantages and disadvantages 'ofthe
water supplies. Water treatment plants employ both major disinfectants are discussed in Section 5.3.
primary and secondary disinfection. Primary Sections 5.4 and 5.5 provide an overview of primary
disinfection achieves the desired level of and secondary disinfection technologies. Appendix B
microorganism kill or inactivation. Secondary provides case histories of water treatment plants
disinfection ensures a stable residual concentration using various disinfection technologies.
of disinfectant in the finished water to prevent
microbial growth in the distribution system.
Chlorine, ozone, chlorine dioxide, and ultraviolet 5.1 The Objectives of Disinfection
(UV) radiation are the major primary disinfectants; According to the Amendments to the SDWA, all
chlorine and monochloramine are the major public water suppliers, including those that rely on
secondary disinfectants. Some disinfectants can be ground water, will have to disinfect drinking water
used for both processes. before distribution. To assure compliance with all
applicable regulations (both current and
The 1986 Amendments to the Safe Drinking Water anticipated), the specific objectives of disinfection are
Act (SOWA) require all public water suppliers to to:
disinfect drinking water. In addition, inorganic and
organic chemicals will be regulated by means of • Assure 99.9 percent (3 log) and 99.99 percent (4
Maximum Contaminant Levels (MCLs). Since some log) inactivation of Giardia LambLia cysts and
disinfectants can produce chemical by-products, the enteric viruses, respectively
dual objective of disinfection is to provide the • Assure control of other harmful microorganisms
required level of organism destruction while • Not impart toxicity to the disinfected water
remaining within the MCLs for by-products set by • Minimize the formation of undesirable
EPA. disinfection by-products
• Meet the Maximum Contaminant Levels (MCLs)
Chlorine has been the most widely used disinfectant for the disinfectants used and by-products that
in the United States; however, it produces may form
trihalomethanes (THMs) and other halogenated
organic compounds in drinking water. Because of Disinfection alone, or a combination of disinfection
this, water suppliers are beginning to utilize other and filtration, can achieve the minimum mandatory
disinfectants, such as ozone, chlorine dioxide, and removals and/or inactivations of99.9 percent Giardia
monochloramine, or combinations of disinfectants, cysts and 99.99 percent enteric viruses. Primary
such as ozone followed by chlorine. Research on the disinfection systems that use ozone, chlorine, or
potential by-products associated with these other chlorine dioxide can achieve greater than the above-
disinfectants and combinations of disinfectants is stated inactivation of enteric viruses when 99.9
ongoing. percent inactivation of Giardia cysts is attained.
Therefore, achieving sufficient Giardia cyst
This chapter discusses the various disinfection inactivation can assure adequate inactivation of both
technologies used today as well as the issues relating types of organisms. This is not the case, however,
to disinfection by-products. Section 5.1 discusses the when using chloramination because it is such a poor
objecti ves of disinfection and Section 5.2, the problem virucide. Pilot-scale tests must be conducted by the

57
58 Upgrading EXisting or Designing New Drinking Water Treatment Facilities

utility to assure both inactivations are achieved SWTR tables to determine the inactivation ratio.
when relying on chloramination as a primary If the sum of the inactivation ratios, or
disinfectant.

Conventional treatment, which includes coagulation,


flocculation, sedimentation, and filtration, along at each point prior to the first customer where CT
with disinfection, can achieve 99.9 percent was calculated is equal to or greater than 1.0, i.e.,
inactivation of Giardia cysts and 99.99 inactivation there was a total of at least 99.9 percent
of enteric viruses when properly designed and inactivation of Giardia lamblia, the system is in
operated. Direct filtration, slow sand filtration, and compliance with the performance requirement of
diatomaceous earth filtration systems, each theSWTR.
combined with disinfection, have also achieved these
reductions. The fmal Guidance Manual for the SWTR is expected
to recommend that systems determine contact time
Ground-water systems that apply disinfection to based on the time it takes water with 10 percent of an
comply with regulations may have to add filtration if approximate tracer concentration (TIO) to appear at
they contain iron and manganese. Insoluble oxides the sampling site at peak hourly flow. For ground
form when chlorine, chlorine dioxide, or ozone are water not under direct influence of surface water, CT
added to these systems; thus, filters would be needed is determined in the same manner using enteric
for their removal. In addition, both ozonation and viruses or an acceptable viral surrogate as the
chlorination may cause flocculation of dissol ved determinant microorganism, since Giardia cysts will
organics, thus increasing turbidity and necessitating not be present.
filtration. The presence of such insolubles will
require the use of secondary disinfection after Table 5-1 presents the CT values required to attain 1-
filtration as well. log reductions of Giardia cysts, for four disinfectants.
As shown, lower temperatures require higher CT
values; with chlorine, an increase in pH also
increases necessary CT values. If more than one
5.1.1 CT Values disinfectant is used, the percent inactivation
"CT values" indicate the effectiveness of achieved by each is additive and can be included in
disinfectants in achieving primary disinfection. They calculating the total CT value.
describe the attainable degree of disinfection as the
product of the disinfectant residual concentration (in
mg/Ll and the contact time (in minutes). For Table 5·1. CT Valun lor Achieving 90 Percent Inactivation
01 Gla,dla Lamblia
chlorine, chlorine dioxide, or monochloramine, the
contact time can be the time required for the water to Temperature
move from the point at which the disinfectant is DiSInfectant pH :S1·C 5°C 10°C 15°C 20°C 25°C
applied to the point it reaches the first customer (at Free 6 55 39 29 19 15 10
Chlonne" 7 79 55 41 28 21 14
peak flow>. This is the total time the water is exposed (2 mgIL) 8 115 81 61 41 30 20
to the chlorinous residual before being used. Ozone, 9 167 118 88 59 44 29
however, has a short half-life in water; therefore, the Ozone 6-9 0.97 003 0.48 0.32 0.24 0.16
contact time is considered the time water is exposed
Chlorine 6-9 21 8.7 7.7 6.3 5 3.7
to a continuous ozone residual during the water dioxide
treatment process only.
Chloramines. 6·9 1.270 735 615 500 370 250
(preformed)
The Final Surface Water Treatment Rule (SWTR)
(US. EPA, 1989b) states: • CT values WIll vary depending on concentrabon of free chlonne.
Values IndlCaled are for 2.0 mgil of free chlonne. CT values for
different free chlonne concentratl()l1s are specified In tables In the
Systems may measure "C" (in mglL) at different Guidance Manual.
points along the treatment train, and may use • To obtain 99.99 percent lnactivahon of enteric viruses With
this value, with the corresponding ''T" (in preformed chloramlnes reqUIres CT values > 5.000 al
minutes), to calculate the total percent temperatures of 0.5.5.10. and 15°C.
inacti vation. In determining the total percent Source: U.S. EPA (1989al.
inactivation, the system may calculate the CT at
each point where "C" was measured and compare When direct filtration is included in the water
this with the CT99.9 value (the CT value treatment process, disinfection credit can be taken by
necessary to achieve 99.9 percent inactivation of the filtration step for a 2-log inactivation of Giardia
Giardia cysts) in the rule for specified conditions cysts and a I-log inactivation of viruses. This means
(pH, temperature, and residual disinfectant that the primary disinfectant must provide an
concentration). Each calculated CT value (CTelle) additional l·log inacti vation of Giardia cysts and 3·
must be divided by the CT99.9 value found in the log inactivation of viruses. In the specific instance of
Disinfection and Disinfection By-Products 59

a conventional treatment process that includes


coagulation, flocculation, sedimentation, and Table 5·3. CT Valu•• lor Achieving Inactivation 01 Virus..
at pH 8 Through 9
filtration, an inactivation credit of2.510gs for
Giardia cysts and 2 logs for viruses may be taken. Log
Temperature
This means that the primary disinfectant must Inac·
tivanDn 0.5"C 5"C 10"C 15"C 20"C 25"C
provide an additional O.510g inactivation of Giardia
Free 2 6 4 3 2 1 1
cysts but a 2-log inactivation of viruses. Chlorine" 3 9 6 4 3 2 1
4 12 8 6 4 3 2
If a water supply system does not use filtration, the Ozone" 2 0.9 0.6 0.5 0.3 0.25 0.15
99.9 percent inactivation of Giardia and 99.99 3 1.4 0.9 0.8 0.5 0.4 0.25
percent inactivation of enteric viruses must be 4 1.8 1.2 1.0 0.6 0.5 0.3
achieved by the primary disinfecting agents alone. Chlonne 2 8.4 5.6 4.2 2.8 2.1
Table 5-2 presents CT values for the four dioxidec 3 25.6 17.1 12.8 8.6 6.4
disinfectants for achieving 99.9 percent reductions of 4 50.1 33.4 25.1 16.7 12.5
Giardia cysts. Table 5-3 presents the CT values for Chlor- 2 1,243 857 643 428 321 214
virus inactivation. Although ground-water amlnesd 3 2,063 1,423 1,067 712 534 356
disinfection regulations have not been finalized, 4 2,883 1,988 1,491 994 746 497
these values will probably apply to systems treating • Data adapted from Sobsey (1988a) for inacbvation of Hepatitis A
ground water determined by the State not to be under Virus (HAV) at pH • 6, 7, 8, 9, and 10 and at 5"C. CT values
direct influence of surface water. include a safety factor of 3.
b Data adapted from Roy et aI. (1982) for inactivation of polioVIrUS at
pH 7.2 and 5"C. CT values Include a safety factor of 3.
c CT values tor ChlOrine diOXide are based on laboratory studies at
Table 5·2. CT Values lor AChieving 99.9 Percent pH 6.0 and 5"C (Sobsey, 1988a). CT values Include a safety
Inactivation 01 Giardia umblla" factor of3.
d Data from Sobsey (1 988a) for Inactivabon of HAV tor pH = 8.0,
Temperature
pH 5"C, and assumed to apply to pH in the range of 6.0 to 10.0.
DiSinfectant :S1°C 5"C 10'C 15"C 20"C 25"C These CT values apply only tor systems using combined chlOrine
Free 6 165 116 87 58 44 29 where chlonne IS added pnor to ammonia In the treatment
Chlorlne b 7 236 165 124 83 62 41 sequence. CT values given here should not be used for estlmabng
(2 mgIL) 8 346 243 182 122 91 61 the adequacy of diSinfection In systems applying preformed
9 500 353 265 177 132 88 chloramlnes, or applying ammonlll ahead of chlonne.
Ozone 6·9 29 1.9 1.4 0.95 0.72 0.48 Source: U.S. EPA (1989a).
ChlOrine 6·9 63 26 23 19 15 11
dioXide
Chioratninesc 6·9 3.800 2.200 1.850 1.500 1,100 750 disproportionation of the disinfecting agent and (2)
(preformed)
reaction of oxidation by the disinfectant with
a These CT values tor free chlOrine. chlorine dioXide, and ozone Will materials already in the water. Reduction, oxidation,
guarantee greater than 99.99 percent InaetJvatlon of entenc VIruseS. or disproportionation can occur when the disinfecting
b CT values Will vary depending on concentranon of free chlOnne.
Values Indicated are for 2.0 mgIL ot tree chlonne. CT values for agent is added to water. Three examples of this
dltlerent free chlorine concentranons are Specified in tables In the reaction are the formation of chlorite and chlorate
GUidance Manual (U.S. EPA, 1989a). ions associated with chlorine dioxide, the formation
C To obtain 99.99 percent InaetJvatlon of entenc VIruses WIth of dissolved oxygen associated with ozone, and the
preformed chlOramlnes requrres CT values > 5,000 at
temperatures of 0.5, 5, 10, and 15"C. formation of chloride ions associated with chlorine.
Source: U.S. EPA (1989a).
Oxidation of humic acids (in the water from organic
materials) produces aldehydes, ketones, alcohols, and
In the final SWTR (U.S. EPA, 1989b), the CT values carboxylic acids upon the addition of ozone, chlorine,
for ozone have been lowered to levels such that the chlorine dioxide, or potassium permanganate.
CT values required to provide O.5-log inactivation of Halogenation of organic materials can occur in the
Giardia cysts at the higher water temperatures are presence of free chlorine to produce trihalomethanes
below those required to provide 2 or 3 logs of and other halogenated organics. Chlorine can also
inactivation of enteric viruses. Consequently, the 2- form organic chloramines by reacting with nitrogen-
or 3-1og virus inactivation CT requirement becomes containing organic compounds (amino acids and
the pacing parameter for the amount of additional proteins). In addition, monochloramine can produce
primary disinfection to be provided by ozone during organic chioramines in the presence of
conventional treatment, rather than the O.5-log organonitrogen compounds.
inactivation of Giardia cysts.
Ifbromide ion is present in the untreated water, it
may be oxidized by ozone or chlorine (but apparently
5.2 Disinfection By-Products not by chlorine dioxide or chloramine) to form
Disinfection by-products are formed by two basic hypobromous acid, which in turn can brominate
mechanisms: (1) reduction, oxidation, or organic materials. Bromine-containing
60 Upgrading Existing or Designing New Drinking Water Treatment Facilities

trihalomethanes, for example, are known to form in common oxidants and disinfectants associated with
this manner. water treatment are listed in Table 5-4.

By-products are also produced when oxidants, like


ozone or chlorine, are used for oxidation purposes
Table 5-4. Oxldll1lon Potential. 01 Water Treatment
other than disinfection. For instance, breakpoint Oxidant.
chlorination is sometimes used early in the water
treatment process to remove ammonia. In the Species Oxidation Potential (Volts)
presence oforganic compounds considered Hydroxyl free radical (OHr 2.80
precursors, the same by-products that are formed Ozone'" 03 2.07
during chlorine disinfection are also formed in this Hydrogen peroxide H20 2 1.76
oxidation step.
Permanganate Ion MnO. 1.68
As another example, ozone is used as an oxidant to Hypochlorous acid' HOCI 1.49
improve turbidity, color, taste, odor, or Chlonne'" CI2 1.36
microflocculation; or to oxidize organic compounds, Hypobromous acid' HOBr 1.33
iron, or manganese. The addition of ozone early in the Bromine'" Br2 1.07
treatment process as an oxidant may produce the Hypoidous acid HOI 0.99
same by-products as when added later in an ozone
Chlonne dioxide'" CI0 2 (aq) 0.95
disinfection process. Potassium permanganate, also
used as an early oxidant, can produce oxidation by- Iodin&' 12 0.54
products as well. The maximum concentration of by- Oxygen O2 0.40
products is usually produced when oxidants are used -Excellent diSlnfecbng agents.
at the point in the treatment process where the
concentration of organics capable of being oxidized is
greatest and/or when large amounts ofoxidizing
agents are employed for long contact times. An agent's effectiveness as a disinfectant is not
always related to its effectiveness as an oxidant. For
Even when oxidants are used in the treatment example, whereas ozone is both a powerful oxidant
process for purposes other than disinfection, some and disinfectant, hydrogen peroxide and potassium
degree of disinfection occurs. In some cases, permanganate are powerful oxidants but poor
especially in treatment processes involving ozone, disinfectants. Chlorine dioxide and iodine are weak
chlorine dioxide, and chlorine under lower pH oxidants but strong disinfectants.
conditions, the primary disinfection requirement
may be satisfied during the preoxidation procedure Oxidation potential does not indicate the relative
(prior to filtration). speed of oxidation nor how complete the oxidation
reactions will be. Complete oxidation converts a
Since oxidation is so important in determining specific organic compound to carbon dioxide and
disinfection by-products, a brief description of the water. Oxidation reactions that take place during
chemistry of oxidation is provided in Section 5.2.1. water treatment are rarely complete; therefore,
This is followed in Section 5.2.2 by a short overview of partially oxidized organic compounds, such as
the occurrence and nature of disinfection by- aldehydes, ketones, acids, and alcohols, normally are
products. Lastly, Section 5.2.3 discusses the produced during the relatively short reaction peI;iods.
strategies for controlling disinfection by-products.
The behavior of a disinfectant as an oxidant will also
depend on the particular organic compounds in the
5.2.1 The Chemistry of Oxidation water supply. The level of total organic carbon (TOC)
The measure of an agent's ability to oxidize organic and the total organic halogen formation potential
material is its oxidation potential (measured in volts (TOXFP), when chlorine is used, indicate the
of electrical energy). Oxidation potential indicates likelihood that undesirable halogenated by-products
the degree of chemical transformation to be expected will be formed. Simply monitoring the reduction in
when using various oxidants. It gauges the ease with concentration of a particular organic compound,
which a substance loses electrons and is converted to however, is insufficient to indicate how completely
a higher state of oxidation. For example, if substance oxidation reactions are taking place. Unless a
A has a higher oxidation potential than substance B, compound is totally oxidized to carbon dioxide and
substance B theoretically can be oxidized by water, the TOC level may not change; therefore, the
substance A. Conversely, a particular substance concentrations of oxidation products must also be
cannot oxidize another with a higher oxidation measured. The TOXFP and the nonvolatile TOXFP,
potential. For example, ozone and chlorine can referred to as the nonpurgeable TOXFP (NPTOXFPl,
oxidize bromide ions to hypobromous acid, but indicate the potential for halogenated by-products to
chlorine dioxide cannot. The oxidation potentials of be formed from a specific raw water source.
Disinfection and Disinfection By-Products 61

5.2.2 The Presence of Disinfection By- Table 5-6. Summary of Health Effect. Auoclated with
Products in Drinking Water Chlorination By·Product.
EPA has recently surveyed 10 operating water Toxicological
utilities for the presence of22 halogenated Chemical Class Example Effect3
disinfection by-products in chlorine-treated water TrihaJomethanes ChlOroform C,H,RT
(Stevens et aI., 1987). Table 5-5 presents the DlChlOrobromomethane H,RT
frequency and range of concentrations of those by- D,l:>romochloromethane H, RT
products of greatest concern. Table 5-6 summarizes Bromoform H,RT
the current knowledge of health effects of selected
chlorination by-products. EPA is currently studying HaJoacetonitriles Chloroacetorutnle G,D
the by-products associated with ozonation. To date, DichlOroacetonitrile M,G,D
however, extensive studies of by-products of Tnchloroacetonitnle G,D
treatment with ozone, chloramination, and chlorine BromochlOroacetonitrile M,G,D
dioxide have not been conducted. Dlbromoacetonitrile G,D
HaJoacld denvatives Dichloroacetic acid MD, C, N, OL, A
Tnchloroacetic acId HPP
Chlorophanols 2-Chlorophenoi F, TP
Tallie 5-5. Occurrence of Chlorinated DI.lnfectlon By- 2,4 -Dichlorophenol F. TP
Product. at 10 Wat... Ulllltl..-
2.4,6-TrichlOrophenol C
Number of Range of Values
Compound Chlorinated ketones 1,l-Dichloropropanone M
Loca~ons !19IL
High Confidence 1,1,1- M
Tnchloropropanone
Chloroform 10 of 10 2.610594 1,1,3,3- M
BromOdichloromethane 10 of 10 2.61077 TetrachlOropropanone
ChlorOdil)romomethane 10 of 10 0.1 to 31 Chlorinated furanones MX M,CI
Bromoform 6 of 10 0.1 to 2.7 Chlorinated aldehydes 2-ChloroacataJdehyde G
D,chloroacetonltrile 10 of 10 0.2 to 9.5 Key to Toxicological Eff9C1S:
D,l:>romoacetOnJInle 3 of 7 0.4 10 1.2 C • Cancioogenic N • NeurotoXIC
H • HepatotoXIC OL • Ocular Lesions
BromOChloroacetonitnle 7 of 7 0.2 to 4.0 RT • Renal ToxIC A • Aspenna~ene$s
Chloropocnn 8 of 10 0.2 to 5.6 G • GenotoXIC HPP • Hepatic Peroxisome Proliferation
Low Confidence
o • DevelOpmental F • Fetotoxic
M • MutagenIC TP • Tumor Promoter
Chloroaceoc aCid 6 of 10 <10 MD· Metabolic CI • Clastogenlc
Disturbance
Dichloroaceoc acid 10 of 10 <10to>100 Source: AkIn el aJ. (1987).
Tnchloroacetic acId 6 of 10 10 to 100
TnchlOroacetaidehyde 10 of 10 10 to 100
(as chloral hydrate) 5.2.3 Strategies for Controlling Disinfection
1,, .1-Tric/llOropropanone 10 of 10 10 to 100 By-Products
2-Chlorophenol o of 10 The formation of halogenated by-products is affected
2,4-DichlOrophenol o of 10 by a number offactors, including the concentration
2.4,6-Trichlorophenol o of 10 and types of organic materials present when chlorine
OualitatiYe Only is added, the dosage of chlorine, the temperature and
pH of the water, and the reaction time. EPA has
, ,,-DichlOropropanone o vI a identified three strategies for controlling formation
1. I-DichlOro-2·l)utanone o of 8 of halogenated materials during chlorination:
3,3-DIChJoro·2·butanone 1018
1.1,1-TncIlIO<o·2-butanone 1. Remove the by-products after they are formed.
o of 8
Cyanogen chlonde 1 of 7 2. Use alternate disinfectants that do not produce
D,chloroacetald9hyOe 00110 undesirable by-products.
- In the first two grouP5, contaminant3 ate grouped accotding to
whether current knowledqe of health effects indicates a hogh or low 3. Reduce the concentration of organics in the water
confidence l!'Iat <l(lverse health effect3 eXIst In the third group, before oxidation or chlorination to minimize the
current knowledge of health effects is only qualitative to date. formation of by-products,
Source: S_ns et aI. (1987).
62 Upgrading Existing or Designing New Drinking Water Treatment Facilities

The first approach, removing the by-products after chloropicrin (nitrotrichloro-methane) can be
they are formed, can be difficult and costly. Section 6 enhanced if chlorination is preceded by ozonation.
discusses the treatment technologies available for Chloramine is known to react with acetaldehyde to
organic contaminant removal. The second approach, produce acetonitrile under drinking water treatment
using alternative disinfectants, is often the most cost- conditions. This and other nitriles might be expected
effective. The third approach, reducing the to be produced upon direct chloramination of
concentrations of organic precursors before adding ozonated waters containing aldehydes.
chlorine or other oxidants, will provide the highest
quality finished water. Chlorine dioxide is effective as a primary and
secondary disinfectant, but some chlorite ion is
produced. (See Section 5.5.1. for a discussion of the
Using Alternative DIsinfectants chemistry of chlorine dioxide.) The use of chlorine
The second approach, using other than chlorine for dioxide has been associated with hematological
disinfection, is sound if the replacements do not effects in laboratory animals, which may result from
produce undesirable by-products of their own and if the production of chlorite and chlorate ions.
they perform equally as both primary and secondary Neurological effects have also been identified. Due to
disinfectants. Cost is also a consideration. these concerns, EPA currently recommends
Alternative disinfectants currently being considered maximum total oxidant levels (total chlorine dioxide
by water'treatment specialists are chlorine dioxide, plus chlorite ion plus chlorate ion) in finished water
monochloramine, UV radiation, and ozone. Both of 1 mg/L. Thus, chlorine dioxide normally can be
ozone and UV radiation do not provide stable used as a primary disinfectant only in very clean
residuals for the distribution system and, therefore, waters, requiring low dosages of no more than 1.2 to
cannot be used as substitute disinfectants by 1.4mg/L.
themselves. l
If a strong chemical reducing agent is added
Although extensive studies of ozone by-products have somewhere in the treatment process after chlorine
not yet been conducted, many immediate oxidation dioxide primary disinfection, then chlorine dioxide
products of naturally occurring organic materials and chlorite ions can be reduced to chloride ion. This
have been identified repeatedly. For the most part, would leave only traces of chlorate ion in the water.
these by-products are organic aldehydes, acids, and This chemical reduction technique will allow much
ketones. Oxidation of raw water containing bromide higher chlorine dioxide dosages to be applied for
ion will produce hypobromous acid, which can oxidation and/or primary disinfection than the 1.2 to
brominate organic precursors. 1.4 mg/L currently recommended

Since ozone is employed only for primary At present, granular activated carbon (GAC) or
disinfection, a chlorinated compound (chlorine or sulfur dioxide are known to chemically reduce
chloramine) must be added for secondary disinfection chlorine dioxide and chlorite ion (but not chlorate
following ozonation, i.e., to provide a residual for the ion) to the innocuous chloride ion. This approach to
distribution system. Consequently, the "secondary" the application ofchlorine dioxide will be discussed
by-products, those formed by the reaction ofchlorine in more detail in Section 5.4.3.2.
or chloramine with the primary by-products of
ozonation, become a concern to water treatment The remaining alternative, monochloramine, is a
specialists. Although some studies have examined weak disinfectant. The very high CT values required
by-products produced by two-step oxidation to inactivate 99.9 perc.ent Giardia and 99.99 percent
sequences of this type, no compounds have yet been enteric viruses make monochloramine impractical
reported that are not produced by one ofthe two for use as a primary disinfectant. Therefore,
oxidation processes acting alone. monochloramine should only be considered as a
secondary disinfectant.
For example, preozonation may affect the yields of
THMs formed by subsequent chlorination. Usually
these THM yields are lowered by preozonation, but in MInImizing Precursor Concentrations
some cases, usually with high ozone dosages or at The third approach for controlling disinfection by-
high pH values, they can be enhanced. The yield of products is to reduce the concentration oforganic
materials before adding chlorine or any oxidant. This
, Ozone has been shown to be an 9t1ec1iYe $9OOl'IdatY disinteclant in approach will minimize the formation of by-products.
a study of European and Canadien appIicalions. However. lour Another option is to use an oxidant that does not
condiliona must be met simultaneoualy, whicll ~ found 10 C09llist contain chlorine, such as ozone, potassium
only rarely. The conditions are: (1) _
coellO slow biological regrow1hs, (2) w_ ten\l)9l'aIUnl8 must be
must be fnIe of iron and
ammonia, (3) lotlII organic cartlon vaiUllS must be less lh8n 1 rngIL,
permanganate, or hydrogen peroxide, before or
during rapid mix and/or filtration to partially oxidize
and (4) reaidence time ,n 1Ile distnbution system must be tess !han organics. This will improve the flocculation and
'2 hours (M~1er 9t aI.• 1978). filtration processes that follow. However, if the water
Disinfection and Disinfection By-Products 63

contains substantial amounts of bromide ion, acids are identical to those produced by potassium
brominated organics may be produced. permanganate, ozone, and chlorine dioxide.

Because all disinfectants and oxidants produce some


types of by-products, the most efficient approach to 5.3.2 Chlorine Dioxide
lowering organic by-product precursors is to optimize Chlorine dioxide is an excellent disinfectant, but is
physical process treatment steps before adding not as strong an oxidant as free chlorine. Because it is
oxidants. For example, if alum is used as the unstable, it must be generated on site. In its pure
coagulant, it should be recognized that the optimum state, chlorine dioxide does not produce THMs in the
pH is about 6.5. In addition, coagulant dosage should presence of organic materials. Some procedures for
be tested to maximize removal oforganics. synthesizing chlorine dioxide (e.g., sodium chlorite
and elemental chlorine procedures) use excess
It is important to note that extensive oxidation chlorine, which can produce THMs. Using less excess
converts natural organic materials (and some SOCs) chlorine in these procedures will lower the resulting
into simpler oxidation products (aldehydes, acids, THM concentrations. Generating chlorine dioxide
ketones, etc.), which are much more biodegradable with mineral acid and sodium chlorite solution
than the initial organic materials. Consequently, a avoids the presence of excess free chlorine.
biological treatment step following oxidation can
mineralize the oxidized materials, removing them Chlorine dioxide treatment of organic pollutants
from solution, thus avoiding the incorporation of generally produces nonchlorinated oxidation
these by-products into the finished water. products and some chlorinated oxidation products,
but in smaller quantities than chlorine. Chlorine
Examples of effective biological treatment steps are dioxide does not oxidize bromide ion to hypobromous
filtration, specifically through sand; dual media acid, as do ozone and chlorine, apparently because of
filters; GAC/sand filters (GAC on top of sand); and its low oxidation potential.
postfiltration GAC adsorbers. The biological
efficiencies of these types of filters increase in the During oxidation and disinfection reactions, as much
order listed. To allow biological filtration, it is critical as 90-95 percent of the chlorine dioxide reverts back
that no residual disinfectant be present in solution. to chlorite ion, which, along with chlorate ion, has
Otherwise, microbial activity present in the filter been associated with undesirable health effects. EPA
media will be eliminated or at least adversely currently recommends that finished water supplies
affected. contain a maximum total of 1 mglL combined chlorite
and chlorate ions, and chlorine dioxide.
5.3 Comparing Disinfectants
This section summarizes the advantages and
disadvantages of the various disinfectants. Since 5.3.3 Monochloramine
most disinfectants also function as oxidizing agents, Monochloramine is a weak cysticidal disinfectant
both oxidizing and disinfecting properties must be and a poorer virucide. Therefore, the contact times
considered when selecting a disinfecting agent. The and concentrations required for adequate primary
important characteristics of chlorine, chlorine disinfection are much longer and higher than with
dioxide, monochloramine, ozone, and UV radiation chlorine, chlorine dioxide, or ozone. When mono-
are described in Sections 5.3.1 through 5.3.5. Section chloramine is generated, dichloramine can be
5.3.6 describes advanced oxidation processes that use present. Water containing chloramines can be fatal
combinations of ozone with UV radiation or hydrogen to individuals on kidney dialysis, so local hospitals
peroxide. Sections 5.4 and 5.5 describe the and treatment centers must be warned against using
technologies used to apply each disinfectant. water containing chloramines for these patients.
Monochloramine is also a weak oxidant, and its slow
dissociation in water to free chlorine produces traces
5.3.1 Chlorine of halogenated organic materials. Monochloramine is
Chlorine is an excellent disinfectant and oxidant, and not recommended as a primary disinfectant because
is also a good chlorinating agent. It provides a stable its inacti vation of Giardia cysts is slow and it is a
residual for the distribution system if the water is poor viricide.
free of chlorine-demanding ammonia and organic
materials. Since chlorine can produce THMs and Monochloramine may be produced in three ways: 0)
other halogenated (TOX) and nonhalogenated adding ammonia to water containing chlorine, (2)
organic compounds, the use of chlorine should be adding chlorine to water containing ammonia, and
minimized, particularly when THM and TOX (3) using a preformed solution ofmonochloramine.
precursors are present. Some of the nonhalogenated Each technique is discussed in more detail in Section
oxidation products of chlorination of humic and fulvic 5.5.2.
64 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Adding ammonia to water treated with chlorine manganese, sulfide, nitrite, and arsenic; or to treat
(Method 1) will form THMs and other by-products organic materials responsible for tastes, odors, color,
associated with chlorination during the chlorine and THM precursors.
contact time used for disinfection or oxidation
purposes. In this case, the benefit from using an Ozone does not directly produce any halogenated
alternative disinfectant to chlorine is negated. organic materials, but if bromide ion is present in the
Adding ammonia will arrest THM generation, raw water, it may do so indirectly. Ozone converts
however, and the THM level will remain as produced bromide ion to hypobromous acid, which can then
from the initial chlorination contact time. form brominated organic materials. The primary by-
products of ozonation are oxygen-containing
Adding chlorine to water already treated with derivatives of the original organic materials, mostly
ammonia (Method 2), assuming proper mixing, will aldehydes, ketones, alcohols, and carboxylic acids.
produce insignificant concentrations of free chlorine Ozone, however, produces toxic oxidation products
and, consequently,lower concentrations of from a few organic compounds. For example, the
halogenated materials. However, disinfection is less pesticide heptachlor forms high yields of heptachlor
effective because the weak disinfectant epoxide upon ozonation. Therefore, when selecting
monochloramine is performing the disinfection ozone for oxidation and/or disinfection purposes, one
function, rather than free chlorine. must know the specific compounds present in the raw
water so as to provide the appropriate downstream
This problem is exacerbated if organic nitrogen treatment to cope with by-products. Researchers are
materials are already in the water. They react with continuing to study ozonation by-products and their
free chlorine and chloramines to form organic potential health effects.
chloramines, and all organic chloramines are even
weaker disinfectants than monochloramine. Even when ozone is used to oxidize rather than
Available field analytical methods do not distinguish disinfect, primary disinfection is attained
between inorganic monochloramine and organic simultaneously provided contact times and dissolved
chloramines in water. A utility with raw water ozone concentrations are appropriate. Consequently,
containing organic nitrogen materials that adds both oxidation and primary disinfection objectives
ammonia and then chlorine to produce can be satisfied with ozone prior to filtration, after
monochloramine for primary disinfection may which only secondary disinfection is needed.
seriously overestimate the degree of disinfection
attained. There are two cases in which this one-step
oxidation/disinfection with ozone is not feasible: (1)
Using a preformed monochloramine solution (Method when high concentrations of iron or manganese are
3), creates the same problem of producing less in the raw water, and (2) when ozone is used for
effective organic chloramines, if organo-nitrogen turbidity control. In both cases, measurement of the
compounds are present. degree of disinfection (dissolved ozone
concentrations) is impractical. When iron or
manganese are in the water, ozonation precipitates
5.3.4 Ozone dark insoluble oxides that interfere with the
Ozone is the strongest disinfectant and oxidizing measurement ofdissolved ozone. When ozone is used
agent available for water treatment; however, it is an for turbidity control, such low dosages of ozone are
unstable gas and must be generated on site. In used that a measurable concentration of dissolved
addition, it is only partially soluble in water, so ozone may never be attained. In these two cases,
efficient contact with the water must be established ozone oxidation and disinfection must occur
and excess ozone from the contactor must be handled separately.
properly. Section 5.4.2 discusses the specifics of ozone
generation and contacting methods. Ozone cannot be After being partially oxidized by ozone, organic
used as a secondary disinfectant because it is unable materials become more biodegradable and, therefore,
to maintain an adequate residual in water for more more easily mineralized during biological filtration.
than a short period of time. Preozonation of water fed to slow sand filters
increases the ease of biodegradation oforganic
Although the capital costs of ozonation systems are materials and enhances biological removal of organic
high, their operating costs are moderate. Because of materials during GAC filtration. The adsorptive
its high oxidation potential, ozone requires short efficiency of the GAC is extended because it only has
contact times and dosages for disinfection and to adsorb the organics unchanged by ozone, while the
oxidative purposes. As a microfiocculation aid, ozone partially oxidized organics are biologically converted
is added during or before the rapid mix step and its to carbon dioxide and water.
usage is followed by coagulation and direct or
conventional filtration. Higher dosages are used to Primary disinfection (or oxidation) with ozone
oxidize undesirable inorganic materials, such as iron, produces a significant amount of assimilable organic
Disinfection and Disinfection By-Products 65

carbon (AOC) comprised of readily biodegradable the amount of ozone generated, which will provide
aldehydes, acids, ketones, and alcohols. Many of some CT value.
these are also precursors ofTHMs and TOX
compounds. Consequently, if ozone disinfection is
immediately followed by chlorination, higher levels 5.3.6 Advanced Oxidation Processes
ofTHMs and TOX compounds may be produced than Ozone used in combination with UV radiation or
without ozonation. hydrogen peroxide can adequately disinfect and, at
the same time, oxidize many refractory organic
compounds such as halogenated organics present in
5.3.5 Ultraviolet Radiation raw water. Although contact times for ozone
UV radiation (254 nm) disinfection of bacteria and disinfection are relatively short, they are quite long
viruses has several important advantages: (1) it is for oxidizing organic compounds. This combination
readily available; (2) it produces no toxic residuals; process accelerates the oxidation reactions.
(3) required contact times are relatively short; and (4)
the equipment is easy to operate and maintain, Advanced oxidation processes involve combining
although maintenance must be performed on a ozonation with UV radiation (UV264 bulbs
regular basis to prevent fouling of certain submerged in the ozone contactorl with hydrogen
components. Section 5.4.3. discusses the specifics of peroxide (added prior to ozonation) or simply by
UV radiation equipment operation and maintenance. conducting the ozonation process at elevated pH
levels (between 8 and 10). Under any of these
UV radiation disinfection is inappropriate for: conditions, ozone decomposes to produce the hydroxyl
free radical, which has an oxidation potential of2.80
• Inactivation of Giardia LambLia cysts V compared with 2.07 V for molecular ozone.
• Water containing high suspended solids However, hydroxyl free radicals have very short half·
concentrations, color, and turbidity Iives, on the order of microseconds, compared with
• Water with high concentrations of soluble much longer half-lives for the ozone molecule.
organic matter that can react with or absorb the
UV radiation, thus reducing the disinfectant's Many organic compounds that normally are stable
performance under direct reaction with the ozone molecule can be
oxidized rapidly by the hydroxyl free radical.
Since UV radiation is ineffective against Giardia Chlorinated solvents such as trichloroethylene (TCE)
cysts, but effective against viruses and bacteria, it is and tetrachloroethylene (PCE) can be destroyed
a good candidate for disinfecting ground water not rapidly and cost effectively by hydroxyl free radicals
directly influenced by surface water. If the amount of (Glaze et al.,1988; Aieta et aI., 1988).
radiation recei ved by a target organism is not a lethal
dose, however, reconstitution of the organism and
reinfection of the water can occur. 5.4 Primary Disinfection Technologies
Primary disinfection is a key step in the water
Since UV radiation disinfection provides no treatment process. Typically, this step occurs either
disinfecting residual, a secondary disinfectant is just before or just after filtration in a conventional
needed. Very little oxidation of organic materials surface water treatment plant. Chlorine dioxide can
occurs with typical UV radiation systems used for be used for primary disinfection. However, the
drinking water disinfection; consequently if maximum recommended residual for chlorine dioxide
oxidation is required (for iron, manganese, sulfide, and its decomposition products (chlorite and chlorat!.'
nitrate, etc.), a strong oxidizing agent may be ions) may limit the use of this chemical as a primary
necessary and can serve as a primary disinfectant as disinfectant, except in cases of very clean waters
well. However, higher energy intensities and lower and/or short distribution systems that need only a
UV wavelengths (184.9 nm) can produce oxidation small amount of disinfectant. Monochloramine
reactions. cannot achieve the required inactivations of Giardia
cysts and enteric viruses within reasonable contact
UV bulbs that produce radiation at 184.9 nm times, and, therefore, is not appropriate for primary
generate some quantities of ozone which, in turn, can disinfection.
provide some oxidation oforganic materials. The
combination of UV radiation and ozone produces the For surface water treatment, chlorine and ozone are
hydroxyl free radical, which is a more powerful the predominant candidates. For ground water (not
oxidizing agent than is ozone itself(see Section 5.3.6, directly influenced by surface water), chlorine, ozone,
Advanced Oxidation Processes). and UV radiation are potential primary
disinfectants. Chlorine dioxide appears to be a good
The 184.9 wavelength radiation is not as effective for primary disinfection candidate for both types of
UV disinfection as the 254 nm wavelength, except by waters, provided that a means of reducing chlorine
66 Upgrading Existing or Designing New Drinking Water Treatment Facilities

dioxide and chlorate ions is used (e.g., GAC or some its efficiency and applicability to specific sites. The
strong chemical reducing agent). equipment, chemical, and operating and
maintenance considerations relevant to the three
Because chlorine produces undesirable by-products physical forms of chlorine are discussed in Sections
regulated by EPA, its use for pretreatment or 5.4.1.2 through 5.4.1.4.
primary disinfection must be carefully scrutinized. It
is anticipated that EPA will lower the current MCL
for THMs (l00 IglL), making chlorine use more 5.4.1.1 Process Description
difficult. In addition, regulation of some of the other Chlorine undergoes chemical reactions when added
halogenated by-products of chlorination listed in to water, and the resulting compounds inactivate or
Table 5-7 may place even greater restrictions on kill undesirable microorganisms. Chlorine gas will
chlorine use. form hydrochloric and hypochlorous acids according
to the following reaction:
Tallie 5·7. DlslnfllCtanll and Disinfectant By-Products
Uated In the Flnlt Drinking Water Priority List
(DWPL) Cl2 + <-----> HCl + HOCI
chlorine hydro- hypo-
DiSinfectants: chloric chlorous
Chlonne acid acid
Hypochlonte ,on
Chlonne dioxide
Chlorite ,on The hypochlorous acid reacts further depending on
Chlorate ,on the pH of the solution. The higher the pH, the more it
Chloramine will react, as shown below:
Ammonia
HaJoacetonltriles:
Bromochloroaceton,trile
Dichloroacetonltnle HOCI < ----- > (OCI) + H+
D,bromoacetoflltnle hypochlorous hypochlorite hy4rogen
Tnchloroacetonltrile acid ion Ion
TnhaJomethanes:
Chloroform
Bromoform The concentration of hydrogen ions in the water
Bromodichloromelhane determines the pH; the more hydrogen ions present,
DtChlOrQbromometl1ane the lower the pH. At neutral pH (pH = 7.0), almost
Halogenated Acids. Alcohols. Aldehydes. Ketones. and Other Nitnles 80 percent of the chlorine is in its most effecti ve
Others: disinfecting form, hypochlorous acid; the remainder
ChlorOllierin (Inchloronltromelhane) exists in the less effective disinfecting form,
Cyanogen chloride hypochlorite ion. Increasing pH reduces the total
Ozone by-products
disinfecting strength of the solution because it causes
an increasing amount of hypochlorous acid to form
more hypochlorite ion. At pH 8.0, for example, nearly
80 percent of the chlorine is present as the
The four primary disinfection technologies, chlorine, hypochlorite ion.
chlorine dioxide, ozone, and UV radiation, will be
discussed in Sections 5.4.1 through 5.4.4. Since most Table 5-2, presented earlier, gives CT values for
of the utilities that are affected by the Surface Water inactivation of Giardia Lamblia cysts with free
Treatment and upcoming Ground-Water Disinfection chlorine (2.0 mgIL). At any given concentration of
Treatment Rules serve less than 10,000 persons, this chlorine, the CT values increase rapidly as the pH
discussion will emphasize smaller utilities. rises above 7.0. This is also true at each temperature
listed. Figure 5-1 shows the relationship between pH
and the concentration of hypochlorous acid. Effective
5.4.1 Chlorine pH control is essential to achieve a desired level of
Chlorine is the most common primary and second!lry disinfection for systems relying upon chlorination.
disinfectant used in the United States. It is available
as a gas, solid, or aqueous solution. Chlorine gas is When sodium hypochlorite (liquid) or calcium
used most frequently, especially by large utilities, hypochlorite (solid) is used for chlorination, the
because of its lower cost. Chlorine in its solid form is resulting chemical reactions produce alkaline (basic)
calcium hypochlorite (Ca[OCI12); the liquid form is compounds as follows:
available as sodium hypochlorite (NaOCll solution.
The resulting hydroxides increase the pH ofthe
Section 5.4.1.1 describes the chlorination process, water, thereby lowering the concentration of
including the physical and chemical factors affecting hypochlorous acid and diminishing disinfection
Disinfection and Disinfection By-Products 67

microorganisms. Since chlorine will react with many


NaOCI + H2 O -----> HOCI + NaOH substances in the water, the "chlorine demand" of
sodium water hvdro- sodium these other substances must be satisfied before an
h~- chlorous hydroxide excess offree chlorine is available for disinfection.
ch orite acid Thus, the amount of chlorine necessary to effectively
disinfect must be greater than the chlorine demand of
the water.
Ca(OClh + H2O -----> 2HOCI + CaOH
calcium water hforo- calcium The "available chlorine residual" is the amount of
hypo- ch orous hydroxide chlorine that remains available for disinfection after
chlorite acid the chlorine demand is satisfied. This is quantified by
an approximate analytical testing procedure. The
100 o residual may be either a free available residual, a
~
~
combined available residual, or a combination of the
two. Free available chlorine refers to the total
90 10
concentration of hypochlorous acid and hypochlorite

80
~ 20
ions. Combined available chlorine is the total
concentration of mono- and dichloramines, plus

~
nitrogen trichloride and organic nitrogen chlorine-
containing compounds. (See discussion of
70 30

60
y 40
chloramines in Section 5.5.2.)

Because of the many complex reactions that take

U
0
:I: 50
\\ 50
U
a
place, the relationship between the amount of
chlorine added and the available residual does not
become linear until a certain minimum amount of
~

40
20· C
-\\. -0· C
60
~
chlorine has been added. In other words, increasing
the amount of chlorine does not resul t in a

30
\\ 70
proportional increase in the available residual until
that "chlorine breakpoint" is reached.

\
,
A series of chemical reactions causes the breakpoint
20 80
phenomenon. Water may contain small amounts of
reduced substances such as sulfides and ferrous iron,

10 ~ 90
as well as organic materials, organic nitrogen
materials (amino acids and proteins), and some
ammonia, all of which exert a chlorine demand. The
~ initial amount of chlorine added is taken up by
o 100 reactions with these contaminating substances,
4 5 6 7 8 9 10 11
pH
leaving no free available chlorine. After the chlorine
demands of the reduced substances have been
satisfied, the hypochlorous acid will begin to react
Figure 5-1. Distribution of hypochlorous acid and with ammonia, organic nitrogen materials, and some
hypochlorite lona In water at different pH valu...
andlemperalures of O'C and 20·C.
of the organics present to yield chloramines, oxidized
organics and chlorinated organic compounds. Next,
efficiencies. Therefore, the ability to adjust and the addition of more chlorine may induce the
control pH is critical when using the hypochlorite hypochlorous acid to oxidize or chlorinate some of the
forms of chlorine. same materials it has just created. At this point, a
decrease in the amount of residual (combined
Hypochlorous acid, a strong disinfecting agent, is one residual) is observed. When these oxidation reactions
of the most powerful oxidizing agents, and an are complete, the breakpoint is reached, and adding
effective chlorinating agent. In addition to acting on more chlorine finally increases the available chlorine
target organisms, it reacts with many substances in measured.
water, as evidenced by the production ofTHMs and
other halogenated compounds associated with Figure 5-2 shows a "chlorine breakpoint curve," with
chlorination. Chlorine also produce~ considerable the amount of chlorine shown on the horizontal scale
quantities of nonhalogenated organic oxidation and the amount of available chlorine shown on the
products, e.g., aldehydes, acids, and ketones. vertical scale. According to the curve, the chlorine
residual will not appear until 3 mg/L ofchlorine is
Adequate chlorine concentration will achieve added. After this point, additional chlorine will result
effective disinfection of currently regulated in an increase in residual. However, at about 6 mg/L,
68 Upgrading EXisting or Designing New Drinking Water Treatment Facilities

ChlOrine R8SlduaJ,
IT¢.

o
2 4 6 8 10 12 14
o

--
1~=,a:e ~
HZS.
Fe' '. etc.
-
ChlOrine and
Ammonia or similar
compounds
l - Free Residuals

F11IunI 5·2. Qnlphlcal repreMnUltlon of the breakpoint Clllorination ruetlon. The "n1lght line .. the left ahowe that chlortne
rwaldual la proportional to clouge In PUnl water. When Impurftt. . ani ~nt, they exert a chlorine demand.
Source: U.S. EPA (1983).

further additions of chlorine actually bring about a


decrease in residual until the breakpoint is reached Factors Affecting the Efficiency of Chlorine
(8 mglL in this diagram). After breakpoint is DIsinfection
achieved, additional chlorine finally results in a Chlorine in its free state (HOCI + [OC\J-) is an
proportional accumulation of residual free available effective disinfectant and inactivates most
chlorine. microorganisms in a matter of minutes. However,
effective disinfection with chlorine requires careful
Actual reactions are considerably more complex than attention to the following factors:
described above because of the time and
concentration dependencies of these processes. For • Free available chlorine concentration. The
these reasons, a breakpoint curve is difficult to concentration must be high enough to always be
recreate and predict; thus, individual tests must be detectable at the farthest points in the
run seasonally, and the data plotted to define the distribution system to effect both primary and
breakpoint for each water supply. secondary disinfection.
Disinfection and Disinfection By·Products 69

• pH. The pH should be maintained as close to 7.0 tank cars. The quantities required by small water
as is practical or consistent with other water systems can be purchased from local chemical or
quality aspects. This is necessary to maintain as swimming pool chemical suppliers.
much of the chlorine residual as possible in the
hypochlorous acid form. New EPA regulatory There are two basic types of gas chlorinators: (1)
initiatives, however, are encouraging utilities to pressure-operated direct gas feed units and (2)
adjust the pH of their product water to 8.0 in vacuum operated solution-feed units. Direct gas feed
order to minimize corrosion effects (U.S. EPA, units supply pressurized chlorine gas to the water
1988a). This higher pH will necessitate higher and are used only when electrical power is
doses of chlorine to attain primary disinfection. unavailable or the water pressure differentials are
insufficient to operate a solution feed unit. The
• Contact time. Contact time must be long enough solution feed units mix the gas with a side stream of
to achieve the desired degree of microbial water to form a solution of hypochlorous acid and
inactivation (Le., attain the CT value that applies hypochlorite ion, which then is mixed with the main
to the concentration of chlorine and the pH and stream. These units operate on a vacuum-controlled
water temperature). basis, automatically shutting off if the side stream
flow is interrupted. The solution feed system is safer
• Mixing. The chlorine contactor should either to operate and, therefore, is preferred by most
contain sufficient baffiing to eliminate the operators. Figure 5-3 shows a solution feed system.
possibility of short-circuiting or an external
mixing device should be added.
EqUipment Costs
Temperature also affects the disinfection rate; the Table 5-8 presents a range of equipment costs for one
higher the temperature, the faster the rate of basic and five increasingly complex solution feed gas
disinfection. Consequently, at higher temperatures, chlorination systems. The basic system includes
the CT values become lower. equipment to handle two 68-kg 050-1b) chlorine
cylinders, two cylinder-mounted chlorine gas
The choice of chlorination system - gas, solid, or regulators, an automatic changeover valve, and a
liquid - depends on a number of site-specific factors chlorine gas flow and rate valve ejector (with system
including: backup).

• Availability of chlorine source chemical The most sophisticated system includes the basic
• Capital cost of the chlorination system system plus two scales, a gas mask, a diffuser
• Operation and maintenance costs of the corporation cock (to allow connection under water
equipment line pressure), a flow-pacing chlorine addition
• Chemical costs system, a flow meter, a booster pump and piping, and
• Location of the facility a chlorine leak detector.
• Operator skills
• Safety considerations The costs are based on a small water treatment
• Local regulations regarding the storage of system sized to treat water volumes up to at least
chlorine gas 0.044 m3 /sec 0 MGD). They include equipment,
installation, safety enclosure, contractor's overhead
Each chlorination method provides the same and profit, and a 10 percent engineering fee. In May
disinfecting power on a pound for pound basis of 1980 dollars. the least expensive gaseous
available chlorine at the same pH. The choice of chlorination system costs were about $9,350. The
method depends primarily on the availability of each most sophisticated gaseous chlorination system with
chemical and the construction and annual operating all the options added cost $16.050.
costs for the different systems.

5.4.1.2 Disinfection with Chlorine Gas Operating and Maintenance Costs


Elemental chlorine is a toxic, yellow-green gas at For small solution-feed systems treating from 9.5
standard temperatures and pressures. It is supplied m3/day to 0.044 m3/sec (2.500 GPO to 1 MGD).
as a liquid in high-strength high-pressure steel operating and maintenance costs for gas chlorination
cylinders, and vaporizes rapidly when released. As systems are approximately the same. About 1,630
the liquid evaporates, its temperature falls and slows kWh each year is required to run the booster pump
evaporation rate, necessitating use of a container and approximately 2,560 kWh annually is required
manifold or vaporizer. for the building housing the system. assuming a 58.1-
m 2 (625 ft2) building. Maintenance materials for
Chlorine gas can be supplied in cylinders with miscellaneous repair of valves. electrical switches.
capacities of 45.4 to 907.2 kg 000 to 2,000 Ib), or in and other equipment cost about $40/yr. Labor
70 Upgrading Existing or Designing New Drinking Water Treatment Facilities

To Vent

Yoke
Clamp
Vacuum Line

Ejec10r and
Check Valve - ........
Assembly
Chlorine
Water Supply - ..1 _ _,..... SOlution

Chlorine Oas
Chlorine Cylinder

- --
...~..,....~

Chlorine Liquid

F1llU" 5·3. Solution IHd-g. . chlorinatIOn 1'f8t*Il.


Source: Capital ContrOls Co.• Inc.

Table 5·8. Clj)ital Coets lor G. . C h ~ ($1180) Chemical Costs


Equipment Costs lor a SYS18m Producing 100 LbIOay The cost ofchlorine for a system can be estimated
or Less using the following formula:
System
C1 cb... (JDWL x water treated (U~) x C1 c.,.t C~b)
Basic System" $1.873 2 2
Complex System 8.579 1.OOOCmw'1l X 4114(lF!bl
InstaJlanon 1.187 • C1 co.t(....Y)
2
Safety EnclOsuns 3.500
Contractor's o-t>ead and Proli1 (20%) 1.889
TatIle 15-1. capital CoeuI 'or an ChlOftnatlonl ($1980)
Engineering Fees (10".) i34
TOlsl Capi/al COSI 118m Costs in $
Ba.9IC System $ 9.343
EIeclricaI EI18f'VY
Pn:lC888 1.630 kWhIyr 114.10
Complex System $18,049
Building 2.560 kWhIyr 179.20
• May 1980 quotes (three vendors).
Sublotal 4,190 kWhIyr 293.3QO
" BasIC system Includes lWO 150·lb chlorine cylinders. lWO cylinder·
mounted regulators, automatic changeover vlllve. Chlorine gas !low Maintenane8 Mel8nals 40.00
rate valve. and "f6Ctor (system WIth backUp).
Ubor 183 hIyr 1.830"
1 Ib • 0.4536 kg
Source: U.S. EPA (1983). Total Annlllll O&M C08t 2,183
• Asaumes $O.07/kWh.
requirements of 1/2 hr/day (or 183 hr/yr) cover daily " Asaumes $lll1hour labor cost.
checks on equipment. Source: U.S. EPA (1983).
In the New York metropolitan area (1989), a 68-kg
Table 5-9 estimates total annual operating and (150-lb) cylinder ofchlorine costs about S1.651kg
maintenance costs of$2.163 for a solution feed gas (SO.7Mb). Applying the above equation, assuming a
chlorination system. The estimate assumes energy dosage of 5 mglL, gaseous chlorine would cost $30/yr
costs of$0.07IkWh and labor costs ofS10lhr, which for a 9.5 m3 /day (2,500-GPD) plant and S11,400/yr for
were the prevailing rates in 1982. a 0.0« m 3 /sec (1-MGDl plant.
Disinfection and Disinfection By-Products 71

provides estimates for basic and complex systems,


5.4.1.3 Disinfection with Sodium Hypochlorite both electrically and hydraulically activated. The
Solution basic liquid hypochlorination systems include two
Sodium hypochlorite solution is usually supplied metering pumps (one serving as a standby), a
commercially in concentrations of 5 and 15 percent solution tank, diffuser, and appropriate quantities of
chlorine. It is easier to handle than gaseous chlorine tubing. The more complex system adds a diffuser
or calcium hypochlorite. Metered chlorinators deliver corporation cock, antisiphon backflow pre venter, a
the solution directly into the water. safety housing enclosure, a flow pacing system, and a
flow meter and signal.
Sodium hypochlorite solutions lose their disinfecting
(oxidizing) power during storage, and thus should be
stored in a cool, dark, dry area. No more than a 1- Table 5·10. Capital Co.,. lor Liquid Chlorinators' ($1980)
month supply ofthe chemical should be purchased at Electrically Hyllraulically
one time to prevent loss of available chlorine. The ACllvated Activated
material is supplied in glass or plastic bottles, EQuipment Cost (basIC system b ) $ 1.800 S 2.266
carboys, or lined drums ranging in size from 1.89 to 500 1.000
208.2 L (0.5 to 55 gal). Bulk shipment by tank truck InS1aJlallon
is also a common form of transport. SiteWor!< 250 250
Contractor's Overhead & Profit 729 1.004
Sodium hypochlorite solution is more costly per (20%)
pound of available chlorine and does not contain the Engineering Fees 364 503
high concentrations of chlorine available from
Add Ons:
chlorine gas. However, the handling and storage
costs are lower than for chlorine gas. Alternate #,: aOd OINuser 165 231
corporatIOn cock anO anllsiphon
backflow preventer
An onsite generation technique for hypochlorite
Alternate 12: adO safety enclosure 6.930 6.930
solutions recently has been developed. This system (hoUSing)
consists of a two-cell unit, in which a brine solution
Alternate #3: add I10w pacing 1.485
(salt in water) is electrolyzed, producing a solution of existing signal
hypochlorous acid in one cell and a solution of caustic
Alternate #4: add I10w meter 1.452
(sodium hydroxide) in the other, according to the SIgnal. 8 in. or less
following equation:
Total Capital CoSl
BasIC System $ 3.643 $ 5.023
Na+ + Cl' + 2HzO + e- -----> MoSl Sophlstlcalled $10,738 $15.121
sodium chloride water electron • May 1980 quotes (two vendors).
b Ba.5IC System Includes two metering pumps (one standby). tubing.
solution tank. and diffuser.
HOCI + NaOH + Hz Source: U.S. EPA (1983).
hypochlorous sodium hydrogen
acid hydroxide
Total capital costs for an electrically activated system
Using onsite generation avoids the purchase and range from $3,643 for the basic system to $10,738 for
storage of large volumes of chlorine gas or the most sophisticated system. The comparable range
hypochlorite solutions, but there are significant for the hydraulically activated systems is $5,023 to
disadvantages. The generation process produces $15,121.
hydrogen, which poses fire and explosion hazards,
and sodium hydroxide, which is a caustic solution
that requires proper disposal. Also, the cost per pound
ofavaiJable chlorine is typically much higher for Operating and Maintenance Costs
onsite generation (e.g., $0.66 to $O.77/kg [$0.30 to As with solution-feed gas chlorinators, operating and
$0.35Ilbl for onsite generation compared to $0.18 to maintenance costs for systems in the 9.5 m3 /day to
$0.33/kg [$0.08 to $0.15Ilbl for chlorine gas). 0.044 m3/sec (2,500 GPO to 1 MGO) size range are
However, certain site-specific considerations may roughly the same. The annual estimated energy
make onsite generation a preferred disinfection requirements for the diaphragm metering pump and
technique. the housing structure, assumed to be 58.1 m 2 (625
ft2), are 570 kWh and 2,560 kWh, respectively.
Maintenance materials for minor component repairs
Equipment Costs are about $20 each year. Approximately 1 hour of
Table 5-10 presents estimated capital costs for labor is required each day to mix the sodium
sodium hypochlorite chlorination systems. This table hypochlorite solution and check equipment.
72 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Table 5-11 summarizes total annual operating and single procurement. However, it is hygroscopic
(readily absorbs moisture), and reacts slowly with
Table 5-11. Operation and Maintenance co.t Summary atmospheric moisture to form chlorine gas.
for Sodium HypocIIlortte SolU1lon Feed
($1982) Therefore, shipping containers must be emptied
completely or carefully resealed. Bulk handling
Item ReqUIrements Costs in S systems cannot be used.
Electrical Energy
Typically, the entire contents of a calcium
Process 570 kWhlyr 39.90
hypochlorite container are emptied into a mixing
Bu,lding 2,560 kWhlyr ill,lQ tank where they are readily and completely dissolved
Subtotal 3,130 kWhIyr 219.10- in water. The resulting corrosive solution is stored in
Maintenance Materials 20.00 and fed from a stock solution vessel constructed of
Labor 365 hlyr 3.650· corrosion-resistant materials such as plastic,
Total Annual O&M Cost 3,889
ceramic, glass, or rubber lined steel. Solutions of lor
2 percent available chlorine can be delivered by a
• Assumes SO.07/kWh. diaphragm-type, chemical feed/metering pump.
• Assumes S' O/hour labor cost
Source: U.S. EPA (1983).

maintenance costs for the sodium hypochlorite EqUipment, Operating, and Maintenance
solution feed system. The total of $3,BB91yr is based Costs
on the same energy and labor cost assumptions of Equipment, operating, and maintenance costs for
$0.07IkWh and $101hr as were used for the gas calcium hypochlorite solution feed systems are
chlorination system. similar to those for sodium hypochlorite feed systems.
The equipment needed to mix the solution and inject
it into the water being treated is the same.
Chemical Costs
Typically, sodium hypochlorite is available as a 15
percent (by weight) solution. Four-tenths of a kg (0.9 Chemical Costs
lb) of sodium hypochlorite solution is equivalent in A 9.5 m31day (2,500-GPD) treatment plant using a 5
oxidation potential to 0.45 kg (1 lbl of gaseous mglL dosage of chlorine needs 0.1041b chlorinelday.
chlorine, and its cost is about 3 times that of gaseous Because solid calcium hypochlorite contains 65
chlorine. These ratios for the two forms of chlorine percent available chlorine, 15.95 kg (0.16 lbllday is
can be used in conjunction with the formula provided required.
in Section 5.4.1.2 to estimate the cost of the solution.
In 1987 (in the BaltimorelWashington D.C. areal, 1
For example, a small water utility treating 9.5 kg of calcium hypochlorite cost $2.09 ($0.95I1bl. For a
m31day (2,500 GPO) with a dosage requirement of 5 9.5 m31day (2,500-GPDI facility, 26.5 kg (58.4Ib)
mglL of chlorine requires 242.2 L (64 ga\)lyr of the 15 costing $55.48 is needed for I year. A 0.044 m31sec (1-
percent solution. If the cost of the solution was MGDl facility, using 400 times that amount of
$0.13IL ($0.50Igall (as it was in the chlorine, would spend $22,192 annually.
BaltimorelWashington, D.C. area in 1987), the
annual chemical cost would be $32.00. A 0.044 m31sec
(l-MGDI plant using the same chlorine dosage would 5.4.2 Ozone
require 400 times the chemical volume and would While ozone is widely used for disinfection and
spend $12,BOO annually. oxidation in other parts of the world, it is relatively
new in the United States. The process of ozone
disinfection, including its chemistry, is discussed in
5.4.1.4 Disinfection with Solid Calcium Section 5.4.2.1. Section 5.4.2.2, system design
Hypochlorite considerations, covers issues related to the essential
Calcium hypochlorite is a white solid that can be components of an ozone treatment system. Lastly,
purchased in granular, powdered, or tablet form. It Section 5.4.2.3 reviews the costs of ozone system
contains 65 percent available chlorine and is readily equipment, installation, housing, and operation and
soluble in water. The chemical is available in 0.9-, maintenance.
2.3-,3.6-, and 15.9-kg (2-, 5-, B-, and 35-1bl cans and
362.9-kg (BOO-lbl drums, which are usually
resealable. Calcium hypochlorite is a corrosive 5.4.2.1 Process Description
material with a strong odor and requires proper Ozone (03) is a powerful oxidizing agent, second only
handling. to elemental fluorine among readily available
chemical supplies. Because it is such a strong
When packaged, calcium hypochlorite is very stable; oxidant, ozone is also a powerful disinfectant. Unlike
therefore, an annual supply can be purchased in a chlorine, ozone does not react with water to produce a
Disinfection and Disinfection By-Products 73

disinfecting species. Instead, when exposed to a This reaction is reversible; once formed, ozone
neutral or alkaline environment (pH above 6), UV decomposes to oxygen. This reverse reaction occurs
light, or hydrogen peroxide, it decomposes in water to quite rapidly above 35°C. Because the reactions that
more reactive hydroxyl free radicals as shown in the convert oxygen to ozone also produce a considerable
equation below: amount of heat, ozone generators have cooling
components to minimize ozone losses by thermal
decomposition.
03 + H2 0 ----- > 02 + 2(OHl
ozone water oxygen hydroxyl Ozone has a characteristic odor that is detectable
radicals even at low concentrations (0.01 to 0.02 ppm by
volume). Higher levels may cause olfactory and other
This reaction is accelerated at pH values above 8. reaction fatigue, and much higher levels are acutely
toxic in some instances. The longer the exposure to
In water, ozone reacts as the ozone molecule, the ozone, the less noticeable the odor.
hydroxyl free radical, or as a mixture of both. For
primary disinfection, CT values for ozone (shown in Ozone is only slightly soluble in water, about 2 to 10
Tables 5-1 and 5-2) have been developed for times more soluble than oxygen, depending on the
molecular ozone, not for hydroxyl free radicals. These temperature and its concentration as it enters the
free radical species are more effective oxidizing ozone contactor. The higher the concentration of
agents than molecular ozone. However, they have ozone generated, the more soluble it is in water.
extremely short half-lives (microseconds) and Increasing pressure in the ozone contactor system
consequently may not be good disinfectants. 2 also increases its solubility. Ozone's half-life in water
ranges from 8 minutes to 14 hours, depending on the
Since ozone is unstable at ambient temperatures and level of ozone-demanding contaminants in the water.
pressures, it must be generated onsite and used
quickly. Ozone is generated by applying energy to 5.4.2.2 System Design Considerations
oxygen (pure oxygen or dried air). A high-energy
electrical field causes oxygen to dissociate according The five major elements of an ozonation system are:
to the equation below:
• Air preparation or oxygen feed
• Electrical power supply
02 + e' ----- > 2[01 • Ozone generation
oxygen electrons oxygen • Ozone contacting
"fragments" • Ozone contactor exhaust gas destruction

These oxygen "fragments" are highly reactive and Air Preparation


combine rapidly with molecular oxygen to form the
triatomic molecule. ozone: Ambient air should be dried to a maximum dew point
of -65°C before use in the ozonation system. Using air
with a higher dew point will produce less ozone, cause
2[01 + 202 ----- > 203 slow fouling of the ozone production (dielectric) tubes,
and cause increased corrosion in the ozone generator
oxygen oxygen ozone unit and downstream equipment. These last two
"fragments"
factors result in increased maintenance and
downtime of the equipment.
The overall reaction that produces ozone is the sum of
the above reactions: Air feed systems can dry ambient air or use pure
oxygen. Using pure oxygen has certain advantages
that have to be weighed against its added expense.
302 + e- < ----- > 203 Most suppliers oflarge-scale ozone equipment
oxygen energy ozone consider it cost effecti ve to use ambient air for ozone
systems having less than 1,587.6 kg/day (3,500
lblday) generating capacity. Above this production
rate, pure oxygen appears to be more cost effective.
Systems that dry ambient air consist of desiccant
dryers, commonly used in conjunction with
2 Very recent data (Wolfe et al., 1989) show that hydroxyl free compression and refrigerant dryers for generating
resIduals produced by the combinabOO of ozone with hydrogen large and moderate quantities of ozone. Very small
perOXIde have lhe ability to inactivate GiBrdilJ cysts and enteric systems (up to 0.044 m 3/sec [l MGD]) can use air
VIruseS rapidly.
drying systems with just two desiccant dryers (no
compression or refrigerant drying). These systems
74 Upgrading Existing or Designing New Drinking Water Treatment Facilities

use silica gel, activated alumina, or molecular sieves • Ifused in a once-through system, gas recovery
to dry air to the necessary dew point (-6S·C). and pretreatment equipment are eliminated.

Ambient air feed systems used for ozone generation • Ozone transfer efficiencies are higher due to the
are classified by low, medium, or high operating higher concentration of ozone generated.
pressure. The most common type is a system that
operates at medium pressures ranging from 0.7 to However, the economic implications of these
1.05 kg/cm 2 (10 to 15 psig).·High-pressure systems advantages must be weighed against the capital
operate at pr:essures ranging from 4.9 to 7.03 kg/cm 2 expenditure required for onsite oxygen production or
(70 to 100 pSlg) and reduce the pressure prior to the operating costs associated with purchase of liquid
ozo?e generator. Low- and high-pressure systems are oxygen produced off site.
typically used in smaH- to medium-sized
applications. Medium- and high-pressure systems Oxygen can be purchased as a gas (pure or mixed
may be used in conjunction with most ozone with nitrogen) or as a liquid (at -183·C or below).
generators and with most contacting systems. Low- Normally the purity of the available oxygen gas is
pressure systems operate at subatmospheric quite adequate, and no particular pretreatment is
pressures, usually created by a submerged turbine or required. Its purity always should be better than 95
other contactors producing a partial vacuum percent, and its dew point consistently lower than
throughout the air preparation and ozone generation -60·C. When liquid oxygen is the oxygen source, it is
system. Creation of this vacuum results in ambient conver:ted to the gas phase in an evaporator, and then
air being drawn into the ozonation system. sent directly to the ozone generator. Purchasing
oxygen as a gas or liquid is only practical for smaH- to
m~diu~-sized systems. Liquid oxygen can be added to
T.he decision to use a high-, medium-, or low-pressure
air prep.aration system often is based on a qualitative dned air to produce oxygen-enriched air (as at the
evaluation of potential maintenance requirements, Tailfer plant serving Brussels, Belgium). A
as well as an evaluation of capital and operating membrane separation method that also produces
costs. High-pressure air pretreatment equipment oxygen-enriched air is being developed.
generally has higher air compressor maintenance
requirements, lower desiccant dryer maintenance There are cl;1rrently two methods for producing
requirements, and lower capital costs. oxygen?n site for ozone generation; pressure swing
adsorption of oxygen from air and cryogenic
At small- to medium-sized instaHations the lower production (liquefaction of air followed by fractional
capital costs may offset the additional ~aintenance distillative separation of oxygen from nitrogen).
required for the air compressors and associated Systems for the production of oxygen on site contain
equipment, such as filters for the oil-type many of the same elements as air preparation
compressors. Typical low- and high-pressure feed gas systems discussed above, since the gas stream must
pretreatment systems are shown in Figures 5-4 and be clean and dry in order to successfully generate
5-5. Figure 5-4 is also representative of a medium- oz~ne. Gaseous oxygen produced on site by pressure
pressure system, but may require a pressure 9wI.ng ~d~rption typically is 93 to 95 percent pure,
reducing valve upstream from the ozone generator. ~h1le hqUid oxygen produced cryogenically generally
IS 99.5 percent pure. In most plants utilizing on site

For many applications, pure oxygen is a more production of ozone, a backup liquid oxygen storage
attractive ozone feed gas than air for the following system is included.
reasons:
At smaller plants, oxygen can be separated from
ambient air.by pressu~e swing adsorption using
• It has a higher production density (more ozone
produced per unit area of dielectric). ~olecula: sieves. Dunng the high-pressure phase,
mtrogen IS adsorbed by the sieves and oxygen exits
• It req~ires lowe.r energy c~nsumption(energy
supphed ~r umt area of dielectric).
the system as product gas. When the pressure is
reduced, the nitrogen desorbs and is removed from
the vessel by purge gas. Precautions should be taken
• Essentially double the amount of ozone can be
g~nerated per unit time from oxygen than from to avoid cont,amination of the oxygen prepared by this
air (for the same power expenditure); this means procedure With hydrocarbons, which are present as
that ozone generation and contacting equipment oils associated with the pressurized equipment.
can be halved in size when using oxygen, to
generate and contact the same amount ofozone. Pressure swing adsorption systems for producing
oxygen are manufactured to produce from 90.7 to
• Smaller gas volumes are handled using oxygen
rather than ~ir, for th~ same ozone output; thu;,
27,216 kg/day (200 to 60,000 Iblday) of oxygen. This
production range would supply ozone for water
costs for anCillary equipment are lower with treatment systems at 6 percent concentration in
oxygen feed gas than with air. oxygen, for an applied ozone dosage of 4 mg/L and
Disinfection and Disinfection By-Products 75

----r_------:J_:
: After-COOler: : Atter-Cooler :

m~m_
r---[---, ,--- ---,
: Refngerant : : Refngeranl :
I
~
Dryer ,
J
,l Dryer I
J

Heat-Reactivated
Dess,cant Dryers
Tower Tower
A A

Tower Tower
B B

TO Ozone Generator To Ozone Generator

Figure 5-4. Low pressure air leed-gas treatment schematic lor ozone generation.
Source: U.S. EPA (19860).

production of water at the rates of about 0.018 m3 /sec (8,OOO-lb/day) ozone system, with cryogenically
to about 5.26 m 3 /sec (0.4 MGD to about 120 MGD). produced oxygen as a feed gas.

For cryogenic oxygen production, low temperature Reuse or recycling of the oxygen-rich contactor
refrigeration is used to liquefy the air, followed by ofTgases is possible, and requires removal of moisture
column distillation to separate oxygen from nitrogen. and possible ammonia, carbon dioxide, and nitrogen
Cryogenic systems are operationally sophisticated, before returning the gas to the generator.
and operating and maintenance expertise is required.
Production of oxygen by the cryogenic technique is
more capital intensive than by pressure swing
adsorption, but generally operation and maintenance Electrical Power Supply
costs are lower. For oxygen requirements of 18,144 to The voltage or frequency supplied to the ozone
18,144,000 kg/day (20 to 20,000 tons/day), cryogenic generator is varied to control the amount and rate of
separation systems are quite practical. This would ozone produced. Varying the power requires
exclude their use for small water treatment plants. specialized supply equipment that should be designed
The Los Angeles Aqueduct Filtration Plant, which is for and purchased from the ozone generator
26.28m 3/sec (600 MGO) is using a 3,628.8-kg/day manufacturer.
76 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Heat·Reactivated
Desiccant Dryers
Tower
A

Tower
B
/

PRY

-----.J'--_.- - - - -
To Ozone Generator To Ozone Generator
Figure 5-5. High pressure air feed-gas treatment ac:hematlc fOf" ozone generation.
Source: U.S. EPA (1986b).

Ozone generators use high voltages (> 10,000 V) or voltage potentials are maintained. Oxygen-enriched
high-frequency electrical current (up to 2,000 Hz), air, pure oxygen, or air that has been dried and cooled
necessitating special electrical design considerations. flows between the electrodes and produces ozone.
Electrical wires have to be properly insulated; high More recent designs use medium and high
voltage transformers must be kept in a cool frequencies, rather than high voltages and low
environment; and transformers should be protected frequencies, to generate ozone.
from ozone contamination, which can occur from
small ozone leaks. The frequency and voltage
transformers must be high quality units designed
specifically for ozone service. The ozone generator
supplier should be responsible for designing and
3 ThiS techniQue produces concentrabOns of ozone suflicienUy
supplying the electrical subsystems. h'llh (above 1 percent by weight) to solubilize enough ozone
and to attain the reQuisite CT values necessary to guarantee
disintection of Giardia cysts. Ozone also can be generated by
Ozone Generators UV radiabOn techniQues. but only at maximum concentrabons
Ozone used for water treatment is usually generated of 0.25 percent by weoghl At such low gas concentrations. It is
nat possoble to solubilize sufficient ozone to guarantee
using a corona discharge cell.3 The discharge cell disinfection of G,arda cysts.
consists of two electrodes separated by a discharge
gap and a dielectric material, across which high
Disinfection and Disinfection By-Products 77

Figure 5-6 depicts the essential components of a arrangement will satisfy peak demands and one
corona discharge ozone generator. Either ambient generator will be on hand during off-peak periods for
air, oxygen-enriched air, or pure oxygen is fed to the standby or maintenance.
generator. Ifambient air is used, the generator
produces dry, cool air containing I to 3.5 percent Ozone Contacting
ozone (b)O weight), which can be mixed with water.
When pure oxygen is used, the concentration ofozone Ozone can be generated under positive or negative air
produced is approximately double that produced with pressure. If generated under positive pressure, the
ambient air (up to 8 to 9 percent by weight). contactor most commonly used is a two- chamber
porous plate diffuser, with a 4.8-m 06-ft) water
The most common commercially available ozone column. With this method, the ozone-containing air
generators are: exits the ozone generator at approximately 1.05
kg/cm 2 (15 psig) and passes through porous diffusers
• Horizontal tube, one electrode water cooled at the base of the column. Fine bubbles containing
• Vertical tube, one electrode water cooled ozone and air (or oxygen) rise slowly through the
• Vertical tube. both electrodes cooled (water and column, the mass of ozone is transferred (dissolves),
oil cooled) and oxidation and/or disinfection takes place. The
• Plate, water or air cooled 4.8-m (I6-ft) height maximizes the amount of ozone
The operating conditions of these generators can be transferred from the bubbles as they rise in this type
subdivided into low frequency (60 Hz), high voltage of porous diffuser contactor.
(> 20,000 V); medium frequency (600 Hz), medium
voltage « 20,000 V); and high frequency (> 1,000 Other types of positive pressure ozone contactors
Hz), low voltage (< 10,000 V). include packed columns, static mixers, and high
speed agitators. An atomizer that sprays water
CurrentlY,low frequency, high voltage units are through small orifices into an ozone-containing
most common, but recent improvements in electronic atmosphere also can be used.
circuitry make higher frequency, lower voltage units When ozone is generated under negative pressure,
more desirable. vacuum action draws the ozone mixture from the
generators, providing contact as the gas mixes with
Operating an ozone generator at 60 to 70 percent of the flowing water as with a submerged turbine.
its maximum production rate is most cost effective. Other common methods ofcreating negative pressure
Therefore, if the treatment plant normally requires use injectors or Venturi-type nozzles. These systems
45.36 kg/day (IOO lb/day) of ozone and 68.04 kg/day pump water past a small orifice (injector) or through
(I50 lb/day) during peak periods, it is wise to a Venturi nozzle, thus creating negative pressure.
purchase three generators, each designed for 27.22
kg/day (60 Ib/day) and operate all three at about 65 The diffuser and packed tower contactors require no
percent capacity for normal production. This energy above that required to generate ozone at 1.05

Electrode
D,electnc
O2 • Discharge Gap

j m
W¥f - Electrode

Heat

FIgure 5-6. Typical ozone generating configuration lor a Corona dlachaNJe cell.
78 Upgrading Existing or Designing New Drinking Water Treatment Facilities

the value for ''1'''. Only when water nowing through


kglcm 2 (15 psig). The high speed agitators, static an ozone contacting system approaches plug now does
mixers, and all the negative pressure contactors the actual ozone contact time approach the hydraulic
require additional energy. residence time.
Ozone reactions are very fast to destroy or inactivate
microorganisms; oxidize iron, manganese, sulfide The greater the number of ozone contact chambers
and nitrite ions, and some organics; and lower that can be connected in series, the closer the water
turbidity levels. However, ozone oxidizes organic flow will approach plug flow. In such cases, the TIO
compounds such as humic and fulvic materials, as (time for 10 percent of an added tracer to pass
well as many pesticides and volatile organic through the ozone contacting system) will approach
compounds, quite slowly compared to these other 50 percent of the hydraulic detention time for water
solutes. passed through the ozone contacting system.
Recently published studies of the 26.28 m3/sec (600·
For disinfection, the initial dose of ozone is used to MGD) Los Angeles Aqueduct Filtration Plant and of
satisfy the ozone demand of the water. Once this the design of the 5.26 m3/sec (120-MGD) Tucson
demand is satisfied, a specific ozone concentration water treatment plant ha ve confirmed that the actual
must be maintained for a specific period of time for hydraulic residence time (TIO) is approximately 50
disinfection. These two stages of ozonation are percent of the theoretical hydraulic residence time.
usually conducted in two different contacting
chambers (see Figure 5-7). Approximately two-thirds
of the total ozone required is added to the first
Destruction of Excess Ozone from Ozone
chamber where the ozone demand of the water is met Contactor Exhaust Gases
and the dissolved ozone reaches a residual level The ozone in exhaust gases from the contacting unit
(typically 0.4 mglL). The remaining ozone is applied must be destroyed or removed by recycling prior to
in the second chamber, where it maintains the venting. The current Occupational Safety and Health
residual ozone concentration for the necessary Administration standard for exposure of workers
contact time with water to attain the required CT during an 8-hour work day is a maximum ozone
value. concentration of 0.0002 g1m 3 (0.1 ppm by volume
time-weighted average). Typical concentrations in
When ozone is added to water, its dissol ved residua I contactor exhaust gases are greater than 1 g/m 3 (500
is not stable. Not only will ozone react with many ppm by volume).
contaminants in water supplies, but its half-life in
water is fairly short, on the order of minutes, due to The four primary methods of destroying excess ozone
decomposition back to oxygen. At the higher pH are thermal destruction (heating the gases to 300· to
ranges (above 8), decomposition of molecular ozone 350·C for 3 seconds), thermal/catalytic destruction,
into reactive intermediates (including the hydroxyl and catalytic destruction (with metal catalysts or
free radical) is accelerated. Consequently, it is not metal oxides). Moist granular activated carbon is
possible to monitor the residual ozone concentration used extensively at European plants treating less
at any single point in the treatment train and expect than 0.088 m3/sec (2 MGD) and for swimming pools
a single concentration level to hold steady from the using ozone.
point of gaslliquid mixing throughout the ozone
treatment subsystem. When ozone is generated from air, destroying ozone
in exhaust gases is more cost elTective than
Therefore, it is important when using ozone for recirculating the gases through the air preparation
primary disinfection to monitor for dissolved ozone at system and ozone generator. When ozone is generated
a minimum of two points. In the event that two ozone from pure oxygen, destroying the ozone and
contact chambers are utilized, the dissolved residual discharging the excess oxygen can be more cost
ozone can be monitored at the outlets of the two elTective than destroying the excess ozone, and drying
chambers. The average of these two numbers can be and recycling the excess oxygen. A number of "once·
used as the "CO for calculation ofCT values. through" oxygen feed systems have been installed
throughout the world since 1980 to generate ozone.
The largest of these is at the 26.28 m3 /sec (600-MGD)
Absolute measurement of ozone contact time (''1'") is Los Angeles plant, which has been operating in this
not simple, because the objective of adding ozone to manner since 1987.
water invoives maximizing the mixing of a partially
soluble gas with the liquid. The more complete the
contacting, the shorter the actual residence time for Other System Design Considerations
the water in the contacting chamber. As a result, the Ozonation system components that directly contact
more completely the gas and liquid media are mixed, ozone-containing gas should be constructed of
the less the hydraulic residence time can be used as corrosion-resistant material, such as reinforced
concrete for the contactors, stainless steel for the
Disinfection and Disinfection By-Products 79

Contact
Chamber

1'~Off'Gas
u~
Water

,/'VV'o.
/VVV
..
Ozonated
Water

.'

Ozone-Rich Flow Meter (Typical)


Air
Valve (TypICal)

Figure 5·7. T~panment ozone contactor with porous dlftu......

piping, and Teflon for gaskets. Piping for dry service discharged from the ozone generator to determine
should be 304-L stainless steel, while piping for wet the ozone production rate.
service should be 316-L stainless steel.
• Power input monitor for the ozone generator.

Proper monitors and control1ers should be supplied


with the ozone system, including: 5.4.2.3 Costs of Ozonatlon Systems
The discussion of ozone system costs is divided into
• Gas pressure and temperature monitors at key four sections covering equipment, instal1ation,
points in the air preparation system. Simple housing, and operating and maintenance costs.
pressure gauges and mercury thermometers are
adequate.

• Continuous monitors/control1ers for the dew Equipment Costs


.point to determine the moisture content of the Ozonation equipment to be purchased includes: air
dried feed gas to the ozone generator. The preparation equipment (drying and cooling), an ozone
monitors should sound an alarm and shut down generator, an ozone contactor, an ozone destruction
the generator when high dew points are unit, and instrumentation and controls. Because of
indicated. Equipment to calibrate the dew point the many differences in air pretreatment methods,
monitor should be provided as well. ozone contacting, contactor exhaust gas handling,
monitoring, and other operational parameters,
e InleVdischarge temperature monitors for the equipment costs presented in this section are general
ozone generator coolant media (water and/or oil, guidelines only.
or air), and a means of determining whether
coolant is actual1y flowing through the generator. For generating large quantities ofozone, 45.36
An automatic system shutdown should be kg/day (l00 Ill/day) and higher, air preparation,
provided if coolant flow is interrupted or if its ozone generation, and contacting equipment costs
discharge pressure exceeds specified limits. run approximately $2,866/kg ($l,300llb) ofozone
generation capacity per day. This figure does not
• Flow rate, temperature, and pressure monitors, include ozone destruction, instrumentation, control,
and an ozone concentration monitor for the gas building, and installation costs.
80 Upgrading EXisting or Designing New Drinking Water Treatment Facilities

For smaller quantities of ozone, costs are higher per Ozone Supplier A provides four monitors with the
kilogram, but vary significantly from site to site. For system: dew point in the air preparation unit, ozone
plants serving less than 10,000 persons per day, 1.4 to output of the generator, ozone in the plant ambient
9.5 kg (3 to 21 lb/day), all items can be assembled into air (in case of leaks), and dissolved ozone residual in
a single skid mounted unit. the water. All are optional (but recommended) for
optimum performance and minimal labor and
Small ozonation systems can use diffuser contactors, downtime.
which are generally constructed of polyvinyl chloride
(PVC) pipe standing on end, or fiberglass reinforced Ozone Supplier B provided estimates for two types of
plastic (FRP) tanks. Tables 5-12 and 5-13 list air preparation equipment. One type operates at high
pressures (5.6 to 8.4 kg/cm 2 [80 to 120 psigJ), the
Table 5-12. Coeta of Ozonl1lon Equipment lor Small Water other at low pressures (0.56 to 0.84 kg/cm 2 [8 to 12
Supply Syatema: SupptlerA ($1982)
psigJ). The high-pressure units have lower capital
Size of Treatment Plant costs, but require more energy for their operation.
500.000 GPO 350.000 GPO 180.000 GPO Supplier B offers two types of devices to monitor
ozone output from the generator. The automatic, in-
Maximum 5 3 5 3 3 line continuous reading monitor costs $4,000; the
dosage of nonautomatic monitor, which requires wet chemistry
ozone calculations to determine ozone output, costs $2,000.
(mgIL) at
peak flow
Daily ozone 21 14 14 7 7 5 Installation Costs
reqUifement
(Ib) Costs to install ozonation equipment include labor
and material for piping and electrical wiring. Piping
Contact 6 6 5 5 4 4
chamber can be extensive - transporting water to and from the
diameter" ozone generators (if they are water-cooled) and the
(ft) contactor, transporting ozone-containing air to the
Equipment contactor chamber, and transporting contactor off-
Costs gases t~ and from the ozone destruction unit.
Air pt'epara- $31.500 $25.000 $25.000 $22.000 $22.000 $19.500
tion and The ozonation equipment suppliers estimate that for
ozone gen· units producing up to 13.6 kg/day (30 lb/day) of ozone,
eration unit
installation costs average from $9,705 to $16,175 for
Contact 11.500 11.500 10.200 10.200 9.900 9.900 Supplier A and $12,750 to $21,250 for Supplier B.
chamber
WIth
dittusers
Housing Costs
Monitonng 15.000 15.000 15.000 15.000 15.000 15.000
Insttumen- Installation of the power supply, air preparation,
tation b ozone generation, and turbine contacting operations
Ozone 6.700 5.000 5.000 4.200 4.200 4.200 require an area of approximately 3 by 5.1 m (10 by 17
OestruclJOn (10 elm) (7 ctm) (7 ctm) (3 ctm) (3 elm) (3 elm) n>. Diffuser contacting units are quite large and high
Unit (5.4 m [18 £til, and are typically installed outside
Totsl $64.700 $56.500 $55.200 $51.400 $51,100 548.600 existing buildings or in the basement of buildings
Equipment constructed for the other ozonation equipment. A
Costs Butler building with the above dimensions can be
POW19r 13.3 10.1 10.1 5.0 5.0 3.65 constructed for about $6,000.
Require-
ment(kWh)

" 14 It. high. four compa/1/llents. four dittusers, Derakane fiberglass Operating and Maintenance Costs
nelnforced plasl\C. Operating costs for ozonation systems vary
b Includes monitors for ozone In generator product, ozone ,n ambl9nt
plant alr, ozone diSSOlved in wat94'. and dew pOlnt monitor In alr
depending on:
preparation unit
1 Ib 0.4536 kg; • Oxygen use or air preparation method - high or
1 ft 0.3048 meters; low pressure, or subatmospheric pressure
1 GPO 2 0.003785 mJlday.
Source: U.S. EPA (1983). desiccant systems with or without an air chiller

equipment cost estimates obtained from two • Generator cooling method - air or water cooled. In
ozonation system suppliers in 1982 for small water northern climates, water at the plant is generally
supply systems. Equipment costs are higher at cold enough to be used as a coolant all year.
higher dosages for a given flow rate. Southern climates must refrigerate cooling water
Disinfection and Disinfection By-Products 81

Table 5-13_ Cosla of Ozonatlon Equipment for Small Water Supply System&: Supplier B (S1982)
Size of Treatment Plant
0.1 MGD 0.2 MGD 0.3 MGD 0.4 MGD 0.5 MGD
MaxJmum ozone 3 3 3 3 3
dosage (mgil). at
peak flow
Daily ozone 3 6 7 12 14
reQUIrement (Ib/day)

Pressure Pressure Pr91!Sure Pre~sure Pre$-c;llrA


EqUIpment Costs
Low'> !:fuI!!c Low b !:fuI!!c Low b !:fuI!!c Low'> !:fuI!!c Low" t!.!ll!lc
Air preparation and $17,500 $33,200 $30,200 $38.500 $35,000 $43.000 $40.000 $49.800 $46.800
ozone generator'
Power reQulfements 10.5 20 10.5 13.5 10.5 13.5 10,5 13,5 10.5 13.5
(kWhilb of 03
generated)
Ozone contractor With $8.500 $12.000 $16.000 $21,000 $29.000
diffusers
Ozone generatIOn 4,000 4,000 4,000 4.000 4.000
monilor
Chamb8f exhausl 2.200 2,200 2,200 2.200 2.200
monrtor
Dew point monitor 3.500 3,500 3,500 3.500 3.500
Total Equipment $35.700 $52,900 $49,900 $62.200 $59.200 $71,700 568,700 $86,500 $83.500
Costs
• Includes <IIr preparauon, ozone generation. ozone destruction. and system controls.
" Alf preparation unit Includes <IIr filters or separators. compressor delivering air al 8 to 12 ps'g to a refngerauve cooler and a dual tower
deSICCant dryer.
c Same as low-pressure air preparation system, except compressor delivers <IIr at 80 to 120 psig. I-'iqh-pressure system takes less space and
reQUlfes less m<llntenance. but reQUIres more energy,
d $4,000 Instrument IS an automauc, continuous reooing In-line monitor. A substrtute IS a $2.000 Instrument that is not automauc and utilizes
wet chemistry.
1 pound ~ 0.4536 kilograms; 1 MGD ~ 0.044 m3/sec.
Source: U.S. EPA, 1983.

most of the year. Medium frequency generators Labor requirements, other than for periodic
require increased cooling. generator cleaning, include annual maintenance of
the contacting basins and day-to-day operation of the
• Contacting method - diffuser contactors do not generating equipment (average 0.5 hour/day).
require electrical energy as do the more compact
turbine diffusers. Table 5-14 summarizes operating and maintenance
costs for equipment from ozone Suppliers A and B.
• Ozone dosage required This table also includes building heating costs
(assumed to be the same up to a 0,22 m3/sec [0.5-
• Contactor exhaust gas handling - thermal, MGDI plant). There are no chemical costs related to
catalytic, or GAC destruction systems. ozone generation, except for periodic changing of
desiccant in air preparation systems (normally after
5 years of use).
Maintenance costs include periodic cleaning, repair,
and replacement of equipment parts. For example, 5.4.3 Chlorine Dioxide
air preparation systems contain air prefilters that
must be replaced frequently, and tube-type ozone Chlorine dioxide is an unstable gas, explosive in air
generators normally are shut down for annual tube in concentrations above 10 percent by volume
cleaning and other general maintenance. Tube (corresponding to solution concentrations of about 12
cleaning can require several days of labor, depending giLl. Because of its instability, chlorine dioxide is
on the number and size of ozone generators in the always generated on site, in aqueous solution, and
system. Tubes, which can be broken during cleaning used shortly after its preparation. Solutions up to 5
or deteriorate after years of operation at high giL can be stored for up to 7 days. As long as care is
voltages (or more rapidly if the air is improperly taken to keep chlorine diolCide in solution and long-
treated), must be replaced periodically. time storage of solution is avoided, there is no
82 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Tlble 5-14. Operating and Maintenance Cost. for Small Ozone System.- ($1982)
ElectrICal Energy (kWhlyr) Energy Maintenance Labor Labor Total Cost
Water Flow Rate (mgdl BUilding Process Total Costs b Material (hr/yr) Cost< ($lyr)
Supplier A
0.18 6.570 6.661 13.231 $926 $120 250 $2.500 $3.546
0.35 6.570 12.775 19.345 1.354 200 250 2.500 4.054
0.50 6.570 51.611 58.181 4.073 300 250 2.500 6,873
Supplier B (High.Pressure Air Preparation)
0,10 6.570 21.900 28.470 $1.993 $120 250 $2.500 $4.613
0,20 6,570 29,565 36.135 2.529 120 250 2.500 5.149
0.30 6.570 34.493 41.063 2.874 200 250 2.500 5.574
0,40 6.570 59.130 65.700 4.599 250 250 2.500 7.349
050 6.570 68.985 75.555 5.289 300 250 2.500 8.089
- Assumes 3 mgtL ozone dosage.
b Assumes 0.07/kWh,
c Assumes 10lhour,
1 mgd • 0,044 mJ/sec
Source: U,S, EPA (1983).

explosion hazard. Chlorine dioxide is readily soluble react with ammonia or form chlorinated organic
in water, but decomposes in sunlight. materials. Depending on the types and quantities of
organic materials present, periodic tastes and odors
Chlorine dioxide is a more powerful biocide than can be produced by chlorine dioxide (Routt, 1989),
chlorine but has a lower oxidation potential. When
prepared in the absence of excess free chlorine, it does Section 5.4.3.1 describes the generation of chlorine
not produce THMs or other chlorinated organic by- dioxide. Section 5.4.3.2 explains how excess chlorine
products of concern. Additionally, oxidation of dioxide and chlorite ion can be removed from
bromide ion to hypobromous acid does not occur solution,and Section 5.4.3.3 discusses system design
except when free chlorine is present. However, some considerations, Lastly, Section 5.4.3.4 discusses
chlorine dioxide generation methods do create system costs.
conditions of excess free chlorine in which by-
products are produced if their precursors are present.
Excess free chlorine can also slowly produce chlorite
and chlorate ions by disproportionation under pH 5.4.3.1 Process Description
conditions below 2 or above 11. (These conditions are For drinking water treatment, chlorine dioxide is
not usually found in treating drinking water.) generated from solutions of sodium chlorite
(NaCl02), which is usually purchased as a 25 percent
Health effects studies have shown hematological aqueous solution or as a solid (80 percent sodium
effects in laboratory animals as a result of exposures chlorite). Historically. chlorine dioxide has been
to chlorate and chlorite ions (Abdel-Rahman et aI., produced by treating sodium chlorite with either
19801. For this reason, EPA currently advises that chlorine gas, sodium hypochlorite solution and
the total concentration of chlorine dioxide and its mineral acid, or mineral acid alone. In all three cases,
decomposition products (chlorite and chlorate ions) the appropriate aqueous solutions of reactants are
be maintained at or below 1 mg/L. which is metered into and mixed in a chlorine dioxide reactor,
equivalent to a maximum applied dosage of 1.2 to 1.4 which is a cylinder containing now distributing
mg/L. packings, such as Raschig rings, glass beads, or
hollow glass cylinders. Residence time of the
Gaseous chlorine dioxide has a strong, disagreeable solutions in a properly sized reactor is only a few
odor. similar to that of chlorine gas, and is toxic to seconds. The resulting yellow solution is pumped
humans when inhaled. Its odor is detectable above directly into the water to be treated.
concentrations of 0.1 ppm.

Chlorine dioxide can be used to preoxidize phenolic In this manner, chlorine dioxide solutions are
compounds and separate iron and manganese from generated as the material is required and used
some organic complexes that are stable to immediately. Reactor operations are automated with
chlorination. Distribution system residuals of appropriate metering and instrumentation that
dissolved chlorine dioxide can be longer-lasting than controls the addition of chlorine dioxide according to
those of chlorine because chlorine dioxide does not the flow rate of the water being treated.
Disinfection and Disinfection By-Products 83

The three historical techniques for generating


chlorine dioxide are discussed in detail by U.S. EPA [CI 202J + H2O
(1983). However, some of these procedures can result
-----> C103- + Cl' + 2W [61
in excess free chlorine being present. Free chlorine
can oxidize chlorine dioxide to form chlorate ions,
which are difficult to remove from solution. and
Consequently, the current recommended approach to
chlorine dioxide generation is to maximize its yield [Cl202J + HOCI
while minimizing the presence offree chlorine (thus -----> C103- + CI + H' [7]
minimizing the formation of chlorate ion).
5100tmaekers et al. (1989) discusses generation Thus, high concentrations of excess chlorite ion favor
techniques to meet these objectives. the second order reactions (Equations 4 and 5), and
chlorine dioxide is formed. At low concentrations,
For water disinfection, chlorine dioxide can be the second order disproponionation process becomes
generated using several reaction schemes, such as unimportant, and reactions 6 and 7 produce chlorate
the reaction of aqueous hypochlorous acid with ion rather than chlorine dioxide. The reasons for the
dissolved chlorite ion: production of chlorate ion are related to the presence
of high concentrations of free chlorine and the rapid
formation of the Cl202 intermediate, which, in turn,
2NaC102 + HOCI [11 reacts with the excess hypochlorous acid to form the
-----> NaCI + NaOH + 2C102 unwanted chlorate ion.

The stoichiometry of the undesirable reactions which


Chlorine dioxide also can be generated by the forms chlorate ion is:
reaction ofsolid sodium chlorite in solution with
mineral acid, with chlorine or with hypochlorous C1 202' + HOCl
acid. The reaction for chlorine and/or hypochlorous -----> CI03- + Cl + W [81
acid with chlorite ion is:
Cl02- + Cl2 + H2O
2C102- + CI2(g) [2a] ..... > Cl03- + 2Cl + 2W [9]
-----> 2C102(g) + 2Cl'
Therefore, the most effective way to minimize
chlorate ion formation is to avoid conditions that
2C102- + HOCI [2bl result in low reaction rates (e.g., high pH values
and/or low initial reactant concentrations, and the
-----> 2C102(g) + Cl + (OH)' presence offree hypochlorous acid). Clearly, the
reaction forming chlorate ion (Equation 6) will be
These reactions involve the formation of the more troublesome in dilute solutions. On the other
unsymmetrical intermediate, C1202: hand, whenever treatment by chlorine dioxide (which
forms chlorite ion in the process) is followed by the
addition of free chlorine (HOC) with a pH of5 to 8),
[31 the unwanted chlorate ion will also be formed.
At high concentrations of both reactants, the
intermediate is formed very rapidly. Elemental 5.4.3.2 Establishing a Chlorine Dioxide
chlorine formed by Equation [4 J is recycled by means Residual
of Equation [3]. Thus, primarily chlorine dioxide is Laboratory studies have shown that about 70 percent
produced as a result: of the chlorine dioxide added to drinking water is
converted to chlorite ion (Singer, 1986). Therefore,
1.2 to 1.4 mglL chlorine dioxide is the maximum
(4) practical dosage to meet the currently recommended
maximum total oxidant residual of 1 mg/L. However,
or Slootmaekers et al. (1989) reports that nearly all of
[51 the chlorine dioxide ion added as a primary
oxidant/disinfectant is converted to chlorite ion.
On the other hand, at low initial reactant Because of differences in the nature of water
concentrations, or in the presence of excess constituents that exert demand for chlorine dioxide,
hypochlorous acid, primarily chlorate ion is formed, this ratio should be individually determined for each
due to the following reactions: water supply.
84 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Recent studies (Pinsky and Coletti, 1988) have shown increase in the rate of the competing sulfite
that when water containing 1 to 5 mg/L of chlorine ion/oxygen reaction. Thus, less acid solutions (e.g., in
dioxide is passed through a GAC column (with 6- the pH 5.5 to 6 range) favor the rapid reduction of
minute EBCT), the residual chlorine dioxide and chlorite ion and minimize the loss of sulfite ion from
chlorite ions are chemically reduced to innocuous the competing sulfite ion/oxygen reaction.
chloride ions. This implies that as long as GAC with
sufficient EBCT is used after primary disinfection or Thus, Slootmaekers et al. (1989) have established
oxidation with chlorine dioxide, the currently that the sulfur dioxide/sulfite ion - chlorite ion
recommended maximum levels for chlorine dioxide, reaction is greater than 95 percent efTective.
chlorite, and chlorate are achievable regardless of the Furthermore, with a 10-fold excess of sulfur
dosages ofchlorine dioxide used. dioxide/sulfite ion and with chlorite ion at the 0.5 to 7
mg/L level, the removal of chlorite ion is complete in
More recently, Howe et al. (1989) have confirmed the less than one minute at pH 5 and below, and
reduction of chlorine dioxide and chlorite ion on completed in 15 minutes or less at pH 6.5. This
passage through GAC. GAC does not, however, means that sulfur dioxide/sulfite ion can be used to
remove chlorate ion. These researchers showed that reduce the level of chlorite ion by-product in drinking
breakthrough of chlorite ion occurred after water to below the 0.1 mg/L without great difficulty.
approximately 94 and 100 days of passage through
GAC columns having 7.5 and 15 minute EBCT, However, the rate of reductive conversion of chlorine
respectively. Thus, the capacity for GAC removing dioxide and chlorite ion to chloride ion decreases with
chlorite ions in these studies was between 14 and 15 increasing pH as illustrated by the following data:
mgofchlorite ion per gram ofGAC.

These data indicate that if GAC is to be installed for


postfiltration adsorption of organics, chlorine dioxide Time required lor
can be employed as a primary disinfectant Beginning Chlonde Beginning Sulfite 99% removal of
somewhere ahead of the GAC. However, if pH Ion Conc. (mgll) Ion Cone. (mgll) chlorite Ion
postfiltration GAC adsorption is installed only to 5.0 0.5 5.0 0.34 minutes
remove chlorine dioxide and chlorite ion, the costs 5.5 0.5 5.0 4.4 mInutes
will be very high (because of the postfiltration GAC 5.5 1.0 10.0 1.1 mInutes
installation and 90-110 days reacti vation needed for 6.5 1.0 10.0 15.2 mInutes
breakthrough of chlorite ion).
6.5 1.0 20.0 3.8 mlnules
Siootmaekers et al. (1989) have shown that it is 7.5 1.0 10.0 15.6 hours
technologically feasible to reduce chlorine dioxide 7.5 1.0 100.0 9.4 mlnules
and chlorite ions to chloride ion with sulfur dioxide. 8.5 1.0 100.0 3.2 days
Once oxidant demand has been satisfied with
chlorine dioxide, excess sulfur dioxide/sulfite ion can
be added to the treated water to remove residual
chlorine dioxide and chlorite ion. This would be In practice, if all of the chlorite ion is not removed in
followed by the addition offree chlorine to remove the the time allotted, three choices are available to the
excess residual sulfur dioxide/sulfite ion. treatment plant operator:

The stoichiometry of the reaction is consistent with • Increase the amount of sulfur dioxide/sulfite ion
the equation: used in the removal process: doubling the
amount will remove the chlorite ion four times
faster since the rate is second order in the
concentration of sulfur dioxide/sulfite ion.

This corresponds to a stoichiometry of two moles of • Decrease the pH: every decrease of pH by one
S03 2-consumed for every mole ofC102' reduced. unit increases the rate of chlorite removal by at
Over a range of conditions stUdied by Siootmaekers et least a factor of ten.
al. (1989), at room temperature, the stoichiometry
deviated by less than 5% from the value shown in the • Increase the length oftime for removal: the rate
equation. Furthermore, the stoichiometry appears to of removal is logarithmic in time, since the
be independent of temperature in the 5° to 30°C removal of chlorite ion is first order in the
range. concentration of chlorite ion.

However, at pH 8 and above and in the presence of The ability to remove chlorite ion by-product allows
air, the overall stoichiometry appears to deviate from considerably higher chlorine dioxide concentrations
that shown by the above equation, probably due to an to be used (easily up to 6 to 8 mg/L), eliminating the
Disinfection and Disinfection By-Products 85

need for a combination of chlorine dioxide and free the flow rate of the water and/or the secondary
chlorine, the most common sources of chlorate ion. disinfectant demand.

It is important to note that chlorate ion is not Sodium chlorite is available either as a solid (80
chemically reduced by sulfur dioxide/sulfite ion percent active sodium chlorite) in 90.72-kilogram
under the conditions studied by Slootmaekers et al. (200-pound) drums or as a solution (25'percent active
(1989). The calculated half-life for the removal of sodium chlorite, 31.25 percent solids in 208.2-liter
mglL levels of chlorate ion with excess sulfur [55 galion] drums). Ifnot used directly from the
dioxide/sulfite ion is in excess of many months. drum, sodium chlorite solution is stored in
polyethylene or fiberglass tanks and transferred by
Chlorate ion in the finished water can arise from one means of PVC, rubber, or Tygon tubing systems.
of two possible sources: Diaphragm pumps with PVC components are used to
transfer the sodium chlorite solution. Provision must
• Improperly tuned older chlorine dioxide be made for immediate washdown of any chemical
genera tors. spills; this precaution is generic to all chlorine
dioxide generating systems.
• As a by-product during the chlorine dioxide
oxidation/disinfection step, especially when free
chlorine is used concomitantly. The CIFEC System for Generating Chlorine
Dioxide
The CIFEC System was developed in France and is
currently used by several treatment plants in the
5.4.3.3 Chlorine Dioxide System Design United States. Figure 5-10 is a schematic diagram of
Considerations this system, which produces chlorine dioxide from
Several types of generation equipment are available. chlorine gas. This system produces high yields of
Each supplier provides recipes for preparing and chlorine dioxide under conditions of minimal excess
metering the solutions to the reactor to produce a free chlorine.
known and constant chlorine dioxide concentration.
For smaller systems treating less than 0.022 m3/sec In the CIFEC system, chlorine gas is injected into
(0.5 MGDl, the chlorine dioxide dosages required are continuously circulating water, referred to as an
very low. In these cases, the generating units may be "enrichment loop." This method produces much
operated intermittently, coliecting chlorine dioxide higher concentrations of dissolved chlorine
solution in an enclosed holding tank from which a (hypochlorous acid) than those achieved in a system
metered flow can be delivered later. Intermittent using a single injection point. The hypochlorous acid
operation is recommended over continuous operation solution (pH below 4) is pumped into the reactor
because the mixing of reactant solutions is less along with a sodium chlorite solution. Since the pH
efficient in the reactor at consistently low flow rates; of the hypochlorous acid solution is below 4, more
thus, the conversion of chlorite ion to chlorine dioxide chlorite ions are converted to chlorine dioxide than in
will be less efficient. Selected generation systems are other chlorine gas systems that operate in higher pH
briefly discussed below. ranges. In addition, chlorine dioxide is produced with
little free chlorine.
Gaseous Chlorine Systems
The most common chlorine dioxide generation Rio Unda Chlorine Dioxide Generator
process uses chlorine gas and sodium chlorite (U.S. Rio Linda Chemical Co., Inc. manufactures a unique
EPA, 1983). This process is convenient when chlorine dioxide generator. In this system, shown
chlorine gas was previously used for primary schematically in Figure 5-11, chlorine dioxide is
disinfection at the plant. Figure 5-8 is a schematic generated by the addition of chlorine gas to a sodium
diagram of such a system. chlorite solution, as in other systems. The novel
aspect of this generator is that chlorine gas is mixed
with concentrated sodium chlorite solution just
Sodium Hypochlorite and Minerai Acid before the reactor. The vacuum action from the water
Systems flowing through an eductor into the reactor brings
A manual feed system that produces chlorine dioxide the two solutions together. Since there is no need for
using sodium hypochlorite solution with sodium a pump, the system is more compact.
chlorite and strong mineral acid is illustrated in
Figure 5-9.This process is well suited to most small This system can be operated manually or
treatment systems. The operator can adjust solution automatically. Additionally, the system can be
strength for each reactant so that feed pumps of equal designed to mix in solutions of hydrochloric acid,
capacities can be used. The chlorine dioxide hypochlorite and sodium chlorite just before the
production and addition rates are paced according to chlorine dioxide reactor. Generator efficiencies with
86 Upgrading Existing or Designing New Drinking Water Treatment Facilities

excess chlorine are claimed to approach greater than should be noted that almost all older chlorine dioxide
98 percent conversion of the sodium chlorite into generators now in use in water treatment plants
chlorine dioxide, with minimal production of chlorate throughout the United States continue to produce
ion at an emuent pH between 6.5 and 7. Production some chlorate ion, unless they are very carefully
rates of these generators can range from a few pounds tuned and properly monitored. A well·tuned
up to 2,721.6 kilograms (6,000 pounds) per day. generator may produce as little as 1 to 2 percent
chlorate ion (Slootmaekers et aI., 1989).

Summary of Commercial Chlorine DloxJde 5.4.3.4 Costs of Chlorine Dioxide Disinfection


Generators The costs of chlorine dioxide systems are presented in
In many older chlorine dioxide generators, three categories: equipment, operating and
hydrochloric acid is fed into the chlorine solution maintenance, and chemical costs.
before reaction with the sodium chlorite. The acid
shifts the chlorine solution equilibria in favor of
molecular chlorine (Slootmaekers et aI., 1989). (The Equipment Costs
hypochlorous acid dissociation and the chlorine Quotes from two suppliers of chlorine dioxide
hydrolysis equilibrium are shown in the equations generation equipment were obtained in 1982 (see
below.) Table 5-15). Supplier A's recirculating loop system
(CIFEC) is the highest priced unit at $34,000. It
operates with a special recirculating pump designed
HOCI .•••• > (OCll' + W to handle hypochlorous acid below pH 4, plus a
Cl2 + H20 --._> HOCI + H+ + HCl' sodium chlorite solution pump and all necessary
instrumentation to allow automatic operation. There
The acid must be carefully controlled so that the pH are shutdown provisions in the event of interruptions
of the chlorine dioxide solution is maintained in water flow.
between 2 and 3. Higher or lower pH values result in
decreased yields. Yields of more than 90 percent The next lowest in price is the system from Supplier
have been reported from a properly pH-adjusted B that generates chlorine dioxide from 33 percent
system (Aieta & Berg, 1986), with approximately 7 hydrochloric acid, 12 percent sodium hypochlorite,
percent excess chlorine remaining in solution (where and 25 percent sodium chlorite solution. This wall-
excess chlorine is defined as the amount ofunreacted mounted unit costs $25,000 (installed), and includes
chlorine remaining in the chlorine dioxide generator three solution pumps, a water flow rate detector, and
effiuent). switches to shut down the unit if the water flow stops.
For chlorine dioxide volumes sufficient to treat flows
High yields of chlorine dioxide, with low levels of free in communities with populations of 5,000 and 2,500,
chlorine in solution, can be produced using a chlorine this unit is capable of continuous operation, with no
solution with a concentration greater than 4 g/L. loss in conversion efficiency of chlorite ion to chlorine
This chlorine concentration is near the upper dioxide. However, to supply the needs of systems
operating limit ofcommercial chlorine ejectors. serving as few as 25 persons, the unit must be
Since these ejectors operate at constant water flow operated intermittently, with chlorine dioxide
rates, the yield of this method of generation is solution being stored in a holding tank for later
dependent upon the production rate, with lower metering into the .water.
production rates resulting in lower yields. This type
of generator normally operates on an intermittent Supplier C provides two types of chlorine dioxide
basis to maintain a high yield when less than generators for small water supply systems. One uses
maximum production capacity is required. acid/sodium chlorite; the other uses chlorine gas and
sodium chlorite. These units cost $3,600 wall-
The most recent development in chlorine dioxide mounted, and $4,320 for a floor· mounted cabinet. The
generator technology is a system that uses the single size unit offered by this supplier is designed to
reaction of chlorine with a concentrated sodium generate up to 63.5 kg/day (140 Ib/dayl. In order to
chlorite solution under vacuum (Aieta & Berg, 1986; produce 3.63 kg/day (SIb/day) or less, a small water
Aieta & Roberts, 1986l. The chlorine dioxide is utility must install a holding tank and operate the
removed from the reaction chamber by a gas ejector, generator intermittently.
which is very similar to the common chlorine gas
vacuum feed system. This technique of generation The chlorine gas/sodium chlorite generator of
produces chlorine dioxide solutions with yields in Supplier C requires a gas chlorinator to feed chlorine
excess of95 percent. The chlorine dioxide gas. Therefore, in new plants considering this type of
concentration is 200 to 1,000 mgIL and contains less equipment, the cost of a chlorinator must be added to
than 5 percent excess chlorine with minimum the cost of the chlorine dioxide generator. In existing
formation ofchlorate ion. However, in this context, it plants currently using gas chlorination, the
Disinfection and Disinfection By-Products 87

G·Gage

r--
[C2]
4;~n~A ---------~--_---j::
Chlonne Oioxide
Generating To_

Raschig ()
Rings ()

Pomt of

W.P.R.V.
+ Apphca~on

Chlorine
Onfice
POSitioner
~
B
Injector Water Supply

Figure 5·8. Schematic diagram of an automatic flHld, automatic flow· proportional chlortne dioxide syatem: Generation from
chlortne and sodium chlorite.
Source: C8j)ltal COntrols Co., Inc.

chlorinator already is in place, and therefore would Chemical Costs


not represent additional equipment cost.
At a production rate of only 3.6 kg/day (8 lbfday)
(maximum for a 0.044 m 3 /sec II-MGD] water
The purchaser of chlorine dioxide generating systems treatment plant at a 1 mg/L applied chlorine dioxide
should consider the inclusion of additional equipment dose), chemical costs are not as significant as
to allow automatic operation, such as microprocessor. pumping costs. Chemical costs in 1982 were as
controlled electronic val ve systems, ratio- follows:
proportioning and flow-paced systems, and remote
start and feed capabilities. In addition, alarm Gaseous chlorine $1.04Jkg ($0.47flb)
facilities for operating parameters such as low
feed/production ratios, low yields, shut-olTs for Sodium chlorite $3.42 to $3.64 ($1.55 to
chlorine dioxide in air, low reactant feed for chlorine $1. 65f1b) (l00 percent
gas or sodium chlorite solution, high or low vacuum, solids)
and low water flows or pressures are advisable
options. Hydrochloric acid $0.22fkg ($O.lOflb)

Sodium hypochlorite $0.25/L ($0.93/gal) (I5


percent solution)
Operating and Maintenance Costs
In general, operating and maintenance costs for 5.4.4 Ultraviolet Radiation
generating chlorine dioxide are independent of the UV radiation is an elTective bactericide and virucide,
quantities generated. Table 5-16 summarizes these but an inelTective cysticide. Consequently, it is
costs on an annual basis. Maintenance material costs recommended as a primary disinfectant only for
are for minor equipment repair only. Labor required ground waters not directly influenced by surface
for preparation of solutions and periodic maintenance waters (where there is no risk of Giardia cyst
of the equipment is estimated to be 1 hour/day. or 365 contamination). UV radiation (254 nm wavelength)
hours/year. Total annual operating and maintenance penetrates the cell wall and is absorbed by the
costs of$4,1241year are estimated. cellular nucleic acids. Radiation absorption prevents
88 Upgrading Existing or Designing New Drinking Water Treatment Facilities

CI0 2

/~
Chlonne Dioxide
Generaong Tower Pointo!
Application
0

Diaphragm Pump

Sodium Chlorite HypochlOrite Sulfuric Acid

Figure 5·9. Manual feed equipment arrangement for generating chlorine dioxide from sodium hypochlorite solution and
minerai acid.
Source: U.S. EPA (1983).

replication, thus killing the celL Since UV radiation emission ofUV radiation. This radiation then
is not a chemical agent, it produces no toxic residuaL destroys the cell's genetic material and the cell dies.

Major advantages of UV radiation are its simplicity, The effectiveness of radiation is a direct function of
lack of impact on the environment and aquatic life, the energy dose absorbed by the organism, measured
and minimal space requirements. In addition, as the product of the lamp's intensity and the time of
required contact times are seconds rather than exposure. Intensity is the rate at which photons are
minutes. The equipment is simple to operate and delivered to the target. The intensity in a reactor is
maintain if the apparatus is cleaned properly on a governed not only by the power of the lamp, but also
regular basis. by the placement of the lamps relati ve to the water,
and by the presence of energy sinks that consume UV
Section 5.4.4.1 describes the process of disinfection radiation. Water with suspended solids, color,
with UV radiation. System design considerations, turbidity, and soluble organic matter can react with
including lamp designs, are provided in Section or absorb the UV radiation, reducing the disinfection
5.4.4 2. Operating and maintenance considerations performance. Therefore, water with high
are discussed in Section 5.4.4.3 and costs of UV concentrations of these substances may recei ve
systems are described in Section 5.4.4.4. inadequate disinfection.

The radiation dose absorbed by the water is the


5.4.4.1 Process Description water's UV demand, which is analogous to chlorine
UV radiation disinfection uses a special lamp to demand and is quantified as the absorption of UV
transfer electromagnetic energy to the target energy (wavelength of253.7 nm) in a given depth of
organism cells. The most efficient and widely used water. The measurement is most commonly
device is the mercury arc lamp. It is popular because expressed by the UV absorbance coefficient alpha:
approximately 85 percent of its energy output is of
the 253.7 nm wavelength, within the optimum alpha = 2.3 absorbance units(a.u.)/cm
germicidal range of 250 to 270 nm. The lamps are
long thin tubes. When an electric arc is struck In addition to intensity and UV demand of the water,
through mercury vapor, the energy discharge the exposure time also affects the energy dosage that
generated by the mercury excitation results in the the target organisms absorb. Exposure time is
Disinfection and Disinfection By-Products 89

Vacuum Line of ChlOOne

® CI0
2
Reactor
@
)iJ'1
r Control EqUipment
®
Enrichment Loop

Sodium
Chlonte
Tank
~
Chionne Cylinder

Recirculating Pump Check Valve ®


Sodium Chlorite Metering Pump
Flow Meter --"""-..-J.....

Make-Up
l!::::==<~ Water Supply

Figure 5-10. Schemlltic of CIFEC chlorine dioxide generating IYltem.


Source: CIFEC, Inc., Pans. France.

Chlonne Dioxide Solution


10 Process

Ejector

Electrical
BallVaJve
Chlorine DioXide
Water Supply
Reaction Column

I
... _------------ ... I
I
I
Sodium Chlonne
Chlorite Manual
Rotameter
Rotameter Control
System

Sodium Chlorine
Chlorite Gas
Supply Supply

F~ure 5-11. Sche....tlc of menUllI chlorine dioxide Iystern.


Source: Rio Linda Chemical Co.. Sacramento, California
90 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Table 5·t5. Costa of Chlortne Dioxide Generating EqulptMnt $1982


Productloo
Capacity Space UOIt CoS!
VendOr System Type (Ib/day) Requirements" Reactants ($)
Supplier A Reclrculabng loop 1-10 2 x 3 x 6 leet Chlorine gas, sodium chlorite solulion 34.000
(CIFEC)
Supplier B Wall·mounted uOlt 4 3.5 x 4 x 1.5 Hydrochloric acid. sodium hypochlonte. and 25,000
feet sodium chlonte solubons
Supplier C FIoot·mounted uOlt 14·140 4 x 3 x 1.5 Chlorine gas and sodium chlorite solulion 4,300·
6.5 inches
Supplier C Wall·mounted unit 14-140 37.5 x 27 x Chlorine gas and sodium chlonte solubon 3.600·
6.5 Inches
Supplier C FIoot·mounted unit 14-140 4 x 3 x 1.5 Hydrochloric acid and sodium chlonte solullon 4,300
feet
Supplier C Wall·mounted unit 14-140 37.5 x 27 x Hydrochlonc acid and sodium chlonte solullon 3.600
6.5 inches
• Space requirements tor SOIulJOn tanks are not included.
• Costs for a chlorinator are not included in these eSbmates.
1 Ib .' 0.4536 kg; 1 It • 0.3 m; linch • 2.54 em.
Source: U.S. EPA (1983).

radiation system should be designed to provide a 4-


Table 5-18. Operation and Maintenance Cost Summary
lor Chlortne Dioxide Generating and Feed
log reduction.
Syatem.

Annual Costs
Item ReqU"emenls (S)
ElectncaJ Energy 5.4.4.2 UV Disinfection System Design
Melering pumps and 1,240 kWhlyr 86.80 Considerations
mixers The basic design considerations for a UV system are:
BUilding 4 , 100 kWhlyr 287.00
Subtotal 5.340 kWhlyr 373.80" • Satisfying the UV demand of the water
Maintenance MatenaJs 100.00
Labor 365 hlyr 3,650·
• Maximizing the use ofUVenergy delivered by
the lamps
Total Annual O&M COlt 4,t24
" Assumes SO.07/kWh. • Maintaining the conditions that encourage plug
• Assumes $101hour labor cost now
Source: U.S. EPA (1983).

controlled by the residence time of the water in the UV lamps are usually submerged in the water,
reactor. Continually maintaining the required perpendicular or parallel to the water flow.
residence time is not always possible, but the system Submerged lamps are inserted into a quartz sleeve to
design should maximize plug-flow operation. minimize the water's fouling effects. The further the
distance between the water and the lamp, the weaker
If the energy dosage is not sufficient to destroy the the radiation dosage delivered because the energy
target organisms' DN A macromolecules, disinfection dissipates or becomes dilute as the space it occupies
is not effective. Photoenzymatic repair occurs if the increases in volume. The UV demand of other
genetic material is only damaged during irradiation. contaminants in the water also consumes radiation.
This repair mechanism, called photoreactivation,
occurs with exposure to light from the sun or most Specific design parameters to consider are:
incandescent and fluorescent lights (wavelengths
between 300 and 500 nm). Photoreactivation does not 1. Residence Time Distribution (RTD) - This
occur with all bacterial species and is difficult to describes the detention time of the water in the
predict. reactor and should be determined for several flow
conditions.
To prevent photoreactivation, the rule of thumb is to
increase the dosage necessary to meet required 2. Plug Flow - The ability to maintain plug now in
reductions in organism levels (U.S. EPA, 1986b!. For the reactor is influenced by the inlet and exit
example, if the disinfection criteria require a 3-log designs. Disturbances at the inlet and exit planes
reduction of microorganism concentrations, the UV of the lamp battery should be minimized and
Disinfection and Disinfection By-Products 91

necessary changes in the flow direction should be size and the lamp placement that achieves the
made outside the lamp battery. desired performance.

3. Dispersion Number - A key goal is to minimize Water flow rate is another key factor in determining
the dispersion number, d (cm 2/s). As a design system size. Both the hydraulic load to the plant and
goal, d should be between 0.02 and 0.05. This the design of the processes preceding disinfection
number represents a plug-flow reactor with low affect flow. The size of the UV system, however,
to moderate dispersion. This value is attained by should be based on peak flow rates and projected
increasing the product of the velocity (cm/s) of flows for the plant's design year rather than on
and distance traveled (em) by the water as it average flows, which are used to predict operating
flows through the reactor while under direct and maintenance requirements.
exposure to UV radiation. However, extended
lengths and higher velocities cause higher head
losses; therefore, adjusting the dispersion
number may be necessary to meet specific 5.4.4.3 UV System Operating and Maintenance
criteria for both full-scale modules or pilot units. Considerations
Head loss is determined over a wide velocity This chapter discusses operation and maintenance of
range and excludes entrance and exit losses. UV lamps and reactor care. The intensity of radiation
in the reactor depends on the lamp output and the
4. Effective Volume - The inlet and outlet designs reactor cleanliness. Therefore, monitoring lamp
should achieve equivalent water velocities at all intensity and properly maintaining the reactor are
points entering and exiting the lamp battery. essential to reliable system performance.
This maximizes the lamp battery use and
improves cost effectiveness. Stilling walls
(perforated barnes) and weirs in the reactor Operation and Maintenance of UV Lamps
design assist in controlling water velocities. Lamp output is influenced by lamp temperature,
voltage potential across the lamp, and age of the
lamp. Lamp temperature cannot be controlled in
UV Lamp Designs submerged lamp systems. In other systems, however,
Lamps used in UV disinfection systems typically lamp temperatures are controlled by regulating
have arc lengths of approximately 0.8 and 1.5 m (2.5 ambient air temperatures using cooling fans or
and 4.9 ft.) and full lengths of 0.9 and 1.6 m (3 and 5.3 recirculating the heat generated by the lamp ballasts
m, respectively. The arc length describes the active, for warming.
light-emitting portion of the lamp. Lamp diameters
typically are 1.52 and 2.0 em (0.6 and 0.8 in). A sleeve Adjusting the voltage will vary the lamp output.
made offused quartz or another material that is Decreasing the voltage reduces current to the lamp
highly transparent to UV light, such as Vycor, and, therefore, lamp output. Voltage regulators
protects lamps that are submerged. Nonsubmerged improve system efficiency by reducing voltage and
lamps are placed near the wall of the water conduit, "dimming" lamps to conserve power during periods of
which is made of a UV light-translucent material. low UV demand. Lamp intensity cannot be reduced to
levels below 50 percent, however, without causing
the lamps to flicker and eventually turn off.
Factors Affecting the Design of the UV
Disinfection System Factors affecting deterioration of performance and
Initial microorganism density, suspended solids (or aging ofUV lamps include electrode failure; mercury
turbidity), UV demand of the water at the plating or blackening in the lamp's glass tube; and
disinfection point, and water flow rate all affect the tube solarization, which results in reduced energy
size and performance of the UV disinfection system. transmission through the glass. These factors can
reduce lamp output by 40 to 60 percent.
The performance of a UV disinfection unit relates
directly to the initial density of the indicator UV lamps used for disinfection are typically hot
organisms. The higher the initial density, the greater cathode lamps, which deteriorate progressively with
the dosage of radiation required. For this reason, each startup. Life expectancy is determined by the
microorganism density should be continually number oftimes the lamp is started. The lamp life
monitored. Turbidity directly affects the performance cited by most manufacturers is 7,500 hours, based on
of the UV disinfection system as well. Particulates a burning cycle of 8 hours; that is, the lamp will last
suspended in water block the UV radiation, thereby 7,500 hours assuming it is restarted every 8 hours.
protecting bacteria and hindering disinfection. The The average UV output after 7,500 hours is
UV demand of the water affects the radiation estimated to be 70 percent of the lamp output at 100
intensity in the reactor and, thus, affects the system hours.
92 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Monitoring Lamp Intensity In Teflon systems, the lamps are on removable racks
and should be cleaned and monitored in the same
Lamp intensity should be measured to monitor lamp manner as the quartz systems. The Teflon tubes
condition and determine the need for maintenance. should also be cleaned with mild soap and swabbed
The monitoring procedure compares current lamp with alcohol. Each tube should be monitored for
intensity to the intensity of new lamps. The operator transmittance, just as with the quartz sleeves.
first measures the intensity, at a fixed distance, of Monitoring may not be as straightforward because of
three to five new lamps that have burned for about the limited accessibility to the tubes and problems in
100 hours (the first 100 hours is considered a ''hurn- obtaining direct measurements with a UV
in" period). The average ofthese values is the radiometer/detector.
benchmark from which to measure the relative
output ofthe lamps. Each lamp is tagged so that it
can be monitored individual1y. 5.4.4.4 Costs of UV Radiation Systems
Table 5-17 summarizes construction costs developed
A similar procedure is used to monitor the for single and multiple UV lamp disinfecting units
transmittance of a quartz sleeve. First, the intensity ranging in water throughput capacity from 54.5 to
of a single lamp is measured with and without a new, 4,251.3 m 3/day (14,400 to 1,123,200 gaVday). UV
clean quartz sleeve. Similar measurements are taken generating units are quite compact; for example, the
ofthe quartz unit in use and compared to the 4,251.3 m 3/day (1,123,400 gaVdayl unit occupies an
transmittance of the new quartz. Before being tested, area less than 2.2 m 2 (24 ft2). Costs listed in Table 5-
the quartz is cleaned to assure that maximum 17 include UV unit equipment costs, and the related
transmittance is restored. piping, electrical, instal1ation, and equipment
housing costs.

Operating and maintenance costs are shown in Table


5-18 for the same size plants as in Table 5-17. Costs
Operation and Maintenance of the Reactor assume continuous 24 hour/day operation, with only
The most important operating factor for the UV occasional shutdown to clean cel1s and replace worn
reactor is the cleanliness of the surfaces through out UV lamps. These costs include process energy for
which radiation must pass. Surface fouling can result the mercury lamps and building energy for heating,
in inadequate performance, so a strict maintenance lighting, and ventilating.
schedule is recommended.
Maintenance materials costs are related to
An operator determines the need for reactor replacement of the UV lamps, which may be required
cleansing by draining and visually inspecting the about every 8 months if operated continuously. Labor
surfaces. Open reactor systems are easily inspected. requirements include occasional cleaning of the
Systems with sealed vessels are inspected through quartz sleeves surrounding the mercury vapor lamps,
portholes or manways in the reactor shell. Surfaces of and periodic replacement of the weak UV bulbs.
submerged quartz systems become coated with an
inorganic scale, very much like boiler scale. This is a
particular problem in areas with hard water. 5.5 Secondary Disinfectants
Additional1y, the inside surface of the quartz and the Secondary disinfectants provide an essential residual
outer surfaces of the Teflon tubes eventual1y develop that prevents regrowth in the distribution system.
a grimy dust layer, primarily from airborne dirt and Although chlorine is the most widely used secondary
water vapor. disinfectant, chlorine dioxide and monochloramine
are appropriate as well. As secondary disinfectants,
Fouling of the reactor's internal surfaces also is chlorine and chlorine dioxide are handled in the same
indicated by reduced performance and intensity manner as for primary disinfectants. (See Sections
measured by in-line probes. While these provide some 5.4.1. and 5.4.3 for information on the process and
indication offouling, operators must stil1 visual1y equipment). Section 5.5.1. discusses the use of
inspect the surfaces. monochloramine as a secondary disinfectant.

The fouled surfaces of lamps and quartz sleeves are


cleaned manual1y with a mild soap solution and then 5.5.1 Chloramination
swabbed with a rag soaked in isopropyl alcohol. The Chloramine is recommended as a secondary
transmittance of the lamps and sleeves is measured disinfectant because it is ineffective as a virucide,
after cleaning and those that have inadequate and is only marginally effective against Giardia
measurements are replaced. An inventory al10ws the cysts. It is formed from the combination of ammonia
plant operator to trace operation of individual and chlorine (hypochlorite or hypochlorous acid). The
components. Quartz sleeves should last between 4 chemical is generated on site, usual1y by injecting
and 7 years, but this varies by site. ammonia gas or adding an ammonium sulfate
Disinfection and Disinfection By·Products 93

solution to chlorinated water, or by adding chlorine to reactions are at least 99 percent complete within a
water containing ammonia. Ammonia gas can be few minutes.
purchased as an anhydrous liquid in cylinders for
small water treatment systems; ammonium sulfate The National Academy of Sciences (NAS)
can be purchased as a powder in bags. recommends adding ammonia to chlorinated water
rather than adding chlorine to water containing
Section 5.5.1.1 describes the chloramination process. ammonia. The recommended process produces a
Sections 5.5.1.2 through 5.5.1.4 discuss establishing residual offree chlorine above that required to
a chloramine residual, system design considerations, oxidize nitrogen (particularly the organic nitrogen
and chloramination systems costs, respectively. compounds>. Organic nitrogen compounds will
compete successfully with ammonia-nitrogen for
chlorine, forming organic chloramines, which are
5.5.1.1 Process Description weaker disinfectants than monochloramine. Normal
When water, chlorine, and ammonia are combined, field analytical techniques cannot distinguish
three different species of chloramine compounds can between inorganic and organic chloramines. Thus,
be generated: formation of inorganic chloramines in the presence of
organic nitrogen compounds can seriously overstate
NH3 + HOCI .. _--> + NH2Cl the actual capability of the chloramine system to
ammonia hypochlorous mono- provide secondary disinfection.
acid chlor·
amine
5.5.1.3 Chloramination System Design
NH2C1 + HOCI _.._- > H20 + NHCl2 Considerations
dichloramine Ammonia is available as an anhydrous gas (NH 3), a
29 percent aqueous solution (aqua ammonia), or as
ammonium sulfate powder ([NH412S04). Gaseous
NHC12 + HOCI ....- > H20 + NC13 ammonia is supplied in 68-kg (l50·lbl cylinders, aqua
nitrogen ammonia in 208.2-L (55-gal) drums, and ammonium
trichloride sulfate in 45.4-kg (lOO·lb) bags (98 percent pure, 25
percent available ammonia).
The mix of species produced depends on the ratio of
chlorine to ammonia and the pH of the water. In the Ammonia gas is injected into treated water using
pH range of7 to 8 with a chlorine· to-ammonia ratio systems and equipment similar to those used for
(by weight) of3 to I, monochloramine is the principal chlorine gas. Aqua ammonia is handled using
product. At higher chlorine-to-ammonia ratios or at systems similar to those used for sodium
lower pH values (5 to 7), some dichloramine will be hypochlorite. This form of ammonia is basic and has a
formed. If the pH drops below 5, some nitrogen strong odor, but is not corrosive. For ammonium
trichloride (often erroneously called "trichloramine") sulfate powder, a 25 to 30 percent solution is prepared
may be formed. Nitrogen trichloride formation in a plastic or fiberglass container and added to the
should be avoided because it imparts undesirable water by means of a chemical metering pump.
taste and odor to the water. Equipment similar to that used for handling calcium
hypochlorite can be used for this process. Solutions of
Figure 5-12 shows the relative percentages of ammonium sulfate are stable, but acidic, and,
monochloramine and dichloramine produced as the therefore, corrosive to some metals. Materials that
pH changes, for different weight ratios of chlorine to withstand dilute sulfuric acid will also withstand the
ammonia. At a pH value of about 5.7, approximately corrosion effects of dilute ammonium sulfate
equal amounts of mono· and dichloramines are solutions.
present in solution.

Care should be taken not to exceed chlorine-to- 5.5.1.4 Costs for Chloramination
ammonia ratios of 5 to 1. This is the "breakpoint" Generation of chloramines requires the same
curve above which all ammonia is removed, equipment as chlorination (gaseous or aqueous
chloramines are absent, and free residual chlorine is hypochlorination), plus equipment for adding
present. ammonia (gaseous or aqueous>. Costs for chlorination
equipment and for its operation and maintenance
were presented earlier (Section 5.4.1). This section
5.5.1.2 Establishing a Chloramine Residual presents the chemical costs for generating
Establishing a chloramine residual involves a period chloramine from ammonia and chlorine gas.
of mixing the chlorine and ammonia with the water,
followed by a short holding time to allow the In the BaltimorelWashington, D.C., area during
reactions to take place. Usually, chloramine-forming 1983, costs for cylinders of liquid ammonia were
94 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Tlble 5-17. Con8truCllon Coa~ for Ultraviolet Light OI,lnfllCtlOn ($1978)


Plant Capacrty (GPO)
Cost Calegory 14,400 28.800 187.200 374.400 748,800 1.123.200
Excavabon and $60 $60 $60 $60 $80 $110
Sltewot1<
Manufactured 800 1,125 4.485 8.685 17,365 26,050
equipment
Concrete 250 250 250 250 280 300
LabO< 110 170 250 300 400 500
Pipe and valves 60 150 350 450 750 1,000
ElectneaJ and 430 430 430 430 480 480
Instrumentation
Housing 1.500 1,500 1.500 1.500 1,800 2.000
SublOtal 3.210 3,885 7,225 11.675 21.155 30....0
Miscellaneous and 470 560 1,010 1,580 2.830 4.060
connngency
Total $3,680 $4,"'5 $8.335 $13,255 $24,085 $34,500
1 GPO • 0.003785 m3/day
Source: U.S. EPA (1979).

$0.881kg ($0.40/lb), $1.54/kg ($0.70Ilb) for drums of minimum of 0.2 mg/L of secondary disinfectant be
contained ammonia in 28 percent solution, and present as the treated water enters the distribution
$1.12/kg ($O.511Ib) for bags of solid ammonium system. This level does not have to be maintained in
sulfate. the distribution system, as long as heterotrophic
plate counts remain below 500/mL. Consequently,
Chemical costs were derived for a 9.46 m3/day (2,500- the costs for chloramination can be lower than the
gaVday) water treatment plant using a maximum costs presented here.
dose of2.5 mg/L monochloramine (the maximum
level currently recommended by EPA). These costs
are presented in Table 5-19. Chemical costs for both
chlorine gas and the ammonia-based chemical range
from $14.75 to $24.55 annually, depending on which
ammonia source chemical is used. Anhydrous
ammonia is the least expensive, while ammonia
sulfate is the most expensive.

The direction of EPA's regulatory initiative under


the SOWA Amendments of 1986 is to reduce the
levels of secondary disinfectants in treated water. For
example, the newly promulgated Surface Water
Treatment Rule (EPA. 1989b) requires that a

Table 5·18. Operatl~ and Mllntenlnce Coati for Ultraviolet Light OI.... fIICtton (Slt71)
Water Flow Electnca! Energy (kWh/yf) Energy M&Ifltenance LabO< Total Cost
Rate (GPO) BuildinlJ Process Total Costs- Matenal ($iyr) labO< (hrtyrl Cost" ($iyr)
14,400 10,260 "'0 10.700 $749 $100 24 $240 $1,089
28,800 10,260 800 11.140 780 140 24 240 1.160
187,200 10,260 5.260 15,520 1,086 600 24 240 1.926
374,400 10,260 10,510 20,770 1.454 1,120 30 300 2.874
748,800 12.310 21,020 33,330 2,333 2.250 36 360 4.943
1,123,200 13,340 31,540 ....880 3,142 3,300 42 420 6,862
- Assumes 0.07!kWh.
• Assumes 10lhour.
I GPO • 0.003785 mJ/day
Source: U.S. EPA (1979).
Disinfection and Disinfection By-Products 95

1. 0 Tible 5·19. ChemlColl Costa lor Gen_lnll Monochloramine


lor 2,500 GPO Water Treltment Pllnt ($1983)

.9 .1 Amount Annual
Source Chemicals Needed (Iblyr) Unit Cost ($/Ib) Cost ($)
Anhydrous ammonia 6.2 0.40 2.48
.8 .2 and chlorine gas 26.1 0.47 12.27
<3 q.
J: J: Total 14.75
Z .3 Z
S .!R Aqua ammonia 8.2 0.70 4.35
~ and chlorine gas 26.1 0.47 12.27
~
~
.4 ~
.c
Total 16.62
.c U Ammonia sulfate 24.1 0.51 12.29
U
., .5 .~ and chlorine gas 26.1 0.47 12.27
.,> u
«u « Total 24.55
'l5 .6 0
1 Ib • 0.4536 kg; 1 GPO • 0.003785 m3 /day.
§ ~
.
~
u-
.7
u:
.8

.1 .9

1.
5 8 7 8 9
pH

FIgU" 5-12. Proportion. ot mono- Ind dlchlOramln. . ln


water wfth equlrnot.r concentration. ot chlorlne
Ind Immonll.
Source: National Academy of Sciences (1980).
6. Treatment of Organic Contaminants

Treatment of organic contaminants in drinking combination of organic compounds. In these cases,


water depends on the nature of the contaminants the resultant total costs are not always additive;
targeted for removal. The two largest categories of installing both systems may result in some cost
organic contaminants are synthetic organic savings.
compounds (SOCs) and natural organic materials
(NOMs). Of special interest are SOCs and NOMs The second category discussed includes treatment
having significant potential health effects, including processes that have some record of performance, but
volatile organic compounds (VOCs), aromatic are not as widely applicable as CAC or PTA to be
hydrocarbons, halogenated aromatic hydrocarbons classified as BAT. These treatments include
(including polychlorinated biphenols [PCBs]), powdered activated carbon (PAC) in a conventional
polynuclear aromatics, halogenated nonaromatics, treatment process train, covered in Section 6.4;
hydrocarbons, and organic pesticides. diffused aeration, covered in Section 6.5; and
multiple tray aeration, covered in Section 6.6.
Natural organic material found in raw water is in a
dissolved or particulate phase; most SOCs are in this The final category, discussed in Section 6.7, consists
dissolved phase. The particulate phase NOM can be of technologies that are emerging from the laboratory
removed using sedimentation and filtration and pilot-level stages as promising alternatives for
processes. To enhance NOM removal, pretreatment the near future. This category embraces a number of
processes preceding sedimentation and filtration can candidates including oxidation (ozone,
be optimized to convert as much of the dissolved ozone/ultraviolet radiation, and ozonelhydrogen
NOM as possible to particulate NOM. Section 6.1 peroxide), reverse osmosis, mechanical aeration,
cor.tains a discussion of these enhancements. catenary grid aeration, Higee aeration, resins, and
oxidation followed by a biological filtration step.
Twelve treatment technologies are discussed in three Appendix C describes case experiences with GAC,
general categories of applicability. The first category, PTA, and PAC.
discussed in Sections 6.2 and 6.3, covers the most
frequently used technologies, which include granular There are so many organic compounds that
activated carbon (GAC) and packed tower aeration performance data concerning each technology's
(PTA), also referred to as packed column aeration removal ability for every organic compound are not
(PCA). GAC and PTA are classified as best available available. Table 6-1 contains a partial matrix offive
treatment (BAT) technologies for removal ofVOCs of the above technologies and their removal
under the U.S. EPA regulations promulgated in July efficiencies for 33 compounds.
1987.

CAC and PTA have different removal efficiencies for


6.1 Pretreatment for Natural Organic
different organic compounds. PTA is only effective for Contaminant Removal
VOC removal. Treatment technologies frequently are Water treatment processes used prior to
chosen based on potential removal efficiencies for a sedimentation and filtration may be designed to
specific organic contaminant. However, if a decision enhance NOM removal. The most important of these
is based only on cost, then PTA is often chosen for is the coagulation process, which is used primarily for
VOCs because of its lower cost. Sometimes, both controlling turbidity and is routinely applied to
technologies are necessary to remove a particular surface water. While the portion of organic

96
Treatment of Organic Contaminants 97

Table 6·1. Perlormance Summary lor Five Organic Technologies


Removal Efliciency"
Granular
Acbvaled Cartxln Packed Reverse Ozone
Adsorption Column Osmosis Thin Oxidabon Conventional
Organic Compounds Filtrasorb 400 b Aerabon Film Composite (2-6 mgil) Treatment
Volatile Organic Contaminants
AlJ<anes
Carbon tetrachloride + + + + + + 0 0
1,2-0lchloroethane + + + + + 0 0
I, I, 1-Tnchloroethane ++ + + + + 0 0
l,2-Dichloropropane + + + + + + 0 0
Ethylene dibromlde + + + + 0 0
D,btomochloropropane + + + NA 0 0
Alkenes
Vinyl chloride ++ NA + + 0
Styrene NA NA NA + + 0
1,1 -DtChloroethylene + + NA + + 0
cis-' ,2-D,chloroethylene + + 0 + + 0
trans-l.2-0ichloroethylene + + NA 0
Trichloroethylene ++ + + + + 0
Aromabcs
Benzene + + + + 0 + + 0
Toluene + + + + NA + + 0
Xylenes + + + + NA + + 0
Ethylbenzene + + ++ 0 0
Chlorobenzene + + + + 0
o-Dlchlorobenzene + + + + + 0
p-D,chlorobenzene + + + + NA 0
PestiCides
Pentachlorophenol + + 0 NA + + NA
2,4-0 NA 0
Alachlor + + + + + + 0
AldicarO NA 0 NA NA NA
Carboluran 0 + + 0
Lindane 0 NA 0 0
Toxaphene + + + + NA NA 0
Heptachlor + + NA + +c NA
Chlordane 0 NA NA NA
2,4,5-TP + + NA NA NA
Methoxychlor + + NA NA NA NA
Other
Acrylamlde NA 0 NA NA NA
EplchlorohydflO NA 0 NA 0 NA
PCBs NA NA NA
~ + + Excellent 70-100%
Average 30-690/.
o Poor 0-29%
NA Data not aV31lable or compound has not been tesled by EPA Drinking Water Research Division_
b Excellenl removal category to< cartlOn Indicates co<opound has been demoosvated to be a<!sortla/:)Ie onto GAC, In
lull· or p,lol-scale applications, or in the labO<ato<y With characteriShCS suggesting GAC can be a cost·eflectrve
lechnology _
c Ozone OXidation of heptachlor produces a high yield of heptaehlo< epaxide. whICh is not su,tal>le to< further oxldabon_

contamination removed by effective coagulation is commonly removes less than half of the organic
generally small, in conjunction with other content of the water, it tends to remove contaminants
treatments, coagulation can significantly add to a that are not removed by GAC_ These contaminants
plant's removal effecti veness. include color, some trihalomethane (THM)
precursors, hydrophobic SOCs, and toxic metals.
In general, coagulation is most effective in removing
organic contaminants that are hydrophobic, acidic, Oxidation is another pretreatment step that can have
and high in molecular weight. While coagulation a pronounced effect upon the removal of organic
98 Upgrading Existing or Designing New Drinking Water Treatment Facilities

contaminants. Ozone or other advanced oxidation gradual increase in removal effectiveness with
processes employed before filtration drastically increasing coagulant dosage that asymptotically
change the chemical structure of many organic approaches maximum removal.
contaminants.
Overdosage is possible with the first pattern of
Section 6.1.1 focuses on how adjusting the coagulant removal, but not with the second. The second pattern
type, coagulant dosage, coagulant aid application is typical for surface waters with moderate to high
procedures, pH, and point of coagulant addition can turbidity and alkalinity. SOC removal tends to follow
improve organic contaminant removal. Section 6.1.2 the second pattern of removal.
describes the use of oxidation pretreatments to
enhance organics removal.
Coagulant Aid Application Procedures
Coagulant aids include substances that improve the
6.1.1 Coagulant Pretreatment nature of the flocculated particles, and reduce
All coagulants will remove some organic turbidity. Among these substances are acids, bases,
contamination, especially contaminants associated anionic and cationic organic polymers, activated
with macromolecules. However, different coagulants silica, and bentonite clay. Except for acids and bases,
with equivalent turbidity removal efficiencies may the mere presence of coagulant aids has not been
have different organic contaminant affinities. found to dramatically affect organic contaminant
Coagulants should be chosen by matching their removal. However, the order of application of
particular affinities to the specific organic coagulant aids has been found to have a significant
contaminants of concern. impact.

Metallic salt coagulants, most commonly alum and Chemical application procedures are subject to many
ferric chloride, are more effective than cationic confounding site specific conditions. In addition, few
polymers in removing lower molecular weight published data exist regarding optimum chemical
organic compounds with acidic functional groups application strategies. However, the following four
(such as carbonyl or carboxyl groups). Newer metallic recommendations were derived from assumptions
salt coagulants, such as polyaluminum chloride and concerning removal mechanisms for anionic
polymeric iron chloride, are also effective at contaminant~:
removing organics. Polymer coagulants are effective
in enhancing the removal of NOMs, such as humic • Add acids before metal-salt coagulants to lower
and color compounds. the pH toward an optimal level and promote the
production of positively charged ions.
Determining the optimal dosage of coagulant for • Add bases after metal-salt coagulants to
treatment of organic contaminants parallels the maintain the lower optimal pH levels and avoid
derivation of the optimal conditions for turbidity encrustation of the mixer or injection nozzle.
removal. The first step is to fully characterize the • Add clay and PAC before metal-salt coagulants to
plant influent using laboratory jar tests and pH allow adsorption of NOMs and sacs before clay
assessments. Care should be taken to avoid errors in and PAC coagulation.
performing the laboratory tests. One common • Add polymers, especially anionic pvlymers, after
laboratory mistake is to overlook the adsorption of metal-salt coagulants.
organic contaminants onto glass and filter paper. In
addition, laboratory procedures should take into
account contaminant volatilization. Under optimum pH
circumstances, laboratory tests should be followed by The optimal pH for the removal of NOM with alum is
full-scale plant tests. See Section 4.1 for a more frequently between 5 and 6, but can be higher with
complete discussion on determining optimum iron salts. While SOC removal is clearly associated
coagulant dosages. with pH, the optimal pH level varies with each SOC.
Experience thus far indicates that negatively
charged sacs generally are removed most easily in
Coagulant Dosage water with a pH of 5 to 7. Furthermore, positively
The most effective method of applying coagulants charged sacs are most easily removed in water with
depends on the characteristics of the raw water. For a pH between 7 and 8.5, but more data are required to
raw water with low pH (about 5 to 6 for alum), confirm this f:nding.
containing high and homogeneous concentrations of
NOM, organic contaminant removal is characterized
by a sharp increase in removal effectiveness around a Point of Coagulation Application
specific coagulant dosage. For raw water with high The point at which coagulants are added can
pH, containing low and heterogenous concentrations significantly affect the final level of organic
of NOM, the removal pattern is characterized by a contaminants. Oxidation typically is used for
Treatment of Organic Contaminants 99

disinfection purposes, yet it also reduces organic discussed in this section. Powdered activated carbon
contaminant levels. Oxidation that precedes the (PAC) is covered in Section 6.4.
coagulation improves its effectiveness by removing
turbidity and changing the nature of the organic
contaminants. However, when chlorine
disinfection/pretreatment is used, it is better to add
the coagulant first to reduce the level of organic
contamination that can form halogenated by-
products during chlorination. This is not as much of a
concern with preoxidation using ozone or chlorine
dioxide, where halogenated by-products are not as
much of a problem.

6.1.2 Oxidation Pretreatment


Several oxidants are available for treating organics
in drinking water, including ozone, chlorine, chlorine
dioxide, permanganate, hydrogen peroxide and
ultraviolet radiation. Organic compounds can be
completely oxidized into carbon dioxide and water, or
partially oxidized into intermediate reaction
products. Complete oxidation is not always possible
because the intermediate products formed may be
more resistant to further oxidation than the original
organic chemical. Pretreatment with an oxidant may
convert dissolved organic material to particulate
matter, thereby enhancing removal by sedimentation
and filtration. In addition, oxidation of many SOCs
may be accomplished during pretreatment.

Because of the concern for halogenated disinfection


by-products, the use of nonchlorinous oxidants may
be preferred in the early stages of water treatment. In
such cases, the nonchlorinous oxidants do not
produce halogenated by-products, except in those
waters containing substantial amounts of bromide
ion. Some waters, when treated with a preoxidant, Granule, activated carbon unit; Smyma, DE.
have been shown to produce more chlorinated by-
products because oxidation of the organics in the raw Activated carbon works on the principle of
water produces precursors to by-product materials. adsorption. Dissolved contaminants (adsorbate) are
transferred from the water solution to the
Recently, studies have been conducted to investigate microporous surface of the carbon particles
the use of advanced oxidation processes (AOPs) (adsorbent). Activated carbon's large internal
where several oxidants are used in combination (e.g., surface area and porosity are the primary reasons for
ozone and hydrogen peroxide, ozone and ultraviolet its excellent adsorption capabilities. One gram of
radiation, ultraviolet radiation and hydrogen activated carbon has a surface area equivalent to that
peroxide). These processes involve the generation of of a footbalilield.
the hydroxyl radical in sufficient quantities to impact
treatment. AOPs may achieve treatment at a lower The adsorption process is primarily a physical
cost than conventional oxidation. The use of AOPs process that can be reversed relatively easily. The
has been applied to the removal ofSOCs and also ease of reversing adsorption is another key factor in
NOMs. activated carbon's usefulness because it facilitates
the recycling or reuse of the carbon.
6.2 Granular Activated Carbon Contaminant characteristics greatly affect GAC's
Acti vated carbon is used for water treatment either adsorption ability. GAC has an affinity for
as a granular adsorption medium or as a powder contaminants that are:
added to the water like a coagulant, which later
settles out in sedimentation basins or clarifiers. • Branch-chained, rather than straight-chained
Granular activated carbon (GAC), the more common • High in molecular weight
method for removing organic contaminants, is • Low solubility
100 Upgrading Existing or Designing New Drinking Water Treatment Facilities

• Nonpolar
• Present in higher concentrations 6.2.1 Process Design Considerations
Key process design considerations include:
CAC's affinity for larger molecular weight
compounds is illustrated by its reduced effectiveness • CAC type
when preceded by ozonation. Ozonation breaks down • Surface loading rate of the CAC filter
contaminants and thus can reduce CAC's ability to • Empty bed contact time (EBCT)
adsorb them. Table 6-2 lists readily and poorly • Contaminant type and concentration
adsorbed contaminants. • Contaminant competition
• Carbon depth and usage
Table 6·2. Readily and Poorly Adsorbed Organics
GAC Type
Readily Adsorbed Organics Various types ofCAC are available for removing
• AromallC solvents (benzene. toluene. nJtrobenzenes) organics from drinking water. The most frequently
• ChlOrinated aromaocs (PCBs. chlorobenzenes. used carbon in U.S. treatment plants is coal-based
chloronaphthalene) carbon because of its hardness, adsorption capacity,
• Phenol and chlorophenols and availability. Some peat and lignite carbons have
• Polynuclear aromatics (acenapthene. benzopyrenes)
been used. Several sizes of carbon are available, and
the size selected for a particular application is based
• Pesocides and herbicides (DDT. aldrin. chlordane. heptachlor)
on backwash and headloss characteristics, rate of
• ChlOrinated nonaromatlcs (carbon tetrachloflde, chloroaJkyl adsorption, and cost. Headloss is lessened with the
ethers)
larger carbons, while the rate of adsorption is
• High molecular weight hydrocarbons (dyes, gasoline. amlnes. increased with the smaller carbons. The relationship
humlcs)
of carbon cost and total production cost for a wide
Poorly Adsorbed Organics range of reactivation frequencies is shown on Figure
• Alcohols 6-1.
• Low molecular weight ketones, acids. and aldehydes
• Sugars and starches ",40
• Very high molecular weight or collOidal organiCS

iii
• Low molecular weight a1iphaOcs <:> a = Carbon Cost $1.001Ib
o b = Carbon Cost $0.80/lb
g 30
c = Carbon Cost $0.601Ib
'iii
c
The outer surfaces of macropores on CAC particles
"
(J
S 20
are large enough to house colonies of bacteria that ;;;
0
feed on biodegradable organics as they pass in and (J
c:
out of the macropores. Consequently, considerable .g 10
mineralization of organic material occurs after a few (,)
::l
weeks of operating unused CAC. Ozonation and other ~
0..
advanced oxidation (ozonelhydrogen peroxide or
~ 0
ozone/ultraviolet radiation) convert organic 0 0 2 4 6 8 10 12
I-
contaminants into more readily biodegradable ReacOvaoon FreQuency. Months
materials.
Figure 6-1. Effect of carbon cost on facility cost.
Raw water characteristics also affect GAC's
adsorption ability. The most important characteristic
is the presence of competing contaminants or Surface Loading Rate
dissolved solids which can adversely affect The surface loading rate of the CAC filter is related
adsorption. to an individual plant's design capacity. Surface
loading rate is the amount of water passing through a
one square foot area of the activated carbon filter bed
Process, least cost, and facility design considerations per unit of time, and it typically ranges from 2 to 10
are covered in Sections 6.2.1, 6.2.3, and 6.2.4, CPMlsq ft.
respectively. Section 6.2.2 contains a discussion of
tests for deriving the optimum activated carbon
usage rates for different applications. The last four Empty Bed Contact Time
sections cover operation and maintenance (6.2.5), Contact between the influent and CAC is the
system performance (6.2.6), and system costs (6.2.7). primary factor in determining the size and capital
Treatment of Organic Contaminants 101

cost of a GAC treatment system. The empty bed regeneration becomes necessary. Determining when
contact time (EBCT), the time required for water to regeneration is necessary, however, is a site-specific
pass through the empty column or bed (absent of decision. It may be necessary either when
GAC), is determined by the following equation: contamination is detected in the finished water or
when the level of contamination exceeds the
EBCT (min) = GAC (m 3)/flow rate (m 3/min) regulated level.

While most EBCTs range from 5 to 30 minutes, Typical carbon usage ranges from 6 to 120 g/m 3 (0.05
EBCTs oflessthan 7.5 minutes have been ineffective. to 1.0 Ib/1,OOO gal) of treated water, although SOCs
Typical effective EBCTs are around 10 to 15 minutes are removed with carbon usage rates as low as 1.2
and 10-minute EBCTs are typical for removal of most g1m 3 (0.01Ib/1,000 gal) of treated water (O'Brien et
organic compounds. GAC is effective in removing aI., 1981). Table 6-3 provides carbon usage rates for
radon with 180- to 200-minute EBCTs (Lowry and the removal of eight organic compounds.
Brandown). However, these long EBCTs are only
practical for point-of-use applications, not for many Carbon usage varies with the type of contamination.
community systems. VOCs utilize the most carbon, and quickly shorten
the carbon's useful life as an adsorption medium.
Pesticides generally use less carbon than VOCs,
Contaminant Type and Concentration however, their demands on carbon usage vary. Of the
GAC's removal efficiency varies for different organic organics listed in Table 6-3, chlorinated aromatic
compounds. To date, removal tests for only a fraction organic compounds use the least amount of carbon.
of the myriad of organic compounds have been VOC carbon usage rates typically translate into
conducted using GAC technology. Table 6-1 carbon replacement intervals of3 to 6 months, while
(presented earlier) shows examples ofGAC's removal other organic compounds require carbon replacement
ability. rates offrom 1 to 2 years depending on contaminant
concentrations. Pretreatment can significantly
Contaminants in the water can occupy GAC extend carbon's longevity. For example, PTA
adsorption sites, whether they are targeted for preceding GAC removes a large portion ofVOCs. In
removal or not. Therefore, the presence of other addition, oxidation can lower the amount of
contaminants may interfere with the removal of the adsorbable organics by converting some organics into
contaminants of concern. A profile of all materials that are biochemically mineralized during
contaminants in the water will help predict the passage through the GAC.
design specifications required to achieve mandatory
effiuentlevels for regulated contaminants. In addition to the specific type of organic
contamination present, an influent's dissolved
Radon requires an extremely long EBCT, ranging organic carbon (DOC) level can increase GAC usage.
from 100 to 200 minutes; therefore, only small DOC is especially a problem for surface water
amounts of radon are adsorbed during typical GAC supplies. Pretreatment of surface water with
operations for organic contaminants. In addition, coagulation and filtration reduces the burden of
radon decays continuously during the removal turbidity and DOC and thus extends GAC's
process (Lowry and Brandown). Consequently, the longevity.
small amount of radon adsorbed during a typical
organic contaminant EBCT also decays during its The relationship between carbon longevity, influent
residency in the GAC, thus the carbon is renewed on concentrations, and effiuent contamination levels for
a continual basis, as shown in Figure 6-2. Section the contaminant trichloroethylene is illustrated in
7.3.1 contains a more detailed discussion of radon Figure 6-3. For this example, the EBCT is 10
removal. minutes. To achieve an effiuent concentration of 1
I1g1L from an influent with 100 I1g1L, the carbon life is
estimated at 130 days. To allow a greater effiuent
Carbon Depth and Usage concentration of 10 I1g1L from an influent of 100 I1g1L,
Carbon depth is related to the amount of carbon the life of the carbon is extended to about 180 days.
necessary to achieve a desired EBCT and filter life
with a specific level of contamination. Typical carbon
depth reaches 3 to 9 m 00 to 30 ft). Figure 6-4 compares carbon longevity in the removal
of three different organic compounds. The curves, for
Carbon usage, expressed in grams of carbon per cubic each compound, illustrate the correlation between
meter (pounds of carbon per 1,000 gallons) of treated influent concentrations and the useful life of the
water, largely determines the system's operating carbon filter. The curves assume an EBCT oflO
expenses. The carbon is considered to be exhausted minutes and an effiuent concentration of 10 I1g1L. At
when the effiuent organic concentration approaches influent concentrations of200 I1g1L, the carbon life
the influent organic concentration, at which time ranges from about 1 month for l,l,l-trichloroethane
102 Upgrading Existing or Designing New Drinking Water Treatment Facilities

InftuenllEffluen~ 'l'.

100

TyptCal GAC
Breal<lhrough CUI'M for
75 Non'Decaying Allsortlat8

Typical Steady-Stale
AdsorptiotVOecay of
Radon on GAC

25
(

o
o 10 15 20 25 30 35 40
Time. days

Table 8-3. Summary 01 Carbon Uuge Rat. . Car1lon Lilv, Compound: Trichloroelhylene
DIIyI EBCT· 10 min
Contaminant
ConcentrabOn
(ll9!ll
,,
Carbon Usage \
Inftuent Ellluent (Ib/l ,000 gal)" \
,,
Volable Organic Compounds ',_ Eflluent Concentrllion
(Voes)
-', 50 IIWL
Tetraetlloroelhylene (PCE) 100 2 0.08 '" 10 IIWL
Trichloroe1hylene (TCEl 100 2 0,16 '. '. .... '"
... .................. _._._.- 1 IIWL
Trich~ (TCAI 100 2 0.1l6
........ -.- _. '-- -----
PesDcides
Chlo<dane 100 0,012
-......... _. _.-._ ..... _.. . .

Dibromochloropropane 100 O.OHI o L..--..............._ ................._ . L . -........_ L . . -.............._ ........._


(DBCP)
Aldicartl 100 0.02 o 100 200 300 400 500 800 700 800 iOO 1,000
Chlormared NomaDcs InlIuent Concentralion, IlO'l-
Dichlorobenzene 100 2 0.01
Flg\n "3. Ertect 01 OOl...... lMt on C8ftxIn IlM.
PCB (Aroclor 1018) 100 2 0.015
., Ib/I,OOO gal • 120 glm J
Treatment of Organic Contaminants 103

removal to over 1 year for tetrachloroethylene The Freundlich Isotherm Relationship yields
removal. isotherms for each contaminant, with residual
contaminant concentrations in the effiuent graphed
Carbon Life. against the ratio of contaminant adsorbed per gram
Days of carbon. Figure 6-6 shows graphs of six isotherms
400 Ellluent Concentration • ~ (Dobbs and Cohen, 1980). Isotherm tests are quick
EBCT· 10 min and cost about one to three thousand dollars. Many
contaminants' isotherms are available from existing
3 literature. l
Tetrachloroethytet1e

~ ~--;,--
200 In contrast to isotherm tests, dynamic column field
tests can take 6 to 10 months and cost tens of
thousands of dollars. These pilot tests are used to
10 model full-scale facility design parameters. The
1.1.1·Trichloroethane
seven most useful design parameters in this test are:
o L.L.:=L::J::::::L::::r:'\::±=~=±='L- • TypeofGAC
o 100 200 300 400 500 600 700 800 900 1.000 • Different EBCTs
Inftuent Concentration. 1l!1'L • Carbon bed depths
Ftvure 6-4. Effect 01 compound on carbOn lila.
• Hydraulic loadings
• Number of vessels
• Carbon exhaustion rates, life, and regeneration
6.2.2 Tests for Deriving Carbon Usage and cycle length
Other Design Criteria • Contaminant loading rates
Carbon usage rates are derived from three standard Typical dynamic pilot column tests use GAC columns
laboratory and field tests: isotherm, dynamic column, that are about 5 feet deep with 4-inch internal
and mini-column tests. The test results are critical in diameters (see Figure 6-7).
deriving and evaluating GAC process design
parameters. Isotherm and mini-column tests are The mini-column laboratory test is between the
laboratory tests, and the dynamic column test is a isotherm and dynamic column tests in accuracy and
field method that determines carbon usage. complexity. The cost is about the same as the cost for
an isotherm test. This test is used to determine the
The isotherm test is the simplest and least costly of feasibility of using GAC, establish preliminary
the three methods. It evaluates the impact of pH, design criteria, and approximate cost. Test
temperature, and the presence of other contaminants procedures involve pumping raw water through a
on adsorption. It can also compare the effectiveness of short column ofGAC about 70 mm deep. A schematic
different carbon types. Isotherms are derived by diagram of a mini-column test is shown in Figure 6-8.
mixing a measured weight of pulverized carbon in
water with a known concentration of contamination.
Following a specific contact time, the contaminant 6.2.3 Least Cost Design Criteria
concentration of the mixture is measured. To derive the optimal design criteria for an effective
GAC system, the design must consider EBCT, carbon
The isotherm test is based on the Freundlich usage, and column configuration. EBCT has greatest
Isotherm Relationship: impact on capital costs, while carbon usage has the
greatest impact on operating costs. The first step in
xJm = Kc l/n determining the most cost-effective E BCT and carbon
usage rates is to examine the breakthrough curves
where: for the organic compounds contaminating the
influent. The breakthrough curves indicate how long
xJm equilibrium capacity (mg of contaminant a GAC filter can produce a desired effiuent
per gram of carbon) concentration for an organic contaminant.
k capacity at 1 mglL of contaminant
concentration The next step is to calculate the carbon usage (CU)
c contaminant effiuent concentration in rate with the following formula:
mglL
lin exponent

Figure 6-5 shows an application of this equation in


making carbon usage estimates.
1 Data for several compounds are conlalned ,n: Carbon Adsorphon
Isolflerms lor ToXIC Organics. EPA·6oo/8-80·023. Apnl 1980.
104 Upgrading Existing or Designing New Drinking Water Treatment Facilities

CU =
M..aofCarbon
-----..:::;;.;~;;.:;;;..=---­
pump systems usually are adequate to maintain
Volume of Water Treated through Broaltthrough pressures in the distribution systems.

Carbon usage is then plotted against EBCT to select Gravity-fed contactors typically are used for systems
the optimal combination. with greater than 0.44 m3/sec (10 MGD) capacity. By
using common wall construction, they are able to
Column configuration is the next important factor to reduce costs. Gravity-fed GAC systems are
consider in determining the length of the carbon constructed either by modifying existing gravity-fed
regeneration period. Parallel column configurations sand filters or by using new concrete contactors.
require carbon replacement or regeneration when a
filter breakthrough occurs, while a series column The foremost consideration in carbon transfer system
configuration allows replacement to take place at the design is minimizing carbon loss due to abrasion. The
point of carbon exhaustion. Carbon exhaustion occurs hydraulics ofthe slurry system, system velocities,
when the saturated carbon can no longer hold and construction materials all affect carbon abrasion.
contamination. Series configuration extends the
operating period between regeneration cycles. Carbon regeneration can take place either on site or
oJTsite, or the carbon may be used on a disposable
Selection of the appropriate system involves a basis, depending on the system size. Systems with
tradeoJTbetween capital and operating costs. Parallel carbon exhaustion rates between 226.8 and 907.2 kg
systems require fewer contactors, and their (500 and 2,000 Ib)/day will generally use oJTsite
associated capital expense. Series configuration has a regeneration, which is routinely performed on a
longer carbon regeneration cycle with lower contract basis. Systems exhausting over 907.2 kg
operating costs. (2,000 lb) carbon/day may consider onsite
regeneration, which involves heating the carbon to
destroy the organics. To avoid unnecessary energy
6.2.4 Facility Design Criteria costs, dewatering of the carbon is required. Systems
The three major components of a GAC system are the using less than 226.8 kg (500 lbllday generally
carbon contactor, carbon transfer system (which dispose of their carbon, rather than regenerate it.
moves the carbon in and out of the contactors), and
carbon regeneration system (see Figure 6-9). Both carbon replacement and regeneration produce
wastes. Disposing of carbon with contaminants
The two important operational characteristics of classified as hazardous waste will dramatically
contactors are flow direction (upflow or downflow) increase disposal costs. Carbon regeneration
and water feed mode (gravity or pressure). Upflow operations must meet all applicable air quality
and downflow carbon contactors are used in either regulations. Gas emissions from these operations are
series or parallel column configuration. Pressure sometimes categorized as significant point sources.
contactors usually cause a minimum pressure drop of
10 psi. 6.2.5 Operation and Maintenance
Upflow contactors are used for contact periods of up to Several operational and maintenance factors aJTect
2 hours (U.S. EPA, 1973). The upflow design permits the performance ofGAC units, including those
suspended solids to pass through without excessive involving the nature of the influent, dynamics of the
drops in pressure. These extended contact periods are GAC process, and management of the resultant
used to remove suspended solids and organic wastes. These factors require consistent and careful
compounds. The disadvantage of upflow contactors is monitoring.
that they sometimes carry fine carbon particles into
their effiuent. Upflow designs are more typical of A significant drop in influent contaminant
wastewater treatment plants than drinking water concentration will cause a GAC filter to desorb, or
treatment facilities. Downflow contactors typically slough oJT, the contaminant, because GAC is an
are used for contact periods of 30 minutes or less, and equilibrium process (McKinnon and Dyksen, 1982).
are able to remove only a small amount of suspended Another operational problem related to influent
solids during back washing. characteristics is the competition among
contaminants for adsorption sites. Adsorbed
Pressure-fed contactors are used for systems with contaminants are displaced by other contaminants
capacities between 0.04 and 0.44 m3/sec (1 and 10 with which GAC has greater affinities. As a result,
MGD); packaged plants using pressure-fed contactors water with frequently changing quality
are rarely over 0.04 m3/sec (1 MGDl. The pressure- characteristics will produce effiuent with
fed contactor operates with higher head loss levels unpredictable levels of contamination.
than the gravity-fed contactor. In the pressure-fed
contactor, pumps move the water from the well to the Bacterial growth on the carbon may be one potential
contactors and into the distribution system. Existing problem. However, the nature and implications of
Treatment of Organic Contaminants 105

Isotherm equation: xJm • Kc Un

where:

K • 28 l!!!!IL--l!J. Vn (From Isotherm data)


(gm) (mg)
Vn 0.62 (From isotherm data)
(C,) · TCE influent concentration 100 I'QIL
(Co)· TCE effluent concentration 51'gIl
SF • Safety Factor • 0.75
Rearranging:

Carbon Usage
Ci Co 8 (Ibl lkL
K (C ,) Un SF X.34 (1,000 gal) (rngl
(llYl.Oaa gal)
(100' 5) 1!9 x 1 '.!l!I
L ~ HQ X 8.34 Ib x..b...
28 UQQ.l082 x 0.75 (1,000 gal) mg
(1.000)

0.16 IlY1000 gal

Note: 1 1tlt'1.000 gal. 120 glm 3

Figure 6-5. Application of Freundlich Isotherm Relatlonahlp,

in Figure 6-10. If the wave of used GAC is


interrupted, then the contactor may experience
100.0
premature breakthrough. Changes in influent
composition and concentration,as well as the method

if lOa
and frequency of filter backwashing all can affect the
movement ofthe wavefront.

The final operational issue concerns the proper


management of GAC wastes. Wastes from carbon
"Ii regeneration and back washing sometimes require
~ 1.0 treatment prior to disposal.
';1-
X
6.2.6 System Performance
0.1 L-_ _--l.. ....L....c..._ _.L..-_ _---'
GAC effectively removes most organic compounds
0.0001 0.001 0.01 0.1 1.0 from drinking water. GAC's precise removal
Residual Concentration. I'gIl
effectiveness for a large number of organic
Numbars in parentheses Indicate the molecular weoght of the compounds is unknown because of the multitude of
compound organic compounds in existence. Table 6-1 presents
removal efficiencies for 33 compounds.

GAC is not effective in removing vinyl chloride from


water. In addition, the long EBCT required for radon
Figure 6-6. Adaorptlon laotherma for several organic removal makes it infeasible at the treatment plant
compounda found In ground-water auppllea.
scale. However, at the residential scale GAC systems
are cost-effective for radon removal.
this growth are not clearly defined and require
further study. Excessive bacterial growth could cause
clogging and higher bacterial counts in the effiuent. 6.2.7 System Costs
However, it is also possible that such growth The cost of removing many organic compounds with
improves GAC's removal efficiencies through GAC is well within the range of general costs for
biodegradation. delivering treated drinking water. Within the United
States, that cost generally ranges from $0.26 to
Another potential problem is the disruption of the $0.40fm 3 ($1.00 to $1.50fl,OOO gall (1989 dollars) of
adsorption zone procession through a GAC contactor. treated water. Most regulated organic contaminants
This zone of saturated GAC typically moves in a are managed with GAC systems having carbon bed
wave formation through the carbon depth, as shown Ii ves of at least 3 months and EBCTs of 15 minutes.
106 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Feed Pretreatmenr Carbon


Tanl<' Feed Carbon COlumns
Tank

Product
Water

• Pretreatment tanks opbonal depending on


suspended solids concentrabon in feed. Backwash
Watsr
Agure 8·7. DI~rlIm of pilot column teet 1ySt......

300 ml Stainless Steel


Pulsabon Dampener Cylinder

/ Pressure Gaoe (0·200 psi)

.12 Neoprene Stopper

318" 00 x 1/4" 10316


Qrade Stainless Stae4 Tubing
)
711 In-line
Fil1llt

Min, Cartlon COlumn


1/4" 00 x 118" 10
Stainless Steel

T elIon Dispensing
Chemical Metenng
Pump 250 p8i Mu Line
Vanspeed

EPA-Approved VOC
Vial (15 or 40 mil

t. Qlass Sample
Re8elWir
1/4" 00 x 118" 10
Tenon Tubing

Figure 8-3. Dlagrllm of dynamic mlnk:olumn adsofptIon IyStlllll.


Treatment of Organic Contaminants 107

To Distribution
System
GAC

. '
........../
Contaminated WeH

-
-

GraVIty GAC
Contaetor

1
GAC

r
1._'-,'o;..o..:.o.:..:.:...-,°o--="o°'i-_-+-H*,+-_ To Distnbubon
I System
--.'- ..1r--

Pressure GAC Contaetor


Figure 8·9. GAC treatment option•.

systems to $0.26 to $0.79/m 3 ($1.00 to $3.00/1,000


Inftuent gall for small water systems (less than 0.04 m 3 /sec or
1 MGDl.

The most costly organic contaminants to remove are


VOCs, followed by chlorinated aromatic compounds.
In general, the least expensive contaminants to
remove with GAC are pesticides. Table 6-4 shows the
costs of removing alachlor (a pesticidel, TCE (a VOCl,
Wavefront and radon using GAC. The costs for alachlor removal
may be considered representative of the costs for
removing typical pesticides using GAC. The costs for
TCE removal are representative of the costs for
removing the typical VOCs found in ground water.
As shown in the table, the costs for radon removal are
extremely high, probably precluding the use ofGAC
for radon removal. Also shown in Table 6-4 are the
approximate costs per home per year for installing
.0 Effluent
GAC for removal of anyone of the organic chemicals.

GAC Contactor The three fundamental cost components of a GAC


F~ure 8-tO. Wavefront within GAC contaetor. system are contactors, GAC, and piping. Carbon
storage and carbon transport facilities are also
Water systems with GAC units having these necessary cost components. In addition, there are
operational characteristics cost from $0.03 to many site-specific cost elements to consider
$0.04lm 3 ($0.10 to $0.15/1,000 gall for very large including:
108 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Table 6·4. Cost for 95 Percent Removal ot Several Organlea ancl Raclon Using GAC AcllOrpllon ($1989)
Costs by System Size

System Capacity (MGD) 0.1 0.5 1.0 10.0 50.0


Average Daily Flow (MGD) 0.032 0.22 0.4 4.32 22.65
Populanon Served 500 1,500 3.000 22,000 100.000
Alachlo'
Capital Cost 69.000 138,000 190.000 1,500.000 5,600.000
Annual 0& M Cost 8.500 19,100 26.500 117,000 329.000
Total Cost ($/1,000 gal)',b 1.41 0.44 0.33 0.19 0.12
CostJHomeIY~ 100 70 50 40 30
rnchlo,oeltlylene
Capital Cost 106,000 212,000 318.000 2.200,000 10.100,000
Annual 0& M Cost 16.000 38,200 54.100 265,000 930.000
Total Cost ($/1,000 gal) 2.44 0.80 0.50 0.30 0.25
Cost/HomelYear 170 125 90 70 60
Radon
Capllal COSl 445.000 1.300,000 2.100.000 d d
Annual O&M Cost 106,000 318.000 530.000 d d
Total Cost ($/1,000 gal) 13.55 6.00 5.30 d d
Cost/HomeJYear 950 900 800 d d
• Total cost IS calculated based on arnornzing the capltal cost over 20 years at 10 percent Interest rate, adding the annual
O&M cost. and diViding by the total annual flow.
b 1,000 gallons 2 3.78 mJ .
c Cost per home per year is calculated based on total cost per year divided by the number of hOmes served uSing 3.0
people per home.
d These costs were not calculated because they are too high for practJcal municipal application,

• Raw water holding tank (for ground-water reactivation systems are cost effective for smaller
systems) systems, and fluid bed reactivation systems are
suited for plants using more than 1,360,800 kg
• Restaged well pump to address excessive (3,000,000 lb) ofcarbon annually. Multiple hearth
pressure drops caused by the contactors reactivation systems are more expensive than either
• Contactor housing, required in cold climates infrared or fluid bed systems, Figure 6-12 compares
• Chemical feed equipment the costs of carbon replacement and two methods of
• Clear well storage tank and pumps carbon regeneration for several carbon consumption
• Backwash storage tank rates.

The capital costs increase directly with system size.


Consequently, the capital costs can be derived from
the estimated EBCT, a direct measure of system size Operating costs do not correlate strongly with system
because of its relationship with the carbon contactor. size, but are dependent on carbon usage rates, which
The predominant elements of capital costs, as well as rely on the nature of the contaminants in the
operating and maintenance costs, are shown in influent. Other operating and maintenance cost
Figure 6-11, components are carbon type, labor, fuel, steam,
power, maintenance, and laboratory analysis.
The two primary capital cost considerations are (1)
construction type and (2) carbon and reactivation
strategy. Steel pressurized GAC units are most
economical for small systems. generally less than
0.44 m3 /sec(10 MGDl, while concrete gravity-fed 6.3 Packed Column Aeration
systems are more appropriate for larger systems. Aeration, also called air-stripping, mixes water with
air to volatilize contaminants. The volatilized
The carbon and reactivation system, including either contaminants are either released directly to the
onsite or offsite carbon reactivation, compose nearly atmosphere or are treated and then released.
half of the GAC system costs. Infrared carbon Aeration is used primarily to remove VOCs.
Treatment of Organic Contaminants 109

c.pIUl Co~

Sitewotk etc. (15".)

CarbOn Transfer and


SlDrage (5%)

Carbon Contaetor (SO".)

CartlOn Regeneration (30".)

Mamtenance Matana! (5".)

Labor (10%)

CartlOn Regeneranon (75".)

Figure 6-11. GAC facility ea.t components.

The aeration process is based on the principles of Kl the coefficient of mass transfer expressed in
mass transfer from liquids to gases. This transfer is Iblhr/ft2
expressed in the following equation:
a = the etTective area in ft2
M = (Kl) (a) (<IlP)
<IlP = the concentration difference or driving force
where: (The driving force is the difference between
conditions in the aeration unit and
M = the mass of the substance transferred from equilibrium conditions for the substance
water to air expressed in Iblhr/ft3 between gas and liquid phases.)
110 Upgrading Existing or Designing New Drinking Water Treatment Facilities

120
Tlble &-5. Henry'l uw Conatantl tor Nine Organic
Chemicall
100 Henry's Law Constant"

§ 80 ~~~-':~tl_ 1'------ VOCs


Type of Organic Chemical (dimensionless Units)

:iI
u y- MullJp4e Heat1h Vinyl chloride
T richloroothytene
265.00"
0.41'
D 60
~c:
.,
U 40
\ Telrachloroethytene
clS-l,2·DlChloroothytene
0.82"
0.32°
FlUidized
Bed Reactor
J~....
Pesticides
Aldicartl 0.00000017°
20
- Chlordane 0.004°
Dibromochloropropane 0.01°
o
100 1000 10,000 100,000 Chlonnated Aromatics
Carbon Requirement (LbsIOay) Polychlorinated biphenols 0059°
Dichlorobenzene 0.081 0
Note: Constants estimated at about 20·C.
Figure &-12. Coat com~rtaon at carbon "AWWA Research Foundation and KIWA (1983).
reg_ratlorv'replacemenl option•. ° U.S. EPA (1988b).
Source: Dyksen et al. (1987).
underneath the medium in the tower or column. A
typical PTA unit is shown in F'igure 6-13.
The Henry's Law constant indicates a contaminant's
volatility and its affinity for the aeration process. The major process elements of PTA are the column
Substances with high Henry's Law constants are (or tower), packing medium, blower, booster pump.
easily aerated, while those with low constants are and instrumentation. Columns can be constructed
difficult to remove with aeration. Henry's Law from fiberglass-reinforced plastic, aluminum,
constant is calculated with the following formula: stainless steel, or concrete. Within the column are
mist eliminators to prevent water from escaping in
H = [(l6.04)·P·MV(T·S) the vents, packing material, support grids for the
packing material, and liquid distributors to separate
where: the influent into many smaller streams. The four
primary designs for liquid distributors are orifice
H the Henry's Law constant in dimensionless plate (see Figure 6-14l, trough-type distributor (see
units Figure 6-14), orifice headers, and spray nozzles.
P the vapor pressure expressed in mm Adding PTA to an existing plan will require (1)
M the molecular weight of the solute changes in the staging of the well pumps and (2)
T the temperature in oK repumping treated water to the distribution system.
S the solubility in mg/L
Housing the tower usually is not necessary because
Table 6-5 lists several compounds and their Henry's the water temperature remains fairly constant
Law constants. As the table indicates, vinyl chloride throughout the PTA-treatment process.
has an extremely large Henry's Law constant Consequently, water rarely freezes during the
relative to any other VQC. process.
The two basic types of aeration systems are diffused PTA systems vary: in some, water cascading over
air and waterfall units. The diffused air system, spillways imparts the necessary turbulence; in
discussed in Section 6.5. involves adding air to a others, several layers of slats mix the water with air.
volume of water. Waterfall aeration units are more Innovations in PTA are reflected in the newest
common and are discussed in the remainder of this additions to the latter type of aeration system.
section. Emerging aeration techniques include the catenary
grid and Higee systems which are discussed in
Packed tower aeration (PTAl. or packed column Sections 6.7.4 and 6.7.5.
aeration (PCA). is a waterfall aeration process that
trickles raw water over a medium within a cylinder The rest of this section profiles aeration technology
to mix water with air. The medium is designed to including design considerations (Section 6.3.1 l,
break the water into tiny droplets. a process operation and maintenance (Section 6.3.4), system
enhanced by the introduction of air blown from performance (Section 6.3.5), and system costs
Treatment of Organic Contaminants 111

• Water loading rate


• Water temperature

Different contaminants require different designs to


accommodate the particular degrees of volatility and,
thus, affinities for aeration. Packing materials are
designed to simultaneously provide a low pressure
drop across the material and maximum air-water
contact area. They typically are made from plastic or
ceramic, and come in the following forms: super
intalox, Tellerettes, Tri-packs, pall rings, ber!
saddles, and Raschig rings. The desired contaminant
removal level and air-to-water ratio determines the
packing material height. The selected column height
and design of air intake louvers must comply with
local zoning regulations concerning structural height
and noise nuisance. These zoning regulations are the
most basic site-specific considerations.

The air-to-water ratio in a column is a function of the


water temperature and the desired level of
contaminant removal. This ratio determines the size
of the system's blower, which is the primary
component of operating costs for PTA systems. Air-to-
water ratios typically range from 30:1 to 100:1. The
water loading rate, the amount of water passing
through the column, routinely ranges from 16.9 to
20.4 Usedm 2 (25 to 30 GPM/ft 2). The column
diameter specification is derived to accommodate the
desired water loading on the column.

All of these factors affect aeration's removal


efficiency. Figure 6-15 illustrates the dramatically
Fiberglass-reinforced plastic packed column with two carbon different removal rates achieved by the same packed
vapor phase treatment units; Battle Creek, MI.
material depth for three different contaminants. PTA
removal effectiveness usually increases as water
<Section 6.3.6). In addition, this section reviews pilot temperature increases. The decrease in column
testing of PTA systems (Section 6.3.2) and managing height required to achieve the same level of
ofVOC emissions from aeration (Section 6.3.3). contaminant removal as the water temperature
increases is shown in Figure 6-16. In most cases,
heating influent is not cost effective, but in
6.3.1 System Design Considerations temporary situations of low flow, it has been used
Aeration provides a fixed percentage of contaminant effectively.
removal regardless ofinf1uent concentration. To
compensate for uncertainty, aeration systems can be Pretreatment to remove iron, solids, and biological
designed to incorporate safety factors of two or three growth may be necessary to prevent clogging of the
times the expected influent contaminant packing material and, thus, the entire system.
concentrations to ensure compliance with regulatory Posttreatment also may be necessary to reduce
standards. corrosive properties that develop in the water due to
the increased presence of dissolved oxygen during the
Aeration system performance is affected primarily by aeration process. One solution is the use of a
column size and airflow. Increases in airflow and corrosion inhibitor.
column height improve removal efficiencies. Typical
design parameters are provided for 13 common VOCs
in Table 6-6. Design considerations include: 6.3.2 Pilot Testing PCA
Figure 6-17 is a schematic diagram of a typical pilot
• Type of organic contaminant(s) column. Pilot columns are used to test various water
• Concentration of contaminant(sl loading rates, airflow rates, air-to-water ratios, and
• Type of packing material packing materials. Typically, 8 to 12 pilot column
• Height of packing material runs are used to test various combina tions of design
• Air·to-water ratio factors. A test run lasts for about 30 minutes to
112 Upgrading Existing or Designing New Drinking Water Treatment Facilities

To Atmosphere

Packed
t
Column

Spray
Header

ttt Plastic
Media

High Service
Vertical
Turbine Pumps

Blower Finished Water

- 1 ~
Assembly

------
~==*=-
to System

Clearwell

Well
Figure 6·13. Packed tower aeration system.

achieve steady state. Influent and effiuent samples Emission rate (in lblhour) = (Cl·C2)·(V)·(S/107)
are collected for each test run.
where:
The following precautions should be taken to ensure
accurate pilot test results: Cl =
Influent concentration of the vac expressed
inllglL
• Carefully place packing material into the column C2 Effiuent concentration of the vac expressed
to avoid channeling and vacant pockets. inllglL
• Level the pilot column to avoid channeling and V = Water flow rate in GPM
wall effects.
• Select representati ve performance points for The emission rate must be evaluated in the context of
sampling. applicable air quality regulations and other site-
• Collect and analyze duplicate samples to check specific factors. These factors include proximity to
results. human habitation, treatment plant worker exposure,
• Verify laboratory results. local air quality, and local meteorological conditions.

Air emission regulations are expressed either in


terms of permissible emission rates (lb/day or lblhr)
6.3.3 vac Emission Control or projected ground level concentrations (mglm 3). If
To meet air emission regulations, PTA exhaust gas the plant's emission rate is unacceptable, then
may require treatment. vac emissions for PTA units column or plant process design may be changed to
are calculated with the following formula: bring the plant into compliance.
Treatment of Organic Contaminants 113

Orlnce - Type Dlstnbulor

Trough - Type D,stnbutor

Figure 6·14. Dlslrlbutor types.

Modifying PTA plants to dilute emissions to an incineration is not used widely because it is not
acceptable level is the least costly method of effective in eliminating low lE:vels of chlorinated
achieving compliance with air regulations. Such organic compounds. To date, ozone destruction used
plant modifications include increases in tower in conjunction with ultraviolet radiation has only
height, airflow rate, and exhaust gas velocity. If these been evaluated on a pilot scale.
steps are insufficient to achieve compliance, then a
vapor phase treatment componen~ may be required.

The four methods used for vapor phase treatment of Carbon adsorption for control of vec off-gases from
vec emissions are (1) thermal destruction, (2) PTA is accomplished with a vapor phase GAC unit.
catalytic incineration, (3) ozone destruction, and (4) Currently, GAC is the most frequently used approach
carbon adsorption. The first three methods are not to controlling these vec emissions. figure 6-19
used extensively. Thermal destruction is effective but shows a schematic diagram of a typica I gas phase
very expensive because of the high energy unit. The heating element eliminates humidity,
requirements. Catalytic incineration, depicted in which competes with organic compounds for the
Figure 6-18, lowers the energy requirements of the available GAC. The specifications of a gas phase unit
thermal destruction process. However, catalytic depend on individual PTA system requirements,
114 Upgrading Existing or Designing New Drinking Water Treatment Facilities

6.3.5 System Performance


PTA is an effective and practical removal process for
several organic compounds applicable to water
treatment plants of all sizes. It is typically used for
systems drawing ground water. Larger plants using
PTA for radon removal have the added task of
properly managing the associated air emissions.

PTA effectively removes from drinking water most


compounds with high Henry's Law constants, which
includes most VOCs. PTA is the BAT for vinyl
chloride, which has an extremely high Henry's Law
constant; GAC and PTA are BAT for other regulated
VOCs. Table 6-7 lists examples of aeration's removal
effectiveness for trichloroethylene, vinyl chloride,
VOCs, and aldicarb.

6.3.6 System Costs


The cost of removing many organic compounds with
PTA, by itse If, is we II within the range of the cost of
delivering treated drinking water in the United
States, which typically costs from $0.26 to $0.40/m 3
($1.00 to $1.50/1,000 gal ofwaterl. The cost of
OrifiCe-type distributor plate lor a packed column aerator. controlling air emissions from PTA units
significantly increases total system costs.

Table 6-8 presents estimated costs of removing 95


including air-to-water ratio, influent concentration of percent of several VOCs and radon using PTA. As the
the VOC, and acceptable level of VOC emissions. table indicates, costs in terms ofcents per 1,000 gal of
treated water drop significantly as system size
Carbon usage for gas phase control of VOCs from increasf's. Of the VOCs listed in the table, the most
aeration units is less than equivalent VOC controls expensive VOC to remove with PTA is
for the liquid phase GAC units. In addition, since gas dibromochloropropane; the least expensive is vinyl
phase adsorption kinetics are faster than liquid chloride. Both vinyl chloride and radon are removed
phase kinetics, the carbon bed size is reduced. Carbon equally as well using PTA, therefore the costs for
exhaustion is estimated either by using a mass removing these contaminants are similar.
balance approach or monitoring with gas Approximate costs per home per year are presented
chromatography or mass spectrometry. in the table for each of the contaminants.

The capital cost elements for PTA systems include


6.3.4 Operation and Maintenance the tower or column, internal column parts, packing
Typically, packed columns are operated material, blower(s), clearwell, booster pump(s), and
automatically. Daily visits assure that all equipment any associated piping. Site-specific costs may include
is running satisfactorily. Maintenance requirements a raw water holding tank, restaged well pump,
generally involve several hours per month to service blower building, chemical facility, noise control
pump and blower motors and to replace air filters on installation, and air emission control.

he blowers, if necessary. Some packed column The cost of vapor phase controls to manage air
installations have reported operational difficulties emissions from PTA operations can strongly
from the plugging of the packing material and influence plant design and total costs. EPA research
inadequate liquid distribution. However, most indicates that adding a vapor phase carbon
installations report no unusual operational adsorption unit will double the costs of PTA. Carbon
problemsthe blowers, if necessary. Some packed adsorption units have significant capital and
column installations have reported operational operation and maintenance cost components,
difficulties from the plugging of the packing material including costs for carbon contactors, initial
and inadequate liquid distribution. However, most activated carbon material, gas heaters, and
installations report no unusual operational problems. installation.
Treatment of Organic Contaminants 115

Table 6-6. Typlcal AIr Strippl"il O. .lQn Parameters for Removal of 13 Commonly Occurring Volatile
Organle Chemal.-
Henry's Law AI< Stnpper Helghl D,ameter 01 Packed
Compound Constant Air·lo-Water Ratio (feetl b Column (Ieet)
Trichloroethylene 0.t16 29.9 38.0338.03 8.10
Tetrachloroethylene 0.295 11.8 43.77 5.97
Carbon letrachlonde 0.556 6.2 44.88 4.95
1, 1,1 -Trichioroelllane 0.172 20.1 40.06 7.07
1,2-DlChloroelllyiene 0.023 150.6 33.47 14.89
Dichioromelllane 0.048 71.59 2861 11.12
cis-l ,2- 0.093 37.10 34.88 8.73
DlChloroelllylene
Vinyl chloride 265.000 0.013< 59.58 1.90
Benzene 0.106 32.69 36.25 8.37
Toluene 0.117 29.62 39.04 8.07
m-Xylene 0.093 37.26 40.49 18.34
Chlorobenzene 0.069 50.29 37.60 22.74
1,2-Dichlorobenzene 0.090 38.67 40.45 8.86
• Water now rate - 2.16 MOO (8.17 mVday). inlet water concentration· 100.0 1Ig!L, water trealmenl oblectlve - 1.0 IIg1L,
3lr-stnpper temperature· 50 of (10°C), 3lr-S1Tlpper packing pressure drop· 50.0 (N/m 2 )/m packing, al<' stnpper
packing - 3-tnch plastic saddles
b 1 It • 0.3 m.
< Theoretical calculation based on lIle extremely high Henry's Law constant
Source: Crinenden et aI. (1988).

"r- \
100
1.2·DlChloroethane PackIng HeIght. It
35
95 Percent Removal
Temperature 55° F
30

!"\\ \ 25
S,le a (55°F)

~::L~----- ------
20

15

10
o 20:1 40:1 60:1 80:1 100:1 120:1
A:W RatIO
5 S,te A (79°F)

o
Figure 6-15. Ellect of compound on packed-column dealgn. 50 75 90 95 97.5 99 99.5 99.8

RemovaJ EffiCiency. %

Figure 6-16. Packing height VB. removal efficiency lor


trichloroethylene.
116 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Effluent Air

Water Distnbubon Plate

3ft
Packing Media

2.5 ft Intermediate Sample Tap (Typ.)

2.5 ft
Support Plate (Typ.)

Water Flow
Metenng 2.5 ft
Valve Water Meter
Connecling Flange (Typ.)

120 Volts Single Phase


From Raw 15 Amps Blower
Tower Suppa" Table
Water
Source
A" Flow Metenng Valve

\'~':~
Raw Water Sample Tap (

Effluent Water Sample Tap

Innuent A"
Effluent Water to Dram

Figure 6·17. Schematic 01 pilot aeration column.


Treatment of Organic Contaminants 117

Filler/Mixer Catalyst Bed

Figure 6-18. Schematic 01 eataly11c Inclneratlon proce...

6.4 Powdered Activated Carbon Plus reactions (as opposed to column-mixed flow
Conventional Treatment reactions).

The PAC process is based on the same principles of The rest of this section discusses PAC application
adsorption as GAC. However, PAC is not a granular techniques, including the Roberts-Haberer process
filter medium; rather it is a powder added directly to and the fluidized-bed adsorber (Section 6.4.1 l, design
the water at one or more points during the treatment considerations (Section 6.4.2l, and system
process. Typically, PAC can be added during performance (Section 6.4.3),
coagulation, flocculation, sedimentation, and
filtration. The addition of PAC:

• Improves the organic removal effectiveness of 6.4.1 PAC Application Techniques


conventional treatment processes Two potential techniques for improving PAC's
• Addresses short-term and emergency problems effectiveness for organic contaminant removal are
with conventional treatment systems the Roberts-Haberer process and the fluidized-bed
• Acts as a coagulant aid PAC adsorber.
• Removes taste and odor

PAC is also an attractive treatment technology Roberts-Haberer PAC Process


because it is less expensive than CAe in addressing The Roberts-Haberer process uses an upflow filter
seasonal problems, is easily started and stopped, composed of foamed polystyrene beads having a
creates no head loss , does not encourage microbial specific gravity of less than 0.1. Figure 6-20 shows a
growth, and has relatively small capital simplified diagram of the three-phased process.
requirements. During the conditioning phase, water recirculates
through the filter media to allow the PAC to adhere.
The chief disadvantage of this process is that some During filtration, the media act as a combination
contaminants require large dosages of PAC for upflow filter and activated carbon adsorber. When
removal. Another disadvantage is that PAC is the PAC is exhausted, it is released to the drain
suitable only for conventional treatment systems. through the back washing process. The Roberts-
PAC also requires specific system hydraulics, space, Haberer process used with coagulation is especially
and sl udge- handling practices. PAC has proven effective with low turbidity influent or combined
ineffective in removing natural organic matter, due with the sedimentation process. It can also provide
to the competition from other contaminants for flocculation for a system.
surface adsorption and the limited contact time
between the water and PAC. In addition, PAC The reported advantages of the Roberts-Haberer
adsorption is not amenable to basin-mixed flow process include:
118 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Treated Air

Conlam,nated Air

Raw Water
GAC
f Heating Elemenl

Packed
Column
Blower

Blower
Treated Water
Figure 8·19. Vapor-pha.. Co1rbon IV-tem tor Ireatment at aeration exheust air.

• Increased SOC removal if used after coagulation,


Table 6·7. Examplea at Removal Eltlelenel. . tor PTA due to the removal of high molecular weight
SV-Iema
substances by coagulation
TCE • Improved filtration efficiency by removingTHM
• A full-scale redwood slate lJay aeratIOn plant With a 3.8 precursors prior to filtration and allowing
MGO capac,ty at an a.r-ta-water ratio of 30:1 achleve<l prefiltration chlorination, which reduces the
50 to 60 percent reductions from Ini1ial TCE Influent microbial presence on the filter
concentrations of 8.3 to 39.5 \1QIl..
• Minimal bacterial growth because of the
• A full-scale multiple tray aeration unit with a 6 MGO frequency of backwashes
capacity achieved 50 percent reduCtIOns from initial TCE
,nfluent concentrations of 150 I'gIl.
• Increased ability to accommodate multiple
objective water treatments
• A full-scale packed tower aeration column plant using
ground water at an air-ta-water ratio of 25: 1 achieved 97
• Eased recovery or modification of PAC
10 99 percenl reduCllons from initial TCE Influent application
conC91ltrations ranging from 1.500 10 2.000 I'gIl.
Vinyl ChlOride
Fluidized-Bed PAC Adsorber
• A pdot packed tower aeratOf, WIth goC Influent. achteved
up to 99.27 percent removal of Vinyl chloride. This experimental method processes water through
• A spray tower aerallon u"" removed vinyl ChlOride from
flocculated carbon, which extends the contact period.
ground water With vec concentrations of 100 to 200 The strength of the flocculated carbon particles
I'gIl. increases the carbon's longevity for treatment
• An In-well aerallon unll With an a.r-lift pump achl6ved 97 purposes. Floc strength must be balanced with
percent removal of Vinyl chloride. particle breakup, which increases carbon surface
Aldlcarb area and, therefore, adsorption of contaminants,
• Aeration was found to be ,neffectlve ,n reducing levels of
Because fluidized-bed PAC adsorbers are relatively
aJdlCafb because of Its low Henry's Law constant. new, more research is required for a full evaluation.
vecs
• A four-stage aeration des.gn With four shower heads and
a pressure drop of lOps, achieved 99.9 percent vec 6.4.2 System Design Considerations
removal.
The primary design considerations for instituting
PAC are dosage, contact time, and points of
• Reduced PAC use because of the increased application. Dosages commonly are less than 100
contact time mg/L but can range as high as 300 mg/L, and the
• Quicker adsorption and increased utilization of minimum contact time is usually 0.25 hour. PAC
carbon, relative to the GAC process, due to the points of application customarily are (1) before the
small size of the carbon particles rapid mix process, (2) after rapid mixing but before
• Decreased carbon use relative to the GAC process flocculation, or (3) after flocculation and before
Treatment of Organic Contaminants 119

Table 6·8. COlIt for!l9 Percent Removal of Several VOCa and Radon Using Packed Tower Aeration (S1989)
Costs by System Size

System Capacity (MOD) 0.1 0.5 1.0 10.0 50.0


Average DaJly Flow (MOD) 0.032 0.22 0.4 4.32 22.65
Populaoon Served 500 1.500 3.000 22.000 100.000

Dibromochloropropsne
Capttal Cost 106.000 420.000 636.000 5.700.000 5.60U.000
Annual O&M Cost 4.500 25.000 50.000 460.000 329.000
Total COSt ($11.000 gall',. 1.90 0.90 0.85 0.60 0.12
Cost/HomelYe3/" 175 165 160 155 30
Trichloroethylene
CapItal Cost 85.000 210.000 318.000 2.100.000 10.100.000
Annual O&M COSt 2.700 10.800 21.200 201.000 930.000
Total Cost ($/1.000 gal) 1.10 0.45 0.40 0.30 0.25
Cost/HomelY ear 75 70 65 60 60
Vinyl/Chloride/Radon
Capital Cost 54.000 148.000 210.000 1.484.000 5.830.000
Annual O&M Cost 1.600 6.800 16.700 106.000 477.000
Total Cost ($/1.000 gal) 0.70 0.30 0.25 020 0.15
CostIHomelYaar 50 45 40 35 30
• Tolal cost IS calculated based on amortiZing the capllal cost over 20 years at , 0 percent Interest rale. adding the annual
O&M cost. and diVIding by the tolal annual flow.
• 1.000 gaflons • 3.78 m3.
c Cost per home per year IS calculated based on lotal cost per year diVided by the number of homes served uSing 3.0
people per home.

Condilionlng Filtration Backwash

,
I

I
I
I

,,
I

I
I
I
I
I
I
I
PAC

~---r---e-
------@--;
I
____ ..J ---4 I
I
I I
I I
I I
I I

Figure 6,20. Schematic of the Roberts,Habel'9r proce...

sedimentation. Figure 6-21 is a schematic diagram of process with and without PAC for 15 organic
PAC usage. compounds, including two vacs and 13 sacs.

6.5 Diffused Aeration


6.4.3 System Performance The diffused aeration system bubbles air through a
Adding PAC to conventional treatment systems can contact chamber for aeration; the diffuser is usually
greatly improve their performance in removing located near the bottom of the chamber. The air
certain organic chemicals. Table 6-9 compares introduced through the diffuser, usually under
removal efficiencies of the conventional treatment pressure, produces fine bubbles that impart water-air
120 Upgrading Existing or Designing New Drinking Water Treatment Facilities

mixing turbulence as they rise through the chamber. resultant bubble size. Air diffusers use porous plates
Diffused aeration units are designed to serve either or tubes, or perforated pipes placed along the bottom
point-of-use or plant situations. Figure 6-22 depicts a or sides of the chamber. Chamber hydraulics affect
plant-scale system. Figure 6-23 shows a home-scale the uniformity with which the aeration process takes
aeration system. place and, therefore, the completeness of removal.
Baffling that achieves plug flow conditions followed
by mixing is the general method of assuring proper
Table 6-9. Typical Performance of Conventional
chamber hydraulics.
Treatment Proceu. . without and with PAC

Conven-
tional 6.5.2 System Performance
Treatment
Without ConventJonal Treatment
Table 6-10 presents removal effectiveness data of
PAC with PAC diffused aeration for five VOCs and nine sacs. The
Percent Oosage Percenl removal rates, ranging from 11 to 95 percent, are
Compound Removal (mgtl) Removal based on diffused aeration units with air-to-water
VOCs ratios of 5: 1 to 15: 1 and contact times of 10 to 15
Caroon tetrachlOride 9.6·30.0 0·25 minutes.
, ,, ,1-Trichloroethane 40·65
SOCs 6.6 Multiple Tray Aerat.lon
Acrylamlde 5 8 13 Multiple tray aeration directs water through a series
Alacnlor <50 4-34 36-100 of trays made of slats, perforations, or wire mesh. Air
CarOOfuran 54-79 9-25 45-75 is introduced from underneath the trays, either with
or without added pressure. Figure 6-24 is a diagram
O-Dlchlorobenzene 8-27 38-95 of a redwood slat tray aerator.
2,4-0 0-3 , 1-306 69-100
Multiple tray aeration units have less surface area
Elhylbenzene 8-27 33- >99
susceptible to clogging from iron and manganese
precipitation than PTA. However, this type of
Heotaehlor 64 11-97 53·97
aeration is not as effecti ve as PTA and can experience
Lindane 10·20 2-34 82·97 clogging problems, in addition to biological growth
MonochlorObenzene 8·27 14·>99 and corrosion problems. Multiple tray aeration units
are generally available as package systems.
Toluene 8·27 0·67
2.4.5-TP 63 1.5-17.0 82-99
6.6.1 System Design
Toxaphene '-44 40-99
The principal design considerations for multiple tray
Xylenes 8·27 60- > 99 aeration are tray type, tray height, pressurized or
unpressurized air flow, and air-to-water ratio. Trays
a Information not aV3Jlable.
Source: Millner and Fronk (1985).
are usually made from wood or plastic and range in
stack height from 3.6 to 4.8 m (12 to 16 ftl.
Pressurized air flow is used to increase the air-to-
The main advantage of diffused aeration systems is water ratio, with the typical ratio being 30: 1.
that they can be created from existing structures,
such as storage tanks. This type of aeration, however,
is less effective than PCA and is generally used only 6.6.2 System Performance
in systems with adaptable existing structures. Slat tray aeration with an air-to-water ratio of 30: 1
and a tray height of 3.6 to 4.8 m (12 to 16 ftl has
6.5.1 System Design Considerations achieved 30 to 90 percent reductions of
trichloroethylene and 20 to 85 percent reductions of
The critical process design considerations for diffused tetrachloroethy leha.
aeration units are:

e Diffuser type and air bubble size


e Chamber depth: 1.5 to 3 m (5 to 10 ftl 6.7 Emerging Applications of Treatment
• Air-to-water ratio: 5: 1 to 15: 1 Technologies for Organic
• Detention time: 10 to 15 minutes Contaminants
• Chamber hydraulics This section discusses seven emerging water
treatment technologies that hold promise of
One of the most important diffuser design becoming BATs for removing organic contamination.
considerations is the air introduction method and the including:
Treatment of Organic Contaminants 121

Chemicals . Powdered Acbvated Carbon DISinfectant

Rapid Mix Flocculation Sedimentabon Filtration


Figure 6·21. Schem~lc of PAC adaorptlon procesa.

Ozone is widely used in Europe and is gaining


substantial interest in the United States. A 600-
Air Supply
MGD direct filtration plant in Los Angeles,
California, is employing preozonation to enhance
Diffuser Gnd turbidity removal and filter longevity. Three
advanced oxidation processes for treating organic
contamination are currently being tested: (I) ozone
with high pH levels, (2) ozone with hydrogen
peroxide, and (3) ozone with ultraviolet radiation.
Each of these processes forms the hydroxyl free
radical, which has an oxidation potential about 30
percent higher than molecular ozone. These processes
combine organic compound removal with
disinfection, and taste and odor control.

Figure 6·22. Schem~lc of a plant-acele diffused aeration Ozone with High pH Levels
procen. Ozone, at low pH levels (less than 7), reacts primarily
as the 03 molecule by selective and sometimes
relatively slow reactions. Ozone at elevated pH
• Oxidation including ozone and advanced (above 8) rapidly decomposes into hydroxyl free
oxidation processes radicals, which react very quickly. ~any organic
• Reverse osmosis compounds that are slow to oxidize with ozone,
• Mechanical aeration oxidize rapidly with hydroxyl free radicals.
• Catenary grid aeration
• Higee aeration The alkalinity of the water is a key parameter in
• Resins advanced oxidation processes. This is because
bicarbonate and carbonate ions are excellent
All of these technologies require extensive laboratory scavengers for free radicals. Consequently, advanced
and field testing before becoming BATs. oxidation processes are incompatible with highly
alkaline water. In addition, carbonate ions are 20 to
30 times more effective in scavenging for hydroxyl
6.7.1 Oxidation Including Ozone free radicals than bicarbonate ions. Therefore,
Oxidation is usually accomplished with either ozonation at high pH should be conducted below 10.3
chlorine, chlorine dioxide, ozone, or potassium at which level all bicarbonate ions convert to
permanganate. Complete oxidation reactions also carbonate ions.
destroy organic contaminants. Incomplete oxidation
reactions, however, produce by-products, most of
which are biodegradable and some of which may pose Ozone with Hydrogen Peroxide (the Peroxide
health risks. Process)
The combination of ozone with hydrogen peroxide
Chlorination, the most widely used disinfection, is much more effectively reduces levels of
used routinely for oxidation (e.g., breakpoint trichloroethylene (TCE) and tetrachloroethylene
chlorination). Unfortunately, THMs and other (PCE) than ozone alone. Table 6-11 shows design
halogenated by-products are known to be formed criteria and assumptions for a full-scale
during chlorination. Other oxidation agents and ozonelhydrogen peroxide plant.
processes that do not pose this problem are being
studied as agents to reduce levels of halogenated A significant advantage of the peroxide process over
organic contaminants. CAC and PTA is the absence of vapor controls
122 Upgrading EXisting or Designing New Drinking Water Treatment Facilities

Air·Stripped Radon to
Outside Vent

Pumpllevet
Conlrtll Box
Aeration
Tank

Pump
Raw Water trom
Radon Generator

To Metered
Household
Water Use

Air Diffuser

~\AirFIOW Hydropneumatic Tank

.\ Regulator

Air Compressor
or Pump
Figure 8-23. Home dlttuHd __ tlon Iv-tem.

because the contaminants are destroyed, not merely


removed from the water. Table 6-12 compares annual
costs for four system types: aeration, aeration with
Tlble S·10. Typical PerlonNlnce of DlffuHd A_Ion gas-phase GAC, liquid-phase GAC, and an advanced
Compound % Removal oxidation process.
VOCS
Trichloroethylene 53-9~
Ozone/Ultraviolet
Tetrachloroethylene 73-9~
In ozone/ultraviolet (UV) treatment, ozone catalyzed
1,2 ·Oichloroethane 42-77
by UV oxidizes organic substances. This process
breaks down the saturated bonds of the contaminant
1.1 -Oichloroelhytene 97
molecules. Typical contact time is 0.25 hours. Table
1.1.1-Tnchloroethane 58-90 6-13 presents system removal efficiencies with 0.25-
hour contact times and varying ozone dosages for 13
SOCs SOCs. A major advantage of this system is that it
Cartlofuran 11-20 does not produce any THMs. These systems also do
, ,2·0IChloropropane 12-79 not require waste disposal because the contaminants
CIS" ,2-Dichloroethy1ene 32-8~
are destroyed.

trans·1,2·Dichloroethylene 37-96 There is some concern about the completeness of the


o-Dlchlorobenzene 14·72
ozonelUV oxidation process and the intermediate
breakdown products. If oxidation is incomplete, some
Ethylt>enzene 24·89
of the compounds produced in the intermediate
reactions may still be available to form THMs. The
lAonochlorobenzene 14-8~
influent contaminant profile also affects the
Toluene 22·89 performance of these systems. However, if oxidation
Xylenes 18·89 is followed by a biological filtration step, particularly
Source: Miltner and Fronk (198~).
GAC on sand or GAC adsorber, these oxidation
products are mineralized into carbon dioxide and
Treatment of Organic Contaminants 123

Mechanical aeration units consume large amounts of


space because they demand long detention times for
effective treatment. As a result, they often require
open-air designs, which may freeze in very cold
climates. These units also have high energy
requirements. Mechanical aeration systems,
however, are easy to operate and are less susceptible
to clogging from biological growth than PTA.

6.7.4 Catenary Grid


Catenary grid systems are a variation of the PTA
process. The catenary grid directs water through a
series of parabolic wire screens mounted within the
column, above which turbulence is created. The
screens mix the air and water in the same way as the
packing materials in PTA systems. Figure 6-26
depicts a sample unit.

These systems can achieve VOC removal rates


comparable to PTA systems. Catenary grid units
require more airflow and, thus, have higher energy
requirements than PTA systems. They also have
shorter aeration columns with smaller diameters.
Their more compact design lowers their capital cost
relative to PTA.

Catenary grid systems, however, have limitations.


Limited data are available concerning this system's
removal effecti veness for a wide variety of organic
compounds. Also, the procedure for scaling systems
up from pilot plants to full-scale operations is not
fully developed.
Multiple tray aerator; Norwalk, CT.
The principal design considerations for catenary grid
water. Consequently, THM formation potential and systems are air-ta-water ratio, number of screens,
TOX formation potential are lowered. and hydraulic loading rate. Removal efficiency
improves with increases in air·ta-water ratios and
increasing number of screens in the column.
6.7.2 Reverse Osmosis
Reverse osmosis is a proven technology for the
removal of inorganic compounds. The process is fully 6.7.5 Higee Aeration
described in Section 7.3.3. Reverse osmosis also is Higee aeration systems are another variation of the
effective in removing THM humic and fulvic acid PTA process. These systems pump water into the
precursors, pesticides, and microbiological center of a spinning disc of packing material ta
contaminants (viruses, bacteria, and protozoal. This achieve the necessary air and water mix (Glitsch,
treatment has been shown to remove VOCs with low Inc.). By design, the packing material has a large
molecular weights. Table 6-14 presents performance surface area per unit volume. Air is pumped
data for reverse osmosis units operating at about 200 countercurrently toward the center from the outside
~ifur7VOCsandI6S0Cs. of the spinning disc. Simultaneously, water flows
from the center of the disc and mixes with the air (see
Figure 6-27).
6.7.3 Mechanical Aeration Higee units require less packing material than PTA
Mechanical aeration systems use surface or units to attain equivalent removal efficiencies. They
subsurface mechanical stirring mechanisms to create require smaller air volumes and can process high
turbulence to mix air with the water. These systems water flows in a compact space. The Higee unit's
effectively remove VOCs but are generally used for compact size permits its application within
wastewater treatment systems. Surface and constrained spaces and heights. Current Higee
subsurface aerator designs are shown in Figure 6-25 systems are best suited for temporary applications of
(Roberts and Dandliker, 19821. less than I year with capacities up ta 6.3 Usec (loa
124 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Inlel A"

c_~.\~seaJ~~~~
Distributor
Nipples

, I • , •• " ,
_1=:Ic::aCO~c::::::Ie::w:::-
.J>, ,'. , .•
_ _ ccr_-=::a

Staggered --~"H~.at=::l:c=a-c:l-=-CZII~.=::ac:=t==:aCD::=.-=:a-=:~c::a1C:ll
Slat Trays c::::a~I;::Ic:::ac=t.c=.c=::ac:::l~C»c=ac=a
c::.c::ac:::lll::=t.llC:::lc:::a"'GZI~~c=a-=--=a
c:::ac:ac:a..c:a~e::tc=::llcacaC:lClO
. . c::a.IIC:::I.:II'::::'.:::Ic:. __ .-iGCac=c::=ac..
c:ac:a.:::aGOc=:aC3_c=-lIl:tc::::a~.::a
1=:Ie::tc=:tc::ac:::a-=ad3-=- _ _ &:3C111"-
c=.c::ll.::=c:a-=ac=ac::::ac:lc:::ae::-==-
c:ac:::ac:::lc=llo:=sc:::a ... ~c=aCDe:::t~c::a
-=:Il_C=-c:.~..::Io-="c:a-.c:::ac:::.c=­
_c::.c=tc:::ac:ac:::a-=a_,..-=ac:ac:::ac:::a
c::c:::.c::::»e=~c::ae=l=::lc::::l-==~c::a
c=c:::ac:::ac:leE:a_~c::ac::._c=::II~-=::a
c=c:::a.~e::tc::JCZI_~~cm=e::­
c::ac.:.:r.~~c::::ac=.~~c=c::.~~c:::a
-=c=c:::ac::::.c=ac=a-=::a':J.::lIc:ac=~
~c::=:Ic:=c::a~c::a_c=a_cac:::ac:l-=­
.:=.~=c::::ac::::::.-:'_C=-CClG:::I~c::I
c=ac:::a~c:ac:::a~c:aCClc::allClc=:ac:::.e::::t
-=-I::lc=ac;l::llc=a c::.-=--=:. _ _
r='=:Ic::ac=:ac:=ac!::a"~~~CDc:ae­
.............j==r..c::Ilc:.~c:lIIbllll::3~~-=-_c::=.
~e:,~~--;~~~~~~
-- ;; i~ :;~.'~i i; ': ':: "
~. " : ~~ ~ "
- ...,~;- '.U ~ - ",,::'.- ":':-~ .
.. --....
.• _-=---~ •. -:.j '>----- -.::.:..:.

Figure 6·24. Schematic 01 a redwood slat tray aerator.

GPMl. Few data are available concerning the organic


compOund remova I efficiencies of Higee systems.

6.7,6 Resins
In this process, synthetic resins are used in place of
GAC to remove organic compounds by adsorption.
Their performance varies with resin type, EBCT, and
regeneration frequency.

The advantages of resins include shorter EBCT


requirements and longer operational life, relative to
GAC. Also, resins can be regenerated on site with
steam. However, wastewater from this process can be
difficult to manage properly because of its high
concentrations of hazardous constituents (Ruggiero
and Ausubel, 1982).

Resins are more costly than GAC, costing up to


$O.02lg ($10Ilbl, compared to less than $O.Ol/g
($O.80/1b) for GAC.
Treatment of Organic Contami nants 125

Table 6·11. Design Paramete,.. for Peroxld.


Ozone Treatment Plant and Table 6-12. Comparison of Annual Treatment Costs for
AlI8umptlona for Coat Comparisona Removsl of PCE and TCE from Ground Water

Parameter Value Aeration Peroxloe


Plant flow·GPM 2,000 With Gas· LiQulo Ozone
Phase Phase Aovanced
TCE concentration . ~ 200 GAC GAC OXloanon
PCE concentration· IJgIl 20 Cost Type Aeration AOsorption Aosorptlon Process
Capital cost (annual- 48,400 108,000 192,500 35,000
Reacnon lank capacity· gal 6,000 ized, 10 years, 8%)

Hydraulic oetention time - min 3 Operating ano 30,200 78,300 13,900 63,900
maintenance costs
Reaction tank stages, number
GAC replacement NA 105,100 210,400 NA
costs
Ozone oosage . mgll 4
Total annualized 78,600 291,400 416,800 98,900
Ozone generator capacity· Ibloay 100 costs
Peroxloe oosage . mgll 2 Dollar cost per 1,000 (0.075) (0.277) (0.397) (0.094)
gallons (based on
Peroxloe storage' gal at 50 percent 1,000 2,000-gallon flow)
concentratton
NA • not available.
1,000 gallons • 3.78 m3
Assumptions Source: A,eta et aI. (1988).
The treatment plant 's to be bUilt for one well With a capacity of
2,000 GPM (126 Lis).
• The TCE concentranon of the grouno water being pumped
remains constant for 10 years. At the eno of thiS time, all Table 6·13. Typical Performance of Ozonatlon Process

the TCE has been extracted ano the treatment plant 's no Compouno Dosage (mgll) Percent Removal
longer needed. SOCs
• The rate of interest remains at 8 percent for 10 years. Carboluran 9 lOa
• The capital cost of the aqueous·phase GAC treatment plant 1,2, Dichloropropane 0.9-6.0 8-22
IS $1.3 million. CIS-l,2 2-10 87·93
• The capital cost of the air·stripping treatment plant alone is Dichloroethylene
$325,000. trans-l,2 0.9-6.0 100
• The capital cost of the treatment plant for air stripping plus Dichloroethylene
gas-phase GAC adsorption IS $725,000. a-DIChlorobenzene 9 88
• The annual operating cost of each system is the power Ethylene oibromide 0.9-6.0 8·9
reQUIrement plus one Quarter of a person-year for Ethylbenzene 1.5-9.0 47·95
mamtenance. Heptachlor 17 100'
• The cost of power 's $1.08Ab. Heptachlor epoxlde 17 26
• The cost of hyorogen peroxloe IS $1 .08Ab. Lindane 0.4-150.0 0-100
1 pouno • 453.6 grams
1 gallon • 3.785 liters Monochlorobenzene 0.4-6.0 86·98
Source: A,eta et aI. (1988a). Toluene 1.5·12.0 49·98
Xylenes 1.5-12.0 54·98
• Oxidation 01 heptachlor produces heptachlor epoxioe, which IS
relatively stable to further OXidation.
Source: Miltner and Fronk (1985).
126 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Drive
Tabl.6·'4. Typiclll PerfonNlne8 of R.v..... a.....oala
Procee.
o
Compound Percent Removal
VOCs
1.2-Dichloroethane 15-70
1,1. 1-Trichloroethane 15-100
Carbon tetrachloride 95

Tnchloroethylene 0-75
MechanICal Surface Aerator
Tetrachloroethytene 70-90
Benzene 2-18 Drive
.fJ.~ rl=compressor
~.
p-D,chlorobenzene 0-10
:
sacs "..
Acrylam,de 0-97
Aldicarb 94·99
Alachlor 100
Turbine ----e~
!Air

Carbofuran 88·99 ~:J


Sparger
'"
1.2-DlChloropropane 10·90
Submerged Turtlil'8 Aerator
C/S·l ,2 -Dichloroethytene 0·30
Source: Roberts and Dand~ker (1982).
Irans-l.2-Dichloroethytene 0·30
2.4-0 1-65
0- DichlOrobenzene 65

Ethytbenzene 30
Ethylene Dibromide 37-84
Lindane 50-75

Methoxychlor >90
Monochlorot>enzene SO·loo
PCBs 95

Xytenes 10-85

Source: Miltner and Fronk (1985).


Treatment of Organic Contaminants 127

FlUidIZed Zone

Treated Water Raw Water


Sample Collector Rotameter

Manometer for
Air Flowrate Measurement

Treated Water
Sample Tap

Treated Waler to Drain

Figure 6-26. Catenary grid s~tem.


128 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Pilot catenary grid aerator unit.

Air In
t Exhaust AI(

Ground Water
Blower

r
Filter

Figure 6-27. Schematic of HigH system.


7. Treatments for Inorganic Contaminants

This chapter describes the technologies available for corrosion by-product. Lead, when found in source
removing inorganic contaminants from drinking water, requires conventional or other treatment such
water. Inorganic contaminants tend to come from as reverse osmosis and ion exchange.
natural sources and, therefore, are more predictable
in behavior and composition than organic Twelve inorganic contaminants are presently
contaminants that come from manmade sources. regulated under SOWA: lead, radium. nitrate,
Inorganic contaminants are classified as cationic, arsenic, selenium, barium, fluoride, cadmium,
anionic, or neutral and occur as atoms, ions, or chromium (total), mercury, silver, and strontium-9C.
molecules. The two most significant properties Lead is not typically found in source water, but
affecting the removal of inorganic contaminants from rather at the consumer's tap as a result of corrosion of
water are valence and solubility in water. the plumbing or distribution system. The other
contaminants are found regionally for a variety of
The most common treatments for inorganic reasons. Radionuclides are of particular concern
contaminants are: because few utilities currently test for other than
radium and EPA has found them to be more
• Precipitation prevalent than previously determined.
• Coprecipitation
• Adsorption This chapter presents two distinct approaches to
• Ion exchange water treatment for inorganic contamination: (1)
• ~embrane separation by reverse 03mosis or preventing inorganic contamination of finished
electrodialysis water and/or (2) removing inorganic contaminants
• A combination of two or more of the above five from raw water. Section 7.1 discusses corrosion
technologies control, the primary method for minimizing
inorganic contamination of finished water. The major
Other treatments for removing inorganic focus of this subsection is lead. Section 7.2 describes
contaminants include: treatment options for removing inorganic
contaminants from raw water. Appendix 0 provides
• Distillation case histories on corrosion control and treatment
• Evaporation methods to remove inorganics.
• Oxidation/Reduction
• Air stripping
• Biological treatment 7.1 Techniques for Controlling
Corrosion
EPA has specified best available technologies (BAT) Lead is found in drinking water chiefly because the
for three inorganic contaminants, including: metals used in water distribution systems, home
plumbing, and appliances, such as water coolers,
• Corrosion control for lead corrode. These metals generally are not found in
• [on exchange or reverse osmosis for nitrate significant amounts in source waters, but rather at
• [on exchange and reverse osmosis for radium the final point of use. Zinc, copper, and iron also are
commonly present in water as corrosion products.
It is expected that aeration will be designated as BAT Cadmium has been found in water systems with new
for radon. Corrosion control targets lead as a galvanized pipe. Lead and cadmium are toxic at low

129
130 Upgrading Existing or Designing New Drinking Water Treatment Facilities

concentrations; copper and zinc are toxic only at For example, pH is lowered with the use of
much higher levels. Because of its ubiquity and coagulants or disinfectants. Other treatment
toxicity, lead is the corrosion product of most con'.:ern. processes affect water chemistry parameters such as
Cadmium also has significant health implications, disinfectant residual, hardness, and alkalinity.
but it is not as widespread in drinking water as lead.
FaC1Of11 AffllC1lng the Corroalvlty of Drinking
Lead levels in drinking water are minimized with Water
corrosion controls. The current lead MCL applies to
finished water from a treatment plant. New lead Factor EHeet on Corrosivlty
regulations may include an MCL at the consumer's pH Low pHs generally accelerate corros,on.
tap. While corrosion is recognized as a critical factor Dissolved Dissolved oxygen In water induces active corrOSIon.
for proper management of any water system, the lack oxygen pancularly of ferrous and copper materials.
of a standard measure of corrosion has thwarted Free chlonne The presence of free chloflne In water promOles
development of uniform controls. reSIdual corrOSion of ferrous metals and copper.
Low There IS InsuffiCIent alkahnlty to limit corrOSion
buHering actiVIty.
7.1.1 The Problem of Corrosion capacIty
Corrosion occurs because metals tend to oxidize when HIgh A molar ratiO of strong m,neral acIds much above 0.5
in contact with potable water and form stable solids halogen and results In conditions favorable to pItting corrOSion
suHate- (mostly in iron and copper pipe).
on metal surfaces. All metals in contact with water alkallfllty
will corrode to some extent. Corrosion has ra~o
implications for health, costs, and aesthetics. Total HIgher concentralions of dIssolved salts ,ncrease
Drinking water contaminated with metals adversely dissolved conducbvlty and may increase corrOSiveness.
affects human health. Corrosion reduces the useful solids ConductiVIty measurements may be used to estimate
life of water distribution systems and household total dIssolved solLds.
plumbing, and is thus responsible for higher costs Calcium Calcium can reduce corrOSIon by forming protective
due to problems with: films WIth dIssolved carbonate. particularly With steel.
"on. or galvanized pIpe.
• Pumping caused by narrowed pipe diameters Tannins Tanmns may form protective organic films over
resulting from corrosion deposits metals.

• Pumping and water production caused by Flow rates Turbulence at hIgh «ow rates allows oxygen to reach
corrosion holes, which reduce water pressure and the surface mere easily. removes protecllve films.
and causes higher corrOSIon rates.
increase the amount of finished water required to
deliver a gallon of water to the point of Metal ions .Certain ions. such as copper. can aggravate
corrOSIOn of downstream materIals. For example.
consumption copper Ions may Increase the corrOSiOn 01 gaJvanlzed
• Water damage caused by corrosion-related pipe pIpe.
failures Temperature High temperature Increases corrOSIOn reaction rates.
• Replacement frequency of hot water heaters, Rates HlQh temperature also lOwers the soIubolity of calcium
radiators, valves. pipes, and meters carbonate. magneSIum slhcates. and calcIum sulfate
• Customer complaints of color, staining, and taste and tt1us may cause SCale lormabon ,n hot·water
healers and popes.
problems
• Repairs for pipe leaks and breaks Source: Adapted from WIlliams (1986).

Lastly, corrosion can produce conditions that promote


microorganisms that cause disagreeable tastes. The type of corrosion products present depends on the
odors, slimes, and further corrosion. metals composing the solder, pipes, valves, meters,
and faucets in distribution and plumbing systems.
All water is corrosi ve to some degree, but water that The most common metals used are steel, iron,
is acidic will have faster corrosion rates. Many galvanized steel, copper, lead, brass, and bronze.
naturally occurring acidic waters are also soft. Soft Zinc, cadmium, and some lead are present in
water is generally defined as water with less than galvanized coatings. The most frequent sources of
100 mglL of calcium as calcium carbonate (CaC03); lead are brass faucets, meters, pipe solder, and
acidic water has a pH of less than 7.0. valves. Lead is no longer generally used for pipes, but
in older cities service connections between houses
The degree of corrosion is determined primarily by and water mains still contain lead pipe. Many cities
the characteristics of the metal and water. and the have initiated lead service line replacement
nature and duration of the contact between the two. programs to address this source.
Table 7-1 summarizes the factors affecting drinking
water corrosivity. Water treatment processes can While lead pipes have long been recognized as
change water quality characteristics that hazardous, lead-based solder was used in the United
significantly affect the water's corrosion potential.
Treatments for Inorganic Contaminants 131

States until it was banned in the 1986 SDWA dissol ved ions, and the absence of corrosion
Amendments. This ban prohibits the use of solder inhibitors.
containing lead and the associated high lead levels
often found in newly constructed homes and in new • Erosion - Occurs due to the removal of protective
plumbing in existing homes. Precise estimates of lead coatings through high water velocities,
levels in drinking water resulting from lead-based turbulence, sudden changes in flow direction, and
solders and fluxes vary. This stems from the abrasive action of suspended solids or gases.
difficulty in consistently measuring lead levels from
the tap, especially if there is a brass faucet, which • Biological- Occurs in mechanical crevices or
contributes to lead levels as well. The method of accumulations of corroded materials due to the
solder application affects the amount of lead interaction between the metal and bacteria,
imparted to the water. Improper application, which is algae, or fungi.
difficult to determine in retrospect, allows solder to
flow on to the inner portion of the pipe, thus Most lead corrosion problems are diagnosed
increasing its area of contact with the water. indirectly. Risk factors that indicate potentially high
lead levels at the tap are:
7.1.2 Diagnosing and Evaluating the Problem • The water distribution system or structure's
There are dozens of types of corrosion, but the two plumbing is made of lead.
broad categories of particular concern in water • The structure's plumbing has solder containing
treatment are uniform and nonuniform corrosion. lead.
Nonuniform corrosion shortens the useful life of pipes • The structure is less than 5 years old.
and plumbing more quickly than uniform corrosion, • The tap water is soft and acidic.
and is thus more ofa problem for water systems. The • The water stays in the plumbing for 6 or more
five most common types of nonuniform corrosion to hours.
affect water systems are: • The structure's e lectrica I system is grounded to
the plumbing system.
• Galuanic - Occurs when two different metals are
joined together; the more electrochc mically The presence of any of these factors justifies further
acti ve of the joined metals will corrode. (See investigation.
Table 7-2 for relative activity of various metals
and Section 7.1.3.1 for discussion of this One direct method of measuring lead corrosivity of
characteristic. ) water is to sample standing water from the
consumer's tap. In addition, consumer complaints,
Table 7·2. Galvanic Series· Order 01 Electrochemical corrosion indices, sampling and chemical analysis,
Activity 01 Common Metals Used In Water pipe scale examination, and measurement of the
Distribution Systems corrosion rate over time are all proxies for corrosion
Melal ACbv,1y contaminants.
Zinc More Active
Mild steel 7.1.2.1 Consumer Complaints
Cast ,ron
Many times a consumer complaint is the first
Lead
indication of a corrosion problem. Table 7-3 lists the
Brass
Copper
Stamless steel
1
Less Active
most common complaints and their causes. In
investigating the extent of the corrosion, complaints
can be plotted on a map of a water service area. Then
investigators can examine the construction materials
Source: U.S. EPA (1984).
used in the water distribution system and in the
plumbing of the complaint areas. Random sample
surveys are commonly used to confirm the extent of
• Pitting - Occurs as uneven pits or holes in the the corrosion problems flagged by these complaints.
pipe surface, which are undetectable until the
pipe fails. Pitting is usually caused by pockets of
corrosion initiated by tiny imperfections, 7.1.2.2Corrosion Indices
scratches, or surface deposits in the pipe. Corrosion related to calcium carbonate deposition can
be estimated using indices derived from common
• Creuice - Occurs locally around gaskets, lap water quality measures. The Langelier Saturation
joints, rivets, and surface deposits. It is caused by Index (LSI) is the most commonly used and is equal to
changes in acidity, oxygen concentrations, the water pH minus the saturation pH. The
saturation pH refers to the pH at the water's calcium
carbonate saturation point, the point where calcium
132 Upgrading Existing or Designing New Drinking Water Treatment Facilities

locations or getting accurate and precise samples


Table 7-3. Typical Cuatomer COmplalnta Due to Corroafon
representing water in prolonged contact with the
Customer Campl3lnt Possible Cause suspected section of pipe (e.g., service line, soldered
Red water or reddish-brown Corrosion of iron PIpes or joints). Experiments show that sample volume also
staining of fixtures and 'aundry presence of natural iron in raw affects the amount oflead detected in tap water.
water
Other critical sampling elements include sampling
Bluish stains on fixtures Corrosion of copper hnes location, amount of water in each sample, flow rate of
Black water Sulfide corrOSIOn of copper or sample, and the contact time of the water with the
iron lines or preciPItations of metal. The amount of time the water remains in the
natural manganese
pipes significantly affects lead levels. Also, the time
Foul taste and/or odors By'products from microbial of day is a critical factor in accurately assessing
actiVIty
sampling results. Early morning sample r~sults will
loss of pressure Excessrve scaling. tubercle reflect water that has been held in the pipes
build-up from pitting corrosion.
leal< ,n system from PItting or
overnight, while evening water samples may assess
other type of corrOSion water intermittently drawn by daily activities.
lack of hot water Buildup of mineral deptQsits in
hot water system (can be The age of the pipe solder has been shown to affect
reduced by S8l\Jng lIlermastats lead levels in water. One study found that 4 to 5
to under 60'C (140'Fll weeks after an application of lead· based solder, lead
Short service hfe of household Rapid detenoraoon of pipeS levels declined by 93 percent from initial
plumbing 'rom Pitting or olller types of measurements. The effects of solder age and the
corrosion
---.'-------------------
Source: Adapted from V,S. EPA (1984).
length of time the water stands in the pipes are
illustrated in Table 7-4, Table 7-5, and Table 7·6 for
low, medium, and high water pH levels, respectively.

carbonate is neither deposited nor dissolved. Calcium Table 7·4. Percentalle 01 Test Slt81 with lead in Drinking
carbonate may precipitate and form a protective Wlter Greater than 20 I'~L at low pH (6.4 and
layer on metals. The saturation pH is related to the '. .a)
water's calcium ion concentration, alkalinity, Age of
temperature, pH, and presence of other dissolved Test
Site First 10 20 30 45 60 90 120
solids, such as chlorides and sulfates. (Years) Draw Sec Sec Sec Sec Sec Sec Sec
0-1 1000/. 1000/. 1000/. 100"" 1000/. 860/. 86% 88%
The Aggressive Index (AI) is a simplification of the
1·2 100 71 86 57 57 29 43 14
LSI that only approximates the solubility of calcium
carbonate and may not be useful. The Ryznar 2·3 86 86 57 57 43 43 43 29
Stability Index (RSI), also a modification of the LSI, 3-4 100 86 100 71 71 71 29 29
uses visual inspections. McCauley's Driving Force 4·5 86 57 29 43 43 43 14 0
Index (DFIl estimates the amount of calcium 6·7 78 44 33 33 11 II 11 0
carbonate that will precipitate based on the same
factors as the LSI. 9·10 71 29 14 14 14 14 a a
15·16 57 14 14 14 14 14 14 14
Riddick's Corrosion Index (Cil is distinct from the 20 and 86 27 29 0 14 a 14 0
LSI. Its empirical equation incorporates different older
factors to predict corrosion, such as dissol ved oxygen,
chloride ion, noncarbonate hardness, and silica.
Appendix E provides the equations, descriptions of
test parameters, and interpretations of each index.
7.1.2.4 SCale or Pipe Surface Examination
7.1.2.3 Sampling and Chemical Analysis Pipe scale and the inner surface of pipes can be
examined by optical or microscopic observation, x-
Corrosion can also be assessed by conducting a rays, wet chemical analysis, and Raman and infrared
chemical sampling program. Water with a low pH spectroscopy. Obviously the most practical and
(less than 6) is more corrosive. Temperature and total economic method is simple observation. The other
dissolved solids can be important indicators of methods require expensive equipment and highly
corrosivity, although this varies case by case. skilled personnel. Even simple observation, however,
requires taking sections of pipe out of service.
Proper sampling and analysis methods are essential Chemical examinations can determine the
to obtain accurate and meaningful test results. composition of pipe scale, such as the proportion of
Determining whether lead is from a service pipe or calcium carbonate present that shields pipes from
internal plumbing requires sampling at multiple dissolved oxygen and thus reduces corrosion.
Treatments for Inorganic Contaminants 133

nat coupons are commonly placed in the middle of a


Table 7·5. Percentage of Te.t Sit.. with Le.d In Drinking
Water Gre.ter thJIn 20 IlWL at Medium pH
pipe in the system. The coupons are weighed before
(7.0·7.4) and after they are put in place and the recorded
Age of weight loss is translated into a uniform corrosion rate
Test with the following formula:
Site First 10 20 30 45 60 90 120
(Years) Draw Sec Sec Sec Sec Sec Sec Sec
534 X W
0-1 1000/. 90% 90% 60% 30% 20% 10% 10% Rate ofCorrosion (MPY) =
1·2 80 60 30 10 20 a 10 a DxAxT
where:
2·3 40 20 10 10 10 a a 0
3-4 50 20 20 30 20 30 30 20 W = weight loss in mg
4·5 30 10 10 0 10 0 a 0 o = density of the specimen in glcm3
6-7 10 0 0 a 0 0 0 0 A = surface area of the specimen in sq in
9·10 20 0 0 a 0 0 a 0 T = exposure time in hours
15-16 40 20 20 10 0 0 0 0
This method can be used to monitor corrosion
20 and 20 0 a 0 10 0 0 0
progress over time or to spot check the corrosion rate.
older

The loop system weight-loss method is similar to the


nat coupon method, except that it uses actual sections
of pipe in the system instead of nat coupons. Loop
system corrosion rates are usually determined by
Tabfe 7-6. Percentage of T..t Sit.. with Lead In Drinking
Wat.r Gr.ater than 20 Il~L at High pH
analyzing corrosion over a period of time.
(8.0 and Great.r)
Age of Electrochemical rate measurement requires
Test expensive and sophisticated equipment beyond the
Site First 10 20 30 45 60 90 120 means of most smaller systems. In this method, two
(Years) Draw Sec Sec Sec Sec Sec Sec Sec
or three electrodes are placed in the corrosive
0·1 100% 100% 60% 10% 20% 10". 20% 0% environment, and instrumentation measures the
1·2 67 22 11 a
2-3 30 10
"
10 "a 0 0
"a 0
0
corrosion rate in MPY.

3·4 25 0 0 0 a a 0 13
7.1.3 Corrosion Controls
4·5 30 10 a 0 a a 0 0
If any of the tests described above reveal
6·7 20 0 0 0 0 0 0 0
unacceptably high levels oflead, then immediate
9-10 10 0 10 0 0 10 0 10 measures should be taken to minimize human
15·16 33 22 11 0 0 0 0 exposure until a long-term action plan is developed,
20 and 20 0
"0 a 0 0 0 0 approved, and implemented. Some short-term
older measures for water consumers are:

• Running the water for 1 to 3 minutes before each


use
Chemical analyses and microscope assistance then • Using only cold tap water for drinking or cooking
should be used to augment these findings. • Using a home-scale reverse osmosis unit or other
treatment process, in extreme cases
• Using bottled water
7.1.2.5 Rate Measurements
The corrosion rate is usually expressed in thousands Longer term solutions fall into six categories:
of an inch (mils) per year (MPYl. The three methods
for calculating this rate are: • Distribution and plumbing system design
considerations
• Water quality modification
• Coupon weight-loss
• Corrosion inhibitors
• Loop system concentration increase
• Electrochemical rate measurement • Coatings and linings
• Use of only nonlead based solder for construction
and repairs
The first two methods are much less expensive than
the third. • Replacement of lead pipes

The coupon weight-loss method uses nat metal pieces


or test sections of pipes called coupons or inserts. The
134 Upgrading Existing or Designing New Drinking Water Treatment Facilities

7.1.3.1 Distribution and Plumbing System Table 7·7. Corrosion Properties of Frequently Used
Design Considerations Matenals in Water Distribution Systems
Many distribution and plumbing system design DistnbutJon ASSOCiated Potential
considerations will reduce corrosion. For example, Matelial CorroSion ReSistance Contaminants
water distribution systems designed to operate with Copper. Good overall corrosion Copper and possibly Zinc.
iron reSistance; sublect to tin, arseniC, cadmium, and
lower flow rates will have reduced turbulence and corrOSive attack 'rom high lead from associated pipes
decreased erosion of protective layers. Other velocities. soft water. chlor· and solder
measures that minimize corrosion include: Ine. dissolved oxygen, and
low pH
• Using only lead-free pipes, fittings, and lead Corrodes In water With low lead (can be well above
components pH and high alkalinl~es MCla). arseniC. and
cadmium
• Selecting appropriate system shape and
geometry Mild steel SubJect to uniform Iron. resulting In turbidity
corrOSion: aHected and red· water complamts
• Avoiding sharp turns and elbows pnmarlly by high dissolved
• Avoiding dead ends and stagnant areas oxygen levelS
• Eliminating shielded areas Cast Of Can be subject to suriace Iron, resulting In lurb/dlty
• Providing adequate drainage duc~1e "on erOSion by aggressive and red·water complaints
• Selecting appropriate metal thickness (unlined) waters
• Using welded ends, instead of rivets Galvanized Sublectto galvaniC carro· Zinc and Iron: cadmium
• Reducing mechanical stresses Iron sian of Zinc by aggressive and lead (Impurilles ,n
• Avoiding uneven heat distribution waters; corroSion IS galvanizing process may
• Providing adequate insulation accelerated by contact exceed plimary MCls)
WIth copper
• Providing easy access for inspection, materrafs;corroslon !s
maintenance, and replacement accelerated at higher
• Eliminating the grounding of electrical circuits temperatures as In hot·
to the system waler syslems
Asbestos· Good corrOSion reSistance: Asbestos frbers
To implement these measures effectively, local cement immune to electrolySIS;
pi umbing codes may need to be modified. aggreSSive walers can
leach calCium from cement
Distribution system designers should base their Plas~c ReSistant to corrOSion
materials selection criteria on system water Brass Farrty good; subject to lead. copper. zmc.
characteristics. For example, water with low pH dezlnCl'icalJOn depending arseniC
on the waler quality and
levels and high dissolved oxygen levels corrodes alloy
metals quickly. Concrete may be more appropriate
than metal in this case because, although concrete "MCl • Maximum contamlnanllevel.
Source: Adapted from U.S. EPA (1984).
dissolves under low pH conditions, the rise in
dissol ved calcium in the water is not objectionable
and since concrete pipes are thick they have a longer lead. Alternative lead-free solders are available at a
lifetime. Table 7-7 presents the corrosion properties higher cost.
of common distribution system materials.
Two alternative solders are made from 95 percent tin
Metal electrochemical activity is the measure of a and 5 percent antimony or silver. The 1984 costs per
metal's tendency to oxidize. When placed together, gram of standard lead-; antimony-; and silver-based
metals with different activities create galvanic solders were $9.92, $17.64, and $44.09, respectively
corrosion cells. Therefore, if different metals must be ($4.50, $8.00, and $20.00Ilb, respectively!. Plumbers
placed together, minimizing the differences in their estimate that plumbing in an average residence
activity will lessen corrosion. Galvanic corrosion is requires less than lib of solder, which makes solder
also avoided by placing dielectric insulating costs relatively insignificant. However, these
couplings between the dissimilar metals. alternative solders do not perform exactly the same
as lead-based solder. For instance, both silver-tin and
While many corrosion control programs address the antimony-tin solders are more difficult to work with
lead content of pipes, solder, and other plumbing than lead solder because of higher and more narrow
components, they do not target lead-bearing melting ranges.
plumbing fixtures, such as brass faucets or valves.
7.1.3.2Water Quality Modifications
Changing the composition of solder used in plumbing
is considered the most important factor in any lead Measures that change water quality to reduce
corrosion control strategy, In the past, solder used in corrosion contamination, especially lead corrosion,
plumbing has been 50 percent tin and 50 percent include pH and alkalinity adjustment, lime-soda
Treatments for Inorganic Contaminants 135

softening, and adjustment of dissolved oxygen levels, water's pH and carbonate levels. Increasing pH levels
although altering oxygen levels is not a common reduces the presence of hydrogen ions and increases
method of control. Any corrosion adjustment program the presence of hydroxide ions. Hydroxide ions
should include a monitoring component that enables decrease lead solubility by promoting the formation
dosage modification in response to changing water of solid basic lead carbonates that passivate the
characteristics over time. surface of the pipe. Similarly, elevating carbonate to
a certain level increases the presence of carbonate
ions, which, in turn, boost the presence of basic lead
pH Adjustment carbonate or lead carbonate. However, continuing to
Frequently corrosion is mistakenly associated merely increase carbonate levels at a pH above about 7.5 will
with water acidity. The pH level plays a central role increase lead solubility (see Figure 7-ll. If carbonate
in determining the corrosion rate and is relatively levels are the only other factors under consideration,
inexpensive to control. Generally, water pH less than then the optimal pH range to avoid lead corrosion is
6.5 is associated with uniform corrosion, while pHs between 9.0 and 10.0.
between 6.5 and 8.0 can be associated with pitting
corrosion. Systems using only pH to control corrosion
should maintain their water above 9.0 pH to reduce
the availability of hydrogen ions as electron
acceptors. However, pH is not the only factor in the
corrosion equation; carbonate and alkalinity levels
may affect corrosion as well.

By adjusting pH, alkalinity, and calcium levels, ~150


operators can promote the precipitation of a §
protective calcium carbonate (scale) coating onto the
metal surface of plumbing. While these protective
coatings build up over time, they are disturbed by:

• Water turbulence due to high velocities and flow


I
(J
~ 100

rates
• Stagnation of water causing prolonged periods of
contact
• Variations in the control of velocities, flow rates, "mgIL. as CaCO,
and stagnation 50 L-_---'-_ _....L..._ _.L..-_---'-_ _""'--_----'
8 9 10 11
Calcium carbonate scaling occurs when water is pH
oversaturated with calcium carbonate. Below the
saturation point, calcium carbonate will redissolve
and at saturation, calcium carbonate is neither Figure 7·1. SOlubility 01 lead aa a lunC1lon 01 pH and
carbonate.
precipitated nor dissolved. The saturation point of Source: Adapted 'rom SChock and Wagner (1985).
any particular water source depends on the
concentration of calcium ions, alkalinity, Usually water carbonate levels are directly related to
temperature, and pH, and the presence of other water alkalinity because alkalinity refers to the
dissolved materials, such as phosphates, sulfates, water's ability to neutralize acids and the most
chlorides, sulfides, and some trace metals. A coating common base is dissolved inorganic carbon
is usually attained through simple pH control, given (carbonate species). At drinking water treatment
adequate concentrations of alkalinity and carbonate. plants, neutralization is performed primarily with
Adding lime or alkaline substances also promotes the calcium and sodium hydroxides, although carbonates
formation of calcium carbonate scale in most may also be added. Carbon dioxide and water form
systems. The Langelier or CCAP Indices are useful bicarbonate as the pH increases; when the pH
for initiating this corrosion inhibiting strategy reaches 10.3, the carbonic species is expressed as
because they estimate calcium carbonate saturation. carbonate. Coordinating pH level with alkalinity
Further pH and/or carbonate and/or calcium supplementation may reduce lead corrosion.
adjustments may be required to reach equilibrium, as Optimum alkalinity levels are 30 to 60 mglL as
established by system monitoring. calcium carbonate (CaC03).

Adjustment of both pH and alkalinity using lime-


Ume Softening soda softening is an effective method for controlling
Lime softening, which is sometimes known as lime- lead corrosion. However, optimum water quality for
soda softening when soda ash is required in addition corrosion control may not coincide with optimum
to lime, affects lead's solubility by changing the hardness reduction. Water hardness is due mostly to
136 Upgrading EXisting or Designing New Drinking Water Treatment Facilities

the presence of calcium and magnesium ions. Lime- indicate that some natural organic r.ompounds
soda softening reduces the presence of these ions by increase lead solubility; however, others have shown
adding hydrated lime (Ca(OH)2), caustic soda that tannins reduce corrosion by forming protective
(NaOH), or soda ash (sodium carbonate) under coatings on the metal surface.
certain water quality conditions. This technique has
been used to address water hardness, rather than
corrosion. Comprehensive water treatment, however, 7.1.3.3 Corrosion Inhibitors
must balance water hardness, carbonate, and Corrosion inhibitors form, or cause to form, protective
alkalinity, as well as corrosive potential. coatings on pipes that reduce corrosion, but may not
totally arrest it. The success of any corrosion
In addition, pH levels that are well suited for inhibitor hinges on the treatment plant operator's
corrosion control may not be optimum for other water ability to:
treatment processes, such as coagulation or
disinfection. To a void this type of conflict, the pH • Apply double and triple ultimate dosages of
level should be adjusted for corrosion control inhibitor during initial applications to build a
immediately prior to water distribution, but after the base protective coat to prevent pitting. Typical
other water treatment requirements have been initial coatings take several weeks to form.
satisfied. • Maintain continuous and sufficiently high
inhibitor dosages to prevent redissolving.
• Attain a steady water flow over all the system's
Oxygen Levels metal surfaces for a continuous application of the
The other major water quality factor affecting inhibitor onto all exposed surfaces in the system.
corrosion is the presence of excessive dissolved
oxygen. Dissolved oxygen increases water's corrosive There are several hundred commercially available
activity by providing a potent electron acceptor. The corrosion inhibitors. Among the most common for
optimum level of dissoIved oxygen for corrosion potable water supply stems are:
control is 0.5 to 2.0 ppm. However, removing oxygen
from water is not practical because of cost. The most • Inorganic phosphates
reasonable strategy is to minimize the presence of • Sodium silicates
oxygen. • Mixtures of phosphates and silicates

Minimizing dissolved oxygen levels is effective for These corrosion inhibitors can be applied with
ground-water supplies. Ground water is sometimes normal chemical feed systems.
aerated prior to treatment to address high levels of
iron, hydrogen sulfide, and carbon dioxide. This
aeration step eliminates free carbon dioxide, which, Inorganic Phosphates
in turn, reduces the amount oflime necessary in Inorganic phosphate corrosion inhibitors include
lime-soda softening or for pH control; however, polyphosphates, orthophosphates, glassy phosphates,
aeration increases corrosion by increasing dissolved and bimetallic phosphates. Zinc, added in conjunction
oxygen. Consequently, excluding the aeration step with polyphosphates or orthophosphates, also helps
and increasing lime softening can effectively reduce inhibit corrosion in some cases. Phosphates can
unnecessary and potentially counterproductive high inhibit excessive calcium carbonate scale formation,
oxygen concentrations. form a protective coating with the pipe metal, or
prevent aesthetically objectionable corrosion by-
Oxygen requirements are also reduced by extending products. Water characteristics that affect the
the detention periods for treated water in reservoirs. efficiency of phosphate corrosion inhibition include:
Longer detention times allow the oxidation of
hydrogen sulfide and organic carbon at the water • Flow velocity
surface. In addition, correct sizing of the water pumps • Phosphate concentration
used in the treatment plant minimizes the • Temperature
introduction of air during pumping.
• pH
• Calcium
• Carbonate levels
Other Modifications
The potential of other substances to affect lead's The effectiveness of potential phosphate inhibitors
solubility has been investigated. Tests show that should be confirmed by laboratory and field tests.
adding sulfate, chlorine, and nitrate to water has
little impact on lead solubility and, thus, corrosion. Glassy phosphates, such as sodium
However, chlorine lowers pH, and chloramines (as hexametaphosphate, effectively reduce iron corrosion
opposed to free chlorine) have been shown to increase at dosages of20 to 40 mg/L. At lower dosages, this
the solubility of lead-based solder. A few studies glassy phosphate may merely mask corrosion by
Treatments for Inorganic Contaminants 137

eliminating the red color associated with iron then corrodes in place of the protected metal.
corrosion. Under some circumstances, adding zinc in Galvanizing is a form of cathodic protection.
dosages of about 2 mg/L improves phosphate's
corrosion control. Zinc phosphate also has been used
effectively to inhibit corrosion. 7.1.3.5 Coatings and LJnlngs
Mechanically applied coatings and linings differ for
Some studies show that orthophosphate is an pipes and water storage tanks. They usually are
effective lead corrosion inhibitor within specific applied prior to installation, although some pipes can
ranges of carbonate and hydrogen ion concentrations. be lined after installation. While coal-tar products for
Polyphosphates have not been shown to be more pipes and tanks have been widely used for linings,
effective than simple orthophosphates, and adding they are under regulatory scrutiny due to the
zinc sulfate has not improved polyphosphates' presence of polynuclear aromatic hydrocarbons in
performance. In addition, polyphosphates can their composition.
increase lead solubility in the absence of
orthophosphate. The most common pipe linings are coal-tar enamels,
epoxy paints, cement mortar, and polyethylene.
Table 7-8 summarizes the advantages and
Silicates disadvantages of these four primary pipe coatings
Sodium silicates have been used for over 50 years to and linings.
inhibit corrosion, yet the process by which they form
a protective layer is not completely understood. The The most common types of water storage tank
effectiveness of sodium silicates depends on pH and coatings and linings include coal-tar paints and
carbonate concentration. Sodium silicates are enamels, vinyls, and epoxy. Eleven of these coatings
particularly effective for systems with high water and linings are described in Table 7-9.
velocities, low hardness, low alkalinity, and pH of
less than 8.4. Typical coating maintenance dosages of
sodium silicates range from 2 to 12 mg/L. They offer 7.2 Treatment Technologies for
advantages in hot-water systems because of their Controlling Inorganic Contaminants,
chemical stability, as contrasted with many Including Radlonuclides
phosphates. Inorganic contamination of raw drinking water
supplies occurs from a wide variety of sources.
While silicates are hypothesized to inhibit lead Contaminants include metal ions, nitrate, fluoride,
corrosion, little data has been gathered to evaluate and radioactive substances; the most common
this theory. contaminants are fluoride, arsenic, and nitrate.
Contaminants that pose significant problems on a
7.1.3.4 Cathodic Protection regional basis are selenium, barium, and radium-
226. Inorganic contaminants are grouped as either
Cathodic protection is an electrical method of naturally occurring or anthropogenic (resulting from
inhibiting corrosion. However, this expensive human activities). Many naturally occurring
corrosion control method is not practical or effective inorganics, such as fluoride, arsenic, selenium, and
for protecting entire water systems. It is used radium-226, are commonly found in ground-water
primarily to protect water storage tanks when a sources. Anthropogenic contaminants are usually
minimum metal surface area is exposed. found in surface water supplies. For example,
nitrates and nitrites are a problem in agricultural
Metallic corrosion occurs when contact between a areas or areas without sanitary sewer systems.
metal and an electrically conductive solution
produces a flow of electrons (or current) from the Contaminant type, valence, and solubility determine
metal to the solution. The electrons given up by the the most appropriate removal mechanism. Each
metal cause the metal to corrode rather than remain removal technology uses a distinct mechanism or
in its pure metallic form. Cathodic protection series of mechanisms to remove inorganic
overcomes this current with an external power contaminants. Valence is particularly important to
source. The electrons provided by the external power the removal efficiency of lime softening and
source prevent the metal from losing electrons, coagulation. The primary factors in selecting an
forcing it to be a cathode. This reaction allows the inorganic removal technology are:
metal to remain in its less stable metallic form.
• Contaminant type(s) and valence(s)
Cathodic protection is accomplished either by • Influent contaminant concentrations, because
inserting electrodes into the metal to induce an many technologies remove fixed percentages of
external current or by placing a sacrificial galvanic contamination
anode in the current system. The sacrificial anode • Desired effiuent contaminant concentrations
138 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Table 7-8. Pipe Wall Linings


Matenal Use Advantages Disadvantages
Hot applied Lining lor steel pipes (used in 50 to Long seMce life ( > 50 years) Need to reapply to welded areas
coal-tar enamel 80% of pipes In distnbution Good erosion resistance to Silt or sand Extreme heat may cause cracking
systems) Resistant to biologteal attachment Extreme COld may cause bnttleness
May cause an Increase In trace
organics In water
Epoxy Lining for steel and dUC1Jle "on Smooth surface results in reduced Relatively expensive
pipes (can be applied in the field or pumping costs Less resistant to abrasIOn than
in a fOUndry) Formulated from components approved coal-tar enamel
by the Food and Drug Administration SeMce life < t 5 years

Cement mortar Standard lining for ductile iron Relatively inexpensive Rigidity of lining may lead to
pipes. sometimes used In steel or Easy to apply (can be applied ,n place or cracking or Sloughing
cast-iron pipeS In pipe manufactunng process) Thickness of coaMg reduces
Calcium hydroxide release may proleCt cross-sectional area ot pipe and
uncoated metal at pipe Joints reduces carrying capacity

Polyethylene Lining used In ductile "on and steel Long service lite (50 years) Relatively expensive
pipe (applied at loundry) Good erosion resistance to abrasives (Silt
and sand)
Good resistance to bactenal corrosIOn
Smooth surface results In reduced
pumping costs

Source: U.S. EPA (1984).

Table 7·9. Water Storage Tank Llnlnga and Coalinga


Malarial Comments
HOI applied coal-tar enamel Most common coal-tar based coating used in water tanks; tends to sag or ripple when applied above
the waterline when tank walls are heated.
Coal-tar paints Most commonly used to reline eXisting water tanks; those paints containing xylene and naphtha solvents
give the water an unpleasant taste and odor and should be used only above the waterline.
Other coal tar paJnts containing no solvent bases can be used below the waterline but should not be
exposed to sunll\lht or ice; service life of 5 to 10 years.
Coal-tar epoxy palnts Less resistant to abr8Sl0n than coal tar enamel; can cause laSte and odor problems in the water;
service life of about 20 years.
CoaHar emulSion paJnt Good adheSIve charactenstics. odorless. and resists sunlight degradahon but not as watertight as other
coal-tar paints, which limits use below waler1lne.
Vinyl Nonreactive; hard. smooth surface; seMce life (about 20 years) IS reduced by soft water conditions.
Epoxy Forms hard, smooth surface; low water permeability; good adhesive characteristics If properly
formulaled and applied.
HOI and cold wax coatings Applied directly over rust or old paint shQr1 seMce life (about 5 years).
Metallic-sprayed lInc coahng Relatively expenSIve process that requores special skills and equipment good rust Inhlblhon, and service
life of up to 50 years.
Zinc-nch palnts Hard surlace; reSistant to rust and abrasion; relatively expenSive.
Chlonnated rubber palnts Used when controlling fumes from application of other linings is difficult or where the" use IS specified.
Asphalt-based linings Use is generally Iimiled to relining existing asphalt-lined tanks.

Source: U.S. EPA (1984).

• Influent levels of dissolved solids and pH • Water flow variability; for example, the
coagulation process may not respond readily to
• Pretreatment requirements, such as filtration flow fluctuations, while ion exchange, reverse
required before reverse osmosis or ion exchange osmosis, and activated alumina are relatively
treatments unaffected by changes in flow
• System size, because of the significant savings • Cost
from economies of scale for most technologies • Sludge management
Treatments for Inorganic Contaminants 139

• Adaptability of existing facilities The average estimated costs for analyzing these
radiation types range from $25 to $100 per sample.
Clearly, no single treatment is perfectly suited for all
inorganic contaminants. Possible treatment Many of the techniques for removing radionuclides
technologies include conventional treatment with are the same as those for inorganic and organic
coagulation, lime softening, cation exchange, anion contaminant removal. Table 7-13 lists the effective
exchange, reverse osmosis and electrodialysis, PAC, treatment processes for radionuclides. The radium
GAC, and activated alumina. Table 7-10 presents the and uranium removal processes can achieve up to 96
principal applications and removal efficiencies of the and 99 percent removal, respectively. GAC and
eight treatment technologies for inorganics. aeration are effective for removing radon. (Complete
discussions ofGAC and aeration are provided in
Many inorganic treatment decisions are based on the Sections 6.2 and 6.3.)
need to reduce the levels of a single contaminant.
However, to treat more than one contaminant,
multiple treatments in series or a membrane 7.2.1.1 Costs
treatment unit, such as reverse osmosis, is most EPA has developed preliminary cost estimates for
appropriate. Reverse osmosi~ and electrodialysis are removing radionuclides (see Table 7-14). The costs for
effective for most contaminants; however, their low aeration are based on experience from two system
tolerance for turbidity makes them inappropriate sizes, those serving 100 to 500 persons and those
without preliminary treatment for most surface serving 100,000 persons. Lower costs for larger
water. The removal effectiveness of the other seven systems refiect economies of scale for the aeration
treatments in Table 7-10 varies from poor to process.
excellent, depending on the contaminant. The
removal percentages provided in the table are Tables 7-15 and 7-16 present the costs of removing
approximations that apply to favorable water quality radon from drinking water. Table 7-15 compares cost
conditions. data for removing radon with aeration for three plant
sizes and Table 7-16 compares packed column
The actual performance of a removal technology aeration and GAC for managing three different
depends on the selection factors listed above. Table 7- levels of radon in emuent. The costs of removing
11 presents a summary of the most effective radon from tap water vary with the radon
treatments for 11 common inorganic contaminants. contamination level and the removal method.
The assumptions in the table include consideration of Aeration is more expensive and effective than GAC
treatment costs. Table 7-12 summarizes advantages for contaminant levels of 30,000 pCilL. However,
and disadvantages of each treatment technology. GAC is more expensive and effective at higher radon
concentrations of 150,000 pCilL.
Section 7.2.1 discusses radionuclide contamination.
Radionuclides present different health risks and
require different analytical evaluation techniques 7.2.2 Conventional Treatment: Coagulation
than other inorganic contaminants, but are removed and Lime Softening
by many of the same technologies that are effective Coagulation and lime softening are traditionally
for inorganic and organic contaminants. Sections used to control turbidity, hardness, tastes, and odors,
7.2.2 through 7.2.5 review the five treatment but also are effecti ve in removing some inorganic
technologies designated primarily for inorganic contaminants. Section 4.1 describes the conventional
contamination. water treatment processes in common practice at
many water utilities. However, employing these
processes to remove specific inorganic contaminants
7.2.1 Removing Radionuclides in Drinking is relatively new.
Water
Radionuclides are radioactive atoms that are Coagulation, discussed in Section 7.2.2.1, is effective
characterized by the number of protons and neutrons in removing most metal ions or colloidally dispersed
in their nucleus and their energy content. The compounds, but is ineffective in removing nitrate,
number that appears with a radionuclide's chemical nitrite, radium, barium, and sulfate. Lime softening,
abbreviation relates to this nuclear composition. The discussed in Section 7.2.2.2, increases water pH,
most common radionuclides in drinking water are which precipitates the polyvalent cations and anions
radium, uranium, and radon. of calcium, magnesium, carbonate, and phosphate;
however, lime softening is ineffective for nitrate
Many methods are used to measure radionuclides' removal.
levels in drinking water. Atoms emit three types of
nuclear radiation: alpha, beta, and gamma radiation. Typically, higher dosages of coagulants are required
for effective inorganics removal than for removal of
turbidity, color, or hardness. For example, typical
140 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Table 7-10. Removal Effec:ttv...... lor Eight Proces_ by Ino'Vanic Contaminant


Contaminant
TrB8tment Ag As ASIII Asv Ba Cd Cr Cr lt CrVl F Hg HglOI Hglll NO] Pb Ra Se SelV SeVl
Conventional lJ'B8tment H M H L H H H L M M L H L M L
Lime sotlsning M H H H H L M L M L H H M L
Reverse osmosis H M H H H H H H M H H H
Cation exchange L H H H L L L H H L
Anoon exchange M M M H H M M H
Activated alumina H L L H L H
Powdered activated L L M L L M M L L
carbon
Granular activated L M L L H H L L
carbon
H • High • > 800/. removal.
M • Medium • 20-80% removal.
L • Low • 20% removal.
- indicates no data were prOVIded.

Table 7-11. Moat Effective Treatment Methoda lor Removal 01 Inorganic Contamlnanla
Contaminant Most Effecnve Treatment Methods
Arsenic As V·"on coagulanon. pH 6-8; alum coagulanon. pH 6-7; excess lime softening; acnvated alumina. pH 5·6
As lII'oxtdation lrealrnent of As III to As V; use same trealrnentlist lor As V
Barium Lime softening. pH "; ,on exchange softening
Cadmium Iron coagulation. above pH 8; lime softenong; excess lime softening
Chromium Cr Ill·iron coagulation. pH 6·9; alum coagulation, pH 7-9; excess lime softening
Cr·ferrous sulfate coagulanon. pH 7-9.5
Fluoride Ion exchange WIth activated alumina or bone char
Lead Iron coagulation. pH 6·9; alum coagulatx>n. pH 6-9; lime or excess "me softening
Mercury InorganIC'ferne sulfate coagulation. pH 7 -8; granular ac1JVated carbotl
Organic·granular activated carbon
Nitrate Ion exchange With anion resin
Radium Lime softening; Ion exchange With canon resm
Selenium Sa IV·ferric sulfate coagulation. pH 6·7; oon exchange With anoon resin or actlvaled alum,na; reverse osmoSIS
Se IV-ion exchange wIlh anion reSIn or activated alum.na; reverse osmosis
Silver Feme sulfate coagulatIOn. pH 7-9: alum coagulanon. pH 6·8; lime or excess lime softening
Source: Sorg (1980).

alum dosages for turbidity control range from 5 to 40 anions. The effectiveness of coagulation also depends
mglL, while the dosage for arsenic removal is about on influent turbidity and/or color levels, the quality
100 mglL. Higher dosages increase costs and create of the flocculated particles produced, and the amount
more sludge. of turbidity or color removed.

Table 7-17 presents potential removal efficiencies


7.2.2.1 Coagulation using alum and ferric salts as coagulants. The
The coagulation processes that remove turbidity as optimum pH level varies depending on the target
well as inorganic contaminants include adsorption inorganic substance and the coagulant used. The
and precipit::.tion. Coagulation with iron or table provides the most suitable pH level for
aluminum salts also removes trace anions of removing each inorganic substance, when available.
selenium, arsenic, and fluoride through As indicated, iron salts are more effective over a
coprecipitation or sorption onto flocculated particles. wider range of pH levels than alum. Ferric salts are
effective at pH levels between 4.0 and 12.0, while
Coagulation is more effective in removing alum is generally effective at pH 5.5 to 8.0.
polyvalent, as opposed to monovalent, cations and
Treatments for Inorganic Contaminants 141

Table 7·12. Advantages and DIsadvantages of lnonjlanlc Table 7·13. Trealment Technologies lor Removing
Contamlnlnt Removal Proc..... Rldlonuclldes

Preclp,ts&on and Coprecipil8&on Used in Reported


Coaguls&onlConventionsl ApprOXimate
Process
Advantages
Treatment Radio· EffiCiency
• LOW cost lOr high volume Technology nuclide (percent) Comments
• Otten Improved by high iorllc strength ConvennonaJ Ra <25 High pH and Mg rllQuored
• Reliable process well SUited to automalJC control treatment U 18·98 High pH (10 -) and high
Disadvantages WIth dosages of lernc chlonde
coagulation· or alum only accomplished
• StoichIOmetric chemICal additions required filtration In lab studies With
• High·water·content sludge dispasal
diatomaceous earth
• Pan· per· billion effluent contaminant levels may reqUIre filtratIon
two· stage prec,pltaOon
• Not readily applied to small or Intermittent flows Lime Ra 75·96 Best chOice lor large plants
• Coprecipitabon efficiency d6ll'lOds on Initial contaminant softening 43·92 Plant· scale results
concentration and surface area of pnmary floc U
80 Plant· scale results
85·90 pH 10.6"'.5
Ion Exchange
99 High pH. high Mg
Advantages
Ion exchange Ra 95 - Best chOice lor small plants;
• Operates on demand 99 cation exchangers
• Relatively InsenSitive to flow variations U 99 Bnne disposal problem
• Essenbally zero level of effluent contamination possible Anion exchangers; largely
• Large variety of specific reSins available expenmental but some lull·
• BenefiCial selectiVity reversal commonly occurs upon scale plants on line
ragenerabon
Adsorption Ra 90 - Adsorption on any solids;
Disadvantages 85·90 experimental
Rn 62·99 Sand adsorphon;
• Potennal lor chromatographIC effluent peaklOg
expenmenta'
• Spent ragenerant d,spasal
GAC adsorption
• Variable effluent quality with respect to background ions
• Usually not leasible at high levels of total dissolved solods Aerahon Rn 20-96 Depends on process
93 + Depends on process
ActIvated Alumlns
Reverse Ra 87·96 Plam-scale data
Advantages
osmosis 87-98 Based on eight plants
• Operates on demand 95- High-volume bnne solution
• InsenSlove to flow and total dissolved solids background U 95+ for diSPOsal
• Low effluent contaminant level possible High·votume bnne solution
• Highly selectlV8 for fluonde and arsenIC for disposal
Disadvantages
• Both acid and base are required for regeneration
• Medium tend to dissolve. prodUCing hne particles
• Slow adsorption kinelJCS
• Spent regenerant disposal
Table 7·14. Radlonucllde Process Treatment Costs
Membranes (Reverse Osmos,s and Electrodia/ys,s)
Range of Costs of Removal
Advantages Process (dOliarsll ,000 gallons
• All contamInant IonS and most dissolved non-Ions are Process of water)
removed CoagulallOnlFiltrabon- 0.07 to 0.28 0
• Relallvely insenSlbve to flow and total dissolved solids
Lime softening- O.lOb
level
• Low effluent concentranon poSSible Aeration 0.10100.75 c
• Bactena and parocles are removed Ion exchange (cabonlc) 0.30 to 0.80 c
Disadvantages
Ion lind manganese treatment 0.30 to 1.10C
• High Capital and operating costs Lime softening (new) 0.50 0
• High level of pretreatment reqUIred
• Membranes are prone to loulong Ion exchange (anionic) 1.60 to 2.10 0
• Relect stream IS 20 to 90% of leed flow Reverse osmoSIs , .60 to 3.20 c
Source: Clifford (1986). 1.000 gallons = 3.78 m 3
• Adding to an eXlsMg facility.
b 1982 dollars.
c 1987 dollars.
Adjusting the pH and using coagulant aids can
optimize inorganic compound removal. Coagulant for coagulants, coagulant aids, and pH adjusters may
aids may reduce the higher doses of coagulant needed be derived using jar tests and pilot studies.
for inorganic contaminant removal. Optimal dosages
142 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Section 7.1.3.2 for discussion of the effect of lime


Table 7·15. Costs lor Removt"il Radon Irom Drinking
Water by Packed Towtlf AeratIon
softening on corrosion.) The pH is an important
(99% Removal) design consideration for inorganic contaminant
Populanon Served
removal using lime softening because it affects:
3,300' 75,000'
• Contaminant species (soluble or insoluble)
100·500 , 0,000 100,000
• Coagulant form
Tolal Capital cost ($1,000) 67 250 2,200 • Type of flocculated particle formed
Operanons and mamtenance cost 1.2 15 230
($ 1,000 per year) Higher dosages of lime create higher pH levels,
Cost (cenlsl1,000 gallons) 75 14 9 which, in turn, increase precipitation of inorganic
contaminants and adsorption of both organic and
1,000 gallons • 3.78 mJ
inorganic species. The harder the water, the more
effective the lime softening process is in removing
inorganic contaminants. This happens because the
Table 7-16. COSls at Radon Treatment at the Plant Scale
softening process depends on the enmeshment of ions
(200 GPO) lor GAC Versus Aeration within the flocculated particles or adsorption on the
surface of precipitates.
Operanng
Effluent Concentrations Capllal Costs Costs In
Innuent
(pCi/L) in Dollars DOllars
Table 7-18 presents the removal efficiency of lime for
Concentra· seven inorganic contaminants. Lime softening is
lions (pCiIl) I' lib I II
effective for cadmium, lead, silver, fluoride, and
15.000 1.350-3.300 750 430-760 900 20 60 cation removal in general. It also effectively removes
30,000 2,700·6,600 1,500 430·760 900 20 80 arsenic(V). barium. and chromium(1II1 at the specific
150,000 1.200 67.500 1.500 1,000 40 80 pH levels shown in the table. Lime-soda softening is
'1 = GAC process.
generally ineffective in removing nitrate,
bll ::::I Aeration process. selenium(VI), mercury(O). or chromium(lVl.

Table 7-18. Removals Possible with lime Softening


Table 7-17. Removal Ertlclency Potential 01 Alum Versus
Ferric Chloride InorganiC Contaminant pH Percent Removal
As (V) 10.0'105 70
Inorganic Removal ENiClency As (V) > 10.8 90
Contammant Alum Coagulant Iron Coagulant As (III) > 10.5 70
Ag (pH < 8.0) 90% Cd NA 90
Ag (pH = 8.0) 70% Cr (III) > 10.5 70·90
~ ~ 90
As V 90%
As V (pH < 90%
~ ~ ro
Ba 95·10.8
75)
NA • Not avatlable.
As V (pH 90%
75)
sg (pH
= 8.01 Figure 7-2 shows a typical lime-soda softening
sg (pH 70% process. Characteristic elements of such a system
= 8.51 include raw water pumps, lime and soda ash addition
Cr (111) 90% facilities, sedimentation, recarbonation, filtration,
Cr 111 (pH = 90% disinfection, storage, and distribution. The
10.51 recarbonation step is designed to lower the pH to the
Cr VI (using 90% point of calcium carbonate saturation after the lime
Fe 11) has had its effect in the sedimentation basin.
Hg 70%
Pb 90% 7.2.3 Reverse Osmosis
Reverse osmosis is an expensive process that uses a
semipermeable membrane to remove contaminants
from solution. Water with different contaminant
concentration is placed on each side of the membrane.
7.2.2.2 Ume Softening The water is directed through the membrane by
Lime softening is a reliable and established hydrostatic pressure to the side with the lower
treatment for "hard" water and corrosion control, and concentration of contaminants. Since the membrane
also removes some inorganic contaminants. (See permits only water, and not dissolved ions, to pass
Treatments for Inorganic Contaminants 143

D'Sinfectant

Rapid M,x Flocculabon Sedimentation Recarbonation Filtration

Figure 7·2. Lime softening treatment system.

through its pores, contaminants are left behind in a removing inorganic substances, it has also removed
brine solution. These membranes are available with the following effectively:
a variety of pore sizes and characteristics.
• Radium
Electrodialysis is a process that also uses
membranes. However, in this process direct electrical • THM precursors including humic and fulvic acids
current is used to attract the ions to one side of the
treatment chamber. • Pesticides
The membranes adjacent to the influent stream are • Microbiological contaminants (viruses, bacteria,
charged either positively or negatively, and this and protozoa)
charge attracts counter-ions toward the membrane.
The membranes are designed so as to allow either Reverse osmosis systems are particularly effecti ve in
positively or negatively charged ions to pass through series. Water passing through multiple units can
the membrane, thus they move from the product achieve near zero effiuent contaminant
water stream through a membrane to the two reject concentrations.
water streams.
Reverse osmosis is also suitable for small systems
The three essential elements of the system are (1) a with a high degree of seasonal fluctuation in water
source of pressurized water, (2) a direct current power demand. Reverse osmosis systems are relati vely
supply, and (3) a pair of selective membranes. The insensitive to changes in flow and operate
average ion removal varies from 25 to 60 percent per immediately, without any minimum break-in period.
stage. Multistage units can increase the efficiency of Their operational simplicity and minimal labor
removal. Many membrane pairs are "stacked" in the requirements make them suitable for small system
treatment vessel. applications. One disadvantage of reverse osmosis
units is high operating and capital costs. Managing
Fouling of the membranes may limit the amount of the wastewater <brine solution) is also a potential
water treated before cleaning is required. Fouling is problem.
caused when membrane pores are clogged by salt
precipitation or by physical obstruction of suspended
particulates. Particulates, suspended in water, can be 7.2.3.1 Design Considerations
removed in pretreatment but salts that exceed their Typical reverse osmosis units include raw water
solubility product at the membrane surface must be pumps, pretreatment, membranes, disinfection,
controlled chemically by pH reduction (to reduce storage, and distribution elements. Figure 7-3 is a
carbonate concentration) or chelation of metal ions schematic diagram of a reverse osmosis system.
(by use of phosphate, for examplel. A recent These units are able to process virtually any desired
innovation has been the occasional reversal of the quantity or quality of water by configuring units
charge on the membranes, a process called sequentially to reprocess waste brine from the earlier
electrodialysis reversal (EDR). This helps to flush the stages of the process. The principal design
attached ions from the membrane surface, thus considerations for reverse osmosis units are:
extending the time between cleanings.
• Operating pressure
Reverse osmosis is a compact system that is well • Membrane type and pore size
suited for treating water with high levels of inorganic • Pretreatment requirements
substances, organic substances, or total dissolved • Product conversion rate (the ratio of the influent
solids. It has effectively treated water with total recovered as waste brine water to the finished
dissolved solids as high as seawater, at 8,600 mglL. water)
While reverse osmosis is a proven technology for
144 Upgrading Existing or Designing New Drinking Water Treatment Facilities

.. ..
0
Generator Set

D l2 9
High Service Pumps

• Brine to
PercolatIOn
Pond

Product to Degas,fier
and ReservoIr

",'
Filter

MixIng
0 SodIum Silicate Feed
Feed Tank CleanIng
Solubon

Figure 7-3. Schematic 01 a reverse osmosis system.

resistance, pH range, temperature range, chlorine


Operating Pressure tolerance, and biodegradation tolerance.
Typical operating pressures range from 5.3 to higher
than 24.6 kglcm 2 (75 to higher than 350 psi). Reverse
osmosis systems rated at less than 17.6 kglcm 2 (250 Pretreatment Requirements
psi) are classified as low-pressure units, while those Assessing pretreatment requirements requires
operating above 24.6 kg/cm 2 (350 psi) are classified careful consideration of the following factors:
as high-pressure units. High water pressure can lead
to noise, vibration, and corrosion problems, but, in • Influent suspended solid concentration
general, these systems are more effective. Recently • Ionic size of the contaminants
developt!d models, however, are able to perform well • Membrane type
at low pressures.
Other influent characteristics such as contaminant
concentration, water temperature, and presence of
Membrane Type competing ions are also important. Some systems
Two distinct membrane designs for reverse osmosis with poor quality influent require extensive
are illustrated in Figure 7-4. The spiral-wound unit pretreatment.
clogs I'!ss frequently when processing influent with
high solids content. The hollow-fiber membrane, Pretreatment is most commonly used to prevent
however, has much greater surface area per unit of fouling of the membrane during operation. Fouling is
space than the spiral-wound design. The hollow-fiber also caused by biological growth during periods of
unit has about 3,280 m 2 of membrane surface/m 3 of disuse. Typical pretreatment for reverse osmosis
membrane module (1,000 sq ftJcu ft of membrane includes particle removal by filtration, and
module), as compared to about 328 m 2 (100 sq ftl for sequestering hardness ions by precipitates and pH
the spiral-wound unit. control to prevent clogging.

Until the 19705, membranes were made a1 most A major problem with reverse osmosis units is proper
exclusively of cellulose acetate. Now membranes are management of the wastewater quantity and quality.
also made from aromatic polyamide and thin-filmed The percentage of treated water recovered from these
polymer composites. Different membrane materials systems typically ranges from 50 to 90 percent of the
will have distinct characteristics, such as hydraulic influent. Systems that produce reject water ranging
Treatments for Inorganic Contaminants 145

Feed
(Raw Water)

Membrane

Filtrate
Retentate

Courtesy of Millipore Corporation

Process Flow Through Spira/-Wound Reverse OSmosiS UOil

[ Hollow Fiber
Casing """'"

Treated
Waler

HOllow-Fiber Reverse OsmOSIS Unit

Figure 7-4. Two types 01 reverse oamosla membranes.


146 Upgrading Existing or Designing New Drinking Water Treatment Facilities

as high as 90 percent of the feed water are generally During ion exchange, the contaminant ions replace
uneconomical. the regenerant ions because they are preferred by the
exchange medium. After the exchange medium
reaches equilibrium with the contaminant ions, the
Ratio of Reject Water to Finished Water medium is regenerated with a suitable solution,
The ratio of reject water, or "brine," to finished water which then resaturates the medium with the
depends on several factors primarily ionic charge and appropriate ions. Because of the required "down
ionic size. The higher the ionic charge and the larger time," the shortest economical regeneration cycles
the ionic size ofthe contaminant, the more easily the are once per day. Ion exchange waste is highly
ion is removed and the more rmished water is concentrated and requires careful disposal,
recovered relative to the amount of brine. The ionic analogous to reverse osmosis reject streams.
charge for some contaminants depends on pH. The
quality of the brine depends on the influent quality The porous exchange medium is covered with tiny
and requires site-specific assessments to develop holes, which clog when significant levels of
environmentally sound waste management. suspended solids are in the influent stream.
Consequently, filtration may be a necessary
pretreatment for ion exchange units. Ion exchange
7.2.3.2 System Performance units are also sensitive to the presence of competing
Reverse osmosis can remove essentially all inorganic ions. For example, influent with high levels of
contaminants from water effectively. It removes over hardness will effectively compete with other cations
70 percent of the following: arsenic(III1, arsenic(V1, for sites on the exchange medium.
barium, cadmium, chromium(III1, chromium(VIl,
fluoride,lead, mercury, nitrite, selenium(lV1, The resin exchange capacity is expressed in terms of
selenium(VIl, and silver. Properly operated reverse weight per unit volume of the resin. The calculation
osmosis units will attain 96 percent removal rates, of the breakthrough time for an ion exchange unit
while similarly operated lime-soda softening requires knowledge of the resin exchange capacity,
operations will attain from 75 to 96 percent removal. the influent contaminant concentration, and the
Comparative ion exchange units achieve between 81 desired effiuent quality.
and 97 percent removal.
The ion exchange process, like reverse osmosis, is
relatively insensitive to flow rate and virtually ful1y
7.2.3.3 System Costs operational upon process initiation. Each resin, of the
The predominant disadvantage of reverse osmosis is many available from suppliers, is effective in
its high cost. Operating costs range from $0.79 to removing specific contaminants. In addition, ion
$1.59/m 3 ($3 to $6/1,000 gal) oftreated water for a exchange is able to achieve very low contaminant
reverse osmosis plant of less than 0.04 m3 /sec (1 concentrations in finished water.
MGDl. For plants with larger capacities and lower
operating pressures, costs may be potentially The primary disadvantage of ion exchange concerns
competitive with other processes. Capital costs are effiuent peaking. Effiuent peaking occurs when
usually high, especially if there is a need for contaminant ions compete with other ions for
pretreatment. In general. capital costs range from exchange medium sites, resulting in unacceptable
$264 to $5281m 3 ($1 to $21gall of capacity. levels (peaks) of contamination in the effiuent. It is
most common with poorly adsorbed contaminants,
such as nitrate. Effiuent peaking necessitates more
7.2.4 Ion Exchange frequent regeneration of the exchange medium.
Ion exchange units are used to remove any ionic
substance from water, but are used predominantly to In addition, ion exchangers that use sodium chloride
remove hardness and nitrate from ground water. to saturate the exchange medium may experience
Typical ion exchange units consist of prefiltration, problems with sodium residual in the finished water.
ion exchange, disinfection, storage, and distribution Sodium is used because of its low cost, but high
elements (see Figure 7-51. sodium residual is unacceptable for individuals with
salt-restricted diets. This problem may be avoided by
Inorganics removal is accomplished through using other saturant materials, such as potassium
adsorption ofcontaminant ions onto a resin exchange chloride.
medium. As the name implies, one ion is substituted
for another on the charged surface of the medium,
which is a resin, usually a synthetic plastic. This 7.2.4.1 System Performance
resin surface is designed as either cationic or anionic. Ion exchange effectively removes greater than 90
percent of barium, cadmium, chromium(lll), silver,
The exchange medium is saturated with the and radium using cationic resins as the exchange
exchangeable ion before treatment operations. medium. In addition, it achieves greater than 90
Treatments for Inorganic Contaminants 147

Salt Loading

Brine

Raw Water

Patable
Water
ta
Brine
System
Blending Valves

Backwash '-------....
Waste ta
Disposal

Alarm and
Shutdown
Distnbutian
System
ConduellVity
Monllar
Well Supply

Figure 7-5. Ion exchange treatment system.

above, ion exchange is especially well suited to


remove barium, radium, nitrate, and selenium.

7.2.5 Activated Alumina


Activated alumina, a commercially available ion
exchange medium, is primarily used for fluoride
removal from ground water. The activated alumina
process is an ion exchange process that consists of the
following six basic elements: raw water pumps,
pretreatment, activated alumina contact,
disinfection, storage, and distribution. Figure 7-6
shows schematic diagrams of the four-phase
operational process.

Before processing, the activated alumina medium is


saturated with hydroxyl ions from a strong sodium
hydroxide solution. During the removal process,
activated alumina exchanges the hydroxyl ions for
Ion exchange lacillty lor nitrate removal; McFarland, CA. fluoride anions. When the medium becomes
saturated, the activated alumina must be
percent removal of nitrites, selenium, arsenic(Vl, regenerated with sodium hydroxide. This increases
chromium(VIl, and nitrate using anionic resins. the pH level of the water in the treatment unit to a
While it is effective for all of the contaminants listed point where buffer solutions of 3 percent sulfuric acid
148 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Raw Water - - - - - - - -.... -----<S Raw Water - - - - - - - - - -...., - - - - - {

Acid ----ot
_ Treated Water ---4----r--{

Treatment Unit Treatment Unit

Waste
Treatment and
Downflow Rinse Backwash and
Upflow Rinse

Raw Water -----------r--< Raw Water ----------.,....---<


Caustic
.... Caustic

Treatment Untt Treatment Unit

Waste Waste
upnow Downflow
RegeneratJon Regeneration

Figure 7·6. Activated alumina IYltema: Operating mode flow schematlea

are required for neutralization, after the water leaves hazardous strongly acidic and basic solutions. The
the exchange unit. process of heating the sodium hydroxide solution to
maintain it as a liquid in cold climates is particularly
While activated alumina requires pretreatment for hazardous.
suspended solids, it can tolerate high levels of total
dissol ved solids. As with other ion exchange
processes, the activated alumina process is sensitive Activated alumina's costs are also higher than ion
to competing ions. It is also sensitive to pH. Optimum exchange costs. The loss of activated alumina to the
removal efficiency for fluoride occurs below a pH of sodium hydroxide solution during processing can
8.2. range up to 20 percent annually. In addition, waste
management may also increase costs because of the
While activated alumina effectively removes several high contaminant and aluminum concentrations in
contaminants, it requires the handling of potentially the waste stream, as well as the high pH.
Treatments for Inorganic Contaminants 149

7.2.5.1 System Performance


Activated alumina is operational immediately
because of its relative insensitivity to flow rates. It
effectively removes over 90 percent of arsenic(Vl,
fluoride, and seleniumUV1; removes 70 percent of
selenium(VI); and also effectively removes iron. It is
ineffective for barium, cadmium, and radium
removal, since these contaminants occur primarily as
cations.
8. Current and Emerging Research

This chapter describes the activities of EPA's In pursuit of these regulatory milestones, research
Drinking Water Research Division (DWRDl, has focused on identifying contaminants that will be
highlighting the division's recent research efforts regulated and evaluating the removal potential of
and future directions. The impetus for the majority of available technologies. This research includes
DWRD's research is to provide technical support for identifying and characterizing disinfection by-
the Agency as it implements the SDWA and its products associated with chlorination and the use of
amendments. At present, there are four aspects of the ozone. Also in support of the rules, DWRD is studying
regulatory process influencing current areas of various methods of meeting mandatory disinfection
research: requirements; investigating factors affecting the
occurrence oflead in drinking water; and evaluating
methods of controlling corrosion, exposure of
1. Development of Final Surface Water and the consumers to disinfection by-products, and microbial
anticipated Ground-Water Treatment Rules, contamination in distribution systems.
which will require disinfection and, in the case of
surface waters, filtration of drinking water in In addition to its research activities, DWRD provides
communities serving more than 25 persons. technical assistance to State agencies, EPA regional
offices, water utilities, and professional
2. The requirement that EPA set Maximum organizations. DWRD also supplies information and
Contaminant Levels (MCLs) for listed provides consultation on the technical aspects of
contaminants, e.g., volatile organic compounds drinking water treatment systems.
(VOCs), synthetic organic compounds (SOCs),
inorganics, radionuclides, microbiological
contaminants, and turbidity according to the 8.1 Current Research on Disinfection
timetable specified in the 1986 Amendments to By-PrOducts
theSDWA. The SOWA and Amendments of 1986 require most
public water supplies to disinfect drinking water
3. The requirement that every 3 years EPA publish before distribution (U.S. EPA, 1986a). In the past,
a Drinking Water Priority List of contaminants chlorine has been the most widely used disinfectant,
known or anticipated to occur in public drinking and while its effectiveness in controlling pathogens is
water systems and pose a human health risk; and proven, unhealthful chemical by-products have been
subsequently regulate at least 25 of the identified.
contaminants listed according to the schedule set
in the act. The first list of 44 contaminants was Research on these by-products is a very important
already published and, therefore, 25 of those area of investigation for EPA. A disinfection by-
contaminants listed must be regulated by product work group was formed to coordinate
January 1990. research efforts between the EPA Office of Research
and Development (ORO) and Office of Drinking
4. The Disinfection By-Products Rule, whose Water (ODW). Research efforts are focused on
proposal is anticipated along with the Ground- identifying and characterizing the by-products of
Water Treatment Rule in 1991, will regulate chlorine disinfection, and evaluating treatment
disinfection by-products that have been, and are techniques that will control by-product generation.
still being, identified. Since a promising strategy for control is to use ozone

150
Current and Emerging Research 151

as an alternate disinfectant, EPA is also researching found to be reduced. Therefore, the overall quality of
the by-products and other issues associated with the water improved under this scenario.
ozone disinfection.
Although the individual concentrations of the
disinfection by-products (DBPs) vary with pH,
8.1.1 Identifying and Controlling temperature, and chlorine concentration, pilot
Chlorination By·Products studies have shown that the removal ofTTHM
DWRD has been conducting experiments, mostly in precursor seems to remove the formation potential for
the laboratory, to develop methods of identifying by- the other individual DBPs. This finding confirms
products associated with the use of chlorine as a earlier studies and opens up the possibility of
disinfectant. In close association with the Technical precursor removal as an effective means of
Support Division ofODW, DWRD is isolating the controlling chlorinated DBP (Lykins et aI., 1988a).
various by-products that have been identified in field
studies (Stevens et aI., 1987). An alternative to chlorine is the use ofchloramines
or chlorine dioxide for minimized DBP. Studies have
To date, six compounds (or in some cases groups of shown that these approaches can be effective in
compounds) have been identified: maintaining microbiological integrity and
minimizing by-product formation (Singer, 1988).
• Trihalomethanes (4 compounds) This issue and the associated technologies will
• Dihaloacetonitriles (3 compounds) become extremely important if the THM standard is
• Chloroacetic acids (3 compounds) ultimately lowered.
• Chloral hydrate
• Chloropicrin
• 1,1,1-Trichloropropanone 8.1.2 Identifying Ozone By-Products
The DWRD is studying several issues related to
These by-products comprise 30 to 60 percent of the using ozone as an alternate disinfectant to chlorine.
total organic halogen (TaX) in drinking water. The The broadest issue concerns the identification of by-
levels of common chlorination by-products found in products formed by ozone reacting with naturally
drinking water are presented in Table 8-1 (Stevens et occurring organics in drinking water (Glaze, 1986).
al.,1987). The health effects of'~ese by-products are being
characterized and compared to the health effects of
by-products of chlorination. Also being investigated
Table 8-1. Common Chlor1natlOn By-Produeta is the effect of ozone on the precursors ofchlorine by-
Number ot
NomIOal ConcentratIOn (1lg!1) products, an issue ifchlorine is used as a secondary
General Compound Individual disinfectant.
IdenbficatKln Compounds < 10 10·100 > lOa
Trihalomethanes 4 X Other research related to ozone as a disinfectant
DihaloacetoMriles 3 x covers its ability to destroy contaminants and its
Chloroacetlc acidS 3 X X X effects on the formation of bromide ion or other
brominated organics when bromide ion is present.
Chloral hydrate X
Some substances formed through the oxidation of
ChlOtoplCt1n X bromide ion are known to be harmful to humans.
1.1 1·TnchlOro-
0 X
p<opanone
Source: Slevens et al. (1987).
8.2 Treatment of Organic and Inorganic
Contaminants
Methods of controlling chlorine disinfection by- The SOWA amendments accelerated the schedule for
products are being studied by DWRD in an onsite setting MCLs for contaminants in drinking water.
pilot plant (Stevens et aI., 1989). (Figure 8-1 is a Contaminant limits are set according to the removal
schematic of the pilot plant.) The pilot plant was potential of existing technologies that can be
constructed with two channels: the first was obtained at a reasonable cost. Traditionally, EPA set
represented by the experiment control; the second contaminant limits by studying treatment
was used to test the effects of changing the point of techniques for each chemical individually. Since
chlorination on the production of by-products. there is such a large number of chemicals to be
considered for regulation, and the deadlines for
Through experimentation, it was found that moving promulgating MCLs are quite stringent, EPA is
chlorination to a point just before filtration studying the effectiveness of various technologies in
minimized the production of total trihalomethanes removing groups of contaminants using bench- and
(TTHMs) while maintaining microbiological quality. pilot-scale experiments. Table 8-2 lists the
The assimilable organic carbon of the finished water, technologies that have been and are being evaluated
which is a measure of the biological quality, was also by DWRD (Clark et aI., 1988b; Feige et aI., 1987).
152 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Floc ~-0-1 Clear Well W


(Phage)

(Acid)
Alum

Filt~~-G-
Chlonne (Base)

Figure 8-1. Pilot plant schematic for disinfection by-product control.

tested for their ability to remove inorganic


Table 8-2. Treatment Technologies Evaluated by the
Drinking Water Research Dlvialon
compounds, including radionuclides. In addition,
technologies for controlling secondary sources of
Contaminant Proven pollution from treatment processes are being
(or Effective Proven investigated. This section summarizes the relevant
Contaminant In Field EffectIve In Being Evaluated as
Classes) Tests Pilot Test.s Promising TechnOlogies
research projects.
Volatile Carbon Ozone OXidation. reverse
organic adsorption. osmosis. ultraVIolet treatment
com..".:Junds packed
8.2. 1 Granular Activated Carbon Systems
tower and GAC is the legal feasible technology for treating
diffused SOCs. DWRD is currently developing data on the
aeration
effectiveness of GAC for removing a variety of
Synthetic Carbon Conventional treaument With organic chemicals.
organic adsorption powdered activated cartIon.
compounds ozone OXidatiOn. reverse
osmoSIS, ultraVIolet treatment One measure of effectiveness is the carbon usage rate
Nitrates Ion Reverse
for each chemical. This indicator describes the ability
exchange osmOSIs of each compound to be adsorbed by the carbon in the
Radium Reverse
treatment system. It also describes how quickly the
osmOSIS. carbon bed will be exhausted (adsorption surfaces of
,on the carbon granules will be used) and, therefore,
exchange dictates the size of the GAC system needed for
Radon Carbon Aeration optimal operation. EPA is currently developing
adsorptIon carbon usage rate data for the 45 chemicals listed in
Uranium Ion Reverse Table 8-3 (Clark et aI., 1989al. Carbon usage rates
exchange osmOSIS are established by laboratory or field experiments.
Source: Clark et al. (1988b); Feige et al. (1987).

8.2.1.1 Establishing Carbon Usage Rates


Granular Activated Carbon (GAC), ozone oxidation, A mathematical model that uses isotherm results to
reverse osmosis, ultraviolet (UV) treatment, simulate full-scale carbon adsorption treatment
ultrafiltration, packed tower aeration, and systems is being used (Speth and Miltner, 1989).
conventional treatment are being studied as methods Pilot testing is a common first step in designing a
of controlling organic compounds, including certain full-scale treatment system and much data are
disinfection by-products. Ion exchange, reverse available correlating pilot-scale test results to actual
osmosis, aeration, and carbon adsorption are being full-scale operating parameters. Therefore, engineers
Current and Emerging Research 153

isotherm predictions (Speth and Miltner, 1989). Some


Table 8·3. Chemlcala lor Which Carbon Uaage Rates are computer models also contain a component that
Belnll Developed
estimates the cost of various sized full-scale systems
2.4-0 DaiaPOn (Adams et aI., 1989). Several of these models are in
Silvax D,noseb the process of being validated by DWRD.
Lindane Glyphosate
Methoxychlor Oxamyl
Toxaphene N.N·Ethylene-thiourea
Chlordane Metribuzin 8.2.1.2 FIeld Tests of Granular Activated
Heptachlor Hexachlorobenzene
CarbOn Systems
Heptachlor ePOxide c;s-I.3-dichlo<opropene DWRD has conducted extensive field studies on the
PCS(s) o ·chlorotoluene
potential ofGAC to remove organic contaminants
from surface water. It was found that carbon
Pentachlorophenol 1.1·D,chlo<oethane
adsorption is an effective means of removing both
Alachlor 1.I·Dichloropropene volatile and synthetic organic compounds. Field
Carooturan 1.1.2.2-Tetrachloroelhane testing sites included Jefferson Parish, Louisiana;
Ald,carb EDB Cincinnati, Ohio; Manchester, ~ew Hampshire;
Aldicarb sulfone DBCP Evansville, Indiana; Miami, Florida; Huntington,
West Virginia; Beaver Falls, Pennsylvania; and
Aldicarb sulfox,de 1.2·D,chloropropane
Passaic, New Jersey (Lykins et aI., 1984).
Ten-Butyl methyl ether c,s·I.2-DlChloroelhylene
Melolachlor Irans-l.2-DlChloroethyJene More recent studies of GAC systems have focused on
Telfachloroelhene Chlorobenzene organics removal from ground water at three sites:
1.3.5·Tnchlorobenzene o·D'chlorobenzene Suffolk County, ~ew York; San Joaquin Valley,
Bromobenzene Toluene
California; and Wausau, Wisconsin. A brief
description of these ground-water treatment studies
D,bromomethane Styrene
follows (Clark et aI., 1988b).
2.2-Dichloropropane Ethyl benzene
1.1.1.2·Telfachloroelhane o·Xylene The purpose of the Suffolk County project was to
2.4.5-Tnchlorophenoxy acelJC acid m·Xylene study the removal of organic compounds, pesticides,
End"n p·Xylene and nitrates under low-now situations (simulating
residential usage). GAC in combination with an ion-
PhthaJate(s) Cyanaztne
exchange system, and a reverse osmosis system were
Ad,pate(sl 1.1.2·TnchIQroethane operated in parallel at this test site. Data on
PAH(s) lsophorone treatment effectiveness and system costs were
Hexachlorocyclopentad,ene Tntluralin collected and analyzed to assist in the design and
D'acamba Methylene chloride testing of large public water supply systems. These
SlmazlOe trans·'.3-Dichloropropene
results will be especially useful to operators of water
treatment systems in farming communities where
Atrazlne p·ChlorOloJuene
pesticides, nitrates, and other organics are a common
D'Quat , .3-D'chloropropane problem.
Endothalf 2,4·0lnllrotoluene
Picioram 1.2.3·T"chloropropane
The field study in the San Joaquin Valley examined
point-of-use GAC units servicing small treatment
conducting pilot tests have sufficient data to systems and private welts of homeowners and
confidently predict full-scale operations. There is not, farmers. The purpose of the project was to develop
however, much information on how well computer cost-effective design and operating guidelines for
models will predict full-scale operations. small GAC systems that are capable of removing
pesticides (including dibromochloropropane) from
DWRD has evaluated microcolumn modeling drinking water supplies.
techniques, focusing on verifying the results of
a vailable computer models. DWRD conducted
microcolumn tests and used existing models to The third site, Wausau. has ground water with high
estimate carbon usage rates and other design levels of contaminants from a nearby Superfund site.
parameters for a full-scale operation. The modeled Air stripping was studied as a companion technology
results were then compared to isotherm model to GAC at this site. Modeling techniques were used to
predictions and to actual results obtained in the field. predict full-scale design criteria, and the modeled
The investigations concluded that microcolumn estimates were verified with actual cost and
predictions offered no particular advantage over performance data obtained from GAC contractors.
154 Upgrading EXisting or Designing New Drinking Water Treatment Facilities

size from 3 m 3 /day to 0.044 m 3 /sec (800 GPO to 1


8.2.2 Ozone Oxidation Systems MGD) were operated in eight locations. Both hollow
Ozone oxidation as a method of removing VOCs was fiber and spiral wound cellulose acetate membranes
studied in DWRD's in-house pilot plant facilities were tested. The results of the study indicate that 82
(Fronk, 1987a). To date, tests on 29 VOCs in distilled to 96 percent ofradium-226 was removed from all
and ground water revealed that ozone is effective in systems, thus meeting the EPA MCL of 5 pCilL.
removing aromatic compounds, alkenes, and certain
pesticides. It is not effective, however, in removing Reverse osmosis systems have also been tested for
alkanes. removing nitrate. In Suffolk County, where research
on removing organic compounds from ground water
Research showed that removal of some alkanes was has been ongoing, seven different commercially
enhanced by increasing pH. Also, increasing the available membranes were tested for nitrate removal
dosage of ozone improves removal of alkenes and (Lykins and Baier, 1985). Removals obtained ranged
aromatics. For most compounds, the composition of from 75 to 95 percent. Polyamide membranes were
the water (the background matrix) did not efficient in reducing nitrate and sacs.
significantly affect the removal efficiencies of ozone.
In general, it was found that, when treated with A field study in Charlotte Harbor, Florida, tested
ozone, VOCs behaved similarly in aqueous solutions reverse osmosis under conditions of high pressure
as in the gaseous or liquid phases, as the literature (18.6 to 25.2 kglcm 2 [265 to 359 psig)) (Huxstep,
concluded. 1981). Removal of nitrate and other inorganic
contaminants added to the ground-water influent
In order to further evaluate the effectiveness of using were monitored. The investigation showed that a
ozone to oxidize VOCs, DWRO funded a cooperative high pressure system could obtain better removals of
agreement with the Los Angeles Department of all substances monitored. For nitrate, the high-
Water and Power (LADWP) (Glaze and Kang, 1988). pressure system attained an 80 percent removal as
The oxidation ofTCE and PCE by ozone and opposed to the low-pressure system, which attained 6
modification of the traditional ozone treatment with to 24 percent removal.
the addition of hydrogen peroxide were studied at the
bench scale. Results from this project were very
promising and indicate that the kinetics of the 8.2.5 Ultrafiltration
H202/03 process warrants further research. Ultrafiltration is being studied as a method of
reducing production ofTTHMs in small water
treatment systems (Taylor et aI., 1987). This study
8.2.3 Ultraviolet Treatment involved treating two ground-water systems that
Ultraviolet light in combination with ozone may also contained large amounts of natural organic
be an effective means of controlling organic contaminants that act as TTHM precursors. Both
compounds. DWRD is conducting a study with the produced more than 400 giL TTHMs when treated
Los Angeles Water and Power Company to determine with conventional systems. A low-pressure
whether these two technologies can successfully treat membrane system to reduce precursors to TTHMs
organics in ground water (Glaze and Kang, 1988). was tested in a pilot plant. During pilot plant tests,
The goal of this treatment technique is to oxidize the ultrafiltration system produced finished water
undesirable chemicals into carbon dioxide and water, that easily met the TTHM MCL (100 gIL). Since the
thus eliminating the need to control off-gases, which tests were successful and the costs of the system were
are a problem with other oxidation processes. considered reasonable, more extensive testing of this
technology is under way.
8.2.4 Reverse Osmosis
DWRD is conducting experim~~ts at the bench and
pilot scale to evaluate the potential for controlling 8.2.6 Packed Tower Aeration
volatile and synthetic organic compounds using Three field studies were recently completed to study
reverse osmosis. Some field tests have been conducted packed tower aeration (PTA) techniques, which have
as well. Preliminary results indicate that, for ground- proven effective in removing VOC. In Wausau,
water treatment systems, certain membranes are Wisconsin, and Baldwin Park, California, research
very effective in removing a wide range oforganic focused on the effectiveness of PTA in removing
chemicals (Fronk, 1987b). VOCs from ground water (Clark et aI., 1989a). Off-
gas control technologies were tested at both sites as
Reverse osmosis has been proven as an effective well (Crittenden et aI., 1988; Lang et aI., 1987). In
treatment technology for radium. Several years ago, Wausau, it was found that air stripping was effective
EPA conducted field studies in cooperation with in removing compounds that normally are expected
Sarasota County, Florida (Sorg et aI., 1980). Six to be difficult to remove (compounds have lower
different types of reverse osmosis systems varying in Henry's Law constants) by air stripping.
Current and Emerging Research 155

At the third site, Brewster, New York, a pill't-scale 2 to 3 months and the brine disposed at a school
project evaluating air stripping techniques was district owned wastewater treatment plant.
completed. Results were used to develop models that
consulting engineers could use to estimate costs and Ion exchange systems, using both weak and strong
performance offull-scale aeration systems (Wallman acid resins in the hydrogen form, have been studied
and Cummins, 1986). as a means of controlling radium from ground water
(Clifford et aI., 1988). Both types of resins are capable
of removing radium to below the MCL of5 pCilL.
However, both systems require the pH of treated
8.2. 7 Conventional Treatment water to be alljusted by either air stripping or the
Although conventional treatment is not very combination of air stripping and caustic addition.
effective in removing organic contaminants, field
studies were conducted to test the effectiveness of EPA has also sponsored tests on the removal of
adding powdered activated carbon (PAC) to water radium with the DOW Chemical Company's Radium-
prior to conventional treatment (Miltner et aI., 1989). Selective Complexer (RSC), a synthetic resin with a
These tests were conducted in Tiffin, Ohio, where high affinity for radium. Field tests have shown the
river water containing seasonalIy high levels of RSC to remove radium from more than 50,000 bed
pesticides from area farms was a problem. The volumes of treated water. However, because the RSC
technique was found to effectively remove SOCs. cannot be regenerated, it must be disposed when its
capacity for radium is reached. For this reason,
studies have been conducted to examine its efficiency
for removing radium from the brine waste of a cation
8.2.8 Ion Exchange exchange system. A field study in Colorado showed
Ion exchange has been studied as a means of the RSC to be very effective for concentrating the
removing nitrate and uranium. In field tests, anion radium from the brine of an ion exchange system for
exchange has proven effective in removing nitrates. over 2 years. When its capacity is reached. this
A demonstration nitrate removal plant operated material will be disposed at a low radioacti ve waste
automatically in McFarland, California, for over 3 disposal site.
years (Lauch and Guter, 1986; Guter, 1982). The
plant has a capacity of 0.044 m3/sec (l MGD) and Recent EPA studies in Illinois have also shown that
consists of three anion exchange vessels designed to activated alumina modified with barium sulfate has
reduce nitrate-N levels to below 10 mglL (the current a removal capability similar to the RSC. Plain
MCL and State standard in California). In this plant, activated alumina also removed radium but its
part of the influent stream bypassed the treatment capacity was about a fifth of the modified alumina.
vessels and was blended with the treated water The advantage of the plain alumina was that it could
before leaving the plant. The blended water meets be regenerated whereas the modified alumina could
the required limits for nitrate and, in addition, meets not.
EPA's Secondary Regulations (40 CFR Part 143) for
chloride (250 mg/L) and sulfate (250 mg/L). The application of preformed manganese dioxide floc
for radium removal was studied at the same field site
Laboratory-, pilot-, and (small) full-scale tests have in Illinois and this process (adsorption)showed
been conducted to examine the effectiveness of anion promising results (Patel and Clifford, 1989). Feeding
exchange in removing uranium (Reid et aI., 1985). In of 1 to 2 mg/L of manganese oxide as manganese in a
the laboratory, more than 9,000 bed volumes were contact tank before a multimedia filter showed 80 to
treated before breakthrough was observed on a 95 percent removal of radium (12 pCilL) from the
ground water containing 300 IIg/L uranium was ground water. Cost estimates of this process indicate
passed through the anion exchange resin. In the field, that it should be competitive with ion exchange
twelve 0.007-m 3 (0.25-ft3) anion exchange systems treatment.
were installed at 10 sites (in New Mexico, Colorado,
and Arizona) where uranium levels in the ground
water exceeded 20 IIg1L. The field results actually 8.2.9 Technologies for Removing
showed better removals than the laboratory tests; Radionuclldes
some systems had treated over 50,000 bed volumes of DWRD has funded a project to study several
water when they were finally shut down. treatment techniques for removing radon gas from
drinking water (Kinner et aI., 1987a). Packed tower
A small 0.63-Usec (lO-GPM) full-scale anion aeration, diffused bubble aeration, and GAC have
exchange system has been shown to be very effective been evaluated at a field site in New Hampshire.
in removing 95 to 98 percent of 47 to 90 IIg1L of Both the packed tower and the diffused bubble
uranium from a water supply serving a school in aeration systems were very efficient at removing
Jefferson County, Colorado (Jelinek and Sorg, 1988). radon (90 to 99 percent). The GAC system flow rate
The system is regenerated with salt about once every and influent radon concentration above design
156 Upgrading Existing or Designing New Drinking Water Treatment Facilities

loadings were attributed to the lower removals by the bacteria. The new regulations, however, specify
GAC. This same study also investigated the removal requirements to control and regulate levels of
of radon by several low cost technologies associated Giardia, viruses, Legionella, and heterotrophic
with storage. The simple techniques consisted of bacteria. In view of these requirements, research
spraying the water into a storage tank, simple (extramural and in-house) is being conducted on
aeration (bubble) in the bottom of the storage tank, methods to achieve adequate disinfection of the
and the combination of these methods (Kinner et aI., specified microorganisms. Inactivation studies using
1987b). These methods showed removals ranging chlorine and chloramine have been completed on
from 50 percent by free fall of the water into the tank particle-associated coliforms, and on inactivation of
to around 90 percent with the combination offree fall three coliphages using chlorine, chlorine dioxide and
and bubble aeration. monochloramine (Berman et aI., 1988; Berman and
Sullivan, 1988). Disinfection studies on inactivation
of hepatitis A virus and model viruses in water are
8.2.10 Secondary Sources of Pollution ongoing; as are studies on inactivation of Giardia
Drinking water treatment technologies are designed (Ohio State University) using chlorine, chlorine
to remove pollutants from water. Once removed, dioxide, and chloramine; and Cryptosporidium
however, the pollutants can create a secondary source (U niversity of Arizona) using chlorine and
problem. Research on GAC systems treating surface chloramines (Sobsey et aI., 1988). Results of these
water found that dioxins were formed as by-products investigations wil\ provide contact time (CT) valwls
of the reactivation process (Lykins et aI., 1988a). for the various organisms and disinfectants based on
Afterburners were installed and operated at 1315.6°C temperature, pH, and disinfectant dosage (Hoff, 1986;
(2,400°F) to eliminate these toxic gases. Clark et aI., 1989b). An in-house project wil\ examine
the influence of strain variations in Giardia LambLia
Harmful compounds in air stripping of waste gases on disinfection inactivation.
can also be a problem. The field studies of air
stripping technologies at Baldwin Park, California, A study is being conducted on the inactivation of
and Wausau, Wisconsin, included testing of gas- viruses and Giardia using chlorine and chloramines
phase carbon adsorption techniques to remove in a flow-through treatment plant (John Hopkins
synthetic and volatile organic compounds from waste University).
gases (Lang et aI., 1987; Hand et aI., 1986). At
Baldwin Park a higher waste gas stack is also being The use of alternative disinfectants, such as ozone
used to increase dispersion of waste gases. applied as a preoxidant, has been shown to increase
the level of readily assimilable organic carbon (AOC)
in treated water. The trend toward the use of ozone
8.2.11 Small Systems Technologies and alternative disinfectants to control THM
Small systems account for most of the MCL violations formation in the United States may result in
under the Safe Drinking Water Act. DWRD has increased problems of bacterial regrowth, and
conducted and is continuing to conduct studies of possible coliform MCL compliance during water
technology that is specifically designed to solve small storage and distribution. Investigations are under
system problems. A recently completed study way to examine AOC levels in treated drinking water
demonstrated the feasibility of using point-of-entry and the relationship between microbial regrowth
devices for controlling arsenic and fluoride in a (coliforms) and AOC, and a recently completed
municipally sound treatment system <Rogers, in project examined modifications of a procedure
press). developed in the Netherlands to determine AOC
levels in drinking water (Kaplan and Bott, 1989).
8.3 Mandatory Disinfection Control of LegioneLla pneumophilia in drinking water
The proposed Surface Water Treatment Rule will treatment through prevention ofthe LegioTU!llaifree-
require communities whose water supply comes from living amoebae association is also being examined.
surface water or is influenced by surface water (e.g.,
ground water that is stored in open reservoirs) and
supplies more than 25 persons to filter and disinfect 8.3.1 Treatment/Distribution Microbiology
drinking water before distribution. Upon Examination offactors involved in microbial water
promulgation of the final rule, all affected suppliers quality have been addressed in other completed
must comply within 3 years. The Ground Water studies. Point-of-use or point-of-entry water
Treatment Rule, anticipated in 1991, is expected to treatment may have possible applications where
apply similar requirements to ground-water drinking central treatment is not feasible for various reasons.
sources. Microbiological studies of point-of-entry treatment
devices, used primarily for aesthetic purposes (taste
Historically, efficacy of disinfection has been and odor removal) have shown that the GAC
measured by testing for the presence of coliform cartridge devices become bacterial generators,
Current and Emerging Research 157

regardless of whether they contain silver as a results of this study indicated that leaching oflead
bacteriostatic agent (Reasoner et aI., 1987a,b). into drinking water is greatly affected by pH and the
age of a home. The leaching was the highest for new
The use of phosphate corrosion control additives on homes using low pH water. Furthermore, first draw
bacterial growth was examined to determine the role samples after overnight detention times had the
of added phosphate on coliform occurrence problems highest amounts of lead. A project report will
(Rosenzweig, 1987). GAC fines were shown to be a summarize the results of the study.
means by which particle-associated bacteria,
including coliforms, can pass from treatment into the In-house research includes studying lead
distribution system (McFeters et aI., 1987). Particle- contamination from kitchen faucets and in water
associated coliforms have also been shown to be less coolers. Of the 12 faucets used in this study,. the tree
susceptible to disinfection than those that are not cast brass faucets contributed significant amounts of
particle-associated. lead to the drinking water. Traces of lead were
obtained from the other metallic faucets and a plastic
faucet was lead free. The results of this study were
8.3.2 Bacterial Detection/Monitoring presented at the Water Quality Technology
A variety of studies, completed and ongoing, ha ve Conference in November 1988, and have been
examined various aspects of the bacteriological accepted for publication in the July 1989, issue of the
quality of drinking water. Studies of detection Jour1UJ1 ortM American Water Works Association
methods for bacteria in potable water, and for (Gardels and Sorg, 1989).
monitoring coliform bacteria in potable water include
the development and evaluation of a composite Twenty-two water coolers supplied by the U.S. Navy
sampling protocol for coliforms in treated and and two from a Portland, Maine, School District were
distribution water, comparison of media for used in the lead studies. Several coolers contributed
recovering total coliforms, detection of Klebsiellain lead to the drinking water in amounts considerably
water systems, the presence-absence coliform test for above the present standard. At the end of the
monitor-drinking water quality, a radiometric leaching period, the reservoirs were cut open and
method for detection offecal coliforms in water, and several had lead-lined tanks. The results of this study
occurrence of pigmented bacteria in treated and appeared in the April 10, 1989, Federal Register
distributed drinking water. An overview that entitled "Drinking Water Coolers that Are Not Lead
summarizes much of drinking water microbiology Free" (U.S. EPA, 1989c). This document and the one
research during the past decade is available entitled "Lead Contamination in School Drinking
(Reasoner, 1988). Water Supplies" were issued pursuant to the Lead
Contamination Control Act (U.S. EPA, 1989d). Some
chemical analyses and x-ray diffraction examina-
8.4 Prohibition of Lead Materials tions of internal deposits were performed on two of
The amendments to SDWA prohibit the use of lead in the reservoirs. The results will be presented as part of
pipes, pipe solder, or any other apparatus that comes a review of lead corrosion research at the AWWA
in contact with drinking water. On August 18,1988, Annual Conference in June 1989.
the Office of Drinking Water published in the
Federal Register a proposed new regulation for Initiation of new in-house research into the effect of
controlling lead and copper in drinking water (U.S. orthophosphate on lead solubility under different
EPA, 1988a). This document has drawn a water qualities, determination of corrosion products
considerable response, and a lot of work and and inhibitor films in lead pipes from field sites,
manpower is being devoted to technical issues integrating new data into corrosion models, and the
surrounding this legislation. On October 31,1988, effect of water softening on corrosivity towards lead
the Lead Contamination Control Act (LCCA) was and copper is planned. DWRD is also participating in
enacted as an amendment to SDWA (Pub. Law 92- a study to gather baseline data and look at the impact
339). The LCCA added Section 1464, which requires of municipal softening on water corrosivity in
EPA to publish and distribute to the States a Oakwood, Ohio, where there are still many lead
guidance document and testing protocol to assist service lines.
schools in determining the source of lead
contamination in school drinking water.
8.5 Systems and Cost Modeling Studies
The Drinking Water Research Division has funded a The DWRD has been modeling the deterioration of
major study in Long Island on the impact of lead and water quality in distribution systems. Subsequent to
other metallic solders on water quality. Ninety test water leaving the treatment plant, conditions in the
sites were selected to provide ten sites in each of nine distribution system sometimes lead to the
age groups ranging from new construction to introduction of contaminants to the water or to the
construction more than 20 years old. During the occurrence of other types of water quality decline.
study, the pH was adjusted to three levels. The The research in this area is focused on computer
158 Upgrading Existing or Designing New Drinking Water Treatment Facilities

modeling to predict and assist in alleviating


distribution system problems (Clark et aI., 1988c,d).

In addition to field and laboratory research, DWRD is


developing a computer model that will predict water
quality based on the known characteristics of the
distribution system. For example, TIHM levels can
be projected throughout the distribution system
based on the time of travel and the mixing of water
from various sources (Eilers and Clark, 1988).

Other modeling efforts include a project being


conducted in cooperation with French scientists
investigating the effects of hydraulics on the
propagation of contaminants in distribution systems.
Case studies have been conducted to develop models
that analyze distribution system characteristics and
identify factors that contribute to the potential for
water main breaks (Clark et aI., 1988a). Some models
estimate costs of repairs and other system
renovations.

DWRD has conducted cost modeling for many years.


Cost data have been collected from design studies and
merged with performance data to develop cost and
performance models. These models are very useful in
estimating the behavior of full-scale systems (Adams
et aI., 1989).

8.6 Future Directions


As the Safe Drinking Water Act matures, areas of
major research will be to complete activities in the
areas mentioned to this point. In addition, increasing
emphasis will be placed in the following areas:

• Use of ozone as a disinfectant and oxidant.


Particular emphasis will be placed on the
formation and control of disinfection by-products
and the maintenance of microbiological integrity
of distribution systems.
• Effects of the use of corrosion inhibitors as a
means of controlling lead leaching from
household plumbing.
• Characterization and treatment of residuals from
drinking water unit processes.
• Development of increasingly sophisticated cost
and performance models.
9. References

Abdel-Rahman, M.S., D. Couri, and J.D. Jones. 1980 Augenstein, H.W. 1974. Use of chlorine dioxide to
Chlorine dioxide metabolism in rats. J. Environ. disinfect water supplies. J. Am. Water Works Assoc.
Pathol. and Toxicol. 3:421-430. 66:716-717.

Adams, J.Q., and RM. Clark. 1989. Cost estimates Berman, D., E.W. Rice, and J.C. Hoff. 1988.
for GAC treatment systems. JA WWA. 81(1):35-42. Inactivation of particle-associated coliforms by
January. chlorine and monochloramine. Appl. Environ.
Microbiol. 54(2):507-512.
Adams, J.Q., R.M. Clark, and RJ. Miltner. 1989.
Controlling organics with GAC: a cost and Berman, D. and R. Sullivan. 1988. Inactivation of
performance analysis. JAWWA. 81(4):132-140. April. three coliiphages by chlorine, chlorine dioxide, and
monochloramine at 5°C. Project Report WERL·Cl-
Aieta, E.M., and J.D. Berg. 1986. A review of chlorine 0807. EPA/2-881039. U.S. EPA Risk Reduction
dioxide in drinking water treatment. J. Am. Water Engineering Laboratory. Cincinnati, OH.
Works Assoc. 78:62-72.
Bryck, J., B. Walker, and G. Mills. 1987. Giardia
Aieta, E.M., K.M. Reagan,J.S. Lang, L. McReynolds, removal by slow sand filtration - pilot to full scale.
J.-W. Kang, and W.H. Glaze. 1988. Advanced Paper presented at the AWW A National Conference.
oxidation processes for treating ground water June.
contaminated with TCE and PCE: pilot-scale
evaluations. JAWWA. 80(5):64-72. May. Chrobak, RS., D.L. Kelleher, and I.H. SufTet. 1985.
Full-scale GAC adsorption performance compared to
Aieta, E.M., and P. V. Roberts. 1986. Application of pilot plant predictions. Paper presented at the 1985
mass-transfer theory to the kinetics of a fast gas- Annual Conference of the American Water Works
liquid reaction: chlorine hydrolysis. Environ. Sci. Association, Washington, DC. June 27.
Techno!. 20:44-50.
Clark, RM., RJ. Miltner, C.A. Fronk, and T.F.
Akin, E. et al. 1987. The U.S. EPA Office of Research Speth. 1989a. Unit process research for removing
and Development Health Research Program on VOCs from drinking water: an overview in
Drinking Water Disinfectants and their By-Products. significance and treatment of volatile organic
Issue paper prepared for a Science Advisory Board compounds in water supplies. Lewis Publishers, Inc ..
Program Review. Chelsea, Michigan.
Clark, R.M., E.J. Read, and J.C. Hoff. 1989b.
American Water Works Association. 1973. Water Analysis of inacti vation of Giardia lamblia by
Chlorination Principles and Practice. AWW A chlorine. Journal of Environ. Eng. ASCE. 115(1):80-
Publication :'-10. M20, p. 12. 90. February.
American Water Works Association Research Clark, R.M., RG. Eilers, and J.A. Goodrich. 1988a.
Foundation and KIW A. 1983. Occurrence and Distribution system: cost of repair and replacement.
removal of volatile organic chemicals from drinking In: Proceedings of the Conference on Pipeline
water. Infrastructure, ASCE, Boston, MA. June 6-7.

159
160 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Clark, RM., C.A. Fronk, and BW. Lykins, Jr. 1988b. Dobbs, RA. and J.M. Cohen. 1980. Carbon adsorption
Removing organic contaminants from ground water. isotherms for toxic organics. Report No. EPA 600/8-
Environ. Sci. Tech. 22(0): 1126-1130. 80-023. Office of Research and Development, MERL,
Cincinnati,OH.
Clark, RM., W.M. Grayman, and RM. Males. 1988c.
Contaminant propagation in distribution systems. J. Dyksen, J.E., K. Raman, RF. Raczko, and R.M.
of Environ. Eng. ASCE. 114(4):929·943. Clark. 1987. GAC treatment costs - minimizing
them. Paper presented at the AWW A Annual
Clark, R.M., W.M. Grayman, RM. Males, andJ.A. Conference, Kansas City, MO. June 14-18.
Coyle. 1988d. Modeling contaminant propagation in
drinking water distribution systems. Aqua. No.3, pp. Eilers, RG. and RM. Clark. 1988. Flow separation
137-151. conditions at pipe walls of water distribution mains.
[11: Proceedings of the AWW A Annual Conference,
Clifford, D., W. Jijjeswaropu, and S. Subramonian. Orlando, FL. June 19-23. pp. 1619-1624.
1988. Evaluating various absorbents and membranes
for removing radium from ground water. JAWWA. Feige, W.A., RM. Clark, B.W. Lykins, Jr., and C.A.
80(7):94. Fronk. 1987. Treatment of water from contaminated
wells. [11: Rural Groundwater Contamination. Lewis
Clifford, D., S. Subramonian, and T. Sorg. 1986. Publishers, Inc.
Removing dissolved inorganic contaminants from
water. Environ. Sci. and Tech. 20(1): 1072-1080. Fronk, C.A. 1987a. Destruction ofvolatile organic
November. contaminants in drinking water by ozone treatment.
Ozone Science and Engineering. 9(3):265-287.
Cline, G.C., T.J. Lane, and M. Saldamando. 1985.
Packed column aeration for trichlorethylene removal Fronk, C.A. 1987b. Removal of low molecular weight
at Scottsdale, Arizona. Paper presented at the 1985 organic contaminants from drinking water using
Annual Conference of the AWW A held in reverse osmosis membranes. [11: Proceedings of the
Washington, DC. June. AWWA Annual Conference, Kansas City,
MO/Denver, CO.
Collins, F. andC.Y. Shieh. 1971. More water from the
same plant. The American City. p. 96. October. Gardels, M.C. and J.J. Sorg. 1989. A laboratory study
of lead leaching from drinking water faucets.
Conley, W.R and S.P. Hansen. 1987. Trident pilot JAWWA.July.
study: Lakeport, California. Unpublished Report.
HDR Engineering. March. Glaze, W. H. 1986. Reaction products of ozone: a
review. Environ. Hlth Persp. Vol. 69, p. 151.
Crittenden, J.C., R.D. Cartright, B. Rick, S.R Tang,
and D. Perram. 1988. Using GAC to remove VOCs Glaze, W.H. andJ.-W. Kang. 1988. Advanced
from air stripper off-gas. JAWW A. 80(5):64·72. oxidation processes for treating ground water
contaminated with TeE and PCE: laboratory studies.
Cryer, E.T. 1986. Preozonation at Kennewick, JAWWA. 80(5):57-63. May
Washington, Case history. Paper presented at IntI.
Ozone Assoc. Workshop on Drinking Water Glaze, W.H., J.-W. Kang, and M. Aieta. 1987. Ozone-
Treatment with Ozone, Perrysburg, OH, April. Hydrogen peroxide systems for control of organics in
municipal water supplies. [11: D. W. Smith and G.R
Culp, R. L. 1976. Increasing water treatment capacity Finch (eds), Proc. Second IntI. Conf. on the Role of
with minimum additions. Public Works. August. Ozone in Water and Wastewater Treatment.
TekTran International Ltd., Kitchener, Ontario,
DeMarco, J. 1983. Experiences in operating a full· Canada. pp. 233-244.
scale granular activated carbon system with on-site
reactivation. [11: I.H. SufTet and M.J. McGuire (ed.), Glitsch, Inc. P.O. Box 660053. Dallas, TX.
Treatment of Water by Granular Activated Carbon.
Guter, Gerald A. 1982. Removal of nitrate for
Advances in Chemistry Series, American Chemical
contaminated water supplies for public use. Final
Society Books.
Report. Project Summary. Report No. EPN600/S2-
82-042. Cincinnati, OH.
DeWalle, F.B., J. Engeset, and W. Lawrence. 1984. Hand, D.W., J.C. Crittenden, J.L. Gerbin, and B.W.
Removal of Giardia lamblia cysts by drinking water Lykins, Jr. 1986. Design and evaluation of an air
plants. Report No. EPN600/S2-84-069, U.S. EPA, stripping tower for removing VOCs from ground
MERL, Cincinnati, OH. ~ay. water. JAWWA. 78(9):87-97.
References 161

Hansen,S.P. 1987. Adsorption clarifier pilot study Report No. EPA/600/2-84-114. C .5. Environmental
evaluation. Unpublished Report. Lake Arrowhead Protection Agency, MERL, Cincinnati, OH June.
Community Services District, September. NTIS: PB-84-212703.

Hoff, J.C. 1986. Inactivation of microbial agents by Lauch, R.P. and G.A. Guter. 1986. [on exchange for
chemical disinfectants. Report No. EPA/600/2-86- the removal of nitrate from well water. JAWWA.
067. Risk Reduction Engineering Laboratory, U.S. 78(5):83-88.
EPA, Cincinnati, OH.
Letterman, R.D. and T.R. Cullen, Jr. 1985. Slow sand
Howe, E.W., E.M. Aieta, S. Liang, and M.J. McGuire. filtration maintenance costs and effects on water
1989. Removal of chlorite residuals with granular quality. Report No. EPA/600/2-85-056. U.S.
activated carbon: a case study. Presented at Environmental Protection Agency, Cincinnati, OH.
International Chlorine Dioxide Symposium, Denver, May. NTIS: PB-85-199669/AS.
CO (Washington, DC: Chemical Manufacturers
Association). Lin, S.D. 1989. Proposed lead rules add chores, costs
to utility burden. Water Engineering and
High-tech water on the green fields oflowa. 1985. Management. pp. 35-47. February.
Water Engineering and Management. January.
Lowry, J.D. andJ.E. Brandown. Removal of radon
Huxstep, Martin R. 1981. Inorganic contaminant from ground water supplies using granular acti vated
removal from drinking water by reverse osmosis. carbon or diffused aeration. University of Maine,
Project Summary. Report No. EPA/600/S2-81-115. Department of Civil Engineering, Orono, ME.
WERL, Cincinnati, OH.
Lykins, B.W., Jr., R.M. Clark, and J.Q. Adams
Jelinek, R.T. and T.J. Sorg. 1988. Operating a full- 1988a. Granular activated carbon for controlling
scale ion exchange system for uranium. JA WWA. THMs. JAWWA. pp. 85-92. May.
8(7):79. July.
Lykins, B.W., Jr., R.M. Clark, and D.H. Cleverly
Kaplan, L.A. and T.L. Bott. 1989. Nutrients for 1988b. Polychlorinated dioxin and furan discharge
bacterial growth in drinking water: bioassay during carbon reactivation. Journal of Environ. Eng.
evaluation. Report No. EPA/600/2-89/030. Risk ASCE. 114(2):300-316. April.
Reduction Engineering Laboratory. U.S. EPA.
Cincinnati,OH. Lykins, B.W., Jr. and M.H. Griese. 1986. l:sing
chlorine dioxide for trihalomethane control. J. Am.
Kinner, N.E., C.E. Lassard, and G.S. Schell. 1987a. Water Works Assoc. 78:88-93.
Low cosUlow technology aeration techniques for
removing radon from drinking water. Report No. Lykins, B.W., Jr. and J.A. Baier. 1985. Removal of
EPA/600/M-87/031. US EPA. agricultural contaminants from ground water. !n:
Proceedings of the AWWA Conference. pp. 1151-
Kinner, N.E., C.E. Lassard, and G.S. Schell. 1987b. 1164.
Radon removal from small community public water
supplies using granular activated carbon and low Lykins, B.W., Jr., E.E. Geldreich, J.Q. Adams, J.C.
technology/low cost techniques. In.: Proceedings of the Ireland, and R.M. Clark. 1984. Granular activated
AWWA Seminar on Radionuclides in Drinking carbon for removing nontrihalomethine organics
Water, Kansas City, MO. from drinking water. Report No. EPA/600/S2-84-165.

Lang, J.S., A.R. Trussel, M.D. Umphries, and C.H. McFeters, G.A., A.K. Camper, M.W. LeChevallier,
Tate. 1987. Gas adsorption of air stripping emissions. S.C. Broadway, and D.G. Davies. 1987 Bacteria
In: Proceedings of AWWA Seminar on Treatment attached to granular activated carbon in drinking
Processes for the Control of Synthetic Organic water. Environmental Research Brief. Report No.
Chemicals, Kansas City, MO. EPA/600/M-87/003. Risk Reduction Engineering
Laboratory, U.S. EPA, Cincinnati, OH.
Lange, K.P., R. T. Rice, and W. Hemphill. 1985.
Integration of contact clarifier and filtration McKeon, W.R., J.J. Muldowney and B.S. Aptowicz.
processes, Microfloc Product. Paper presented at the 1986. The evolution ofa modified strategy to reduce
19a5 Rocky Mountain AWW NWPCA Conference, trihalomethane formation. In: Proc. Annual Meeting,
Albuquerque, NM. Am. Water Works Assoc., Denver, CO. pp. 967-997

Lange, K.P., W.D. Bellamy, and D.W. Hindricks. McKinnon, R.J. andJ.E. Dyksen. 1982. Aeration plus
1984. Filtration of Giardia cysts and other carbon adsorption remove organics from Rockaway
substances. Vol. 1. Diatomaceous Earth Filtration. Township (NJ) ground water supply. Paper presented
162 Upgrading Existing or Designing New Drinking Water Treatment Facilities

at the 1982 Annual Convention of the American Randtke, S. 1988. Organic contaminant removal by
Society of Civil Engineers, New Orleans, LA. October coagulation and related processes. J AWWA.
25-27. 80(5):40-56. May.

Miller, G.W., RG. Rice, C.M. Robson, RL. Scullin, Reasoner, D.J. 1988. Drinking water microbiology
W. Kuhn and H. Wolf. 1978. An assessment of ozone research in the United States: an overview of the past
and chlorine dioxide technologies for treatment of decade. Water Sci. Techno!. 20(11/12): 101107.
municipal water supplies. EPA Report No.
EPN600/2-78-147. Reasoner, D.J., J.C. Blannon, E.E. Geldreich, and
J .A. Barnick. 1987a. Microbiological studies of
Miltner, R.J. and C.A. Pronk. 1985. Treatment of activated carbon point·of-use systems: an interim
synthetic organic contaminants for Phase II report. Report No. EPN600/2-88/004. Risk Reduction
Regulations. Internal Report. DWRD, U.S. EPA. Engineering Laboratory, U.S. EPA, Cincinnati, OH.
December.
Reasoner, D.J., J.C. Blannon, and E.E. Geldreich.
Miltner, RJ., T.F. Speth, and C.A. Fronk. 1989. 1987b. Microbiological characteristics of third-faucet
Treatment of seasonal pesticides in surface waters. point-of· use devices. JAWWA. 79(10):60-66.
JAWWA. 81(1):43-52. January.
Reid, George N., P. Lassovszky, and S. Hathaway.
Morand, J.M. and M.J. Young. 1983. Performance 1985. Treatment, waste management and cost for
characteristics of package water treatment plants. removal of radioactivity from drinking water. Health
Project summary. Report No. EPN600/S2-82-1 01, Physics. 48(5):671-694.
Municipal Environmental Research Laboratory,
Cincinnati, OH. March. Roberts, P. V. and P. Dandliker. 1982. Mass transfer
of volatile organic contaminants during surface
Morse, D. 1988. Lead: taking it from the tap. Civil aeration. Paper presented at the 1982 Annual
Engineering, pp. 71-73. Conference of the AWWA, Mia!"i Beach, FL. May.

Myers, G.L., A. Thompson, D.M. Owen and J.M. Rogers, K.R. in press. Arsenic and f1uoride removal
Baker. 1986. Control oftrihalomethanes and taste from drinking water at San Ysidro, N.M. Risk
and odor at Galveston County Water Authority. In: Reduction Engineering Laboratory, Office of
Proc. Annual Meeting Am. Water Works Assoc., Research and Development, U.S. EPA, Cincinnati,
Denver, CO, 1986, pp. 1667-1675. OH.

National Academy of Sciences. 1980. Drinking Water Rosenzweig, W.D. 1987. Influence of phosphate
and Health. Vol. 1, Chapter 2. Washington, D.C., corrosion control compounds on bacterial growth.
National Academy Press. Report No. EPN600/2-87/045. Risk Reduction
Engineering Laboratory, U.S. EPA, Cincinnati, OH.
O'Brien, RP., D.M. Jordan, and W.R Musser. 1981.
Trace organics removal from contaminated ground Routt, J.C. 1989. Chlorine dioxide trihalomethane
waters with granular activated carbon. Calgon control and distribution system odor problem: a
Corporation, Pittsburg, PA. Paper presented at the utility'S 5-year experience. Paper presented at the
National American Chemical Society Meeting, AWWA Annual Meeting, Los Angeles, CA. June.
Atlanta, GA. March 29 - April 3.
Roy, D., RS. Englebrecht, and E.S.K. Chian. 1982.
Patel, Rand D. Clifford. 1989. Radium removal from Comparative inactivation of six enteroviruses by
water by manganese dioxide adsorption and ozone. JAWWA. 74(12):660-664.
diatomaceous earth liItration. Draft Report. U.S.
EPA.
Ruggiero, D.D. and R. Ausubel. 1982. Removal of
Pinsky, D.E. and V.F. Coletti. 1988. Rehabilitating a organic contaminants from drinking water supply at
:-.lew England water treatment plant. J. New Glen Cove, New York, Phase II. Report No.
England Water Works Assoc., pp. 132-157. June. EPN600/2-821027, U.S. EPA. March.

Pyper, G.R 1985. Slow sand liIter and package


treatment plant evaluation operating costs and Ruggiero, D.D. and R Ausube!. 1980. Removal of
removal of bacteria, Giardia, and trihalomethanes. organic contaminants from drinking water supply at
Report No. EPN600/2-85-052. U.S. Environmental Glen Cove, New York, Phase I. Report No. EPA600/2-
Protection Agency, Cincinnati, OH. April. 80-198.
References 163

Schock, M.R 1980. Response oflead solubility to water regulations for inorganics, Part II: arsenic and
dissolved carbonate in drinking water. JA WWA. Vol. selenium. JAWW A. p. 379-393. July.
72, p. 695.
Sorg, T.J., R Forbes, and D. Chambers. 1980.
Schock, M.R and I. Wagner. 1985. Corrosion and Removal ofradiu-226 from Sarasota County,
solubility of lead in drinking water: Internal Florida's drinking water by reverse osmosis.
corrosion of water distribution systems. AWW A JAWWA. pp. 230-237. April.
Research FoundationJDVGW.
Sorg, T.J., M. Csanady, and G. Logsdon. 1978.
Sequeira, J.A., L. Harry, S.P. Hansen, and RL. Culp. Treatment technology to meet the interim primary
1983. Pilot filtration tests at the American River drinking water regulations for inorganics, Part III:
water treatment plant. Public Works. 114: 1:36. cadmium, lead, and silver. JAWWA. pp. 680-691.
December.
Singer, P.S. 1988. Alternative oxidant and
disinfectant treatment strategies for controlling Speth, T.A. and RJ. Miltner. 1989. Effect of
trihalomethane formation. Report No. EPA/600/S2- preloading on the scale-up ofGAC microcolumns.
881044, Water Engineering Research Laboratory, JAWWA. 81(4):141-148. April.
Cincinnati, Ohio. October.
Stevens, A.A., L.A. Moore, and R.J. Miltner. 1989.
Singer, P.S. 1986. THM control using alternate Formation and control of nontrihalomethane by·
oxidant and disinfectant strategies: an evaluation. products. Tn.: Proceedings of the AWWA Water
Tn: Proc. 1986 Annual Conference, Am. Water Works Quality Technology Conference, St. Louis, Missouri.
Assoc. Denver, CO. pp. 999-1017.
Stevens, A.A., L.A. Moore, C.J. Slocum, B.L. Smith,
Slootmaekers, B., S. Tachiyashiki, D. Wood, and G. D.R. Seeger, and J.C. Ireland. 1987. By-products of
Gordon. 1989. The removal of chlorite ion and chlorination at ten operating utilities. In.:
chlorate ion from drinking water. Presented at Proceedings of the Sixth Conference on Water
International Chlorine Dioxide Symposium, Denver, Chlorination: Environmental Impact and Health
CO, November (Washington, DC: Chemical Effects. Oak Ridge Associated Universities, Oak
Manufacturers Association). Ridge, Tennessee. May 3-8.

Sobsey, M.D., T. Fugi, and P.A. Shields. 1988. Tanner, S. 1988. Evaluation of Slow Sand Filters in
Inactivation of hepatitis A virus and model viruses in Northern Idaho. Idaho Division of Environmental
water by free chlorine and monochloramine. Tn: Quality. May 9.
Proceedings of International Conference on Water
and Wastewater, Microbiology, Newport Beach, Taylor, J.S., D.M. Thompson, J.K. Carswell. 1987.
California. 2(64): 1-7. Applying membrane processes to ground water
sources for THM control. JA WW A. 79(8):72-82.
Sobsey, M.D. 1988a. Detection and chlorine
disinfection of he patitus A in water. CR-813-024, U.S. EPA. 1989. National Primary and Secondary
EPA Quarterly Report. December 1988. Drinking Water Regulations. Proposed Rule. Fed.
Reg. 54(97}:22064. May 22.
Sorg, T.J. 1979. Treatment technology to meet the
interim primary drinking water regulations for U.S. EPA. 1989a. Guidance manual for compliance
inorganics, Part IV: chromium alid mercury. with the filtration and disinfection requirements for
JAWWA. pp. 454-466. August. public water systems using surface water sources.
Draft. Science and Technology Branch, Office of
Sorg, T.J. 1978. Treatment technology to meet the Drinking Water, U.S. EPA, Washington, DC,
interim primary drinking water regulations for October.
inorganics, Part I: nitrite and fluoride. JA WWA.pp.
105-112. February. U.S. EPA, 1989b. Drinking water; national primary
drinking water regulations; filtration, disinfection,
Sorg, T.J and G. Logsdon. 1980. Treatment turbidity, Giardia lamblia, viruses, LegioMlla, and
technology to meet the interim primary drinking heterotrophic bacteria. Final Rule. Fed. Reg.
water regulations for inorganics, Part V: barium and 54(124):27485-27541. June 29.
radionuclides. JA WWA. pp. 411-422. July.
U.S. EPA. 1989c. Drinking water coolers that are not
Sorg, T.J. and G. Logsdon. 1978. Treatment lead free. Proposed List and Request for Comments,
technology to meet the interim primary drinking Fed. Reg. 54(67): 14320-14322. April 10.
164 Upgrading Existing or Designing New Drinking Water Treatment Facilities

U.S. EPA. 1989d. Lead contamination in school U.S. EPA. 1984. Corrosion manual for internal
drinking water supplies. Guidance Document and corrosion of water distribution systems. Report No.
Testing Protocol. Notice of Availability and Request EP N570/9-84-00 1. Office of Drinking Water. April.
for Comments. Fed. Reg. 54(67):14316-14318. April
10.
U.S. EPA. 1983. Microorganism removal for small
U.S. EPA. 1988a. Drinking water regulations: water systems. Report No. EPN570/9-83-012. U.S.
Maximum Contaminant Level Goals and Natural EPA, Office of Drinking Water, Washington, D.C.
Primary Drinking Water Regulations for lead and June.
copper. Proposed Rule. Fed. Reg. 53(160):31516.
August 18. U.S. EPA. 1979. Estimating water treatment costs.
Volumes 1-4. Report No. EPAl600/2-79-162a.
Municipal Environmental Research Laboratory,
U.S. EPA. 1988b. Technologies and costs for the Office of Research and Development, U.S. EPA,
removal of synthetic organic chemicals from potable Cincinnati, OH. August.
water supplies. Draft. U.S. EPA, Office of Drinking
Water, Washington, DC. September.
U.S. EPA. 1973. Process design manual for carbon
adsorption. Technology Transfer. October.
U.S. EPA. 1988c. Drinking water; substitution of
contaminants and drinking water priority list of Wallman, H. and M.C. Cummins. 1986. Design scale-
additional substances which may require regulation up suitability for air stripping columns. Report No.
under the Safe Drinking Water Act. Federal Register EPN600/S2-86-009. U.S. EPA, Drinking Water
53(14): 1891-1902. Research Division, Cincinnati, OH.

Westerhoff, G.P. 1971. Experiences with higher


U.S. EPA. 1987. Technologies and costs for the filtration rates. J AWWA. 63:6:376.
removal of microbiological contaminants from
potable water supplies. Draft. Office of Drinking Westerhoff, G.P. and R. Miller. 1985. Design of the
Water, Criteria and Standards Branch. nation's first major GAC facility for drinking water
treatment at Cincinnati, Ohio. Paper presented at
the 1985 Annual Conference of the AWWA held in
U.S. EPA. 1986a. Safe Drinking Water Act 1986 Washington, DC. June.
Amendments. Report No. EPN570/9-86-002. Office
of Drinking Water. August. Williams, R.B. and G.L. Culp. 1986. Handbook of
Public Water Systems. New York, New York, Van
Nostrand Reinhold Company.
U.S. EPA. 1986b. Design manual: municipal
wastewater disinfection. Report No. EPAl625/1- Wolfe, R.L., M.H. Stewart, K.N. Scott and M.J.
86/021 (October 1986). U.S. EPA, Office of Research Mcguire. 1989. Inactivation of Giardia muris and
and Development, Water Engineering Research indicator organisms seeded in surface water supplies
Laboratory, Center for Environmental Research by peroxone and ozone. Env. Sci. Technol. 23(6):744-
Information, Cincinnati, OH 45268. 745.
Appendix A-Experience Modifying Existing
Filtration Systems

This Appendix provides brief case histories of settlers when preceding either dual- or mixed-media
modifications resulting in plant expansions (Section filters.
A.I) and installation of different filtration systems,
including slow sand (Section A.2), package plant The area of the basin covered by tube modules was
(Section A.3), and diatomaceous earth (Section A.4) calculated using the allowable tube capacity to
filters. Experience selecting a filtration system is achieve the desired overflow rate. The two-step
provided in Section A.5. calculation is based on acceptable plant capacity and
tube rate parameters:
A.1 Upgrading Existing Treatment
Facilities 1. Quantity capacitY,GPM
of tubes =
A. 1.1 Horizontal Flow Basin Example (area) allowable tube rate, GPMlsq ft
This example involved a modification of existing
facilities. The plant equipment includes a
rectangular sedimentation basin and rapid sand 8 MGD X 700 GPM/MGD
filters with a capacity of 0.18 m3/sec(4 MGD). The =
basin is 9 meters (30 ft) wide, 40.5 meters 033 ft) 3 GPMlsq ft
long, and averages 4.5 meters (15 ft) in depth. The
design surface overflow rate is 40.7 m 3/day/m 2 0,000 1,870 ft2 or 173.7 m 2
GPD/sq f't). The sedimentation basin has a single
overflow weir across the outlet end. Sedimentation is
preceded by mechanical flocculation with a 40-
minute detention period. Coagulant aids are added 2. Area length x width
during periods of high turbidity and low water
temperature conditions. Influent water 1,870 length x 30 ft
characteristics include a maximum turbidity ranging
from 25 to 30 NTU and a temperature rarely below length = 1,870/30 = 62.5 ft or 19 m
10°C (50°F).

The plan was to expand the plant capacity from 0.18 The tube lengths were rounded oITto the nearest 10 ft
to 0.35 m 3 /sec (4 to 8 MGDl. At a 0.35 m 3 /sec (8- to permit use of standard tube module dimensions.
MGDl capacity, the sedimentation basin overflow
rate is 81.5 m 3/day/m 2 (2,000 GPD/sq ft) or 0.95 The tube modules were installed in an area extending
Useclm 2 0.4 GPM/sq ft). This basin overflow rate, from the outlet end of the basin to an isolation baffie
coupled with the plant's influent turbidity, are the wall separating the tube modules from the area of
primary factors used to derive the appropriate inlet turbulence. Three new effiuent launders
overflow rate for the tube modules. According to the extending from the exit end wall were required to
recommended overflow rates for systems with ensure uniform flow through the tube modules. The
influent temperatures generally over 10°C (50°F), an launders were installed at 3-m OO·ft) intervals. The
overflow rate of2.04 Useclm 2 (3 GPMlsq ft) is tube modules were submerged 1.2 m (4 ft) because of
acceptable for the basin area covered by the tube the 4.8-m 06-ftl depth of the basin.

165
166 Upgrading Existing or Designing New Drinking Water Treatment Facilities

A.1.2 Upflow Solids Contact Clarifier media; each pilot filter was monitored for turbidity
Example and headloss.

This example also involves the modification of Figure A-I shows the pilot filtration equipment
existing facilities. The plant has two 3.9-m 2 (42-sq ftl setup. Influent came either directly from the river or
upflow clarifiers of 189.2 Usee (3,000 GPM),each after coagulation, flocculation, and sedimentation in
with peripheral collection launders. The total surface the plant. Pilot filter influent turbidities ranged from
area of each clarifier is 163.5 m2 (1,760 sq ft.). The 3 to 33 NTU and pilot filter rates were set at 3.4,4.4,
influent center well consumes 18.6 m2 (200 sq ft.) of and 6.8 Usedm 2 (5, 6.5, and 10 GPM/sq ft.). Alum
settling area. dosages ranged from 12 to 21 mglL and polymer
dosages ranged from 0 to 0.3 mglL. The City of
Peak flows reach 1.3 Usedm 2 (1.92 GPMlsq ft.), which Sacramento allows only 0.1 NTU of turbidity in
is high enough to impede clarifier performance, finished water, so the pilot tests were terminated
especially at colder temperatures. Influent turbidity when turbidity levels achieved that figure. Both
ranges from 30 to 70 NTU. Influent temperature mixed- and dual-media filters performed well at all
rarely falls below 10·C (50·F). filter test flow rates, but mixed-media filters were
more reliable in achieving the desired turbidity goal.
The plan is to increase the capacity of each clarifier to
252.3 Usee (4,000 GPM). At this flow, the clarifica- The recommendations from the pilot test data
tion rate based on the available overflow rate endorsed both dual- and mixed-media filters. A
allowable for tube settlers would be 1.77 Usedm 2 (2.6 polyelectrolyte preparation and feed system were also
GPM/sq ftl. This rate is within the recommended recommended. During qualifying tests of the new
limits for the expected influent turbidity levels and filters. the plant produced effiuent turbidities ofless
could be achieved by totally covering the basin with than 0.1 NTU (generally between 0.06 and 0.08
tubes. Total coverage of each basin is the simplest NTU).
solution because of the ease of supporting the tube
modules. This would result in 144.9 m2 (1,560 sq ft.) of A.1.4 Erie County, New York (Westerhoff,
tubes, with radial launders added to provide equal
flow distribution. At the 1.77 Usedm 2 (2.6-GPMlsq 1971)
ftl clarifier overflow rate, the probable effiuent This case involves upgrading a 1.93 m3/sec (44 MGDl
turbidity would fall within 3 to 7 NTU. An effiuent plant to 2.63 m3/sec (60 MGD). The Erie County
with this turbidity level is acceptable for filtration Water Authority built this plant in 1961 using Lake
with dual- or mixed-media filters. Erie as its water source. The steps of the plant
treatment train are:

A.1.3 Sacramento, California (Sequeira • Aeration


• Chemical addition
et aI., 1983) • Rapid mixing
This case involves upgrading the capacity of a • Flocculation and sedimentation
municipal plant from 2.6 to 4.4 m3/sec (60 to 101 • Filtration
MGDl. This plant, built in 1963, uses the American • Chlorination
River as its water source. Originally the plant
provided for 25 minutes of flocculation followed by The Authority decided to add six new mixed-media
115 minutes of sedimentation detention. The filters to the existing four rapid sand filters as the
expansion reduced flocculation detention to 14 first stage of facility expansion. The decision to use
minutes and sedimentation time to 65 minutes. mixed-media filters for the expansion was based on a
Filter flow rates increased from 2.0 to 3.6 Usedm 2 year-long study in which 48 parallel filters were
(3.0 to 5.3 GPMlsq ft.) at the expanded flow. tested. Sand filters were tested at a flow rate of 1.36
Usedm 2 (2 GPMlsq ftl. while mixed-media filters
The feasibility study found that a change in filtration were tested at rates of 1.36 to 6.79 Usedm 2 (2 to 10
media alone from rapid sand to either dual or mixed GPMlsq ftl. Tests were terminated at head losses of
media, would compensate for higher turbidities 2.4 m (8 ftl. Test influent turbidities ranged from 2 to
anticipated from the increased plant flow rate. It was 4 NTU, clarifier overflows ranged from 25.5 to 38.3
concluded that the modification could be m3/day/m 2 (625 to 940 GPD/sq ftl, and alum dosages
accomplished without structural changes. ranged from 14 to 16 mglL.

Extensive pilot tests of various filter media were The test results indicated that:
conducted fromFebruary to May, 1982, to determine
the optimal filter medium in terms of efficiency and • Sand filter effiuent turbidity was 0.10 NTU or
economics. The tests involved four types of filter less for greater than 80 percent of the time.
Appendix A-Experience Modifying Existing Filtration Systems 167

Plant Filler Influent Water


(FollOWing Alum Addition, MIXIng,
Flocculation. and Settling)

NOTE:
Fine Filler Media Depth IS 24
Inches Throughout

} Pilot Filters

} Effluent Rate of Flow Controllers


and Meters

} Turbldimeters

Figure A-t. Flow diagram 01 the plio! Ilitration equipment.

• Mixed-media filter effiuent turbidity was 0.10 • Maximum turbidity of less than 0.5 NTU
NTU or less for greater than 88 percent of the • Average total microscopic count ofless than 200
time. sulmL
• Maximum total microscopic count of less than
• Mixed-media filters produced consistently less 300sulmL
turbid finished water at flow rates of 1.36 to 6.79 • Less than 1 unit of color
UsecJm 2 (2 to 10 GPMlsq mthan did sand filters • Less than detectable odor
at flow rates ofjust 1.36 Useclm 2 (2 GPMlsq fl.). • Less than 0.05 mglL of aluminum
• Less than 0.05 mglL of iron
• Mixed-media filters had lower microscopic counts
of particulate matter at flow rates of up to 4.1
UsecJm 2 (6 GPMlsq m than did sand filters at
1.36 UsecJm 2 (2 GPMlsq m.
A.l.S Corvallis, Oregon (Collins and
Shieh, 1971)
• Mixed-media filters at flow rates of 3.4 UsecJm 2 This case involved upgrading an 0.35 mJ/sec (8 MGD)
(5 GPMlsq mhad an average run length of29 plant to 0.92 mJ/sec (21 MGD). This municipal plant,
hours. which uses water from the Willamette River, was
planned in phases offour MGD subunits, with the
• Mixed-media filters at flow rates of 4.1 UsecJm 2 first phase built in 1949. The second phase was built
(6 GPMlsq fl.) had an average run length of20 in 1961, but the third and fourth were never built.
hours. Instead, the 0.35 mJ/sec (8 MGDJ plant was expanded
with nonstructural modifications during the early
• Mixed-media filters at flow rates offrom 3.4 to 1970s. Willamette River water has turbidity ranging
4.1 UsecJm 2 (5 to 6 GPMlsq fl.) used from 15 to 30 NTU, with surges to 1,000 NTU. The
proportionately less wash water than sand filters original facility included flocculation and
at a flow rate of 1.36 UsecJm 2 (2 GPMlsq fl.). sedimentation basins, with two rapid sand filters per
set of flocculation and sedimentation basins.
The finished water goals for this modified plant,
using mixed-media filters at a rate of 4.1 UsecJm 2 (6
GPMlsq m, include: The expansion involved the replacement of the sand
media with mixed media in the filters. It also added
• Average turbidity of less than 0.1 NTU shallow-depth sedimentation using tube settlers and
168 Upgrading Existing or Designing New Drinking Water Treatment Facilities

coagulation control techniques. Typical chemical coliform MPN values range from 5 to 72,400 per 100
dosages used to treat the raw water include: mL and its color ranges from 25 to 50 units. Although
the existing system could manage the high raw water
• 20 to 40 mglL of alum turbidity and coliform count, the system had a
• 10 to 20 mglL ofJime clarification problem when the water exhibited low
• 2 mglL of chlorine turbidity and high color. Also, high plankton
• 0.1 to 0.2 mglL of polymer as coagulant aid populations caused taste and odor problems. The
• 5 to 10 mglL of activated carbon for taste and odor plankton and other biological contaminants
control significantly shortened filter run length.

Almost all treatment plant piping was enlarged to The characteristics of the existing treatment system
accept the expanded flows. Settling tubes covering 60 included:
percent of the 325.2 m 2 (3,500 sq ft) rectangular
sedimentation basins were positioned at the • Disinfection with chlorine and coagulation with
discharge end of the basins. The tube modules were alum in a downflow hydraulic mixing chamber
supported on a simple "I" beam extended across the • Flocculation, with additions of recirculated
basin width. ;-'<ew effiuent weirs and launders were sludge and lime, followed by clarification
also installed to ensure proper flow distribution. The • Additions of activated carbon immediately prior
overflow rate in the settling basin with tubes was 2.8 to rapid sand filtration
Usedm 2 (4.2 GPMlsq ft) compared to 0.71 Usedm 2 • Dechlorination through the addition of sulfur
(1.05 GPMlsq ft) prior to the tube installation. dioxide in the filtered water clearwell
• Elevated water storage for backwashing
Experience has proved that the modified plant, at • Discharge of backwash water to a recovery pond
full-scale operation, consistently produced water and recycling to the plant influent for
with less than 0.2 NTU turbidity. Also, filter run reprocessing
lengths have increased from 40 to 65 hours due to the • Spraying of flocculated sludge on land for
improved turbidity removal achieved by tube disposal
settling. The filtration rate is now 5.1 Usedm 2 (7.5
GPMlsq ft). The plant uses a pilot filter to establish The expansion required three major changes:
coagulant dosages and has turbidimeters on each
filter to continuously monitor effiuent turbidity. The 1. Modifying the clarifier by adding settling tubes
increased filter cycle length reduced operating costs
through backwash water savings. 2. Changing the filter media from sand to mixed
media
The cost of this expansion in 1969 was $430,000. In
contrast, the original expansion plan of constructing 3. Installing a coagulant control center and
new flocculation, sedimentation, and filter units to turbidity monitoring unit
0.7 m 3/sec (16 MGDl capacity was estimated at
$650,000, in 1969 dollars. The nonstructural These changes are discussed below.
expansion included:
Settling tube modules were added to the annular
• New 0.22 m3/sec (5-MGD) reservoir outer settling compartment of the circular clarifier.
• New high service pump station As a general rule the addition of settling tubes
• A cross-town 40.6-cm 06-in) transmission pipe to permits processing two to four times as much water
the new reservoir with no loss in clarification effectiveness. However,
• 219 Usee (5 MGD) larger capacity than the due to the presence of light alum floc, it was decided
original expansion to increase capacity to only 263 from 164 Usec (6.0
from 3.75 MGD). This change increased the surface
In addition, the expansion resulted in improved overflow rate on the clarifier from 1.03 to 1.7 Usedm 2
finished water quality. 0.52 to 2.5 GPMlsq ft) and decreased the detention
time from 1.3 to 0.8 hours. The weir loading increased
from 1.47 to 2.3 Usedm 2 (7 to 11 GPMlft).
A.1.6 Novato, California (Culp, 1976)
This case involved the expansion ofa North Marin In addition, capacities of certain plant equipment
County Water District plant from a capacity of 0.16 to were increased to accommodate the increased plant
0.27 m3/sec (3.75 to 6.2 MGDl during 1973 and 1974, flows. In addition to expanded pump capacities,
without major additions to the plant. pipelines, meters, chemical feeders and valves, and
controls were modified to accommodate the increased
Stafford Lake is the District's water source. The plant capacity. The simple hydraulic mixing chamber
turbidity ranges from 2 to 35 NTU, with the usual was equipped with a vertical mechanical rapid mixer.
turbidity between 10 and 20 NTU. Stafford Lake's Provisions were made to add polymer to aid
Appendix A-Experience Modifying EXisting Filtration Systems 169

nocculation. Also, a system to store and feed sodium • Microscopic particulate analysis (MPA) can be a
hydroxide solution for pH control replaced the lime good indicator of slow sand filter performance.
system.
• The easiest method of projecting slow sand
The filter bottoms were replaced along with performance is to examine design and
replacement of sand media with mixed media. The operational specifications.
mixed-media filter was supported by graded silica
gravel with an upper layer of coarse garnet gravel. • Raw water quality can have a large impact on
Gate valves were replaced with butterny valves. slow sand performance.

A coagulant control center and turbidity monitoring • Slow sand filters can have significant operational
equipment were installed to ensure compliance with problems.
the California Department of Health Services
standard of 0 5 NTU This study made the following recommendations in
considering the slow sand filter option:
The changes resulted in achieving a 107 Usec (2.45
MGD) increase in capacity at a cost of$337,445, in • Use pilot filters to guide design and construction.
1974 prices. The expansion took only 8 months to • Use experienced design professionals.
complete. The plant has performed efficiently and • Use multiple beds for the filters.
reliably at the maximum design capacity, • Use properly educated operators.
consistently achieving less than 0.15 NTU turbidity • Reduce filtration rates during winter.
levels and effecti ve bacteriological removal. • Include proper watershed management in
operation and maintenance procedures.
A.2 Slow Sand Filter Systems The researchers also recommended the use of routine
sanitary surveys, which would include monitoring
A.2.1 Idaho State (Tanner, 1988) influent water quality, operating factors, and plant
In ajoint I-year project in cooperation with Idaho performance. Specifically, water quality should be
State, the EPA and University of Washington's monitored as to water source, watershed
Department of Environmental Health focused on management and status, turbidity, and temperature.
three typical slow sand filter plants. In the course of Operating factors monitored should include filtration
this study, 13 additional slow sand filter plants were rates, cleaning procedures, and bypassing. Finally,
surveyed, and the following plant performance data monitoring plant performance should include the
were collected: recording of turbidity, cleaning frequencies, and
freezing frequencies.
• Turbidity
• Total and fecal coliform removal
• Giardia samples
• Microscopic particle evaluation A.2.2 New York State (Letterman and Cullen,
• Temperature 1985)
A recent study of seven slow sand filter plants in New
In addition, the study collected and examined the York State examined the effects of cleaning
following operating data: procedures on the effiuent water quality. The water
quality parameters examined included turbidity,
• Filtration rates total particle count, standard plate count, and total
• Sand depth coliforms. Characteristics studied at the seven plants
• Effective sand size included average operating flow, design filtration
• Cleaning method and frequency rate, average operating filtration rate, filter sand
• Problems uniformity coefficient, and effective sand size (see
Table A-I).
From the results of the study, the researchers
concluded that: The operational data from these plants are contained
in Table A-2. Even during the ripening periods, all
• Slow sand filters are very effecti ve and reliable, if but one plant achieved effiuent turbidities ofless
they are properly designed, operated, and than 0.43 NTU. The one plant that exceeded 0.43
maintained. NTU had submicron particle size turbidity, which is
ineffectively controlled with slow sand filtration. All
• Slow sand filter performance is not reliably seven plants performed well in removing particles
determined through examination of turbidity and greater than 2 flm. The removal rate for particles
coliform removal. ranging from 2 to 60 Ilm was from 90 to 99.8 percent.
170 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Table A·l. Charac:tertltlca of Slow Sand Filter In81allallonl In New York


Slow Sand Filtration Average Operahng
Average Operating Flow Design Filtration FiltratIOn Rate Filter Sand
Location Rate (MGD) Rate (mgad) (mgad) Uniformity Coefficient Effective Size (mm)
Aubum 6.0 2.83 3.6 2.4 0.45
Geneva 2.5 4.9 4.9 1.9 0.37
Hamiltoo -0.3 1.0 2.4 0.27
llton 1.5 4.1-4.6 2.2 0.37
Newar1< 2.0 4.1 4.1 1.7 0.35
Ogdensburg 3.6 5.1 4.6 1.7 0.35
W~ 1.2 4.1 4.1 2.4 0.15
1 MGD - 0.044 mJ/sec; 1 MGAD - 0.935 mJ/m 2/day
Source: Letterman and Cullen (1985).

Total plate count removals varied widely from 0 to 50 The evaluation of this slow sand filter plant revealed
percent. the following:

• Effiuent turbidity ofless than 1 NTU was


A.2.3 Mcindoe Falls, Vermont (Pyper, 1985) achieved 99.19 percent of the time.
This was a 2-year study of a municipal slow sand
filter plant with two 37.2-m 2 (400-sq tt) filters, sand • Effiuent turbidity of less than 1 NTU was
depth of106.7 cm (42 in), effective silica sand size of achieved 99.68 percent of the time, after the first
0.33 mm, and a design filtration rate of 1.92 100 days of operation.
m3/m 2/day (2.05 mgad).
• Effiuent turbidity of less than 0.2 NTU was
The raw water was obtained from two shallow spring- achieved 72 percent of the time (during this time
fed ponds with several beaver dams and lodges. influent turbidity was 1.45 NTU or lessl.
Influent water turbidity ranged from 0.4 to 4.6 NTU.
with a seasonal average of2.1 NTU. The study • Influent total coliform count was reduced from
examined the removal of turbidity, bacteria, 1,300 per 100 mL or less to 10 per 100 mL or less,
coliforms, and Giardia cysts. 86 percent of the time.

Table A-2. Filter Ripening Data - Summary

Raw Water Filter Turbidity Approximately 5 Hours


Type of TUrbidity Water Tem- After Filter Stanup (NTU) Evidence of ApproXimate Length

LocabOn
Operation
Visit-
Date 0'
Site Visit
Dunng Site
Visit (NTU)
perature Dur-
ing Site Visit
ScrapedlRe-
sanded Filter Control Filter'>
Ripening
Pencid
01 Ripentng Pencid
(days)
Aubum (1) Jul83 1.2-2.0 -19'C 0.43 0.27 Yes 0.25
Aubum (1) Jul83 1.2-2.0 -19'C 0.28 0.27 None
Aubum (1) Jul84 2.0-2.8 -18'C 0.22 0.23 None
Geneva (1) Jul83
Hamilton (1) May 84 1.0·1.5 -12'C 0.28 None None
Iiton (1) Jul83 2.0-4.0 -23'C 0.30 0.40 Minimal 0.5
(particle
count only)
Newar1< (1) Aug 83 1.2·3.5 - 13'C 0.35 0.35 None
Newar1< (2) Jan 84 0.6-2.7 -4'C 0.41 0.12 Yes 2
Ogdensburg (3) Aug 83 0.3·0.6 -15'C 0.12 0.10 None
Ogdensburg (3) Feb 84 1.0-1.2 -2'C 0.22 0.24 None
Waverfy (1) Jun 84 6.0-11.0 -15'C 2.3 1.6 Yes 10
• (1) Scraping operanon.
(2) Resanding operallOn.
(3) Scraping combined With resanding.
b Control filter· Filter on·line at least 1 month. except Ogdensburg where the filter was on-line 1 week.
Source: Letterman and Cullen (1985).
Appendix A-Experience Modifying Existing Filtration Systems 171

• Influent occurrences of high total coliforms ranged between 0.2 and 0.3 mm, with an average of
(spikes) and plate count bacteria were removed in 0.25 mm. The sand media uniformity coefficient
water temperatures offrom 5 to 10·C. ranged between 3.3 and 3.8, with an average of3.5.

• Biological removal efficiencies were lower at Each of the filter cells are 43 m (141 ft) long, 6 m (20
temperatures ofless than 5·C, especially from 0 ftl wide, 3.75 to 3.9 m (12.5 to 13 ft) deep, and have a
to 1·C; for example: total surface area of262 m2 (2,820 ft2). The design
filter rate is 0.11 Usedm 2 (0.16 GPM/ft 2). All three
Giardia removal efficiency was lowered to filters are operated to the accepted maximum unit
93.7 percent in one test. flow of 2,422.4 m 3/day (640,000 GPO) per filter for a
total of7,255.8 m3/day (1,917,000 GPO). To protect
- At l·C total coliform count removal was against freezing, the filters are covered with precast
reduced from 98 to 43 percent and the panels. Filter walls have rigid insulation and are also
standard plate count bacterial removal was insulated with backfilled soil.
reduced from 98 to 80 percent over a 9-day
monitoring period. The performance results of the new plant are
presented in Figures A-2 and A-3. Influent sample
• Influent standard plate counts of 500 per mL Giardia cyst counts for the period of November 1985
were reduced to 10 per mL or less, 94 percent of through November 1986 are contained in Figure A-2.
the time. No cysts were detected in the effiuent during that
period. Influent and effiuent sample turbidity data
• Influent total coliform counts were reduced from are contained in Figure A-3. Effiuent turbidity
440 organisms per 100 mL to 4 per mL. ranged from 0.15 to 3.5 NTU ,and was lower both
when influent turbidities were lower and after the
• Influent heterotrophic organisms were reduced filter ripened.
from 520 organisms per 100 mL to 15 organisms
per 100 mL, based on the standard plate count. Operating cycles for the filters ranged from 52 to 215
days, with the longer filter cycles occurring in the
• Influent Giardia levels were reduced by 99.98 winter and spring months. Increased loadings of
percent under warm water conditions. algae during late spring and summer months
contributed to shorter filter cycles during those
• Influent Giardia levels were reduced by 99.36 to periods.
99.91 percent in water temperatures of less than
7·C. The construction cost of the new plant, in 1984
Canadian dollars, was $780,000. The average annual
operating costs were $20,700, including costs for
chlorination; energy requirements; media
A.2.4 Village of 100 Mile House, British replacement; cleaning; and labor for daily inspection,
which consumed from 0.5 to 1.0 hours. Cleaning
Columbia, Canada (Bryck et aI., required 16 person hours per filter and cost about
1987) $225 per cleaning. The total operating cost was
This example concerns a water supply operation that estimated to be $0.25/m 3 ($0.96/1,000 gall.
only chlorinated their water before distribution to a
2,000 person service area. Giardia in the raw water
intake from beavers and muskrats upstream became
a concern. In 1984,the village decided to construct a A.3 Package Plants
slow sand filter system to augment their chlorination
operation after Giardia occurrences in 1981, 1982, A.3.1 Conventional Package Plants (Morand
and 1983. The new plant included the following and Matthew, 1983)
components: The EPA surveyed the elTecti veness of six
conventional package plants, most of which were
• Surface water intake built during the 1970s. The results of this survey
• Raw and treated water pumping station were published in March 1983. The plants were
• Chlorine equipment chosen for study because they operated throughout
• Contact tank the year, used surface water as their source, and
• Clear well served small populations. Profiles of the six plants
• Three slow sand filter beds are contained in Table A-3; treatment processes are
described in Table A·4. Design capacities ranged
The three filters used filter media derived from local from 545.0 to 3,028.0 m3/day (144,000 to 800,000
sand that was washed, dried, and sieved. The GPO) and average daily flows ranged from 45.4 to
resultant elTective sand size for their filter media 1,249.0 m3/day (12,000 to 330,000 GPO).
172 Upgrading EXisting or Designing New Drinking Water Treatment Facilities

Village 01 100 Mile House

500

400

300
Number
of Cysts

200

100

o
27·NoY·85 21·Jan·86 20·Mar·86 29·Apr·86 21·Jul·86 26·Aug·86 03·NoY·86

Time (Months)

Figure A·2. Giardia cysts In the raw water.

The survey examined samples for turbidity, total


coliforms, and chlorine residuals through the A.3.2 Adsorption Clarifier Package Plants
collection of grab samples from the influent, effiuent, This section contains descriptions of six facilities that
and distribution systems. Influent and effiuent have used adsorption clarifier package plants. The
turbidities are shown in Table A-5. Plants C, T, and plant profiles and effiuent data are from facilities in
W consistently met the 1 NTU standard, while plants Lake Arrowhead, California; Greenfield, Iowa;
P, V, and Yl met the 1 NTU standard less than half Lewisburg, Virginia; Philomath, Oregon;
the time. This discrepancy in the latter plants was Harrisburg, Pennsylvania; and Red Lodge, Montana.
due to:

• Inadequate design detention time Lake Arrowhead, California (Hansen, 1987)


• Inadequately trained operators This case history invol ves a pilot study where
• Inadequate time allocated for the operation Giardia cysts were introduced into a water supply to
• Periods of high influent turbidities for plants V test the removal effectiveness of an adsorption
and M clarifier. The adsorption clarification/filtration pilot
plant had a 1.26 Usec (20·GPM) capacity. The
concentration of Giardia introduced to the raw water
After the appropriate adjustments were made to the was 2,100 cysts per liter. The results of the pilot test
equipment and operations, plants P, V, and M met revealed that filtration removed 100 percent of the
the I "ITt.: standard. cysts. Plant effiuent turbidity ranged from 0.05 to
Appendix A-Experience Modifying Existing Filtration Systems 173

Village 01 100 Mile House


5..,....--------------------------------,

3
a-
S
?:
'ti
~
"
~

O+-r--r-,-,.-.,.....-,-.....--r--r-,-.---.,.....-,--r--r--,-,-,.-.,.....-,--r--r--.---,-,.-.,.....-,-.....,....--!
Q4·Nov·85 11·Feb-86 Q4·Jun·86 29·Sep·86
Date (monlhs)

C Raw Water + Filtered Waw


Figure A-3. Average raw and lU1erlld water turbidity.

Table A-3. Water Treatment Facilities SUrv9yed in Field Study


Manufacturer Design Flow PopulallOn Served! Average Volume per Group Type of Distribution
Site Model. Year Rate GPM (Us) No. of Meters Day. GPM (mUd) Served Pipe Used Source
W Neptune Mlcro/loe 200 (12.6) 1.500/552 110,000 (0.42) City PVC Suriace
AQ-40. 1973 Impoundment
T Neptune M«:rofloe 200 (12.6) , .0001360 78,000 (0.30) City PVC, cast Iron, Suriace
AQ-40, 1973 asbestos cement Impoundment
V Neptune Mlcro/loe 200 (12.6) .. ·/423 72.000 (0.27) PSDa PVC RIver
AQ-40. 1976
M Neptune Mlcrofloe 560 (35.3) "'/1,680 330.000 (125) PSD PVC River
AQ·t 12. 1972
P Neptune Mlcro/loe 100 (6.31) .. ·/411 82,000 (0.31) PSD PVC Suriace
Water Boy. 1972 ImpOundment
C PermulO 200 (12.6) State pari< 57,000 (0.22) Stale pari< Asbestos cement River
Perm3lel. 1971
aPSD • Pubhc Service D,stnet
Source: Morand and Matthew (1983).
174 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Tallie A-4. Treatment Process Charactertstlcs

Rapid Mix Flocculation Sedimentation Filtration


Detentron Detention
PrechemlcaJ Time Time Loading Rate
Site Addition Type (sec) Type (min) Type (GPMIft2) Media (GPM/ft2) Notes
W Cl. alum. In pipe 3 Paddle 12.8 Tubes 100 Mixed 5 Polymer
soda ash anthracite. 18 in. added before
SIlica sand. 9 In. tubes
garnet sand. 3 In.
T Cl. alum. In pipe 3 Paddle 12.8 Tubes 100 Mixed 5 Post· soda
soda ash. antllracite. 18 In. ash
pOly silica sand. 9 in.
garnet sand. 3 in.
V Cl. alum. In Pipe 3 Paddle 14 Tubes 100 Mixed 5 Post· sodium
soda ash. anthracite. 18 In. hexameta·
poly SIlica sand. 9,n. phosphate
garnet sand. 3,n.
M Cl. alum. Chamber 30 Paddle 10 Tubes 100 Mixed 5
soda ash. anthracite. 18 in.
pOly silica sand. 9 In.
gamet sand. 3 in.
P Cl. alum. Chamber Paddle 10 Tubes 150 Mixed 5
soda ash. not used anthracite. 18,n.
cartx>n sihca sand. 9 In.
(summer) garnet sand. 3 In.
C C1. alum. In pipe Upllow solids contact 2 hr detention trme Silica Sand. 24,n. 2 SOda ash
soda ash. osa rate - 1 GPMlft2 added belore
poly 1,ltrallon

GPMlft2 • 0.679 UsecJm 2 ; 1 inch • 2.54 em


Source: Morand and Matthew (1983).

0.06 NTU. Clarified effiuent contained cysts when


effiuent turbidity ranged up to 0.3 to 0.4 NTU. Philomath, Oregon (Lange et al., 1985)
A 0.044 m 3 /sec (l.O-MGD) plant utilizing an
adsorption clarification unit began operation in
Greenfield, Iowa (WEM, 1985) February 1986. Seven months of turbidity data and
This case h~story presents operating performance chemical feed data are contained in Table A·8.
data from a 0.044 m 3/sec (1.0 MGDl plant between Influent turbidity ranged from 3.02 NTU to 26.7
July and December 1984. The plant's influent comes NTU in February 1986. The system performed
from Lake Greenfield which contains turbidity, taste, effectively with filter effiuent turbidity ranging from
color, and iron contaminants. At the Greenfield 0.17 to 0.26 NTU. Net water production did not fall
plant, powdered activated carbon is added to control below 88 percent during this period.
odor and taste problems. Table A-6 presents the
operating data for the plant. Pilot test turbidities
were similar to those from regular operational Harrisburg, Pennsylvania (Lange et al., 1985)
experience, typically ranging from 0.3 to 0.5 NTU. This example involves an existing plant that was
Alum dosages in the full-scale plant range from 7 to built in 1964 without sufficient clarifying capacity.
20 mglL, which are slightly lower than those The original plant contained a circular upflow sludge
established in pilot tests. blanket clarifier and stored backwash gravity filters.
The system was operated in the direct filtration mode
at 2.7 Llseclm 2 (4 GPMlsq ttl because clarification
capacity was inadequate.
Lewisburg, West Virginia (Lange et al., 1985)
This case involves a 0.087 m3/sec (2-MGD) plant that Three adsorption clarifier packaged units were added
has operated since December 1983. Three months of to expand the plant to 0.394 m 3/sec (9 MGD) with
records on turbidity removal are contained in Table each adsorption clarifier operating at between 6.79
A-7. The clarifier generally achieved 90 percent and 10.19 Llseclm 2 (10 and 15 GPMlsq ft). Three
turbidity removal. While turbidity ranged as high as months of turbidity and chemical feed records are
50 NTU in May 1985, the plant was able to produce found in Table A-9. Overall turbidity removal ranged
effiuent with turbidities ofless than 0.5 NTU at all from 70 to 90 percent. The length of filter runs
times. doubled from 13 to 26 hours with the addition of the
Appendix A-Experience Modifying Existing Filtration Systems 175

Table A-5. Turbidity Values (NTU)

PlantC Plant W Plant T Plant V Plant M Plant P


Clearwell Clearwell Clearwell Clearwell Clearwell Clearwell
Raw Effluent Raw Effluent Raw Effluent Raw Effluent Raw Effluent Raw Effluent
8.5 0.3 0.9 to.O 1.9 4.0 1.8 0.2 12.0 0.8
6.2 0.2 5.0 0.3 8.0 0.2 12.0 2.8 39.0 3.8 4.4 2.4
1.2 0.3 4.2 0.4 6.0 0.4 9.6 40.0 2.6 70
1.6 0.1 19.0 0.8 3.2 1.1 35.0 1.5 27.0 2.4 3.5 1.5
2.2 0.1 9.2 2.0 3.2 0.2 42.0' 2.0 6.0 1.2 2.0 O. ,
4.0 0.1 11.5 0.3 3.2 0.2 10.0' 2.4 3.8 0.1 1.2 as
12.6 0.7 12.0 0.2 5.8 0.2 90.0' 8.5 73.0 11.0 15.6 9.7
5.2 0.2 11.0 0.3 10.4 3.2 28.0' 5.4 3.6" 0.1 3. I 2.2
2.2 0.2 29.7 0.9 3.4 0.7 19.0' 0.3 3.8 0.3 177
.2 1.9 12.8 0.2 47.0" 1.2 6.0 0.5
13.0' 0.8 70.0' 16.0
8.0' 0.3 25.0' 3.4
6.0' 0.3 > 100.0' 55.0
> 100.0' 0.5 > 100.0' 31.0
60.0' 0.5 8.5" 2.2
24.0 1.2 4.3 0.4
13.0 0.3 4.0" 1.0
2.7 1.2 9.6" 1.9
1.2 1.0 19.1 1.1
3.3 0.5 64.0 6.9
8.2 1.0
-Averaged values for day.
Source: Morand and Matthew (1983).

Table A-6. Operating Data· Greenfield, IA


TUrbidity, NTU Iron, mglL Chemical Feed, mgiL
Month Raw Clan/ier Filter Raw Filter Alum Polymer Chlorine
Jul, 1984 Mean 5.9 0.51 0.13 0.03 15.5 0.32 7.0
_________ ~~g~ ~'~:.'~:E ~:..3!:0:•.7~ 3:..0~:0:..2! __ 3:..0~ :O:..O! '2~ :'~:..6 3:..2~:0:..3! ~~~! _
Aug, 1984 Mean 3.9 1.26 0.46 0.18 0.03 12.2 0.32 3.9
_________ ~~,:g~ !:..5~~~ ~:..3~:0:..6_2 3:..0?:.0:..3! O:..~-~.~ ~~.~~~ 3:..'~:0:..4~ ~.~?:..6 _
Sep,1984 Mean 4.4 1.6 0.39 0.14 0.02 8,4 0.35 7.8
Range 3.6-5.3 0.36-0.60 0.10-0.24 0.01-0.03 6.9-' 2.9 0.32·0.37 74-8.1
._----------------------------------------------------------------------------------
Oct, 1984 Mean 6.3 2.3 0.40 0.21 0.02 6.7 0.20 5.9
_________ ~~g~ ~:..O~~~ ~ ~.:.2.:~ ~}~:0:..5_6 3:..'~:.0:..~__ ~:..O~ :O:..O} 3:..'~:.0:..2~ ~-:~! :..4 _
Nov. 1984 Mean 6.4 2.3 0.33 0.20 0.Q1 '3. I 0. I 8 3. I
_________ ~~g~ ~~ __' ~.~ 3:..0.:.2.:~ ~:..'!:0.:4! 3:..1~:.0:..3: __ 3:..0~ :O:..O} ~~-~~~ 3:..'::.0:..2.! ~'~-'::..' _
Dec, 1984 Mean 2.8 1.7 0.35 0.11 0.01 10.2 0.22 2.5
Range 1.8-7.5 1.6-2.0 0.27-0.48 0.02-0.17 0.00-0.03 8.3-14.6 O. I 7·0.25 2. I -2.8

adsorption clarifiers, in spite of the 20 to 30 percent of its high quality. Because the raw water turbidity is
increase in flow. so low, the plant is equipped with a bentonite feed
system to assist in the treatment of the plant's
typically cold, low turbidity (less than 1 NTU)
Red Lodge, Montana (Lange et al., 1985) influent. Bentonite is added to the influent before the
This case involves installing a new 0_061 m3/sec 0.4- alum and polymer feed points. The treatment goal is
MGDl plant starting up in January 1984. Prior to to produce effiuent with less than 0.1 NTU turbidity
construction, the water was only chlorinated because to guard against Giardia breakthrough. As Table A-
176 Upgrading Existing or Designing New Drinking Water Treatment Facilities

• Influent concentrations of Giardia cysts, ranging


Table A·7. Opel'lltlng Data· Lewisburg, WV
from 50 to 5,000 cysts per liter
TUrbidity. NTU • Filter headloss rates
Clanfier Clanfier • Filter run times
Month Influent Effluent Filler Effluent • Alum coating of the diatomaceous earth
Dec. 1984 Mean 5.0 12 0.16
Range 18-18 0.6'2.5 0.10-0.20 The results of the study indicated that Giardia
May. 1985 Mean 9.6 1.0 0.18 removal exceeded 99 percent for all grades of
Range 2.0-5.0 0.4-6.0 0.10-0.50 diatomaceous earth at filter hydraulic loadings of
0.68 to 2.72 Useclm 2 (1.0 to 4.0 GPM/sq ft), and at all
Jun. 1985 Mean 1.7 0.60 0.16
study temperatures.
Range 0.70-5.0 0.1-1.1 0.10-0.50
In addition, the grade of diatomaceous earth affected
the removal rates of other contaminants. The
coarsest grade of material achieved removal of99.9
Operating Data· Philomath, OR
percent of the Giardia, 95 percent of the cyst-sized
Table A·8.
particles, 20 to 35 percent of the coliform bacteria, 40
(Monthly Average Values)
Net to 70 percent of the heterotrophic bacteria, and 12 to
Turbidity. NTU ChemiCal Feed. mglL Proouc- 16 percent of the turbidity. The finest grade of
Month Raw Clanfier Filter Alum Soda Ash Polymer bona
diatomaceous earth achieved the removal of99.9
Feb. 26.7 3.40 15.5 21.2 029 percent of the bacteria and 98 percent of the
1986
turbidity.
Mar. 6.50 \.67 0.17 11.4 4.5 0.08 910;0
1986
Apr. 6.27 1.49 0.20 90%
1986 A.4.2 Mcindoe Falls, Vermont (Pyper, 1985)
May. 5.82 1.45 0.23 11.7 1.2 0.08 88%
This case involves an evaluation of the parallel
1986
operation of a diatomaceous earth and a slow sand
Jun. 4.25 \.21 0.14 1\.3 4.8 0.07 93'1'.
filter. The diatomaceous earth filter rate ranged from
1986
0.68 to 1.22 Useclm 2 (1.0 to 1.8 GPM/sq ft), using a
Jul. 343 0.22 10.0 5.7 007 93%
0.63 to 1.26 Llseclm 2 (10 to 20 GPM) pressure unit.
1986
Aug. 3.02 0.26 16.6 0.07 93%
The pressure diatomaceous earth filters dependably
1986
removed 99.97 percent of the Giardia cysts and 86
°Net prooucnon IS the percentage of raw water that was turned Into percent of the coliforms. They also achieved 80
lintShed water.
percent standard plate count bacteria removal in 70
percent of the samples. The average influent
10 indicates, the plant achieved its turbidity goal for contained 271 total bacteria count per 100 mL and 30
the entire 6-month period over which data were standard plate count bacteria per mL. The average
collected. It should be noted that the turbidity of the effiuent contained 38 total bacteria count per 100 mL
clarifier effiuent is sometimes higher than that of the and 6 standard plate count bacteria per mL
raw water influent due to occasional carryover of
alum-coagulated bentonite particulates.
A.S Selecting a Filtration System
A.4 Diatomaceous Earth Filters
A.5.1 Lake County, California (Conley and
A.4.1 Colorado State University Study Hansen, 1987)
(Lange et aI., 1984) This case study concerns the process of selecting a
The study examined Giardia inactivation in addition centralized new treatment facility for an area that
to removal of turbidity, total coliform bacteria, was previously served by individual supplies. Clear
standard plate count bacteria, and particles. Eight Lake, the source of raw water, experiences severe
operational parameters evaluated for their influence seasonal taste and odor problems caused by blue-
on removal effectiveness were: green algae blooms (Microcyst is and Anabaena).

• Seven grades of diatomaceous earth Clear Lake is classified as eutrophic. The taste and
• Hydraulic loading rates of 0.68, 1.36, and 2.72 odor problems typically start in May and finish in
Llseclm 2 0, 2, and 4 GPM/ft 2 ) November with lake water taste and threshold odor
• Influent temperatures, ranging from 5 to 19°C numbers (TON) ranging as high as 10 during this
• Influent concentrations of bacteria, ranging from period. Clear Lake's turbidity ranges from 3 to 90
100 to 10,000 per 100 mL NTU. Clear Lake has excellent mineral qualities, as
Append ix A-Experience Modifying Existing Filtration Systems 177

Table A-g. Operating Oata - Harrisburg, PA


Turbidity, NTU Chemical Feed mgiL Clantier Rate
Month Raw ClarIfier Filter Alum Polymer GPMlft'
Jul. 1985 Mean 8.1 1.36 0.17 NA NA NA
__________ ~~~ __ ~:?:.~5..: ~ ~~~~.~ ~ ~~~.~~ _
Aug. 1985 Mean 8.6 0.94 0.18 11 0.12 14
Range 4.8'14.3 0.7,1.5 0.13-0.25 8.1·13.4 0.10-0.15 11.9-16.5
---------------------------------------------------------------------------
Sap. 1985 Mean 9.3 1.2 0.18 9 0.07 13
Range 5.9-37 0_5-4.9 0.09-0.34 4.9-18.7 0-0.31 11.7-14.9
GPM/ft' 0.679 Useclm 2
2

NA not aV3Jlable
2

Table A-la_ Operating Oata - Red lodge, MT


Turbidity. NTU Chemical Feed mglL Water Temperature
Month Raw Clanfier Filter Alum Polymer BentOnite °F
Jan. 1985 Mean 0.20 0.29 0.04 2.5 0.48 0.93 36
__________ ~~~ __ ~:.! ~~~3: 0..:2..1:.~~5 ~.~~-~~~ 1.:~:~.~ __0..:~8:.~~8__ ~~~~~: ~~-~l!. _
Feb. 1985 Mean 0.19 0.28 0.05 2.3 0.48 0.93 34
Range 016-0.21 0.22·0.39 0.30-0.07 1.4-3.6 0.48-0_48 0.93·093 33·38
------------------------------------------------------------------------------
Mar. 1985 Mean 0.20 0.29 0.05 3.0 0.48 0.93 36
Range 0.18-0.22 0.25·0.32 0.03·0.06 1.9-3.9 0.48-0.48 0.93-093 35·37
------------------------------------------------------------------------------
Apr, 1985 Mean 1.1 0.49 0.08 5.0 0.60 a 40
__________ ~~g~ 0..:2_2:.~~ ~.~:-~~ ~.~.~:.!~ 2.:~-:~ __0.:~8:.~:5 ~l!..~~ _
May. t985 Mean 2.1 0.41 0.07 7.3 0.48 a 44
Range 0.16-4.5 0.16-0.80 0.03·0.14 3.6-13.4 0.48-0_48 43-47
------------------------------------------------------------------------------
Jun. 1985 Mean 3.5 0.41 0.07 10.4 0.60 a 45
Range 1.5-6.2 0.21·0.75 0.04·0.18 7.4-12.8 0.48·0.75 43·49
°C • (OF· 32) x 519

shown in Table A-II, but the lake water requires


turbidity removal, taste control, and odor control. Table A-l1. Clear lake Water Quality Analysla
Concenlra~ons. mglL
Several other jurisdictions in addition to Lake Mineral ConS1Jtuents High low Melllan
County use Clear Lake for a water supply. Calcium 30 17 23
Consequently, there were data available from other Magn8Slum 21 98 15
operations using the same influent. One of these Sodium 14 4 10
neighboring jurisdictions, the City of Lakeport,
PO!asSlum 2.8 0.1 2.0
installed a plant equipped with both ozone and
activated carbon, which effectively manages the odor Bicarbonate 212 96 145
and taste problems. Sulfate 35 5.1 9
Chlonde 10 3.2 6
A feasibility study recommended a complete Nitrate 11 0.0 1.6
treatment system rated at O.088m 3/sec (2 MGD) with
Fluonde 0.4 0.0 0.1
preozonation, chemical coagulation, pH control,
flocculation, sedimentation, filtration, and BOton 1.2 0.1 0.7
postfiltration. A suitable package plant was selected Silica 34 0.7 14
as the most cost-effective choice. Hardness 158 78 115
Turbidity 90 TU' 3 TV 15 TU
The selected package plant required pilot tests for
'Turbidity UnitS
establishing process feasibility and design criteria.
Four utilities were contacted with regard to their
experiences before pilot testing commenced. The floc problems due to air bubbles introduced by algae.
utilities reported that laste and odor had been In addition, algae blooms had interfered with the
increasingly troublesome, and they described rising clarification process. Water quality data from
178 Upgrading EXisting or Designing New Drinking Water Treatment Facilities

January 1977 to November 1980 for a sampling The tests were performed during September and
station on Clear Lake is summarized in Table A-12. October of 1986 and verified by a design consultant.
There were no major algae blooms during the test
The odor and taste problems had been successfully period.
treated by the other utilities with potassium
permanganate and powdered activated carbon. The pilot test report indicated that the new plant
However, the two utilities using ozone received the could remove a high percentage of turbidity.
fewest complaints. Also, some utilities had recently Furthermore, low concentrations of potassium
installed postfilter granular activated carbon permanganate could be used to control taste and odor
contactors. substances. In addition, the powdered activated
carbon assisted in the removal ofTHM precursors.
As a result of reviewing the existing data, Lake The report indicated that the adsorption clarifiers
County decided to install adsorption clarifiers as a were effective at dosages of up to 25 mglL, with no
prefiltration step. The adsorption clarifier uses an filter breakthroughs or short filter cycles.
upflow design with a buoyant medium to assist the
clarification process. Based on the pilot test results, an adsorption clarifier
package plant and chemical feed system were
The manufacturer was requested to perform the pilot recommended. The adsorption clarifier will have a
tests. The two primary test objectives were to: design loading of6.79 Usec/m 2 00 GPMlsq ftl, a
filtration rate of 3.4 Usec/m 2 (5 GPMlsq ftl, and a trio
• Measure turbidity removal capabilities mixed filter media. The chemical feed system wil\ use
the following chemical dosages:
• Determine the effectiveness of adding potassium
permanganate or powdered activated carbon
prior to clarification to remove trihalomethane • 1 to 2 mglL of ozone
(THMl precursors, taste substances, and odor .10to20mglLofalum
substances • 0.2 to 1.0 mglL of potassium permanganate
• 1 to 10 mglL of cationic polymer
If the pilot tests were successful, the utility would be • 10 to 20 mglL of powdered activa ted carbon
able to avoid the expensive option of using ozone • 0.1 to 0.5 mglL of nonionic polymer
treatment to control taste and odor. • 20 to 40 mglL of sodium hydroxide
Appendix A-Experience Modifying Existing Filtration Systems 179

Table A·12. Clear Lake Water Quality at DWR Sampling Station NO.1 at Lakeport
Dale Turbidity Sp. Cond. DO Sodium Alkalinity Chloride Nitrates Or1ho Po.
Sample<l (NTU) (IlMHOS/Cm) pH (mgll) (mgll) (CaCO,) (mgll) (mglL) (mglL)
lf7m 21 307 8.3 9.9 13 142 68 025 0.05
2/3f77 9 308 8.1 11.8 13 144 7.0 0.14 0.00
3110f77 25 308 7.7 9.2 12 144 6.6 0.12 001
4f7f77 18 316 8.1 7.7 12 147 6.4 0.69 0.00
5/5m 14 321 8.0 9.1 14 156 6.5 0.01 0.00
6116/77 35 337 7.8 4.5 14 154 7.9 0.05 0.00
7/14m 14 346 8.0 5.0 14 160 8.8
8/11m 302 8.0 4.3 16 138 9.6 0.04 0.09
9/22177 60 364 8.2 2.6 17 165 8.8 0.05 008
10/14177 50 385 8.1 3.6 17 170 10 0.01 0.03
11/4m 33 373 8.3 5.7 16 174 7.7 0.14 002
12J9177 36 366 8.2 6.8 16 169 8.5 0.17 0.02
1/6f78 a7 193 7.3 9.7
2/10f78 49 197 7.6 10.0
319f78 20 216 76 9.1
4/6178 13 223 7.6 9.1
5/11f78 6 227 8.2 10.4
618f78 3 225 8.2 6.9
7/13f78 3 238 8.1 8.1
8/10f78 7 258 78 5.9
9/14178 18 271 78 7.0
1015178 12 275 8.4 10.6
1119178 22 290 7.5 4.3
12f7f78 20 294 7.7 9.5
1/5/79 21 299 7.8 10.4 13 138 8.1
2/16179 18 255 7.7 10.4
3/15/79 12 281 72 6.7
4/5f79 20 260 8.1 10.2
5/10179 8 259 8.2 9.9
618179 5 284 7.8 7.0
7/12179 6 291 8.1 7.0 12 130 6.0
8/10179 4 304 8.0 5.8
9f7f79 13 311 8.1 5.3 13 143 6.0 0.08
1014f79 17 320 8.6 7.0
,,18179 28 299 7.7 8.5
12f7/79 19 299 7.5 7.1
1/11180 28 260 8.3 10.2
2/6180 17 245 79 10.3
316180 3 228 7.7 9.8
4/1Q/80 19 235 7.7 9.6
518180 3 236 8.1 8.8
616180 6 246 8.1 8.5
7/11180 5 263 8.1 7.7 10 N.D.
8/22180 5 268 7.9 6.7
9/18180 12 280 8.3 4.8
10/15/80 9 185 8.3 7.1
11/13180 13 256 8.4 9.1
Appendix B-Case Histories of Emerging
Disinfection Technologies

Several case histories of water supply systems that the water supply, local officials found that the lake's
have used disinfectants other than chlorine are watershed had a large muskrat population.
presented in this appendix. The descriptions of each
case include the modifications in disinfection/oxida- At the time of the Giardiasis outbreak, treatment of
tion techniques necessary to (1) attain the desired North Andover water included chlorination without
degree of disinfection while minimizing the filtration. Two pumping stations distributed the
formation of disinfection by-products; and (2) cope water through the piping system, each station
with other common problems such as taste, odor, handling 0.12 to 0.13m 3/sec (2.5 to 3 MGD) Over the
color, and algae. years, the distribution pipes have formed substantial
biofilm and scale, due to high humic concentrations
The case histories cover ozone, ultraviolet radiation, in the water.
chlorine dioxide, and chloramine. Section B.l
presents case histories invol ving the use of ozone as a Increased chlorination overcame the immediate
primary disinfectant and as a preoxidant for lowering Giardia problem, but subsequent descaling of the
THM levels. Section B.2 illustrates the use of distribution pipes released coliform organisms to the
ultraviolet radiation as a primary disinfectant for water. The State Department of Environmental
surface water followed by chlorination. Three case Quality Engineering (DEQE) had to continue the
studies involving the uses of chlorine dioxide as a "boil water" notice. In addition, THM levels rose to
preoxidant for THM control, as a primary and above 200/11glL due to the increased chlorination, and
secondary disinfectant, and in combination with residents began complaining about the strong
chloramines to lower THM production are presented chlorine concentrations. Thus, ozone was
in Section B.3. Lastly, chloramine case histories are investigated as an alternative primary disinfectant.
presented in Section B.4.
Interim Solution
B.1 Ozone Case Histories In October 1986, ozonation systems, rented from an
equipment supplier, began operating at both Lake
B.1.1 Primary Disinfection with Ozone: North Cochichewick pumping stations. One system had a
Andover, Massachusetts 22.7-kglday (50-1b/day) ozone generating capacity
Under normal circumstances, total coliforms in and the second had a 68.04-kglday (l50-1blday)
North Andover's raw water supply, Lake capacity. When the efficacy of ozonation was proven,
Cochichewick, are between 50 and 500 per 100 mL the town purchased and installed two 68.04-kglday
and turbidity is between 1 and 2 NTU. However, in (150-lblday) ozonation systems, one at each pumping
early 1986, 18 cases of Giardiasis were reported. station. At four points in the distribution system,
Residents were instructed to boil tap water for at chlorine was added to provide a residual disinfectant.
least 5 minutes before use, while public health
officials sought to locate and correct the problem. These two ozonation systems were installed as a stop-
gap measure to control Giardia cysts until a proposed
The raw water was found to contain Giardia cysts in $10.5 million 0.53 m 3/sec (l2-MGDl plant was
samples taken in April 1986. In addition to the designed and constructed. The new plant, to be on
concern that residential septic systems were affecting line in 1991, will provide complete treatment,

180
Appendix B-Case Histories of Emerging Disinfection Technologies 181

including ozonation, filtration, and postfiltration percent lower) and taste and odor of the finished
GAC adsorption with residual chlorination. water greatly improved.

The State Department of Environmental Quality


Engineering (DEQE) provided emergency funding of Permanent Solution
$2.5 million for installing the interim ozonation The planned 0.53 m 3 /sec (l2-MGD) treatment plant
systems, connecting pipelines to the three will include ozonation, filtration, and postfiltration
neighboring communities, relining pipes with GAC adsorption, followed by secondary disinfection.
cement, and replacing water mains. Annual rental Preozonation will be applied before rapid mix and
fees for the two ozonation systems were filtration. After dual media filtration, GAC
approximately $90,000. The two 68.04-kg/day (150 adsorption is incorporated, followed by chlorination
lb/day) ozonation systems were purchased for for secondary disinfection. In addition to providing
$325,000 (total). This price included the air additional removal of contaminants, the GAC step
preparation systems, ozone generators, diffuser will allow mineralization of much of the
contactors, and ozone destrl1ction equipment; plus the biodegradable organic fractions of the water.
appropriate instrumentation, including a residual
ozone monitor for the outlet of the contact chamber.
8.1.2 Preozonation for THM Controi:
Each of the purchased contacting chambers is 3 m (10 Kennewick, Washington (Cryer, 1986)
Il) wide, 6 m (20 ft) long, and 4.8 m (16 Il) deep. The Prior to 1977, the City of Kennewick drew its water
contactors have five barned sections to which equal supply from a system of five Ranney collectors located
amounts of ozone are applied. Plug flow is in the Columbia River. The water was chlorinated
maintained throughout the ozone contactors. Applied before distribution. When installed, this system was
ozone dosages are 5 mglL and the dissolved ozone capable of producing approximately 0.876 m 3 /sec (20
concentration at the outlet of each contact chamber is MGD); however, its output deteriorated to about
between 0.9 and 1.0 mg/L. For design purposes, the 0.657 m 3/sec (15 MGD) by 1977. By 1978, peak water
average ozone concentration in each contactor was demand reached the capacity of the system. It was
assumed to be 0.5 mglL. Total residence time of water determined that direct utilization of the Columbia
in the contactors is 10 minutes during the summer at River would be the only long-term reliable water
full pumping rate. During the winter, when pumping source.
rates are reduced by 50 percent, the residence time is
20 minutes. Since the raw water would be drawn directly from the
river, additional treatment was necessary to
Water temperatures of the lake vary from about 5°C maintain finished water quality. A pilot plant study
(41°F) in winter to just under 20°C (68°F) during the was undertaken to test alternative water treatment
summer. Therefore, the appropriate CT values for processing steps. This study included the use of
99.9 percent inactivation of GiardiaJamblia range preozonation and conventional postfiltration GAC
from 1.9 to 0.72 mglL-min. During the summer, the adsorption, in addition to conventional and direct
period when the shortest contact time is experienced, filtration procedures.
at an average dissolved ozone concentration of 0.5
mgfL, a TIO contact time of 5 minutes (50 percent of Average values of the raw water quality parameters
the peak-flow hydraulic detention time) would attain of the Columbia River are:
a CT value of2.5 mgfL-min, which is more than
adequate. During winter (TIO = 10 min), the CT will • TIHM· 1 IlgfL
be 5 mgfL-min. • TIHM Formation Potential- 136 Ilg/L (7 days
chlorine contact time)
• TOC - 2.4 mglL
• No. Particles - 1l,650/mL
The Results • Particle Volume - 160,700 nUL
After approximately 90 days of ozone treatment, the • Turbidity - 1.7 NTU
State DEQE unconditionally lifted the "boil water"
order, which had been in effect for 9 months. Both In the pilot study, the preozonation and
Giardia cysts and coliform organisms were coagulation/filtration steps each provided
eliminated from the North Andover water supply. approximately 30 percent reduction in TTHM
formation potential (TTHMFP) levels and 10 percent
In addition to the microorganism control, several reduction in TOC levels. The combined processes
other benefits resulted from use of ozonation. Prior to gave approximately 60 percent reduction in levels of
the outbreak problem, THM values ranged from 8 to TTHMFP and 20 percent reduction in TOC levels.
120 IlgfL. Since ozonation, measured THM levels The preozonation, coagulation, and filtration steps
range from 1.1 to 2 Ilg/L. In addition, the color of the combined were determined to be operationally
treated water has improved significantly (65 to 95 equivalent to activated carbon adsorption for the
182 Upgrading Existing or Designing New Drinking Water Treatment Facilities

removal ofTHM precursors; the process was also complaints, the treated river water has a higher
determined to be the most cost-effective method of customer acceptance than the water from the Ranney
treating organics in the Columbia River water collector. Water quality from the collector system is
supply. not very different from the raw water from the
Columbia River.
Solution THM analyses indicate that the treated water from
A 1.314-m3 /sec (30-MGDl water treatment plant was the new plant has average TTHM concentrations of
constructed in four stages, adding 0.329m 3/sec (7.5 approximately 141lg/L, while the water from the
MGDl capacity in each stage. The new plant includes Ranney collector averages approximately 107 Ilg/L
preozonation, flash mix, coagulation, flocculation, TTHM.
filtration, and postchlorination. Design criteria for
the preozonation facilities are given in Table B-l.
ApplIed Ozone Dosages
Applied ozone dosage rates have ranged from 1.7 to
Table B·l. O..llIn Cotena for Kennewick Water
Treatment Plant Preozonatlon Facilities
2.5 mg/L. Until 1983, ozone residual levels at the
contactor outlet were maintained at approximately
DeSign Cntena 0.5 mg/L.ln 1983, the city installed a dissolved ozone
Item Umts Initial Ultimate analyzer to control the ozone dosage, which has
Plant CapacIty lowered the dissolved ozone residual concentrations
Design MGD 7.5 30.0
to 0.1 mg/L, thus saving ozone, and still controlling
biological growth in the filters and basins prior to
cIs 11.6 46.41
chlorination.
GPM 5.200 20.800
Ozone Contact Assuming an ozone concentration of 0.1 mg/L is
Basms present throughout the 10-minute hydraulic contact
No. 01 baSIns 4 16 time (TIO = 5 min), a CT value of 0.5 mg/L-min is
Detennon bme at min 10 10 obtained. Since the treatment process includes
desilln flow filtration, which provides a 2-log reduction of Giardia
Basin dimenSions - 14 x 8 14 x 8
inside
Average water dep1h
" 16 16
cysts, only 1 additional log reduction is required to
meet the disinfection requirements of the Surface
Basin volume "3"
gal
1.792
13.400
1.792
13.400
Water Treatment Rule. The CT value attained by
ozonation (0.5 mg/L-minl is less than the I-log
Total basin volume "3 7.168 28.672
reduction requirement at 5°C (0.63 mg/L-min) but
Chemical Feed Rate IMlay 250 1.000 provides at least 80 percent of the required
(max. dosage at
design now) disinfection. In addition, the 60-minute chlorine
2 x 125
contact time (at pH 8.0l for secondary disinfection,
ChemIcal Feeders Iblday 3 >l 250
2 x 125 provides an additional 36 percent of the required
disinfection. Consequently, the combined disinfection
Chemical Storage 250 1.000
Capacity with ozone and free chlorine provides at least 116
percent of the primary disinfection required by the
Number 01 Ozone 2 5
Generators Surface Water Treatment Rule.
1 pound • 0.4536 kilograms: 1 MGO • 0.044 m3lsee:
I GPM • 0.063 Usee: , foot • 0.3 meters: I "3 • 0.0283 System Operation
m3 .
Operationally, the ozone generation equipment has
At the point of application, ozone dosages averaged performed very well. The compressors have required
1.5 mg/L and peaked at 4.0 mg/L. The contactors only preventive maintenance. The ozone generators
provide 10 minutes of detention time. Raw water required the replacement of only three burned out
total coliform levels are consistently less than 50 per tubes during the first 6 years of operation. The major
100 mL. Raw water turbidities range from 1.5 to 2.0 maintenance problem appears to be tube fouling
NTU. caused by excessive moisture in the feed gas. This
situation was caused by two factors. First, after
The new 0.329-m 3/sec (7.5-MGDl direct filtration several years of operation, the refrigerant air dryer
treatment plant operates from May through October, unit had developed a small leak, which reduced the
when system demand exceeds the capacity of the effectiveness of the air preparation system. Second,
Ranney collector system. The collector system is still the absorptive medium in the desiccant drier should
used because it operates at lower cost than the new have been replaced when its regeneration capacity
treatment plant. Generally, the service area was reduced to 40 percent of its original capacity.
customers have been satisfied with the quality of Cleaning of dielectric tubes has become an annual
water provided. Judging from the limited number of maintenance procedure.
Appendix B-Case Histories of Emerging Disinfection Technologies 183

The only other significant operational problem designed to reduce concentrations of Escherichia coli
concerned the ozone contactors. Excess foaming and organisms by a minimum of 5 logs 00 5 reduction!.
scum production can occur during spring and late
summer, primarily caused by algae destruction. This The system is equipped with a telemetry control
problem was remedied by installing surface system and fully automated backup system. Each
skimmers and froth spray equipment. In addition, bank ofthree irradiation chambers has two units on
the stainless steel tubes holding the ceramic diffusers line at all times, with the third unit serving as
corrode after about 2 to 3 years of use, and must be backup. In the event that the UV intensity drops
checked and occasionally replaced when the diffusers below acceptable limits (20,000 I-lW sedcm 2 ) in any of
are cleaned. the chambers, the automatic butterfly valve will
close, stopping flow through the chamber; at that
time, the automatic butterfly valve on the standby
System Costs unit will open. The alarm system also is activated if
Ozone generation efficiency has averaged UV intensity drops below acceptable limits in any of
approximately 35.3 to 37.5 kWhfkg 06 to 17 kWMb) the chambers. The UV alarm system is interfaced
of ozone produced. Based on qualitative observation, with the automatic dialer and alarm system.
the addition of ozone prior to flocculation and
filtration results in a 10 percent reduction in the In 1987, total equipment costs for the six-unit UV
amount of flocculants. In 1987, the treatment plant irradiation system with butterfly valves was $74,587.
was producing finished water at a cost of
approximately $3.73/1,000 m3 ($14.10/mil gal). EPA's latest Draft Guidance Manual for compliance
with disinfection requirements (EPA, 1989a)
contains "CT Values" for inactivation of viruses by
B.2 UV RadIation Case Histories UV radiation independent of temperature:
B.2.1 Ultraviolet Radiation for Primary
Disinfection: Fort Benton, Montana
CT Values by UV
The city of Fort Benton, Montana, obtains drinking Log Virus Inactivation (mW-sedcm 2 )
water from the Missouri River. The then current
filtration plant (20 to 30 years old) was in need of 2.0 21
upgrading. Rather than building a new filtration 3.0 36
plant, the city built a new 0.088m 3/sec (2-MGDl
treatment plant in 1987. Water is drawn through
Ranney collectors installed 6 to 7.5 m (20 to 25 fi) For the UV facility at Ft. Benton, the initial UV
below the river bed, a system that allows the river dosage of 41 mW-sedcm 2 provides well in excess of 3
bed to naturally filter the raw water. Turbidities of logs inactivation of viruses. However, after 4,500
water entering the treatment plant average 0.08 hours of UV tube operation, the anticipated decrease
NTU. No Giardia cysts have been found in the water in UV dosage (to 25 mW-sedcm 2 ) will provide only 2
from the Ranney collectors. logs of viral inacti vation.
The water is treated with UV radiation for primary
disinfection, then chlorinated for secondary
disinfection. An applied chlorine dosage of only about B.3 Chlorine Dioxide Case Histories
1 mglL is necessary. The entire water treatment
system is housed in a 2.97-m 2 (32-sq fi) building. B.3.1 Predisinfectlon for THM Control:
Evansville, Indiana (Lykins and Griese,
The UV disinfection system consists of six irradiation 1986)
chambers, two control cabinets with alarms, chart In order to comply with the November 1979
recorders, relays, hour-run meters, lamp and power Amendments to the National Interim Primary
on-lights, six thermostats, electrical door interlocks, Drinking Water Regulations, the Evansville Water
mimic diagrams, and six UV intensity monitors and Sewer Utility had to reduce THM levels in
measuring total UV output. Each irradiation drinking water. At the time, raw water was
chamber contains one 2.5-kW mercury vapor, chlorinated, treated with alum prior to primary
medium-pressure arc tube, generating UV radiation settling, treated with lime to raise the pH to 8, passed
at 253.7 nm. through secondary settling, fluoridated, filtered
through rapid sand, and finally postchlorinated.
The initial UV dosage is 41,000 I-lW sedcm 2 at These processes were conducted in two 1.31-m3 /sec
maximum water flow 104 Usee 0,650 gal/min) (30-MGD) treatment trains. THM levels exceeded the
through each irradiation unit. Expected arc tube life THM standard of 100 I-lgiL. Prechlorination doses
is 4,500 operating hours, providing a minimum UV averaged 6 mglL and distribution system residence
dosage 0£25,OOO I-lW sedcm 2 . These conditions are time averaged 3 days.
184 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Pilot Plant Study Operating Experience


With EPA's assistance, research evaluating the use During the first 5 months of use, various chlorine
of chlorine dioxide as an alternate disinfectant was dioxide dosages were used to determine the optimum
initiated in a 6.3-Usec (100-gaVmin) pilot plant. One dosage for maximum reductions in THMs and the
train of the adjacent full-scale treatment plant served portion of the total dosage that would appear as an
as the control for the pilot plant study. oxidant residual in the finished water. A general
review of system operations was conducted at the
In the first phase of the study, chlorine dioxide was same time.
substituted for postchlorination. The water treated
with chlorine dioxide was stored 3 days to simulate During this period, operators encountered only one
the distribution system residence time. TIHM major problem. The concentrated disinfectant
concentrations were much lower using chlorine oxidized the brass corporation cocks used to connect
dioxide postdisinfection, 1.211g1L (1.9 mglL residual) the PVC chlorine dioxide feed lines to the water
as compared with 14111g1L (2.5 mglL residual) using influent piping. This oxidation and subsequent
chlorine postdisinfection. leaking of chlorine dioxide solution temporarily
disrupted operations. The problem was resolved by
Predisinfection with chlorine dioxide to maintain a sliding a section of PVC pipe through new
residual through the pilot plant did not increase the corporation cocks into the main stream of the raw
THM concentration and provided adequate water lines. This modification permitted the PVC
disinfection. The chlorine dioxide residual decreased piping to serve as an inductor, while preventing
from 4 to 0.3 mglL through the pilot plant. direct contact of the brass corporation cocks with
concentrated chlorine dioxide solution.
In the second phase of the study, the pilot plant
procedures were evaluated throughout the year to Since the implementation of the new chlorine dioxide
determine the extent of seasonal effects. In this system, total oxidant levels from chlorine dioxide
phase, predisinfection with chlorine dioxide (1.1 have been maintained consistently below the 1.0
mglL average applied dosage) reduced the amount of mglL recommended by EPA. With an average applied
THMs formed during postchlorination by chlorine dioxide pretreatment dosage of 1.2 mglL,
approximately 60 percent. The idea ofpostdisinfec- total oxidant concentrations in the finished water
ting with chlorine dioxide was abandoned because of have averaged 0.5 mg/L. These data show that
the difficulty in maintaining an adequate residual in approximately 42 percent of the chlorine dioxide
the distribution system while meeting the current dosage remains as total oxidant. In addition, the new
EPA recommendation for total oxidant concentra- system maintains TIHM levels in the distribution
tions from chlorine dioxide (1 mglL maximum). The system between 50 and 80 I1g1L.
average chlorine residual concentration in the
c1earwe II was 2.1 mglL. 8.3.2 Primary and Secondary Disinfection
with Chlorine Dioxide: Hamilton, Ohio
(Augenstein, 1974; Miller et al., 1978;
Plant Modifications U.S. EPA, 1983)
Based on data obtained from the pilot plant study, In 1956, a 0.657-m 3 Isec (15-MGDllime-softening
changing the primary disinfectant from chlorine to ground water treatment plant began operating in
chlorine dioxide was judged to be the most effective, Hamilton, Ohio (18 wells, 60-m [200-ft] deep).
lowest cost procedure for meeting the THM standard. Chlorine was used as the sole disinfectant when the
plant opened. Because of customer complaints of
A separate building was constructed to house the chlorinous tastes and odors, however, chlorine
chlorine dioxide generator am! 1."'0 907.2-kg (1-tonl dioxide was tested and then substituted for chlorine
cylinders of chlorine. The installation is capable of in 1972 as a primary and secondary disinfectant.
generating 6.46 kg (14.24 Ibl of chlorine dioxide per
hour, which can be divided in any proportion between Hamilton's treatment process now includes aeration,
the two halves of the treatment plant. Gaseous lime addition, flash mixing, sedimentation,
chlorine and 25 percent sodium chlorite solution are recarbonation (with food-grade carbon dioxide),
delivered to the chlorine dioxide reactor under partial filtration, fluoridation (sodium silicofluoride),
vacuum generated by an eductor. Both reagent flows chlorine dioxide for primary and secondary
are controlled by flow rate meters, and the system is disinfection, and c1earwell storage. Raw water
designed to shut down if the eductor water supply turbidities are below 1 NTU, and raw water total
fails or if chemical feed lines break. coliforms are less than 1 per 100 mL.
Appendix B-Case Histories of Emerging Disinfection Technologies 185

Generation of Chlorine Dioxide chlorine dioxide accounting for only a fraction of the
total. The operating and maintenance costs were less
Chlorine dioxide is generated by mixing 37 percent than $50 annually. The plant-fabricated reactor,
aqueous sodium chlorite and aqueous chlorine in a piping, hardware, and installation, which was
ratio of 1:1 by weight, 0.24 g of each reagent/m 3 of performed by plant personnel, cost roughly $400 in
water to be disinfected (2 Ib of each reagent/mil gal 1977. The peristaltic pump for sodium chlorite
water). This provides an applied chlorine dioxide solution cost less than $200. The two chlorinators,
dosage of 0.25 mg/L. The residual leaving the each worth $600, were already on line at the plant.
clearwell is approximately 0.15 mg/L, and is 0.10
mg/L at the extremities of the distribution system.
Implications of the SWTR CT Values for
The generation system consists of one plant- Hamilton
fabricated reactor vessel for chlorine dioxide Although the Hamilton raw water is ground water,
production, one peristaltic pump for sodium chlorite and therefore probably will not be subject to the
solution, and two chlorinators (one serves as a requirements of the Surface Water Treatment Rule,
standby). Aqueous sodium chlorite solution (37 it is interesting to consider the effects if such
percentl is delivered in 90. 72-kg (200-lbl drums. disinfection requirements as listed in the latest EPA
Chlorine gas is delivered to the site in 68-kg (150-1b) Guidance Manual (U.S. EPA, 1989a) were to be
cylinders, and two 68 kg (150-lbl liquid chlorine levied on this water supply system.
cylinders are positioned next to the chlorinators. The
weight of the cylinder contents is measured by a Chlorine dioxide is added to the Hamilton water in
scale. Switching over from one cylinder to the other is applied doses of 0.25 mg/L as it enters the clearwell.
conducted manually. PVC tubing connects the The water temperature is about 20°C (68°F) year
chlorinator and chlorine dioxide reactor vessel; heavy round, and the pH is 9.4 to 9.5. Hamilton's first
Tygon tubing transports the sodium chlorite solution customer is located about 0.5 miles from the plant.
from the drum to a small plastic day-tank and the Thus there is very little contact time in the
reactor. After about 1 month, the Tygon tubing loses distribution system.
its rigidity and must be replaced. The semi-
transparent day-tank allows visual inspection of the The plant filters efficiently, and therefore only I-log
sodium chlorite lev-lI, thus enabling the operator to additional inacti vation of Giardia cysts and 2-logs
maintain an acceptable suction head on the inactivation of viruses need to be provided by the
peristaltic pump. chlorine dioxide. The latest EPA Guidance Manual
shows that at 15°C (59°Fl CT values of 5 mglL-min
The chlorine dioxide reactor vessel is constructed of and 2.8 mg/L-min will provide the required degrees of
Schedule 80 PVC piping, 45.72 cm (18 in) high and disinfection for Giardia cysts and viruses,
approximately 15.24 cm (6 in) in diameter. The vessel respectively.
is filled with PVC rings, 2.5 cm (1 in) in diameter.
The chamber is opaque except for the sight glass The Hamilton clearwell holds 1,892,500 L (500,000
mounted in-line on the discharge piping. A white gal). During periods of high water use, water is
card is positioned behind the sight glass for better produced at the rate of 0.83 m 3/sec (19 MGDJ. In
observation of the chlorine dioxide color. periods oflow water use, only 0.35 m 3/sec (8 MGD)
are produced. Thus the contact time in the Hamilton
clearwell ranges from 30 to 90 minutes, at the high
Other Benefits of the Chlorine Dioxide System and low production rates, respectively. Assuming
Prior to the installation of the chlorine dioxide that the average concentration of chlorine dioxide in
system, customers complained about brown iron the clearwell is 0.2 mg/L, then the CT values
stains on laundry articles. The switch to chlorine provided are 6 and 18 mglL-min, respectively.
dioxide treatment loosened the brown slimes from the
mains. The distribution system was flushed, and Thus the current disinfection conditions using
shortly thereafter the staining problems were chlorine dioxide meet the CT requirements for both
resolved. Plant personnel attributed the source of the Giardia cysts and viruses.
problem to crerwthrix and leptothrix bacteria (iron
bacterial that had been present in the extremities of
the distribution system before chlorine dioxide was Secondary Disinfection with Chlorine Dioxide
introduced. at Hamilton
The recently promulgated Surface Water Treatment
Rule requires that a minimum secondary
Costs of Chlorine Dioxide disinfectant concentration of 0.20 mg/L be present as
In 1977, the costs of chlorine dioxide disinfection for the water enters the distribution system or that the
Hamilton were about 3.6e1capitalyear. Total HPC level be less than 500 mL in the distribution
chemical costs averaged $0.05/m 3 ($0.19/1,000 gall, system. Since Hamilton's chlorine dioxide
186 Upgrading Existing or Designing New Drinking Water Treatment Facilities

concentration entering the distribution system is Ilg/L, but bacterial counts for coliforms indicated
0.15 mg/L, the city will either have to increase this an inadequate residual in the distribution
residual to 0.20 mg/L or rely on HPC analyses to system. The experiments were repeated using a
comply. 7: 1 chlorine-to-ammonia ratio. Acceptable
bacterial quality was achieved, but numerous
B.3.3 Preoxidatlon with Chlorine Dioxide, taste and odor complaints were received during
this period.
Postchlorlnatlon with Chlorine Dioxide
and Chloramine: Galveston, Texas • Chlorine dioxide/chlorine - Chlorine dioxide was
(Myers et al., 1986) installed in May 1984, using a generator with a
The Galveston County Water Authority (GCWA) conversion efficiency (chlorite ion to chlorine
owns and operates an 18-MGD water treatment plant dioxide) ofapproximately 80 percent. After
in Texas City, Texas. In 1986, the plant produced preoxidation, the clarified water showed no
0.526 m3/sec (12 MGO) from the Brazos River. Raw traces ofTHMs. No taste and odor complaints
water had high color; variable turbidities (68 to 111 were received, despite very high raw water algae
NTU); high organic content; high iron (2.7 to 3.8 counts (up to 5,000 blue-green algae/mLl.
mg/L); and seasonally high algae content, sometimes However, finished water THM levels sometimes
reaching levels of 5,000 blue-green algae per mL. persisted above 100 Ilg/L. In November 1984, a
Total THM formation potentials for Brazos River chlorine dioxide generator with a greater than 98
water between September 198J and April 1984 percent conversion efficiency was installed. THM
ranged from 400 to 650 Ilg/L. Finished water THM levels averaged 1021lg/L, and no taste and odor
concentrations during the same period ranged from complaints were received.
180 to 350 Ilg/L. During periods of intermittent raw
water algae blooms and associated high organic • ChloriM dioxide/chlorine dioxide - Chlorine
carbon contents, numerous taste and odor complaints dioxide for preoxidation and postdisinfection was
were received. These problems prompted an tested in March 1985. The Texas Department of
investigation of alternative disinfectants. Health required that a maximum chlorine
dioxide residual of 1.0 mg/L be maintained and
The original treatment process included the addition that finished water quality be monitored
of cationic polymers for primary coagulation, lime for throughout the distribution system. During the
pH adjustment, prechlorination for taste and odor test period, finished water THM concentrations
control, and ferric sulfate as a flocculent aid prior to averaged 60 Ilg/L and turbidities were the lowest
upflow reactors and clarifiers, which provided of any of the alternative disinfectant scenarios.
flocculation and sedimentation. Dual media Bacterial counts generally were excellent, but
filtration, chlorine disinfection (2.4 to 5.0 mg/L total intermittent elevated counts were noted at the
available chlorine residual), and fluoridation clearwell and at two locations 3.2 and 8 km (2 and
followed. 5 mil from the plant. Additionally, bacterial
counts displayed a shift from orange to yellow-
staining gram negative (-) rods to white-staining
PI/ot Study gram positive ( + ) rods, similar to slime-forming
Chloramines and chloride dioxide were selected for Bacillus sp.
further study and pilot plant testing. Several
preoxidationlpostdisinfection combinations involving • Chlorine dioxide/chlorine and chloriM dioxide -
chlorine, chloramine, and chlorine dioxide were The above test was repeated, adding chlorine in
studied. The studies indicated that preoxidation with conjunction with chlorine dioxide. Excellent
chlorine dioxide and postdisinfection with chlorine bacterial quality was obtained with plate counts
dioxide in combination with either chlorine or' at or below the 500 colonies per 100 mL for all
chloramines provided effective taste and odor control, monitoring locations. The shift in bacterial
maintained an active disinfectant residual, and species distribution continued as the plate counts
minimized THM formation (to a level of about 68 decreased, so that over 95 percent of all colonies
Ilg/L). A brief summary of the results obtained from examined were either yellow gram negative rods
the combinations studied follows: or white gram positive rods. THM levels ofthe
finished water rose to an average of81llg/L.
• Chlorine/chlorine - Finished water exhibited
intermittent algae-related tastes and odors; THM • Chlorine dioxide/chloramine and chlorine dioxide
levels were in excess of Federal standards; - These tests were conducted in December 1985.
bacterial quality was excellent. THM levels of the finished water averaged 68
Ilg/L, and the bacterial quality remained
• Chloramine/chloramine - Chlorine-to-ammonia excellent. No coliforms were found in the
weight ratios of3:1 and 7:1 were used. With the clearwell or in the distribution system and
3: 1 ratio, THM levels were lowered to about 60 bacterial counts had ranged from < 1 to 30
Appendix B-Case Histories of Emerging Disinfection Technologies 187

colonies per 100 mL. Of those colonies identified, Thus, as MCLs for halogenated organic materials
over 95 percent were white-staining, gram other than THMs are promulgated, utilities using
positive Bacillus sp. with less than 1 percent postchloramination should plan to determine the
belonging to the orange-staining gram negative makeup of their TOX fraction.
type.
Table B·2. Summary of THM Data at Bloomington, Indiana,
Chlorine dioxide preoxidation has proved an with Fr. . Chlorination, August 16, 1984
excellent algicide and biocide. The effectiveness of
chlorine dioxide in removing algae in flocculation Chlorine Residual Toc Tthm Tox
and sedimentation has resulted in a decrease in filter Sampling Point (1TlQiL) (mgiL) (..gil) (l'gIl)
fouling and improved finished water turbidities. Raw water 4.3 1 23
Odors experienced at the flocculating clarifiers, and SetUed water 0.25 3.6 48 127
taste and odor complaints were reduced. Filtered water 1.0 2.4 81 205
Dlst system "1 1.8 110 291
Bacterial counts, using the membrane filter method
Dist system "2 0.65 151 363
for coliforms, have greatly improved as well. With
the new system, all bacterial counts obtained from Source: Singer (1986).
samples collected at the distribution point have been
below 30 colonies per 100 mL, and all but two
samples have been below 10 colonies per 100 mL.
Samples from the distribution system are continually Table B-3. Summary of THM Data at Bloomington, Indiana,
with Poatchloramlnatlon, August 26, 1984
below the guideline of 500 colonies per mL, and often
are below 5 colonies per 100 mL. These counts Chlonne Residual Toc Tthm Tox
indicate an adequate and stable residual. Sampling Pomt (1TlQiL) (mgll) (~gIl) (~)
Raw water 4.1 0 17
settled water Trace, tree 2.8 53 94
8.4 Chloramine Case Histories
Filtered water 1.2 combined 2.8 55 91
B.4.1 Prechlorination, Postchloramination: Dist system '" 1.0 combined 52 115
Bloomington, Indiana (Singer, 1986) Oist system n 0.9 combined 57 116
Bloomington obtains raw water from a lake with Source: Singer (1986).
TOC levels of 4 to 6 mglL. The water is treated with
alum coagulation, flocculation, sedimentation, and
filtration through pressure filters. Prior to Since switching to postchloramination, the utility
September 1984, chlorine was applied to the raw has experienced no adverse effects in operations or in
water and just before the pressure filters. Average finished water quality. According to distribution
chlorine dosages were 1.8 and 1.0 mglL at each point, system monitoring records, the microbiological
respectively. According to quarterly compliance quality of the water has been maintained.
monitoring reports, average TTHM concentrations
were exceeding the 100 }lglL limit a majority of the
time.
B.4.2 Prechlorine Dioxide, Prechlorination,
In September 1984. the Bloomington water utility and Postchloramination: Philadelphia,
changed from post chlorination to postchloramina- Pennsylvania (McKeon et aI., 1986)
tion. Before the pressure filters, an average 0.54 The Baxter Water Treatment Plant, a 12.35-m3 /sec
mglL of ammonia was applied along with 1.5 mglL of (282-MGD) conventional treatment plant built in
chlorine. The desired residual chlorine concentration 1960, supplies drinking water from the Dela ware
leaving the plant of 1.0 mglL of free chlorine was River to a population of over 800,000. Chemicals used
changed to 1.5 mglL of combined chlorine. After the in treatment include chlorine, ferric chloride or
change, quarterly THM levels ranged from 24 to 57 ferrous sulfate, lime, fluoride, and ammonia.
}lglL. Powdered activated carbon is used on demand for
control of taste and odor, and chloride dioxide is used
Table B-2 summarizes THM and TOX (total organic for control ofTHMs, tastes, and odors. The chlorine
halide) data for samples collected at points in the dioxide system was left over from the previous water
treatment train when chlorine was used for both pre- treatment plant on that site. In the 1950s, it was used
and posttreatment. The data show that TOX levels to oxidize phenolic compounds found in the
increase with TTHM levels. Table B-3 summarizes watershed, which have since been eliminated.
similar data after postchloramination was instituted.
These data show that although the TTHM formation Prior to 1976, the Baxter plant practiced breakpoint
ceases after the addition of ammonia, the production chlorination at the raw water basin and maintained
ofTOX continues, but at a greatly reduced rate.
188 Upgrading EXisting or Designing New Drinking Water Treatment Facilities

free chlorine in the distribution system. A total of96 basins. Sufficient chlorine was then added at the new
hours offree chlorine contact time was typically application point to achieve a free chlorine residual of
achieved. 1.5 to 2.0 mglL. This residual was converted to
chloramines 1 hour later as the water left the filter
In 1978, analyses ofTHMs showed peak building.
concentrations above 300 IlgiL with an annual
average of 140 IlgiL. In light of these results, the This treatment regimen gave adequate control of
Philadelphia Water Department began to reevaluate taste and odor, again, except during spring algae
its disinfection strategies. Factors that affected the blooms, which forced a reversion back to free chlorine
selection of an alternative included THM, bacterial, at the intake. THM levels, with only 1 hour of free
taste, odor, algae, and corrosion control; residual chlorine contact time, resulted in summer values
duration; and economics. averaging 100 Ilg/L with an annual average of60
IlgiL.
Process Modfflcatlons In November 1982, a 10-minute chlorine contact time
Modifications were implemented between 1976 and was tested. Results indicated that satisfactory
1983. Chloramination of the finished water was disinfection could be achieved with only chloramines
introduced in 1976 to reduce free chlorine contact carried through the flocculationlsedimentation
times and THM levels. Ammonia was added to basins when the water temperature was below 15.6°C
convert free chlorine to monochloramine. The (60°F). This strategy was initiated on a plant scale in
monochloramine reduced contact time from 96 to 24 December 1982. Adequate disinfection was achieved,
hours, provided a stable residual in the distribution but periodic taste and odor problems persisted,
system, improved the organoleptic properties of the especially after storms. Average annual THM levels
water, and reduced the corrosion rates associated were reduced from 60 to 50 IlglL.
with the use of free chlorine. Adequate disinfection
was assured by maintaining a free chlorine residual The 10-minute chlorine contact time trial was
of 2 to 3 mglL throughout the plant treatment terminated in December 1983, because the
process. THM formation potential within the disinfection scheme did not adequately address taste
treatment process was reduced by 40 percent (from and odor problems. The treatment regimen returned
231 to 1741lglL with the 96-hour contact time) under to a I-hour free chlorine contact time.
summer conditions.

Between 1976 and 1979, the addition of chlorine at Costs


the raw water basin inlet was gradually phased down Over the 10-year period, a 70 percent reduction in
from 7.19g(60 I/) in 1975 to between 2.4and 3.6g THM concentrations was realized. In 1978,
chlorine/m 3 (20 to 30 Ib/mil gal) in 1979. While algae disinfection cost $1.3211,000 m3 ($5.01lmil gal); in
blooms, taste, and odor problems were controlled, 1986, the cost was $1.46 ($5.52) 0977 dollars). Cost
THM levels were still too high (200 giL in the increases were minimized because the reduced
summer, and 140 IlgiL annual average). chlorine contact times resulted in less evaporative
losses of chlorine, which netted a 20 percent decrease
In early 1980, routine use of chlorine at the raw in the amount of chlorine needed.
water intake basin was abandoned. In its place,
chlorine dioxide was added in dosages between 0.5
and 1.0 mglL. Free chlorine contact times were Future Considerations
reduced to 5 hours. Summertime THM values were The near-term goal of the Philadelphia Water
reduced from 200 IlgiL to 140 IlgiL. This treatment Department is to reduce the annual average THM
was sufficient to control taste and odor problems from concentrations to below 50 IlglL, a 15 percent
algae at all times except during the spring algae reduction from 1986 levels. This can be achieved at
bloom. For that period of time, breakpoint minimum expenditure by installing a pH adjustment
chlorination of the intake water and/or 12 to 24 g point at posttreatment. The existing treatment
powdered activated carbonlm 3 000 to 200 Ib/mil gal) scheme calls for raising the raw water pH to 8.4
were added to eliminate vegetative tastes and odors. during rapid mix and carrying this high pH through
the distribution system for corrosion protection.
In the fall of 1980, a chlorine application point was Addition of a pH adjustment point at posttreatment
installed in the "applied to filters" channel which will allow a pH of7.5 to be used through the
allowed for increased flexibility in the use of chlorine. flocculationlsedimentation basins and filters, moving
Free chlorine contact time was reduced from 5 hours adjustment to a 8.4 pH to just after chloramination.
to 1. Chlorine was added at the rapid mix to barely Plant-scale trials of this strategy yielded a 20 percent
achieve breakpoint and provide a residual, which reduction in THM formation (to about 40 IlgiLl.
dissipated within a few minutes. Chloramines were
carried across the flocculation and sedimentation
Appendix B-Case Histories of Emerging Disinfection Technologies 189

[fthe THM MCL is reduced to below 50 ~g/L, ozone


and/or GAC become the likely alternatives at the
Baxter plant. Extensive laboratory and pilot plant
evaluations have developed conceptual full-scale
plant designs incorporating these two treatment
techniques.

Estimated annual amortized capital and operating


costs for ozone at the Baxter plant are estimated to be
about $13.21/1,000 m 3 ($50/mil gall. The associated
costs for GAC postfiltration (IS-minute EBCT) with a
75-day regeneration frequency would be about
$56/1,000 m 3 ($2121mil gall. This design
configuration (ozone plus postfiltration GACl is
capable of producing THM concentrations of less than
10 ~g/L.
Appendix C-Experience with Treatment Technologies
for Organic Contaminants

This appendix is a collection offield experiences with Iblmil gal) for PCE and l,l,l-trichloroethane,
organics treatment technologies. Section C.l includes respectively.
descriptions of GAC systems in use, C.2 describes
packed tower aerators, and C.3 describes powdered The capital cost for the system was $508,500 in 1981
activated carbon. dollars, while the operating costs have been about
$15,000 annually based on 9 hours of operation per
C.1 Experience with Granular Activated day.
Carbon
C.1.2 GAC for Contaminant Control:
C.1.1 GAC for VOC Removal: Washington,
New Jersey (Chrobak et aI., 1985) Cincinnati, Ohio (DeMarco, 1983)
(Westerhoff and Miller, 1985)
A well with a 0.04 m 3/sec (0.8-MGD) capacity in this This case involves a 9.6-m 3/sec (220-MGDl plant that
community was contaminated with four VOCs. The uses Ohio River water. The proposed addition ofGAC
VOC levels fluctuated throughout the well's 9-hour to the original process train is shown in Figure C-2.
daily operating cycle according to the following Figure C-3 shows TOC removal during a GAC pilot
concentration ranges: study conducted at the plant. The chief goals of
adding GAC to this plant were to (1) reduce TOC in
• Tetrachloroethylene (PCE): 50 to 500 lIg/L the effiuent to less than 1.0 mg/L, (2) maximize the
• Trichloroethylene (TCE): 1 to 10 lIg/L use of existing plant facilities, (3) maximize the
• l,l,l-Trichloroethane: 1 to 20 lIg/L flexibility to accommodate future requirements, and
• Carbon tetrachloride: 1 to 5 lIg/L (4) keep costs reasonable.

The town selected a GAC system with two downflow The primary design elements for the addition were a
pressure contactors in parallel. The filters have 2.1-m postfiltration GAC adsorption unit that used
(7 oft) diameters, 3-m (1 O-ft) carbon depths, hydraulic downflow deep-bed contactors, post-GAC
loadings of 4.8 Usedm 2 (7.1 GPMlft 2 ), and EBCTs of chlorination, and onsite carbon regeneration using a
10.5 minutes. Filter backwashing is performed once a fluidized-bed furnace. In addition, the design was
month due to solids in the influent. Wash water is intended to keep carbon losses to a minimum.
filtered with sand media and recycled. Figure C·I
shows a schematic diagram of the plant. The addition of the GAC process resulted in the
following system capacities:
Carbon usage rates for meeting the PCE standard
averaged about 10.8 g GAC/m 3 (90 lb/mil gall of • 7.7 m3 /sec (l75 MGDl maximum daily plant flow
treated water. Meeting the 1, 1,I-trichloroethane rate
standard required carbon usage rates of25.2 g • 5.4 m 3/sec (l24 MGDl average daily average
GAC/m 3 (210 lblmil gal) of treated water. GAC plant flow rate
consumption for the two compounds before • 15-minute EBCT- 3-m (l0-ft) GAC bed depth
breakthrough (any contamination of the effiuent, as • 3-meter (lO-foot) GAC bed depth
opposed to emuent concentrations over the regulated • 24,494.4 kg/day (54,000 lblday) average carbon
standards) was about 12 and 32.4 g/m 3 (l00 and 270 usage

190
Appendix C-Experience with Treatment Technologies for Organic Contaminants 191

fO.
------- .~ ~

,rCarbon
Carbon~
T'~
.~ .. ~"'\.~ V
"'';~.;;Z
1 p..c~: '.,,. I

-
(;:

- ,
chIOnne_\

I
1
~ 1~
r~;:ri"';~;~~
~~:~~

P:JI-
p

..
.' .. ~
-. Sand ··.... t·
. .... ::- .-
r ~I- - . """1
l i'Z-
~

t~ • ·11 t
• II
I' .
Wastewater Recycle

Filtered Water 10 \
D,stnbunon System

FlQure C·l. GAC lreat....nl planlschemalic Vannatta Sl,eel Stalion.

• 41,731.2 kg/day (92,000 lbldayl peak carbon demands of 0.18 m 3 /sec (4.2 MCDl. Table C-l
usage rate presents a profile of this 1957 system.

The carbon transportation system employed for The influent for the system is characterized in Table
regeneration uses schedule 10 316L stainless steel C-2. The presence of vinyl chloride required use of
pipe and maintains water velocities of91.4 to 152.4 PTA because CAC is ineffective for that compound.
cm/sec (3 to 5 £t/sec). The layout design required that However, CAC was also needed to remove
pipe bends in the transportation system have trichloroethylene and aldicarb.
minimum radii to allow the free flow of the carbon
and minimize abrasion. As a result, pipe bends for Costs for three technological alternatives were
7.6-cm (3·in) diameter pipe used 60.96 em (24-in) developed. The alternatives included CAC and PTA
minimum radii, 10.16-cm (4-in) pipe used 91.4-cm individually and CAC and PTA together. PTA proved
(36-in) radii, and 20.32-cm (8-in) pipe used 121.9·cm to be almost one fourth as expensi ve as CAC.
(48-inl radii. Figure C-4 shows the regeneration However, combining the two treatments proved less
system. expensive than adding both treatments separately.

The capital cost of adding the CAC system was $57.7 C.2 Experience with PTA: Scottsdale,
million (1988 dollars). This included the CAC
contactors, regeneration equipment, intermediate Arizona (Cline et aI., 1985)
pumping equipment, outside piping, and The only application of the PTA system described in
modifications to the existing plant. The operating this appendix is located in Scottsdale, Arizona. In the
and maintenance costs are estimated to average $3 to Scottsdale system, PT A was added to a system with
$4 million annually, including costs for labor, power, 24 weBs and the combined capacity of 1. 75 m3 /sec (40
natural gas, and replacement carbon. These MCDl. Two of the wells were contaminated with TCE
combined costs resulted in a 30 to 40 percent increase at levels of 18 to 200 I-lg/L and 5 to 431-lglL,
in the consumers' average annual water charge from respectively.
$80 to about $110 per year.
Both PTA and CAC were considered as potential
treatment solutions; PTA was chosen for its cost
C.1.3 EPA Health Advisory Example effectiveness. For this application, GAC was
This case involves a system serving 30,000 persons. estimated to cost from $0.04 to $0.10/m 3 ($0.17 to
The plant had a capacity of 0.22 m3/sec (5.1 MCDl, an $0.3811,000 gall, while PTA was estimated to cost
average demand of 0.13 m3 /sec (3.0 MCDl, and peak only $0.02lm 3 ($0.07/1,000 gall.
192 Upgrading Existing or Designing New Drinking Water Treatment Facilities

lamella~
Settler Pumping
Sla~ons

I East
Chemical
FlocculaborV
Sedimenlanon
Basins Filters Clearwell DiS1nbutlon System
BUilding

Proposed Pumping
Slanon

Proposed GAC
Facilities

Figure C-2. Cincinnati treatment train with addition 01 GAC.

The P'I'A unit was designed to manage a 75.7 Llsec Haberer units operating in Africa and Europe,
0,200 GPM) flow rate with 3.6 m 02 ft) of packing including a 0.42 m3/sec (9.S-MGDl plant in
material, a 50: 1 air-to-water ratio, and a column Wiesbaden, West Germany, that has operated since
diameter of 3 m (10 ft). Figure C-S is a schematic 1970.
diagram of this facility.
In Newport News, Virginia, a Roberts-Haberer pilot
The P'I'A achieved over 98 percent reductions from plant reduced TOC by 90 percent during initial
the initial influent levels, producing effiuent levels of operations. Within a 24-hour period, TOe removal
TCE ranging from 0.5 to 1.211g1L. The system's decreased to only 20 percent. During that same
capital cost was $300,000, and its operation and period, this unit significantly reduced THM
maintenance costs average $25,000 annually. precursors.

In Shreveport, Louisiana, a Roberts-Haberer unit


C.3 Experience with PAC was used as a roughing filter to reduce THM
Currently only pilot-scale units operate in the United formation potential. The unit succeeded in reducing
States. There are, however, full-scale Roberts- THM formation potential from about 265 lIg/L by 7 to
Appendix C-Experience with Treatment Technologies for Organic Contaminants 193

Concentratron
IJ9Il
4000

3500

/ 1\ Influent

/
/ \
1\ . /K
f.- -
3000 V"
/
~ \
V
1/ 1\
2500

'" "-

"'"
/ \

I"-
1\
2000
\
TOC Removed "\
\ V
'\ ./
V
1500

1000

:r
""
~

"" ..........
/ ""-
--- / ~
500

r-.. / V
~

l"- I-- -- ./
V V --- Contaetor Effluent

I I I I I I
o 20 40 60 80 100
Time. Days

Figure C-3. Typical TOC rec:luetlon curve during pilot Iludy.

31llgiL with PAC dosage of26 to 29 mglL. This


operation, however, did not use the optimum PAC
dosage.
194 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Dryer Off-Gas
Dewatered
Spent Cartlon

To Scrubbers
and Stack

Gnnder

Regenerabon
Furnace
Furnace

J
Off·Gas

Recuperator

Recycle

Afterburner

Dned Spent
Cartlon

Figure C-4. RlIlIeneratlon Iyltem schematic.


Appendix C- Experience with Treatment Technologies for Organic Contaminants 195

Table C-l. Prolile 01 System used in EPA Health Advisory

Source:
• Three wells approximately 500 feet deep
• Capacity of each well IS 1.8 million gallons per day
• Screened between 400 to 500 feet With gravel pack
• 18-lnch steel casing from 0 to 400 feet
• Portland cement grout from 0 to 200 feet
• All wells are pumped to a common manifold that Nows to
the water treatment plant
• 5011 profile: 0 to , 00 feet sandy SOil; 100 to 400 feet.
sand clay mrxture: 400 10 500 feet wet sand and gravel:
500 feet. bedrock
Slorage:3.5 million gallons
Treatment: Iron removal uSIng chlonne OXIdation, alum
coagulation. sedimentation, and rapid pressure
sand filtra~on: diSinfection (chlonne), fluondation.
and corrOSion control (lime and metallic
phosphates) are also practiced.
Constructed: 1957
Mechanical Structural Condition: Excellent

1 foot = 0.3048 meters.


1 Inch = 2.54 centimeters.
1 gallon = 3.785 Illers.

Table C-2, Inlluent Characterization lor System Used In


EPA Health Advisory

Well #1 Well #2 Well #3


Parameter Raw Treated Raw Treated Raw Treated
Iron (mgJL) 3.0 0.05 2.2 0.05 2.0 005
pH 6.0 78 59 78 62 78
Alkalinity Img/L) 10 110 14 110 12 110
Vinyl chlonde 40 20 14 20 6 20
(~gJL)

Tnchloroethylene 50 60 30 60 100 60
(~g/L)

Aldrcarb (total) 30 30 30 30 30 30
(~g/L)

TOlal organic 30 1.0 2.1 1.0 1.0 1.0


carbon (mgJL)
196 Upgrading Existing or Designing New Drinking Water Treatment Facilities

'.

I
I
Exhaust Air
-------- - ----- ...

I
I
I I

Well NO.6
Treated Water
to Reservo"

Figure CoS. Schematic diagram 01 a Scottsdale packed column.


Appendix D-Experience with Treatment Technologies
for Inorganic Contaminants

This appendix describes a number of field


experiences with inorganics treatment technologies. Table 0·1. Metropolitan Dlslr1et Commission Water Quality
Data
Section 0.1 covers two cases of corrosion control;
Section 0.2, the use of coagulants; Section 0.3, Shaft 4 Norumbega
reverse osmosis; Section 0.4, ion exchange; and (Southborough. Reservoir
MAl (Weston. MAl
Section 0.5, activated alumina. Parameters Raw Water Finished Water
Hardness (as caC0 3 ) 12 12
0.1 Corrosion Control Alkalinity (as CaC0 3 ) 8 12
Tolal Dissolved Solids 37 46
0.1.1 Controlling Lead: Seattle, Washington Calcium 3.2 3.4
Seattle, Washington, ha~ instituted a successful Sodium 5.5 9.7
corrosion control program. The program has a Su~ate <15 <15
component that assists residents in materials
ChlOnde <10 <10
selection, in addition to a water treatment
modification program. The city encourages the use of Specific conductance 59 78
(m,cromhOs)
plastic piping, types K and L copper piping, and
solder with less than 0.2 percent lead. In addition, the pH (Unlls) 6.7 8.5
city requires dielectric insulators to mitigate Coppa< <0.02 <0.02
corrosion from joining dissimilar metals. Iron <0.10 <0.10
Zinc <0.02 <0.02
0.1.2 Controlling Lead with pH Adjustment: Lead <0.005 <0.005
Boston, Massachusetts Note: All values In mgIL unless otherwise specified.
Source: U.S. EPA (1984).
A 1975 survey found significant levels of lead in tap
water supplied to Boston residents by the
Metropolitan District Commission (MDC). The solution for corrosion control. The monitoring
protected watersheds of the MDC's main water program was also designed to measure progress in
sources precluded contamination of the raw water. reducing corrosion with the chosen solution. The
Instead, the lead contamination came from a MDC selected two treatment alternatives.
combination of corrosive water and extensive use of
lead in service lines and plumbing. The corrosive The first treatment, plan adding zinc
water was characterized as acidic and low 'in orthophosphate, was implemented from June
hardness and alkalinity. Table 0-1 presents water through December 1976. While reductions in lead
quality characteristics for the MDC's raw and contamination were attained, levels were still above
finished water. Prior to discovery of the lead problem, the MCL.
the finished water underwent only chlorination and
ammoniation treatments, as shown in the column for The second treatment plan data shown in the second
"raw water.· column in Table 0-1, which proved more successful,
consisted of adding sodium hydroxide to adjust the
The MDC, in conjunction with U.S. EPA, developed a pH. During 1979, pH levels dropped below 8.0 and
water monitoring program to ascertain the optimal then were increased to the target level of9.0. Lead

197
198 Upgrading Existing or Designing New Drinking Water Treatment Facilities

levels in the water rose and then dropped with the Total capital costs were $2,366,000 (1980 dollars) to
resumption of the higher pH level. The pH increase address the barium contamination. The costs to
also reduced copper contamination levels, but had address the barium contamination covered 10 capital
little effect on iron corrosion. components, including aerator, rapid mix tank,
flocculation basin, gravity filter, recarbonation
The treatment consisted of applying a dosage of 14 system, transfer pumps, potassium hydroxide
mglL of a 50 percent sodium hydroxide sol ution to system, gypsum system, polymer system, and
about 13.18 m3/sec (301 MGDl. This dosage, in 1981, appurtenances. Total construction costs were
cost $2.54/1,000 m 3 ($9.64/mil gal of treated water). $1,068,100. Annual operating and maintenance costs
Chemical costs for that year were $900,000 and were $155,900.
operation and maintenance costs were $161,000.
0.3 Reverse Osmosis: Sarasota, Florida
0.2 Coagulation to Control Barium:
This case involves a study of eight existing ground-
Illinois water supplies in Sarasota County, Florida. Water
This case involves a conventional treatment plant in treatment plants ranged in capacity up to 3.03
northeastern Illinois with high barium m 3 /day to 0.044 m 3/sec (from 800 to 1 MGD) and
concentrations in its source groundwater. Barium served five mobile home/trailer parks, one school, and
concentrations in the plant's influent ranged between two communities. The study was performed
0.4 and 8.5 mglL; the MCL for barium is 1 mglL. cooperatively by U.S. EPA's Division of Water
Contamination was found in small areas of the water Supply Research and the Sarasota County Board of
utility's watershed and the contaminated water was Health between January and June 1977.
drawn through Cambrian/Ordovician soils. No
barium contamination occurred in wells with sulfate The influent for the eight plants was naturally
levels greater than 50 mglL, because of the low contaminated with concentrations of radium-226
solubility of barium sulfate. ranging from 3.4 to 20.2 pCilL due to the presence of
phosphatic limestone. In addition, many supplies had
Plant operators conducted many jar tests to evaluate high levels of total dissolved solids.
alternative chemical additives to address the barium
contamination. Influent samples with 7.4- mglL Design parameters for the eight treatment plants are
levels of barium were used. The tested additives and summarized in Table D-2. These plants added reverse
their operational purposes include: osmosis units with either hollow-fiber or spiral-
wound membrane designs from six different
Alum as a coagulant-precipitant manufacturers. Their pretreatment processes
Ferrous sulfate as a coagulant-precipitant included cartridge filtration, pH adjustment, and ion
Calcium hydroxide as a precipitant-pH adjuster sequestration. Posttreatment processes included pH
Sulfuric acid as a pH adjuster-precipitant adjustment, degassification, and chlorination. Table
Hydrochloric acid as a pH adjuster 0-3 lists the pre- and posttreatment processes for
Sodium hydroxide as a pH adjuster each plant.
Potassium hydroxide as a pH adjuster
Calcium sulfate (gypsum) as a precipitant All eight plants achieved compliance with the 5 pCilL
Commercial gypsum as a precipitant standard for radium-226, with operating pressures
Sodium bisulfate as a precipitant ranging from 14.1 to 29.9 kglcm 2 (200 to 425 psi).
Anionic polymer as a flocculent-filter aid Finished or product water consisted of28 to 54
Diatomaceous earth as a filter precoat percent of the influent volume, and finished water
concentrations ranged from 0.14 to 2.0 pCilL. Waste
The laboratory jar tests determined that the most or reject brine water contained from 7.8 to 37.8 pCilL
effective additive was calcium sulfate. Optimum ofradium-226.
dosages ranged between 75 and 175 mglL with a pH
of 11.0. A 0.066-m 3/sec (l.5 MGD) pilot plant Table D-4 presents chemical analyses for the eight
provided precipitation, direct filtration, and polymer plants. The table contains analyses of the raw water
additions for a variety of barium concentrations, and product water collected after at least 1 hour of
which confirmed the laboratory approximation. operation. Analytical data were examined at several
different points during operation to assess the effect
At full-scale operation, the plant uses dosages of 100 of operating duration on removal efficiency.
mglLofgypsum and 0.25 mglL of polymer to achieve Although the eight systems attained 76 percent
a 91 percent reduction in barium from 6 mglL to 0.5 efficiency upon initiation of operation, two systems
mglL. The plant's filter operates at 1.0 Usedm 2 (l.5 examined attained peak operational efficiency after
GPMfft2) at a pH level of 11.0. about 5 minutes of operation.
Appendix D-Experience with Treatment Technologies for Inorganic Contaminants 199

Table 0-2. Design Criteria of Rev..... Osmosis Equipment


Product Water
Design Capacity Recovery
Reverse Mem-
Pumps Operaong Pressure Percenl
Popula· GPO OsmOSIS System brane
System tlon kLJday x 103 Manufacturer Type kw hp kPa psi Stages Design AClual
Venice 15,000 3,800 1,000 Polymetncs. Hollow 4,900 500 2,800 400 2 50 54
Santa Clara CA fiber·
SorrenlO 1,300 780 200 PermuOI. Spiral 1,470 150 2.900 425 75 39
Shores Paramus. NJ wound b
Spanish 800 265 70 Universal Oil Sprral 490 50 2,800 400 2 66 31
Lakes MHP Products (UOP), wound b
San 01990, CA
Bay Lakes 340 150 40 Pun'ication Hollow 290 30 2.800 400 2 55 NA
Estates Techniques, ribera
MHP Avon·by·
the· Sea NJ
Kings Gate 800 115 30 Punficatlon Hollow 390 40 2.800 400 2 70 NA
TP Techniques. ribera
Avon-by-
the-Sea NJ
Sarasota 135 19 5 Polymetncs. Hollow 50 5 2.800 400 50 50
Bay MHP Santa Clara CA fiber-
Bay Front 39 6 1.6 Continental Hollow 314 1,400 200 28 28
TP Waler tiber-
ConditiOning,
EI Paso, TX
NokomiS 800 3 0.8 BasiC Tech- Sprral 5 1/2 1,400 200 35 NA
school nolegles. West wound c
Palm Beach, FL

-DuPont Wilmlnglon, Delaware.


bUrTlV9fSaJ Oil Products (UOP). San Diego. Califo<nla
'Basic Technologies, Wes1 Palm Beach, Florida
NA • not available.
1,000 gallon - 3.78 m3
1 PSI • 0.0703 kglcm 2
Source: Sorg (1 980a).

The operating costs for these units ranged from $0.16 • Aerated lagoons and spray irrigation for brine
to $0.4l1m 3 ($0.60 to $1.54/1,000 gal) of treated water waste treatment
(see Table 0-51. These costs include chemicals,
electrical power, filter cartridge replacement, and The ion exchange units used three reaction basins,
labor. However, these data are not complete or each measuring 1.8 m (6 ft) in diameter and 3 m (10
comparable because they were derived from ft) in height. The standard operational height of the
interviews with operators and owners, and do not reaction basin is only 0.9 m (3 ft), with operational
necessarily use the same basis for each cost element. maximums of 1.5 m (5 ftl. According to system
design, one of the basins undergoes regeneration,
while the other two operate. The plant uses a 2.5-
0.4 Ion Exchange: McFarland, California minute EBCT; the treated water f10w rate was 15.77
This case involves a 0.044 m3/sec (1-MGDl plant with Usec (250 GPM) with surface loading rates of6.13
four ground-water wells contaminated with nitrates Usedm 2 (9.03 GPMfft2). Treated water was blended
from agricultural application of fertilizers and with raw water in a 7-to-3 ratio.
manure. Influent concentrations of nitrate ranged
from 6.8 to 22.1 mglL and averaged 16 mg/L. The regeneration process used a 6 percent sodium
chloride brine. Regeneration involved quick rinse,
To address this contamination problem, ion exchange slow rinse, and resin reclassification procedures that
units were selected for wellhead application because required 981 kg (2,1621b) of salt daily during periods
they are effective and easy to operate. The treatment of continuous operation. The process produced
process incl uded: saturated brine at a rate of2.27 Usec (36 GPM) and
diluted brine at 12 Usec (190.5 GPM). Brine was
• Anion exchange with A-101-D, Duolite resin discharged to a municipal wastewater treatment
• Sodium chloride regeneration with slow rinse plant, where it was diluted by the other waste
and declassification streams and then placed in aeration lagoons. This
200 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Table [)O3. Pre- and Posttreatment Reverse Osmolls Procesaes

Pretreatment Posttreatment
pH Adjustment Sequestering Degasslficaoon pH Adjustment Chlonn- Blending Raw and Relect Water
System Chemical Agent Filters I'Il1 or AerabOn Chemical ation Treated Water D,spcsaJ Method
Venice H SO 4 NllJ(P0 3)e 5 Yes a Yes a Creek
2
Sorren1O H SO 4 Na3(P03)e 10 No Na3C03 Yes 75% Bay
2
Shores reverse osmOSIs
Spanish H SO 4 NllJ(P0 3)e 25 Yes None Yes No Pond
2
Lakes MHP
Bay Lakes H SO 4 NllJ(P0 3)e 10 Yes None Yes 84% Storm sewer
2
Estates reverse osmOSIS
MHP
Kings Gate H SO 4 NllJ(P0 3)e 10 Yes None Yes No Ditch
2
TP
Sarasota HCI NllJ(P0 3)e 5 Yes None Yes No Subsurtace
BayMHP dralnlield
Bay Front None pcly-stabilizer 10 Yes None Yes 55% Subsurtace
TP A-5 reverse osmOSIS drainfield
NokomiS CO 2 None Yes None Yes No Subsurtace
school dralntield

• Blended with lime-softened water In approximate ratio of 2 parts lime-softened water \0 1 pan reverse osmosis water.
Source: Sorg (1980a).

aerated solution was spray- irrigated onto animal finished water averages 0.7 mglL of fluoride, with a
feed crops and cotton. maximum of 1.4 mglL.

The treated water had nitrate levels of 2 to 5 mglL. The alumina medium was placed in two vessels 3 m
The blended finished water nitrate levels ranged 00 ft) in diameter and height. The alumina takes up
from 6 to 10 mglL with a 7 mglL average. 1.5 m (5 ft) of vessel height, and expands about 50
percent of its original height during backwash
The total construction costs for the ion exchange operations. Approximately 15 cm (6 in) of basin
units were $354,638 in 1983 dollars. The operating freeboard is provided. The water's superficial
and maintenance costs totaled $0.03/m3 02.8 residency time is 5 minutes. The maximum
centsl1,OOO gal) of treated water. Table 0-6 shows the operational flow rate in the basins is 4.75 Useclm 2 (7
components of both types of costs. These costs include GPMlft 2 ), while the backwash rate is 7.47 Useclm 2
the annual loss of20 percent of the resin. 01 GPMlft 2 ).

For every 13.2 to 15.1 thousand cubic meters (3.5 to 4


0.5 Activated Alumina: Gila Bend, million gallons) of water treated, a 10-hour
regeneration cycle is required. Annual losses of
Arizona alumina due to regeneration range from 10 to 12
This case involves a ground-water supply with percent. Regeneration of the alumina medium is
undesirable levels of fluoride ranging from 4 to 6 accomplished with a 1 percent solution of sodium
mglL, with a 5 mglL average. The plant, with an hydroxide. The 1 percent solution flows through the
average capacity of37.8 Usec (600 GPM) and a basins at a maximum rate of 1.70 Usec/m 2 (2.5
maximum capacity of 56.77 Usec (900 GPM), was GPMlft 2), with a detention time of24 minutes. The
equipped with the following elements to manage the regeneration process uses 757 L (200 gal) of sodium
fluoride contamination: hydroxide solution/lb fluoride removed.

• Activated alumina: Alcoa activated alumina, The caustic water from the regeneration process
grade F-1, -28 to +48 mesh requires a 0.04 percent solution of sulfuric acid for
• Caustic regeneration neutralization. The acid solution is derived by
• Acid neutralization diluting a bulk 93 percent acid solution. (The
• Evaporation pond for regenerant waste neutralization process flow rate is 4.75 Useclm 2 [7
treatment GPM/ft 2 j at most.) The goal of the neutralization
process is to produce acceptable pH levels from 6.5 to
The treatment process produces 90 percent treated 8.5 for disposal. The backwash and neutralization
finished water and 10 percent waste water. The rinse water wastes are discharged to the sewer. The
Appendix D-Experience with Treatment Technologies for Inorganic Contaminants 201
202 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Table 0-5. Reverse Osmosis Treatment System Coats regenerant waste is discharged to a lined evaporation
Caprtal EStImated OperaMg pond which is 73 m (240 ftl by 134 m (440 ftl by 2.7 m
Year Costl Costb (9 ftl.
Con- $11,000 $11,000
System structed gal $/kL gal
The construction costs for the unit totaled $285,000
Venice 1975 1,000 0,16 0.60
in 1978 dollars, and included treatment facility, well,
Sorrento Shores 1975 250 0.29 1.08 1,892,500-L (500,OOO-gall steel tank, evaporation
Spanish Lakes 1974 150 0.41 1.54 pond, booster pumps, standby generator, and
Bay Lakes Estates 1973 25 0.33 1.06 chlorination facilities. The plant's operating costs
MHP were 7.1 to 7.4 cents/m 3 (27 to 28 cents/l,OOO gall of
Kings Gate MPH 1974 30 c c treated water. These costs included salaries, power,
Sarasota Bay MPH 1975 14 0.32 1.22 chemicals, and media replacement.
Bay Front TP 1976 11.2 0.28 1.06
Nokomis school 1976 5.5 c c

'CapItal COS1S include reverse osmQSJS uOilS and pumps. pre- and
post- treatment eQuipment, filters. flushing eQuipment, booster pumps,
and shelter.
bQperaling costs Include chemicals. power (0.4/kwhr). filled cartridge
replacement. and labor. Excluded were amortization costs and
membrane replacement. Blending was not taken Into account.
clnsuffiClent data.
1,000 gallons = 3.78 mJ
Source: Sorg (1 980a).

Table 0-6. Ion EXChange Cost Components for McFar1and,


California ($1983)

Construction Costs Components in (expressed inS):


Ion exchange unit vessels 111,741
OnSl1e Construction 81 ,154
Resin 56.610
Engo-ring 46.388
Brine tank 18.700
Other 40.045
Total 354,638

Operauon and Maintenance Costs Components (expressed ,n


centsl1 ,000 gallons):
Salt 3.4
Resin replacement 3.2
Power 2.2
Normal operanng and m<lJntenance 1.9
Ope<aling labor 1.3
MlsceI1eneous 0.8
Total 12.8
1,000 gallons = 3.78 mJ
Appendix E-Summary of Corrosion Indices

Index Equation Parameters Meaning

Langelier LSI = pH· pH, Total alkalinity, mgIL as GaGO, LSI > 0 = Water 's super-
Saturaoon Index Galcium, mglL as GaGO, saturated; tends to precipitate
(LSI) Hardness.mgIL as GaGO, GaGO,
Total dissolved solids. mgtL
Onsile pH LSI = 0 = Water 's saturated (in
Onslte temperature eqUilibnum); GaGO, scale 's neither
dissolved nor depOSited

LSI < 0 = Water IS under'


saturated: tends to dissolve Solid
GaGO,

Aggressive Index AI = pH • log IIA)(H)/ Total alkalinity. mglL as 'GaGO, AI < 10 = Very aggressive
(AI) (lor use w,th Hardness.mglL as GaGO,
asbestos cemenll Onslte pH AI 10 ·12 = Moderately
aggressive

AI > 12 = Nonaggresslve

Ryznar Stability RSI = 2pH, . pH Total alkalinity, mgIL as GaGO, RSI < 6,5 = Water IS super-
Index IRSI) GaiClum, mglL as GaGO, saturated: tends to precipitate
Hardness. mgtL as GaGO, GaGO,
Total dissolved sol,ds, mglL
Onslte pH 6.5 < RSI < 7.0 = Waler IS
Onslte temperature saturated (In eqUilibnum): CaGO,
scale 's neither dissolved nor
depoSited

RSI > 7,0 = Water's under·


saturated: tends to dissolve solid
CaGO,

Riddick's G0 2 , mgIL CI = 0·5 Scale forming


Corrosion Index
IGI)
:,~ [C0 2 + 1/2 tardness - AI) Hardness.mglL as CaCO,
Alkalinity, mglL as GaGO,
6-25 Noncorrosive
26·50 Moderately corrOSive
CI " mgIL 51 -75 Gorroslve
+ CI • 2~ x
N,mgIL
DO.mgIL
76·100 Very corrosive
101 + Extremely corroSive
SaturatIOn DO· Ivalue lor

[~ ~ oxygen saturaOon), mgtL

[ DO • 2
Sa + DO )

Driving Force CalCium, mgIL as GaGO, DFI > 1 = Water super·


Index (DFI) Ga' • (ppm) x GO, = (ppm) GO, ~ = mglL as CaGO, saturated: tends to precipitate
GaGO, K so x 10'0
Kso- = solubility product of DFI = 1 = Water saturated I,n
GaGO, equllibnuml: GaGO, scale IS
neither dissolved nor deposited

DFI < 1 = Water under-


saturated: tends to dissolve
CaGO,

·DO = dissolved oxygen


Source: U.S. EPA (1984).

203
Part II

Water Treatment Technologies


for Small Communities

The information in Part II is from Environmental Pollu-


tion Control Alternatives: Drink ing Water Treatment for
Small Communities, prepared by James E. Smith, Jr. of
the U.S. Environmental Protection Agency, Center for
Environmental Research Information, April 1990.

205
Acknowledgments

Appreciation is expressed to those Management of the preparation of this


individuals who assisted in providing document was provided by James E.
technical direction and resource Sm~h, Jr. of EPA's Center for Environ-
materials and reviewing drafts of this mental Research Information. Techni-
pUblication: cal wr~ing, ed~orial work, produaion,
and design was provided by Jenn~er
John R. Trax, National Rural Water Helmick. lynn Knight, Susan
Association, Washington, DC Richmond, and Karen Ellzey of
Fred Reiff, Pan American Hea~h Eastern Research Group, Inc.,
Organization, Washington, DC Arlington, MA.

Deborah Brink and Elizabeth


Kawczynski, American Waler Works
Association Research Foundation,
Denver, CO
John Dyksen, Department of Water
Supply, Ridgewood, New Jersey

U.S. Environmental Protection Agency


OIIice of Drinking Water, Washington,
DC:

Steven Clark
Peter L Cook
Jane Ephremides
Ken Hay
AW. Marx
Marc J. Parrolla
Peter Shanaghan

U.S. Environmental Protection Agency


Risk Reduaion Engineering
laboratory, Cincir,:1ati, OH:

Jon Bender
Wa~er Feige
Kim Fox
Benjamin lykins. Jr.
Donald J. Reasoner
Thomas Sorg
Alan A. Stevens
James Westrick

U.S. Environmental Protection Agency


OIIice of Technology Transfer and
Regulatory Support, Washington, DC:

Ronnie levin

206
Introduction

Small drinking water systems lace a Chapter 3 provides an overview of


difficult challenge: to provide a sale, how to select drinking water treatment
sufficient supply 01 water at a technologies and discusses special
reasonable cost. Our growing aware- management issues for small systems.
ness of the biological and chemical Chapter. 4 through 7 describe tech-
contaminants that can affect the safety nologies that can enable small sys-
01 drinking water has led to the need tems to meetlederal drinking water
for more frequent monitoring and regulations covering filtration, disinfec-
reporting and, in some cases, addition- tion, removal of organic and inorganic
al or upgraded treatment by water sup- contaminants, and corrosion contro/.
pliers. These chapters describe established
This document provides information technologies, which are commonly
lor small system owners, operators, used in the water treatment industry.
managers, and local decision makers. They also describe several emerging
such as town officials, regarding drink- technologies suitable for small sys-
ing water treatment requirements and tems. These technologies have not
the treatment technologies suitable for been widely used, but have proven
small systems. It is not intended to be effective on the pilot scale and are
a comprehensive manual for water emerging as viable full-scale options
treatment and protection of public for treating water supplies.
water supplies from sources of con-
Chapter 8 lists organizations, publica-
tamination. Rather, it is designed to tions, and other resources that can as-
give an overview 01 the problems a sist small systems in their efforts to
small system may face, treatment op- provide safe drinking water to con-
tions that are available to solve sumers.
specific problems, and resources that
can provide further information and For the purpose of this document.
assistance. small systems are defined as systems
that serve 25 to 1,000 people, or that
Chapter 1 discusses why we need have a flow 01 9,500 to 380.000 liters
drinking water treatment and gives an (2,500 to 100,000 gallons) per day.
overview of drinking water treatment They include small community sys-
processes. tems as well as noncommunity sys-
Chapter 2 provides a summary 01 ex- tems, such as campgrounds and
isting and new federal drinking water restaurants.
regulations and explains how these
regulalions affllCl small systems.

207
1. Drinking Water Treatment: An Overview

Why Do W. Need Drinking Water


Tl'Ntment?
For thousands of years, people have
treated water intended for drinking to
remove particles of solid matter, re-
duce hea~h risks, and improve aes-
thetic qual~ies such as appearance,
odor, color, and taste. As early as
2000 B.C., medical lore of India ad-
vised, "Impure water should be puri-
fied by being boiled over a fire, or
being heated in the sun, or by dip-
ping a heated iron into it, or it may
be purified by filtration through sand
and coarse gravel and then allowed
to cool."
Ancient medical lore of India ad-
Early in the nineteenth century, scien- vised that Impure water should be
tists began to recognize that spec~ic purified by heating or lIIterlng
diseases could be transmitted by through sand and coarse gravel.
water. Since that discovery, treatment
to eliminate disease-<:ausing microor-
ganisms has dramatically reduced the Although water treatment processes
incidence of waterborne diseases have greatly improved the quality and
(diseases transmitted through water) safety of drinking water in the United
such as typhoid, choler'll, and hepa- States, there are still over 89,000
t~is in the United States. For example, cases each year of waterborne
in 1900, 36 out of every 100,000 diseases caused by microorganisms-
people died each year from typhoid bacteria. viruses. protozoa, helminths,
fever; today there are almost no and fungi (Figure 1-1).' Water can be-
cas'ts of waterborne typhoid fever in come contaminated with these or-
the United States. ganisms through surface runoff (water

FUNGI W

VIRUSES
~
... : .• ,..
.•..

•.... "

.
...
.
PROTOWA
(micrOKOpic OGe-
ullC'd uim~ls)

' .

HELMINTHS
(puuitic worms)
fl· .'

BJ\C1CRIA
.

Flgur.1-1. Disease-causing microorganisms that might be found In


watar supplies.

'Source: EPA estimate. Federal Register. June 19. 1989 (54 FR 27522).

208
Drinking Water Treatment: An Overview 209

that travels over the ground during Table 1-1. Waterborne Diseases
storms), which often contains animal
wastes; failures in septic or sewer
systems; and sewage treatment plant Source of
effluents (outflow). Microbiological con- Waterborne Causative Organism
Disease Organism In Water Symptom
tamination occurs most often in sur-
face water. but it can also occur in
ground water, usually due to improp- Gastroenteritis Rotavirus Human feces Acute diarrhea
erly placed or sealed wells. Contamina- or vomiting
tion can also occur after water leaves Salmonella Animal or Acute diarrhea
the treatment plant, through cross con- (bacterium) human feces and vomiting
nection (connection between safe
Enteropathogenic Human feces Acute diarrhea
drinking water and a source of con- E. Coli or vomiting
tamination), backflow in a water supply
line. or regrowth of microorganisms in Typhoid Salmonella Human feces Inllamed intes-
typhosa tine, enlarged
the distribution system.
(bacterium) spleen, high
temperature-
Table ,-, lists some of the diseases
sometimes fatal
caused by microorganisms found in
water supplies. The protozoan Giardia Dysentery Shigella Human feces Diarrhea - rarely
(bacterium) fatal
lamblia is now the most commonly
identified organism associated with Cholera Vibflo comma Human feces Vomiting, severe
waterborne disease in this country. (bacterium) diarrhea, rapid
This organism causes giardiasis, dehydration.
mineral 10ss-
which usually involves diarrhea.
high monality
nausea, and dehydration that can be
severe and can in some cases last for InfectIOUs hepatitis Hepatitis A Human feces, Yellowed skin,
months. Over 20.000 water-related (vorus) shelllish grown enlarged liver,
cases of this disease have been ir, .'olluted abdominal pain -
walers low mortality. lasts
reported in the last 20 years, 2 with
up to 4 months
probably many more cases going un-
reported. Another protozoan disease, Amoebic dysentery Enramoeba Human feces Mild diarrhea.
histolytica chronic dysentery
cryptosporidiosis, is caused by Cryp-
(protozoan)
tosporidium. a cyst-torming organism
similar to Giardia. Other common GiardiaSIS GIardia lamblia Animal or Diarrhea, cramps,
waterborne diseases include viral (protozoan) human feces nausea, and
hepatitis. gastroenteritis, and legoonel- general
losis (Legionnaires' Disease). weakness-not
fatal. lasts' week
Chemical contaminants, both natural to 30 weeks
and synthetic, might also be present in Cryptospolldiosls Cryptospofldium Animal or Diarrhea, stomach
water supplies. Contamination (protozoan) human feces pain-lasts an
problems in ground water (used by 85 average of 5 days
percent of small systems) are frequent-
ly chemical in nature. Common sour- Source: Adapted from American Water Works ASSOCIation, Introduction to
ces of chemical contamination include Water Treatment: Principles and Practices 01 Warer Supply Operations.
minerals dissolved from the rocks that Denver, CO, 1984.
form the earth's crust; pesticides and
herbicides used in agriculture; leaking
underground storage tanks; industrial promote corrosion of materiais in the when chlorine reacts with natural
effluents; seepage from septic tanks. distribution system. possibly introduc- organic materials and other chemical
sewage treatment plants. and landfills; ing lead and other materials into the contaminants-into the drinking water.
and any other improper disposal of drinking water. The water treatment (" should be noted, however, that
chemicals in or on the ground. In process might also introduce while the potential for chlorination
some systems. the water quality can trihalomethanes- chemicals formed by-product formation cannot be

2 Source: Dave Ryan. "Water Treatment to Combat Illness: EPA Journal, Decembet 1987.
210 Upgrading Existing or Designing New Drinking Water Treatment Facilities

neglected. adequate disinfection is of mun~ies add fluoride to drinking water ground-water supplies from con-
paramount importance to protect the to improve dental health. taminants reduces the extent of treat-
public from microbiological contamina- ment needed to protect public health.
tion of drinking water.) To protect the public from the hea~h The Wellhead Protection (WHP)
risks of drinking water ccntaminants, Program for public water supplies is
Drinking water can be treated for the U.S. Environmental Protection an example of a protection program.
reasons other than to reduce hea~h Agency (EPA) has issued regulations Publications on the WHP Program are
risks from microorganisms and chemi- covering the qual~y and treatment of listed in Chapter 8, Resources.
cals. A system might treat water to im- drinking water. These regulations are
prove its color, odor. or taste even if ~ discussed in Chapter 2. How Is Drinking Water Treated?
is safe to drink. For example, some
EPA has also issued guidance for Table 1-2 shows the types and goals
systems remove iron and mangan-
protecting public drinking water from of water treatment processes typically
ese, which can stain laundry and
sources of contamination. Protecting used by small systems (including
plumbing fixtures. Some com-
preliminary treatment and main water

Table 1-2. Water Treatment Processes

Process/Step Purpose

Pr.llmlnary Tr.atm.nt Prac....a·


Screening Removes large debris (leaves. sticks, fish) thaI can foul or damage
plant equipment
Chemical pretreatment Conditions the water for removal of algae and other aquatic nuisances

Presedimentation Removes gravel, sand, silt, and other gritty material


Microstraining Removes algae, aquatic plants, and small debris

Main Tr.atm.nt Prac....a


Chemical feed and rapid mix Adds chemIcals (coagulants, pH adjusters, etc.) to water
Coagulation/flocculation Converts nonseltleable to settleable particles
Sedimentation Removes settleable particles
Softening Removes hardness-causing chemicals from water
Fillration Removes particles of solid matter which can include biological
contamination and turbidity
Disinfection Kills disease-causing microorganisms

Adsorption using granular Removes radon and many organic chemicals such
activated carbon (GAC) as pesticides, solvents, and trihalomethanes.
Aeration Removes volatile organic chemicals (VOCs), radon, HzS, and other
dissolved gases; oxidizes iron and manganese
Corrosion control Prevents scaling and corrosion
Reverse osmosis. electrodialysis Removes nearly all inorganic contaminants
Ion exchange Removes some inorganic contaminants, including hardness,causing chemicals
Activated alumina Removes some inorganic contaminants
Oxidation filtration Removes some inorganic contaminants (e.g .. iron, manganese, radium)

'Generally used lor treating surface water supplies.


Source: Adapted from American Waler Works Association, In/roduction /0 Water Treatment, Vol .. 2, 1984.
Drinking Water Treatment: An Overview 211

Filtration might produce potentially harmful by-


products, such as trihalomethanes.
Fi~ration is the process of removing
Small systems can successfully use
particles of solid mailer from water,
ozone and u~raviolet radiation as
usually by passing the water through
primary disinfectants, but chlorine or
sand or other porous materials. Fi"ra-
an appropriate substitute must also be
tion helps to control biological oon·
used as a secondary disinfectant to
tamination and turbidity. (Turbidity is a
prevent regrowth of microorganisms in
measure 01 the cloudiness of water
the distribution system. Disinfection is
caused by the presence of suspended
discussed in Chapter 5.
mailer. Turbidity can she"er harmful
microorganisms and reduce disinfec- Treatment of Organic Contaminants
tion effectiveness.) Filtration tech-
nologies oommonly used in small Many synthetic organic chemicals
systems include slow sand filtration, (SaCs). manmade compounds that
Chemicals from leaking under· diatomaceous earth fi"ration, and oontain carbon, have been detected in
ground storage tanks mIght water supplies in the United States.
package filtration systems. Filtration is
migrate to ground water and/or
discussed in Chapter 4. Some of these, such as the solvent
surface water.
trichloroethylene, are volatile organic
Disinfection chemicals (VaCs). VOCs easily be-
treatment processes). This document oome gases and can be inhaled in
Disinfection is a chemical and/or physi-
showers or baths or while washing
discusses the water treatment proces- cal process that kills disease~ausing
dishes. They can also be absorbed
ses designed to protect the oonsumer organisms. For the past several
from waterborne disease. Chapter 3 through the skin.
decades, chlorine (as a solid, liquid, or
discusses how to select appropriate gas) has been the disinfectant of Water supplies beoome contaminated
processes and technologies lor a par- choice in the United States because it by organic compounds from sources
ticular water system, and Chapters 4 is effective and inexpensive and can such as improperly disposed wastes,
through 7 discuss treatment tech- provide a disinfectant residual in the leaking gasoline storage tanks, pes-
nologies in more detail. distribution system. However, under ticide use, and industrial effluents.
certain circumstances, chlorination Technologies that can be used effec-
tively by small systems to remove
these contaminants include activated
carbon and aeration. These tech-
nologies are discussed in Chapter 6.

Treatment of Inorganic
Contaminants

The inorganic oontaminants in water


supplies consist mainly of naturally oc-
curring elements in the ground. such
as arsenic, barium. fluoride. sulfate.'
radon. radium, and selenium. In-
dustrial sources can contribute metal-
lic substances to surface waters.
Nitrate, an inorganic substance fre-
quently found in ground-water sup-
plies, is found predominantly in
agricu"ural areas due to the applica-
tion of fertilizers. High levels of total
dissolved solids (IDS) might. in some
instances, require removal to produce
Runoff from agricultural areas can IntrOduce microlliological can·
tamlnants. pesticides, and nitrates Into drinking water sources. a potable supply.

Inorganic chemicals might also be


present in drinking water due to oor-
rosion. Corrosion is the deterioration
212 Upgrading Existing or Designing New Drinking Water Treatment Facilities

or destruction of components of the


water distribution and plumbing sys-
tems by chemical or physical action,
resu~ing in the release of metal and
nonmetal substances into the water.
The metals of greatest hea~h concern
are lead and cadmium; zinc, copper,
and iron are also by-products of cor-
rosion. Asbestos can be released by
corrosion of asbestos-eement pipe.
Corrosion reduces the useful I~e of the
water distribution and plumbing sys-
tems. It can also promote microor-
ganism growth, resulting in dis-
agreeable tastes, odors, and slimes.

Treatment alternatives for inorganic


contaminants include removal tech-
niques and corrosion controls.
Removal technologies - coagulationl
fi~ration, reverse osmosis, ion ex-
change, and activated alumina - treat
source water that is contaminated wilh
metals or radioactive substances
(such as radium). Aeration effectively
strips radon gas from source waters.
Corrosion controls reduce the
presence of corrosion by-products
such as lead at the point of use (such
as the consumer's tap). Treatment
technologies for inorganic con-
taminants are discussed in Chapter 7.
2. New and Proposed Drinking Water Treatment
Regulations: An Overview

In 1974, CongrllSS passed the Safe to fi~er waler from surface waler
Drinking Water Act (SDWA). setting up sources. ~ must also develop pro-
a regulatory program among local, cedures for states to deterr.line
state, and federal agencies to help en- which systems have to fi~er.
sure the provision of sale drinking
• DisinfectIOn. EPA must develop
water in the Un~ed States.
rules requiring all public water sup-
Under the SDWA, the federal govern- plies to disinfect their water.
ment develops national drinking water • UN of l6ad materials. The use of
regulations to protect public hea~h solder or flux containing more than
and we~are. The states are expected 0.2 percent lead, or pipes and pipe
to administer and enforce these regula- fittings containing more than 8 per-
tions for public water systems (sys- cent lead, is prohibited in public
tems that either have 15 or more water supply syslems. Public
service connections or regularly serve not~icalion is required where there
an average of 25 or more people daily is lead in construction materials of
for at least 60 days each year). Public the public water supply system, or
water systems must provide water where the water is suHiciently cor-
trealment, ensure proper drinking rosive to cause leaching of lead
water qual~y through mon~oring, and from the distribution system/lines.
provide public notification of con-
tamination problems. • Wellh8ad protection. The 1986
SDWA amendments require all
Congress sign~icantly expanded and states to develop Wellhead Protec-
strengthened the SDWA in 1986. The tion Programs. These programs are
1986 amendments include provisions designed to protect public water
on the following: supplies from sources of con·
tamination.
• Maximum conUlml".nt Level..
The Safe Drinking Waler Act re- The impact of these new regulations
quired EPA 10 set numerical stan- on small systems will generally con-
dards, referred to as Maximum cern some fundamental aspects of
Contaminant Levels (MCLs), or water treatment. Many syslems will be
treatment technique requirements required to improve trealment for
for contaminants in public water removal of microorganisms (through
supplies. The 1986 amendments the addition of fi~ration and/or disinfec-
established a strict schedule for tion processes). Most small systems
EPA to set MCLs or treatment re- do not face contamination by organic
quirements for previously unregu- and inorganic chemicals at levels ex-
lated contaminants. ceeding the MCLs, and therefore will
not need to install treatment for
• Monitoring. EPA must issue
removal of these chemicals. 3 Small
regulations requiring monitoring of
systems will be required to conduct
all regulated and certain unregu-
periodic monitoring, however, to docu-
lated contaminants, depending on
ment whether chemical contaminants
the number of people served by the
are present in their water supplies.
system, the source of lhe water
Future regulations covering radioac-
supply, and the contaminants likely
tive substances (particularly radon)
to be found.
and disinfection by-products could
• FlltrIItlon. EPA must set cr~eria also have a sign~icant impact.
under which systems are obligated

3 G. Wade Miller, John E. Cromwell, III, Frederick A. Marrocco. "The Role of the
States in Solving the Small System Dilemma: JAWWA, August 1988. pp 31-37.

213
214 Upgrading Existing or Designing New Drinking Water Treatment Facilities

The rest of this chapter explains the


major provisions of EPA's new and Community, Nontransient Noncommunity,
proposed drinking water regulations as
they apply to small systems. In addi- and Transient Noncommunity Systems
tion to th9 f9d9ral rsgulations dis-
cuss9d h9r9, th9 wat9r suppli9r should
check with th9 stat9 ag9ncy respon- The drinking water regulations distinguish between community water
sible for drinking wat9r (S99 Chapter 8, systems (CWS), nontransient noncommunity water systems (NNWS),
Resourc9sj to find out about stat9 and transient noncommunity water systems (TNWS).
r9gulations that apply to drinking wat9r
• A community syst9m is a public water system that serves at least
tr9atment facilities.
15 service connections used by year-round residents, or regularly
serves at least 25 year-round residents. Community systems include
Compliance SChedules
mobile home courts and homeowner associations.
Most of the regulations contain com-
pliance schedules that affect large sys- • A nontranslent noncommunIty system regularly serves at least 25
tems initially and small water systems of the same people over six months of the year. Examples are
2 to 4 years later. This means that schools and factories.
small systems have additional time to • TransIent noncommunIty syst9ms, such as restaurants, gas sta-
plan for their specific compliance tions, and campgrounds, serve intermittent users.
requirements.
The regulations governing each 01 these systems are slightly different.
The SDWA recognizes that meeting This is because certain contaminants cause health problems only when
drinking water standards might place a consumed on a regular basis over a long period of time, and are there-
large burden on small systems. The lore of greater concern in systems that regularly serve the same people
law therefore provides for variances, than in those that serve transient users.
allowing small systems to meet a less
stringent standard if an organic or inor- Only the MCLs lor turbidity, nitrate, and bacteria apply to transient non-
ganic contaminant cannot be removed community systems. (The new surface water treatment requirements, ex-
due to the quality of the raw water or plained below in this chapter, will replace the currently MCL for turbidity
other good reasons, as long as the lor TNWS). Most MCLs are set at levels designed to prevent health el-
less stringent standard poses no un- lects caused by long-term consumption of drinking water from a system.
reasonable health risk. A small system However, the presence of nitrate, bacteria, and turbidity indicate the
may also be granted a temporary ex- potential of the water to cause illness even from short-term consump-
emption if economic conditions tion, so MCLS lor these contaminants apply to transient noncommunity
prevent the system Irom making as well as other systems.
necessary corrections, provided no un·
reasonable risk to public health
results. No exemptions are allowed for (MCLGs), which are the maximum The SDWA amendments direct EPA to
disinfection of surface supplies or the levels of contaminants at which no establish MCls for 83 specrtic contami-
colilorm rtole tor all public waler known or anticipated adverse health ef- nants and to develop a list of contami-
supplies. Monitoring requirements may lects will occur. MCls are set as close nants every 3 years to be considered
also be reduced in some cases if the to the MCLG as is leasible. In se"ing for regulation. The Agency must prom-
public water supply has a program an MCl, EPA takes into account the ulgate at least 25 MCls from each of
under an EPA-approved state technical feasibility 01 control systems these lists starting in t 991. EPA may
Wellhead Protection Program. for the contaminant, the analytical set treatment technology requirements
detection limits, and the economic im- instead of MCLs when it is difficult or
Maximum Contaminant Levels pact of regulating the contaminant. An expensive for water suppliers to test
A Maximum Contaminant Level, or MCL is usually expressed in mil- for specific contaminants.
MCL, is the highest allowable con- ligrams per liter (mgll), which is
centration of a contaminant in drinking equivalent to parts per million (ppm) Whenever EPA establishes an MCL
water. In developing drinking water lor water quality analysis. lor a particular contaminant, the
regulations, EPA establishes Maxi- Agency must also identrty the Best
mum Contaminant Level Goals Available Technology (BAT)' lor

• EPA delermines lhe BAT based on high removal efficiency of contaminant concentration; general geographic applicability; service
Iile; compatIbility with other water treatment processes; and ability to achieve compliance at a reasonable cost.
New and Proposed Drinking Water Treatment Regulations: An Overview 215

removing that contaminant. To comply Table 2-1. Volatile Organic Chemicals: Final MCLs (in mgJL)
w~h the MCL. public water systems
are required to use the BAT or an a~er­ Chemlce' Fine' Mel
native treatment technology deter-
mined by the state to be at least as
Trichloroethylene 0.005
effective as the SAT.
Carbon tetrachloride 0.005
MCls lor Volatile Organic Vinyl chloride 0.002
Compounds
1.2-Dichloroethane 0.005
EPA has issued final MCls lor 8 Benzene 0.005
volatile organic compounds. These
para-Dichlorobenzene 0.075
are shown in Table 2·1. Nontransient
noncommunity water systems as well 1,1-Dichloroethylene 0.007
as commun~ systems must meet t, 1,1-Trichloroethane 0.2
these MCls.
Source: Federal Register, July 8, 1987 (52 FR 25690).
MCls lor Inorganic and Synthetic
Organic Compounds
The final MCl for Iluoride has been
For surface water (or ground water future. The anticipated MCL for radon,
set at 4.0 mgiL (see Federal Register
under the direct influence 01 surface a naturally occurring radionuclide,
April 2. 1986 - 41 FR 11396).5 EPA
water)6 EPA has set treatment require- might affect many small public water
has proposed MCLs for lead and cop-
ments instead 01 MCLs lor Giardia. supplies. Table 2-6 shows the current
per (Table 2-2), and lor 8 other inor-
viruses, heterotrophic bacteria, MCls for radiological contaminants.
ganic compounds and 30 synthetic
Legionella, and turbidity, These re- Note that only systems serving popula-
organic chemicals (Table 2-3). Table
quirements are explained below under tions greater than 100,000 people are
2-3 shows proposed MCLs along with
Surface Water Treatment Require- required to meet MCls lor manmade
maximum allowable levels under
ments. EPA intends to issue disinfec- radionuclides.
regulations currently in effect.
tion regulations for ground water,
MCls lor the inorganic chemicals and including regulations to control the MCLs lor Dlslnlectants and
level of viruses, Legionella, and Dlslnlectlon By-Products
synthetic organic chemicals in Table
2-4 will be proposed in the near luture. heterotrophic bacteria, at a later date. In 1979, EPA established an MCl for
total trihalomethanes-chloroform,
MClslor Microbiological MCls lor Radlonuclldl bromoform, bromodichloromethane,
Contaminants Contaminants and dibromochloromethane-of 0.1
EPA has sellinal MCLs for total New MCls for radionuclides (radioac- milligram per liter. This MCl applies
colrtorms (Table 2-5). Colijorms are tive elements) will be proposed in the only to systems serving populations
usually present in water contaminated
w~h human and animalleces and are
Table 2-2. Lead and Copper: Proposed MCls
often associated with disease out-
breaks. A~hough total coliforms in-
(Measured as water leaves the treatment planl or enters the dist"butlon
clude microorganisms that do not
system)
usually cause disease themselves,
their presence in drinking water might Lead 0005 mglL
mean that disease-<:ausing organisms Copper 13 mglL
are also present. All public water sys-
tems must meet the MCl for total Source: Federal Regisrer, August 18. 1988 (53 FR 31571).
colilorms; mon~oring requirements are
discussed below.

5 See Chapter 8, Resources, for information about the Federal Regisrer.


e Any water beneath the surface of the ground with (i) significant occurrence of inseC1S or other microorganisms, algae, or large-
diameter pathogens such as Giardia lamblia. or (iil significant and relatively rapid shifts in water characteristics such as turbidity,
temperature, conductivity, or pH which closely correlate to climatological or surface water conditions. Direct influence must be deter-
mined tor individual sources in accordance with criteria established by the state.
216 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Table 2-3. Proposed MCLs for Synthetic Organic Chemicals and Inorganic Chemicals

Exlatlng Proposed
Contaminant NPDWR a (mg/L) MCL(mg/L)

Acrylamide ~
Alachlor 0,002
Aldicarb 0.01
Aldicarb sulfoxide 0.01
Aldicarb sulfone 0.04
Atrazine 0.003
Carbofuran 0.04
Chlordane 0.002
cis-' ,2-Dichloroelhylene 0.07
Dibromochloropropane (DBCP) 0.0002
1.2·Dichloropropane 0.005
o-Dichlorobenzene 0.6
2,4-0 0.1 0.07
Elhylenedibromide (EDB) 0.00005
Epichlorphydrin Tr
Ethylbenzene 07
Heptachlor 0.0004
Heptachlor epoxide 0.0002
Undane 0.00<1 0.0002
Methoxychlor 0.1 0.4
Monochlorobenzene 0.1
PCBs (as decachlorobiphenyl) 0.005
Pentachlorophenol 0.2
Styrene' 0.005/0.1
Tetrachloroethylene 0005
Toluene 2.0
2.4.5-TP (Silvex) 0.01 0.05
Toxaphene 0.005 0.005
trans-l,2-Dichloroelhylene 0.1
Xylenes (Iotal) 10.0
Asbestos 7 F/L d
Barium 1.0 5.0
Cadmium 0.010 0.005
Chromium 0.05 0.1
Mercury 0.002 0.002
Nitrate (as nitrogen) 10.0 10.0
Nilrile (as nilrogen) 1.0
Selenium 0.01 0.05

"NPDWR ~ National Primary Drinking Waler Regulations.


bTl = Treatment Technique.
'EPA proposes MCLs of 0.1 mgiL based on a group C carcinogen classification and 0.005 mg/L based on a B2
classification.
d7 million fibers/liter (only fibers longer than 10 m).

Source: Federal Regisrer, May 22, '989 (54 FR 22064).


New and Proposed Drinking Water Treatment Regulations: An Overview 217

greater than 10,000 people. EPA Table 2-4. Inorganic and Synthetic Organic Chemicals
plans to propose new rules for ground- to be Regulated
water disinfection and for disinfection
by-products; small systems might be Arsenic· Sulfale
included in these new requirements. Methylene chloride Hexachlorocyclopentadiene
Disinfectants and disinfection by- Antimony Nickel
products that might be included in Endrin" Thallium
Dalapon Beryllium
these rules are shown in Table 2-7.
Diquat Cyanide
Endothall 1.1.2-Trichloroethane
Monitoring Glyphosate Vydate
New monitoring requirements for Andipates Simazine
chemical contaminants under the 2.3.7.8- TCDD (Dioxin) PAHs
Trichlorobenzene Atrazine
1986 SDWA amendments could have
a major impact on small systems.
These new requirements are ex- 'Current MCl is 0.05 mg/L.
plained below. "Current MCl is 0.0002 mg/L.
Source: U.S. Environmental Protection Agency. Fact Sheet. "D"nking Water
Volatile Organic Chemicals Regulations under 1986 Amendments 10 SDWA.- February 1989.
All systems must monitor for the regu-
lated VOCs in Table 2-1 and the un-
regulated VOCs in Table 2-8. The
required monitoring is shown in Table Table 2·5. Maximum Contaminant Level for Total Collforms
2-9. Small systems serving fewer than
3,300 people must complete initial
• Compliance is based on presence/absence of total coli forms in sample. rather
monitoring for these VOCs by Decem- than on an eSlimate of colilorm density.
ber 31, 1991. Nontransient noncom-
munity systems, as well as community • MCl lor systems analyzing at least 40 samples/month: no more than 5 a per-
systems, must meet the requirements cent of the monthly samples may be total coliform-poslllve.
for VOCs.
• MCl for systems analyzing fewer than 40 samples/month: no more than 1
sample/month may be total colilorm-poSilive.
Fluoride
Monitoring requirements for fluoride • A public water system must demonstrale compliance With the Mel for total
are shown in Table 2-10. colitorms each month IllS required to monilor

• MCl Violations must be reported to the state no later than the end 01 the next
Other Inorganic and Synthetic business day after the system learns 01 the Violation.
Organic ChemicalS
EPA has proposed monitoring require- Source: U.S. Env:~nmental Protection Agency. Fact Sheet. -O"nklng Water
ments for 38 regulated chemicals and Regulations under 1986 Amendments to SOWA.- February 1989.
111 unregulated contaminants (inor-
ganic and synthetic organic chemicals).
In addition, EPA will propose monitor-
ing requirements for chemicals in ments for total coliforms, effective fecal coliforms. Alternatively, the
Table 2-4. December 31, 1990. Tables 2-11 system may test for the bacterium
and 2-12, respectively, show the Escherichia coli instead of fecal
Radlonuclides minimum number of routine and coliforms. The requirements for
Currently, community systems must repeat samples required. monitoring fecal colitorms and E.
monitor for natural radiological chemi- coli are effective December 31,
• Fecal collforms! Escherichia coli.
cals every 4 years. EPA will be propos- 1990.
The presence of fecal coliforms in
ing new monitoring requirements for drinking water is strong evidence of • Heterotrophic bacteria. Hetero-
radionuclides, including radium-226, recent sewage contamination, and trophic bacteria can interlere with
radium-228, uranium (natural), and indicates that an urgent public total coliform analysis. Effective
radon. health problem probably exists. December 31, 1990, public water
Therefore, EPA requires that public systems must follow specific proce-
Microbiological Contaminants
water systems analyze each dures to minimize this interlerence.
• Total collforms. In June 1989, sample that is positive lor total
EPA issued new monitoring require- coliforms to determine if it contains
218 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Table 2-6. Maximum Contaminant Levels for Radiological Chemicals

Natural Radlonuclldea Manmade Radlonuclldes

Combined
Radium
Gross 226 & Gross Slrontium
Alpha 228 Beta Tritium 90

Community 15 pCi/L 5 pCilL 50 pCi/L b 20,000 pCi/Lb 8 pCi/Lb


Systems

Noncommunity State State State Slate State


Systems option option option option option

'Picocuries per liter (pCi/L) is a measure of the concentration of a radioaclive subslance. A level of 1 pC ilL means that
approximately 2 atoms of the radio nuclide per minute are disintegrating in every liter 01 waler.
bApplies only to surface water systems serving populations greater than 100,000 people.

Source: Adapted from National Rural Waler Association, Water System Decision Makers: An Introduction to Water System
Operation and Maintenance, Duncan, OK, t 988.

Table 2·7. Disinfectants and Disinfectant By.Products

Disinfectants and Residuals

Chlorine, hypochlorous acid. and hypochlorite ion


Chlorine dioxide, chlorite, and chlorate
Chloramines and ammonia
Ozone

Disinfectant By·Products

Trihalomethanes: chloroform. bromoform. bromodichloromelhane. dibromochloromethane


Haloacetonilriles: bromochloroacetonllrile, dibromochloroacetonilrile, dichlorobromoacelonitrile, trlchloroacetonitrile
Haloacetic acids: mono-. di-. and trl·chloroacetlc acids: mono- and dibromoacelic acids
Haloketones' 1.1 ·dichloropropanone and 1.1. l-tri-chloropropanone
Other chloral hydrate, chloropiCrin

Cyanogen chloride
Chlorophenols (2-chlorophenol. 2.4-dichlorophenol. 2.4.6-trlchlorophenol)
N·organochloramlnes

MXI3-chloro-4-(dichloromethyl)-5-hydroxy-2(5H)-furanone)1

Ozone by-products

Source: U.S. EnVironmental Proleclion Agency, Fact Sheet. -Drinking Water Regulations under 1986 Amendments to
SDWA: February 1989.
New and Proposed Drinking Water Treatment Regulations: An Overview 219

Table 2-8. Monitoring for Unregulated VOCs

Required for All Systems:

Chloroform 1,2-Dichloropropane
Bromodichloromethane 1,1,2,2-Tetrachloroethane
Chlorodibromomelhane Ethylbenzene

Bromoform 1,3-Dichloropropane
trans-1,2-Dichloroethylene Styrene
Chlorobenzenll Chloromethane
m-Dichlorobenzene Bromomethane
Dichloromethanll 1,2,3-Trichloropropane
cis-1,2-Dichloroethylene 1,1,1,2-Tetrachloroethane

o-Dichlorobenzene Chloroethane
Dibromomethane 1,1,2-Trichloroethane
1,1-Dichloropropane 2,2-Dichloropropane
Tetrachloroethylene o-Chlorotolullne
Toluene p-Chlorotoluene
p-Xylene Bromobenzene
o-Xylene 1.3-Dichloropropane
m-Xylene Ethylene dibromide
1,1-Dichloroethane 1,2-Dibromo-3-chloropropane

Required for Vulnerable Systems· Only:

1,2-Dibromo-3-chloropropane (DBCP)
Ethyilinedibromide (EDB)

At Each State's Discretion:

1,2,4-Trimethylbenzene 1,3,5-Trimethylbenzene
1,2,4-Trichlorobenzene p-Isopropyltoluene
1,2,3- Trichlorobenzene Isopropylbenzene

n-Propylbenzene tert-Butylbenzene
n-Butylbenzene sec-Butylbenzene
Naphthalenll Fluorotrichloromethane

Hexachlorobutadiene Dichlorodifluoromethane
Bromochloromethane

°A system's vulnerability to contamination is assessed by evaluating factors such as geological conditions, use patterns
(e g., pestiCides), type of source, location of waste disposal facilities. historical monitoring record, and nature of the dis-
t"bulion system.

Source: Adapted from U.S. Environmental Prolectlon Agency, Fact Sheet, ·D"nking Water Regulalions under 1986
Amendments to SDWA: February 1989.
220 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Table 2-9. Compliance Monitoring for Regulated and Unregulated Volatile Organic Chemicals

Initial monitoring: All community and nontransient noncommunity systems must monitor each source at least
once within 4 years.

• Surface waters: four quarterly samples.

• Ground water: four quarterly samples; state can exempt systems from subsequent monitoring if no
VOCs are detected in the first sample.

• Composite samples of up to five sources are allowed.

Repeat monitoring: varies from quarterly to once every 5 years. The frequency is based on whether VOCs
are detected in the first round of monitoring and whether the system is vulnerable to contamination.

Source: Adapted trom U.S. Environmental Protection Agency, Fact Sheet. "Drinking Water Regulations under 19S6 Amend-
ments to SDWA: February '9S9.

A water system manager should ob-


• Sanitary surveys. Periodic • Samples must be collected in
tain the following information from a
sanitary surveys are required for all proper containers and preserved as
laboratory he or she plans to use:
systems that collect fewer than five necessary. Discuss sample collec-
coliform samples per month. (A tion procedures for the contaminant
• References of similar work
sanitary survey is a comprehensive with the laboratory in advance.
review of a system's operations, in- (See Appendix A for bacteriological • Copy of applicable accreditations
cluding watershed control, the disin- sample collection procedures.) or certifications
fection system, raw water quality,
• Sample chain of custody must be • Information about availability and
and monitoring, to determine
maintained (to ensure that some- cost of sample containers, preser-
whether operational requirements
one is always accountable for the vatives, and shipping containers
are being met.) The schedule for
sample).
conducting sanitary surveys is • Commitment or estimate of project
shown in Table 2-13. • Analysis must be performed within turnaround
specified holding times (the period
Laboratory Analysis and Sampling • Definition of analytical methods,
of time between sample collection
Requirements detection limits, and cost of analysis
and analysis).
To meet the monitoring requirements • Type of quality assurance data that
• ApproVed analytical procedures
for some contaminants, small systems will be reported
must be used.
will need the services of a commercial
laboratory. For compliance monitoring • State-approved reporting forms, ij
• Adequate quality assurance data
purposes, analyses must be per- applicable
must be generated wrthin the
tormed in an EPA or state-approved laboratory. • Fees
laboratory. Contact the Sate Drinking
Water Hotline (see Chapter 8, Resour-
ces) tor assistance in locating a cer-
titied drinking water laboratory in your Table 2·10. Monitoring Requirements for Fluoride, , '
area. (These laboratories must suc-
cessfully analyze performance evalua-
tion samples within limits set by EPA.) Surface waters: 1 sample each year
Ground waters: 1 sample every 3 years
To ensure proper sampling and
analysis?' Minimum repeat: 1 sample every 10 years

Source: Federal Register, April 2, 1986 (41 FR 11396).

7 From Metcalf and Eddy, A Guide to Water Supply Management in the 19905, Wakefield, MA. November 19S9.
New and Proposed Drinking Water Treatment Regulations: An Overview 221

Typical costs of laboratory analyses


Table 2-11. Total Coliform Sampling Requirements According to
Population Served are shown in Table 2-14.

Contact your state drinking water


Minimum Number
agency for additional information
of Routine Samples
about requirements for sample collec-
Population Served Per Month'
tion and analysis and reporting.

25 to 1.000 b Surface Water Treatment


1,001 to 2,500 2 Requirements

2,501 to 3,300 3 EPA has set treatment requirements to


control microbiological contaminants
al n lieu oflhe frequency specified in this table. a noncommunity waler system in public water systems using surface
using only ground water (excepl ground water under the direct influence of sur- water sources (and ground-water sour-
face water) and serving 1,000 persons or fewer may monitor al a lesser fre- ces under the direct influence of sur-
quency specilied by the state (in wriling) until a sanitary survey is conducted face water). These requirements,
and the slale reviews Ihe results. Thereafter. such systems must monitor in
effective December 31. 1990, include
each calendar quarter during which the system provides water to lhe public. un-
less the state delermines (in writing) thaI some other frequency is more ap- the following:
propriale. Beginning June 29. 1994, such systems must monitor at least once
every year • Treatment must remove or inac-
tivate at least 99.9 percent of
A noncommunlly waler system using surface water, or ground waler under the Giardia lamblia cysts and 99.99 per-
direct influence of surface water, regardless 01 the number 01 persons served, cent of viruses.
must monitor at the same frequency as a like-sized community waler system,
i.e., the frequency specified in the table. A noncommunity water system using • All systems must disinfect. and also
ground water (which is not under the direCI influence of surface water) and serv- might be required to fi~er rt certain
ing more lhan 1.000 persons during any month must monitor at the same fre- source water qualrty crrteria and s~e­
quency as a like-sized community waler system. i.e., Ihe frequency specified in specrtic crrtena are not mel.
lhe lable. except that the state may reduce lhe monitoring frequency (in writing)
for any monlh the syslem serves 1,000 persons or fewer. However, in no case • The regulations set criteria for
may the state reduce the sampling frequency to less than once every year. determining rt treatment, including
turbidity removal and disinfection
blncludes public waler systems that have al least 15 service connections, but requirements. is adequate for fil-
serve fewer than 25 persons.
tered systems.

Source: Federal Register. June 29, 1989 (54 FR 27545). • All systems must be operated by
qualified operators as determined
by the states.

Systems using surface water must


Table 2·12. Monitoring Requirements Following a Total Coliform- make certain reports to the state
Positive Routine Sample documenting compliance with treat-
ment and monitoring requirements.
Number of Routine Number of Repeal Number of Routine
Samples/Month Samplea' Samples Next Month b Detailed guidance on surface water
treatment requirements is provided in
1/monlh or fewer 4 5/monlh EPA's Guidance Manual for Com-
21monlh 3 5/monlh pliance with the Filtration and Disinfec-
tion Requirements for Public Water
3/monlh 3 5/monlh
Systems Using Surface Water
Sources .
• Number of repeat samples in lhe same monlh for each tOlal coliform-posillve
routine sample.
b Excepl where state has mvalidated the origmal routine sample. or where the
slate substitutes an onslle evaluation of the problem, or where the state waives
the requirement on a case-by-case basis. See 40 CFR 141.21a (b) (5) for more
details.

Source: Federal Register. June 29, 1989 (54 FR 27546).


222 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Table 2-13. Sanitary Survey Frequency for Public Water Systems Collecting Fewer than Five Samples/Montha

Frequency 01
Initial Survey Subsequent
System Type Completed by Surveys

Communlly water system June 29, 1994 Every 5 years

Noncommunity water system June 29, 1999 Every 5 years·

'Annual onslte inspection of the system's watershed control program and reliability of disinfection practice is also required by
40 CFR 141. 71(b) lor systems using unfiltered surface water or ground water under the direct influence of surface water.
The annual onSlte Inspection, however, is not equivalent to the sanitary survey. Thus, compliance with 40 CFR 141.71(b)
alone does not constitute compliance with the sanitary survey requiremenlS 01 this coliform rule (141.21a(d)), but a sanitary
survey during a year can Substllute for the annual onsite inspection for that year.

bFor a noncommuOlly water system thai uses only protected and disinfected ground water, the sanitary survey may be
repeated every 10 years, instead of every 5 years.

Source: Federal Register, June 29, 1989 (54 FR 27546).

Table 2-14. ApproxImate Commer·


clal Laboratory Costs Per Sample
Analysis ($1989)

Turbidity $ 20
Coliform Baclena $ 20
Copper $ 20

Lead $ 20

Radium 226/228 $120


8 VOCS (Table 2-10) $200
Table 2-8 Contaminants $500

Source: Adapted from Metcalf and


Eddy. A Guide to Warer Supply
Management In the 19905,
WakelJeld, MA. November 1989.
Systems using surlace water are required to Illter unless stringent
criteria are met.
3. Solutions to Drinking Water Treatment Problems:
An Overview

This chapter presents an overview of improgerly constructed or located


the technologies that a small system wells. Corrosion control is a concern
should consider for meeting its treat- for systems using both surface and
ment needs. In addition, it discusses ground-water supplies.
administration and other issues that
can be important for small systems, Figure 3-1 presents an overview of
including financial and capital improve- steps that a small system can follow to
ments, cooperative arrangements, determine treatment needs. Tables 3-1
operator capabilities, and selection of and 3-2 present the contaminants like-
a consu~ing engineer or equipment ly to be found in surface water and
vendor. Appendix B presents a check- ground water, and the most suitable
list of factors that can affect water treatment technologies for each.
treatment system performance. While
this list is not all-inclusive, it may help The treatment options available to
a water system operator or manager small systems for @ration, disinfec-
in determining improvements that may tion, organic and inorganic con-
be needed. taminant removal, and corrosion
control are also listed in Table 3-3.
The treatment needs of a water sys- This table presents the major ad-
tem are likely to differ depending on vantages and disadvantages of each
whether the system uses a ground- technology. Costs are shown in Table
watllf or surface water source. Com- 3-4. Chapters 4 through 7 contain
mon surface water contaminants more detailed information about each
include turbidity, microbiological con- of these technologies, including their
taminants (Giardia, viruses, and bac- effectiveness in removing specific con-
teria), and low levels of a large taminants in water.
number of organic chemicals. Ground-
water contaminants include naturally Questions to Consider In
occurring inorganic contaminants Choosing Treatment Technologies
(e.g., arsenic, fluoride, radium, radon) When selecting among the different
and nitrate, and a number of specdic treatment options, the water system
organic chemicals (e.g., trichloro- manager must consider a number of
ethylene) that sometimes occur in rela- factors: regulatory requirements, char-
8
tively high concentrations. Bacteria acteristics-of the raw water, configura-
and viruses can also contaminate rela- tion of any existing system, cost,
tively shallow ground water (for ex- operating requirements, availability of
ample, from sewage overflow or nontreatment aiternatives, com-
seepage into wells and springs. or patibility of the processes currently
from surface runoff). Giardia cysts are being used or to be used, waste
less likely to be found in ground water. management, and future needs of the
but they have contaminated ground- service area. Each of these factors is
water supplies where sewage or con- discussed below.
taminated surface water entered

8 Thomas J. Sorgo 'Process selection for Small Drinking Water Supplies," Proceedings
of the Twenty- Third Annual Public Water Supply Engineers Conference: New Direc-
tions lor Water Supply Design and Operation. University of Illinois. April 21-23. 1981

8Gunther F. Craun, "Review of the Causes of Waterborne Disease Outbreaks."


Surveillance and Investigation 01 Waterborne Disease Outbreaks. Health Effects
Research Llboratory. U.S. Environmental Protection Agency, November 1989.

223
224 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Systems using surface water sour-


ces must disinfect, and must filter
unless stringent requirements are
met; systems using ground-water
sources will have to meet future
ground-water disinfection rules
(see Chapter 2)
Identify and evaluate
.'.: .. treatment technologies avail-
able to control contaminants
If filtration and/or
of concern (see Tables 3-1
disinfection must
through 3-4 and Chapters 4
be added or
through 7)
upgraded

Contact vendors or consulting engineer for information about


treatment technology selection, costs, and design (see Appendix C)
and/or contact organizations that assist small systems
(see Chapter 8, Resources)

Figure 3-1. Steps to determine treatment needs.


Solutions to Drinking Water Treatment Problems: An Overview 225

Table 3·1. Common Problems and Suitable Treatment Technologies: Ground Water a

MIcro-
biological Fluoride Barium Nitrate Radium Radon Arsenic Selenium Organics

Chlorination •
Ozonation •
Ullraviolet
radiation •
Aeration • •
Ion exchange • • • • •
AClivat",d alumina • • •
Coagulation/Filtration • •
Membranes
(reverse osmosis
and electrodialysis)
• • • • • •
Granular activated
carbon (GAC) • •
Point-of-use/
POint-of-entry systems Might be suitable for some very small systems to remove organic or inorganic contaminants.

Package Planls Package plants might be available to solve specific inorganic or organic contamination problems.

'In general. "." indicates the principal function of the treatment technology listed Many of these technologies.
however. have secondary effects in addition to those shown here. (For example. ozone can remove some organic chemicals
as well as provide disinfection.)

What Are the Requirements for tions), and American Water Works As- protection is an appropriale nontreat-
Drinking Water Supplied by the sociation water quality goals. ment alternative.
System?
Federal and state drinking water Are Nontreatment Alternatives As another option, a small system can
regulations are the most important fac- Available? consider cooperating with other sys-
lors 10 consider in developing a water Small water systems might not always tems, such as by buying trealed water
system's treatment goals. The supplier have the resources to install a com- from a larger Ulility (see Multicom-
may also choose to consider con- plete treatment system to solve a con- munity Cooperative Arrangements
sumer preferences and non mandatory tamination problem. In such situations, below).
guidelines developed by regulatory however. a small system might be
agencies and professional organiza- able to find a new water source. For What Are the Characteristics of the
lions. such as EPA's Secondary Maxi- Raw Water?
example. a well can be located in an
mum Contaminant Levels (federally area distant from the source of To determine treatmenl needs. the sys-
nonenforceable goals for controlling contamination. tem manager must know the quality of
contaminants that affect the aesthetic the source water-what biological and
qualities of drinking water), Health Ad- The development of a new well, chemical contaminants are in the
visories (guidance values developed however, is only part of the solution. water and at what concentrations they
by EPA to address immediate or emer- The area around the well must be are present. Knowledge of other water
gency concerns associated with acci- managed to protect it trom future con- characteristics, such as pH, tempera-
dents, spills, or newly detected taminaiion. Establishing a wellhead ture, alkalinity, and calcium and mag-
drinking water contamination situa- nesium content is useful because of
226 Upgrading Existing or Designing New Drinking Water Treatment Facilities

a
Table 3-2. Common Problems and Suitable Treatment Technologies: Surface Water

Turbidity Microbiological Control Corrosion Control Organics

Chlorination •
Ozonation •
Ultraviolet radiation •
Package plants b • •
Slow sand filtration • •
Diatomaceous
earth filtration • •
Ultrafiltration
(membrane filtration)c
• •
Cartridge fiIlratlon c •

Aeration •
Granular activated carbon (GAC)

pH control •
Corrosion inhibitors

Point-of-usel
Point-of-entry systems d • • •
'In general_ ••- indicates the principal function of the trealmenttechnology listed Many of these technologies. however.
have secondary ellects in addition to those shown here. (For example, ozone can remove some organic chemicals as
well as provide disinfection)

bWhlle package plants are most Widely used 10 remove turbidity. color, and microbiological contaminants, package plants
are also available lhat can remove organic andlor inorganic conlamlnanlS.

CEmerging technology.

dMight be suitable lor some very small systems thai cannot inslall centraltrealmen!. POlnl-of·use/Point-of-enlry systems
use a variely of trealment processes. including reverse osmoSIS. Ion exchange. and aClivated carbon.

their impact on aesthetics and the ef· new technology. In addition. an exist- pipe. copper, polyvinyl chloride, gal.
ficiency of treatment processes. ing system's ability to blend treated vanized, lead) can also be an impor-
water with raw water can be important. tant factor in selecting a treatment
What Is the Configuration of the A system might be able to economize option. For example, corrosion would
Existing System? with an expensive technology by treat- not be a great concern in systems
The configuration of the existing sys- ing only part of the total flow. and still using polyvinyl chloride (PVC) pipes.
tem can be an important consideration meet regulatory requirements that limit The length of the distribution system
in selecting a treatment option. For ex- the concentration of a contaminant in and how quickly water moves through
ample, rt a supplier is considering add- finished water. rt can affect requirements for secon-
ing a new treatment technology, he or dary disinfection (to prevent regrowth
she must know rt the existing system The method of water distribution and of microorganisms in the distribution
Its composition (e.g .. asbestos-cement system).
is compatible wilh or adaptable to the
Solutions to Drinking Water Treatment Problems: An Overview 227

What Are the Costs of the How Compatible Are the Processes What Are the Future Needs of the
Treatment Options? Used? Service Area?
The total costs of treatment include one- To achieve overall treatment goals, all The future of the service and supply area
time capital costs and annual operating the treatment processes must be com- is another important factor in selecting a
and maintenance costs. Each treatment patible. For example, a lower pH is treatment technology. The supplier can
technology has a different mix of capi- desirable for efficient chlorine disinfec- evaluate future demands using popula-
tal and operating and maintenance tion; however, lower pH increases cor- tion and economic forecasts of the ser-
costs. Technologies with high capital rosion in the water distribution system. vice area. Present and potential water
costs often have lower operating and Therefore, a system might maintain a supplies should also be examined to
maintenance costs (and those with lower lower pH but use a corrosion inhibitor determine their vulnerability to natural
capital costs often have higher operating (described in Chapter 7) to minimize and manmade contamination.
and maintenance costs). Thus, small sys- corrosion (or the system might elevate
tems that cannot alford appropriate capi- the pH before distribution). All the ele- Speclallssu8S for Small Systems
tal equipment can become saddled with ments of treatment should be chosen FinanciaVCapital improvement
higher operating and maintenance costs. so that they interact as efficiently and
Financing water system improvements
effectively as possible.
can be an obstacle for a small system.
What Are the Treatment
User charges must be high enough to
Technology'. Operating In addition, using one treatment tech-
Requirements? cover the actual costs of water treat-
nology to meet more than one
ment, analytical work, and distribution.
The most important operalional con- regulatory requirement reduces costs
Because costs are spread over fewer
sideration is the consistency of the raw and operating complexity. For ex-
people, rate increases have a greater
water. The less consistent the raw water ample, a system might use reverse os·
impact on the individual customer than
quality, the greater the need for monitor- mosis (described in Chapter 7) when
those for large systems.
ing, and the greater the operating com- both organic and inorganic con-
plexity of most systems. Thus, a less taminants are present in raw water.
Most small systems are also at a dis-
consistent influent requires a higher Ozone (described in Chapter 5) can
advantage when they anempt to raise
level of operator training and anention, remove organic chemicals as well as
funds in the local and national capital
and might require greater instrumenta- provide primary disinfection. Packed
markets, since their credit base,
tion, controls, and automation. tower aeration (described in Chapter
market recognition, and financing ex-
6) can remove both volatile organIC
pertise are usually limited. They might
Other important operating considera- chemicals and radon in ground water.
be able to obtain financial assistance
tions include: It also removes carbon dioxide, there-
through state and federal loan and
by raising the pH to a more desirable
grant programs. Many states currently
• Energy requirements level for corrosion control. (Aeration,
have drinking water financial assis-
however, can increase dissolved
• Chemical availability, consumption tance programs. Some states assist
oxygen levels, which can contribute to
rate, and storage small systems in gaining access to
corrosion. )
capital through low-interest loans from
• Instrumentation and automation state revolVIng !oan funds, state bond
What Waste Management Issues
• Preventive maintenance pools, and state-funded bond in-
Are Involved?
surance.
• Noise Waste management can be a significant
issue for water treatment systems. Most Sources of federal grants and loans
• Aesthetics treatment processes concentrate con- for small systems include the Farmers
• Backup/redundant systems taminants into a residual stream (bone Home Administration (FmHA) and a
or sludge) that requires proper manage- proposed federal grants program that
• Requirements for a startup phase ment. For example, removal of radon blends federal grant money with state
before full removal capacity is with granular activated carbon can bond money to provide Iow-intQrQst
achieved produce a low-level radioactive waste. loans to small water systems.
• Cleaning and backwashing require- Water treatment systems must follow
mllnts federal and state regulatIOns covering Other financing options for small sys-
the management of wastes. In some tems include fQderal revenue sharing
• Distribution system cases, this can sign~icantly increase dis- and revenue bonds (for municipal sys-
posal costs for the treatment system. lems), loans through the United States
• Staffing needs
Small Business Administration (SBA),
• Opllrator training requirements and use of tax uempt industrial

• Process monitoring requirements


228 Upgrading Existing or Designing New Drinking Water Treatment Facilities

a
Table 3·3. Overview of Water Treatment Technologles

Technological
Options to Meet
Treatment Regulatory Stage of
Requirements Requirements Acceptability Comments

Filtration of surface Slow sand filtration Established Operationally simple: low


water supplies to operating cost; requires
control turbidity relatively low turbidity source
and microbial waler
conlaminatlon
Package plant filtration Established Compact; variety of process
combinations available
Ultrafiltration
(Membrane filtration) Emerging Experimental. expensive

Ca, .ridge fiIlration Emerging Experimental. expensive

Disinfection Chlorine Established Most widely used melhod:


concerns aboul health effects
of by-products

Ozone Established Very effective but requires a


secondary disinfectant. usually
some lorm of chlorine

Ultraviolet radiation Established Simple, no established harmful


by-producls, bul requires
secondary disinfectant,
usually some form 01 chlorine

Organic contamination Granular activated carbon Best Available Highly effective: potential waste
conlrol Technology (BAT) disposal issues: expensive

Packed column aeralion Best Available Highly effective for volatile com-
Technology (BAT) pounds; potential air emissions
issues

Diffused aeration Established Variable removal effectiveness

Multiple tray aeralion Established Variable removal effectiveness

Higee aeration Experimental Compact, high energy require-


ments: potential air emissions
issues
Mechanical aeratIon Experimental Mostly lor wastewater treatment:
high energy requiremenls.
easy to operate

Catenary grid Experimenlal Performance data scarce:


pOlenlial air emissions issues

(conrinued on next page)

revenue bonds by a private contractor water supply authority. Multicommunity


• CentraliZing functions. A group of
supplying service to a municipality. cooperative arrangements can im-
small systems working together
prove cost eHectiveness, upgrade
can centralize functions such as
Multicommunity Cooperative water quality, and resu~ in more eHi-
purchasing, maintenance,
Arrangements (Regionalizatlon) cient operation and management.
laboratory services, engineering
In some cases, a small community
services, and billing. Sev'lral small
can share resources with other small A wide range of cooperative ap-
systems together might be able to
communities or a larger community proaches is available to the small sys-
tem, including:
aHord resources, such as highly
through a cooperative or a regional
Solutions to Drinking Water Treatment Problems: An Overview 229

a
Table 3-3. Overview of Water Treatment Technologies (continued)

Technolog leal
Options to Meet
Treatment Regulatory Stage of
Requirements Requirements Acceptability Comments

Inorganic contam- Membranes Established Highly effective: expensive;


ination control (Reverse osmosis potential waste disposal issues
and eleclrodialysis)

Ion exchange Established Highly effective; expensive;


potential waste disposal issues

AClivated alumina Established Highly effective; expensive;


potential waste disposal issues

Coagulation/Filtration Es tablished May be difficult for very


smail systems

Aeration Established Preferred technology for radon


removal

Granular Activated Carbon Established Highly effective for radon


removal; potential waste
disposal issues

Corrosion pH control Established Potential to conflict with other


controls treatments

Corrosion inhibitors Established Variable effectiveness


and emerging depending on type of inhibitor

'A variety of package plants are available thaI can perform one or several treatment functions (i.e., filtration, organic con-
taminant conlrol. etc.).

Source: U.S Environmental Protection Agency, Office of Drinking Water and Cenler lor Environmental Research Information,
Technologies lor Upgrading Existing or Designing New Drinking Water Treatment Facilities, March 1990. EPA 625/4-69-023.

skilled personnel, on a part-time effectiveness when distribution ces and/or physically connect sys·
basis. lines become too long. tems, so that one or two facilities
supply water for the entire district.
• Physically interconnecting exist· • Creating a satellite utility. A satel-
By forming a water district, privately
ing systems. Two or more small lite utility taps into the resources of
owned systems become eligible for
systems can be connected, or a an existing larger facility without
public grants and loans.
small system can join a larger sys- being physically connected to, or
tem, to achieve the economies of owned by, the larger facility. • Creating county or state utilities.
scale available to large systems. Resources provided by the larger A county or state government can
This approach might not be utility can include technical, opera- create a board to construct, main·
feasible in some situations, tional, or managerial assistance; tain, and operate a water supply
however, such as in locations wholesale treated water: or opera- within its district, Construction
where supplies are isolated from tion and maintenance responsibility. and/or upgrading of facilities may
each other by long distances or be financed through bonds or
• Creating water districts. Water
rugged terrain. It is also important property assessments,
districts are formed by county offi-
to w9lgh potential disadvantages
cials and provide for the public
such as loss of local autonomy, Operator Capabilities
ownership of the utilities. The
complexities of ensuring equity in
utilities in a district combine resour- The level of understanding and techni-
each community, and loss of cost
cal ability of small systems operators
230 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Table 3·4. Estimated Costs of Drinking Water Treatment Technologies for a 100,000 GPO Planta ($1989)

Capital Total Cost b Per


Technology Cost AnnualO&M 1,000 Gallons

Package Planl Flltrallon

Coagulalion/Fillralion with lube selliers $176,000 $11,000 $1.73

Pressure depth clarifier/Pressure filter $206,000 $10,400 $1.90

Pressure depth clarifier/ $246,000 $16,300 $2.47


Pressure filter with GAC adsorber

Other Fillrallon

Dialomaceous earth vacuum filter $103,000 $1t,100 $1.27

Diatomaceous earth pressure filler $106,000 $10,600 $1.26

Slow sand filter: covered $580,000 $ 7,700 $4.15

Slow sand filler: uncovered $335,000 $ 7,100 $2.55

Inorganic Conlamlnanl Control

High pressure reverse osmosis $275,000 $4t,300 $4.03

Low pressure reverse osmosis $275,000 $29,800 $3.40

Cation exchange $151,000 $ 8,500 $1.44

Anion exchange $115,000 $10,300 $1.46

Activated alumina $104,000 $14.600 $1.47

Organic Conlaminanl Conlrol

GAG in pressure $175,000 $14,400 $1.92


vessel ( 6-mo carbon replacemenl)

$ 9.800 $1.67
(12-mo carbon replacement)

Packed lower aerator $ 45.100 $ 2,900 $0.45

(continued on next page)

is crucial to the success of the Safe


• Knowledge of all aspects of the dis- • Skill to maintain drinking water
Drinking Water Act (SDWA), The
tribution system (including main- quality (including water treatment
operator's basic knowlQdge. skills, and
tenance) where necessary. plus state and
training might include the following '0:
fQderally required sampling
• Knowledge of the source water
routines)
• Sufficient training to protect public supply (including pump operation)
health

10 From National Rural Water Association, Water System Decision Makers: An Introduction to Water System Operation and Main-
tenance, Duncan, OK, 1988.
Solutions to Drinking Water Treatment Problems: An Overview 231

Table 3·4. Estimated Costs of Drinking Water Treatment Technologies for a 100,000 GPO Planta (continued)

Capital Total Cost b Per


Technology Cost AnnualO&M 1,000 Gallons

Olslnfecllon

Gas feed chlorination $ 10,465 $ 3,520 $0.26

Hypochlorite solution $ 4,080 $ 5,558 $0.33

Pellet feed chlorinalors $ 1,670 $ 4,010 $0.23

Ullraviolet IightC(57,600 GPO) $ 25,990 $ 2.090 $0.49


c
Ozonation-high pressure $ 39,270 $ 5.074 $0.53

Gallons x 3.785 • Iilers.


Sources: G.S. Logsdon, T.J. Sorg, and R.M. Clark, Cost and Capability of Technologies for Small Systems, Drinking Water
Research Division. Risk Reduction Engineering Laboratory, EPA, Cincinnati, OH, May 1989.
R.C. Gumerman et aI., Estimation of Small System Water Treatment Costs, Final Report, CulplWesner/Culp, Santa Ana,
CA, November 1984.
U.S. Environmental Protection Agency, Office of Drinking Water, Microorganism Removal for Small Water Systems,
Washinglon. DC. June 1983. EPA 570/9-83-012.

'Construclion coSIS generally include manufactured equipmenl, concrete. sleel. labor. pipes and valves, electrical equipment
and instrumentation, housing, site evacualion. some other site work, general conlractor's overhead and profit, engineering
costs, financial and administrative costs, and interest costs during construclion. Construction costs do not include land
costs, legal fees, interface piping, roads, and certain other site work. O&M costs generally include annual energy. labor, and
chemical costs. Construction COSIS can vary depending on specific data characteristics. O&M costs can vary, up to plus or
minus 100 percent lor some technologies. depending on such variables as feed water characteristics. flow rate, and chemi·
cal dosage requirements.

bCosts include cap;lal costs annualized at 10 percent interest over 20 years plus annual O&M costs. Average flow assumed
at 50 percent of design flow.

CCosts for ullraviolet and ozone disinfection reflect those lor primary disinfection only. A secondary disinfectant is necessary
to maintain a reSIdual in the distribution system. The costs 01 secondary disinfection are not included in the table.

cause of economic constraints. In addi· ing firm. Small systems can also use
• Knowledge of sources of con-
tion, many small system operators the National Rural Water Association
tamination and methods used to
have multiple duties, such as maintain- for technical assistance (see Chapter
manage these sources.
ing the grounds or performing other re- 8, Resources).
• Knowledge and understanding of lated public works duties, and might
energy sources not have the opportunity to specialize A "circuit rider· approach, in which ser-
and develop expertise in drinking vice is provided to several systems
• Understanding of emergency proce- that cannot individually afford a trained
water treatment.
dures
operator, can also be used. The circuit
• Knowledge of state and federal Operator capability can also limit the rider anends to a number of trealment
regulations technology options available to a systems, and his or her salary is
small system: a technology that works shared among them. The circuit rider
• Recordkeeping skills well in a large city might require more can directly operate the plants or
operator training than the smal! sys- provide technical assistance to in-
• A willingness 10 participate in con-
tem can obtain. dividual plant operators through on·
tinuing education programs
the-job training and supervision.
Small systems might be able to obtain
Without it properly trained operator,
qualijied plant operators by contract- Another source of training is the treat-
system operation and water quality will
ing the services of personnel from a ment equipment manufacturer. When
suffer. The small system might have
larger neighboring utility, government treatment equipment is purchased,
difficulty anracting skilled staff be-
agency, service company, or consult- vendors should supply startup assis-
232 Upgrading Existing or Designing New Drinking Water Treatment Facilities

tance and training, as well as detailed office to determine whether POUIPOE


operation and maintenance manuals. devices are appropriate.

Additional training resources for small


systems are listed in Chapter 8.

SBlectlng 8 Consuftfng Englneer/


Equipment Vendor
In some cases, a small community
might need to use the services of a
consuhing engineer or equipment ven-
dor to design a treatment system.
Consuhants should have proven ex·
perience in solving problems for small
systems. Appendix B provides some
guidelines for selecting a consultant.

UsIng 8 Polnt-of-UselPolnt-of-Entry
(POUIPOE) System
A number of point-<lf·use (POUl and
point-<lf·entry (POEl systems are avail·
able from a large number of manufac·
turers. Types of systems include those
using reverse osmosis, activated
alumina, and ion exchange. In cerlain
situations, POUIPOE devices can be a
cost·effective solution when a very
small community cannot afford central
treatment for a contaminant. such as
an organic chemical or fluoride. For ex·
ample, with state approval, several
small communities (25 to 200 people)
in Arizona installed home systems
using activated alumina to remove
fluoride. A manufacturing/engineering
company on contract with one com·
munity provides and maintains all the
U
syslems

In addition to home devices, some


very small systems (such as trailer
parks) might be able to install a treat·
ment system at the point of entry and
blend resulting treated waters with
water not treated with the POE device.

A public water supplier must monitor


and ensure the quality of water treat·
ment, whether it provides central treat·
ment or decentralized treatment
through POUIPOE devices. The sup-
plier should check with the state drink-
ing water agency or regional EPA

11 Thomas Sorgo 'Process Selection lor Small Drinking Waler Supplies.' Proceedings of the Twenty- Third Annual Public Water
Supply Engineers' Conference: New DirectIons for Supply Design and Operation. Universily of Illinois. April 21-23. 1981.
4. Filtration Technologies for Small Systems

Filtration is the process of removing them between the grains 01 the filter
suspended solids from water as the medium (such as sand). A more impor-
water passes through a porous bed of lant process is adhesion, by which
materials. Naturalfillration removes suspended particles stick to the sur-
most suspended matter from ground face of filter grains or previously
water as the water passes through deposited material. Figure 4-1
porous layers of soil into aquifers illustrates these two removal
(water-bearing layers under the mechanisms. Biological processes are
ground). Surface waters, however, are also important in slow sand filters.
subjeclto runoff and other sources of These filters form a filter skin contain-
contamination, so these waters must ing microorganisms that trap and
be filtered by a constructed 'reatment break down algae, bacteria, and other
system. organic mailer before the water
reaches the filter medium ~se~.
The solids removed during filtration in-
clude soil and other particulate mailer Processes PrecedIng Filtration
from the raw water, oxidized metals,
Even when treating low turbidity water.
and microorganisms. Filtration can be
filtration is preceded by some form of
used to remove many microorgan- pretreatment. Several processes may
isms, some of which might be resis- precede filtration (Figure 4-2):
tantto disinfection. Filtration also
prevents suspended material • Chemical feed and rapid mix.
(measured as turbidity) from interfer- Chemicals may be added to the
ing w~11 later treatment processes, water to improve the treatment
including disinfection. Filtration com- processes that occur later. These
bined w~h disinfection provides a chemicals may include pH ad-
"double barrier" against waterborne justers and coagulants. (Coag-
disease caused by microorganisms. ulants are chemicals. such as
alum. that neutralize pos~ive or
The filtration process usually works by negative charges on small par-
a combination of physical and chemi-
ticles. allowing them to stick
cal processes. Mechanical straining
together and form larger, more easi-
removes some particles by trapping ly removed particles.) A variety of

t_
8. Ada0'l'"on
RAW '.'VA rER RAW WATEFl

••11
large parllcles become lOdged and cannol
continue downward fhrough the me(ha ••••
Particles slick 10 the meOl3 and cannot
continue downwanj lhrough the medl"

Figure 4·1. Filtration primarily dapands on physical and chemical


mechanism. to remove particle. from waler. (Reprinted from Introduction
to Water Treatment, Vol. 2, by permission. Copyright 1984, American Water
Works Association.)

233
234 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Figure 4-2. Several processes can precede filtration to Improve treatment processes that occur later.

devices, such as baffles. hydraulic Filtration processes can include only removal capacities (percentages that
jumps, static mixers, impellers. and one of these pretreatment procedures are effectively removed) of Giardia
in-line jet sprays can be used to or all of them. cysts and viruses for these four tech-
mix the water and distribute the nologies. Filtration technologies for
chemicals evenly. Choosing a Filtration Technology small systems are described in more
Conventional filtration. which includes detail below.
• Flocculation. In this process,
which follows rapid mixing, the coagulation with the addition of chemi·
cals, rapid mixing, flocculation and Slow Sand Filtration
chemically treated water is sent into
a basin where the suspended par- sedimentation, and granular media Slow sand filtration, first used in the
ticles can collide and form heavier filtration. is the most versatile system United States in 1872, is the oldest
particles called floc. Gentle agita- for treating raw water that is variable in type of municipal water filtration. A
tion and appropriate detention quality. However, a conventional filtra· slow sand filter consists of a layer of
times (the length of time water tion plant is usually neither appropriate fine sand supported by a layer of
remains in the basin) facilitate this nor economically feasible for very graded gravel. Slow sand filtratio'n
process. small systems. Package plants are does not require exlensive active con-
one available cost-effective alternative trol by an operator. This can be impor-
• Sedimentation. Following floccula- when automatic chemical feed control tant for a small system in which an
tion, a sedimentation step may be systems simplify operation. operator has several responsibilrties.
used. During sedimentation, the
velocity of the water is decreased Other filtration technologies that can Slow sand filters require a very low ap-
so that the suspended material (in- be more surtable for small systems are plication or filtration rate (.022 cubic
cluding flocculated particles) can slow sand filtration and diatomaceous meters per hour per square centimeter
settle out of the water stream by earth filtration. Membrane filtrarion and [0.015 to 0.15 gallons per minute per
gravity. Once settled, the particles cartridge filtration are two emerging square foot of bed area], depending
combine to form a sludge that is technologies that are surtable for small on the gradation of the filter media and
later removed from the clarified su- systems. Table 4·1 presents the ad- the quality of the raw water). The
pernatant water. vantages and disadvantages of these removal action includes a biological
technologies. Table 4-2 shows the process in addrtion to physical and
Filtration Technologies for Small Systems 235

Table 4-1. Advantages and Disadvantages of Filtration Technologies

Flhratlon Technology Advantages Disadvantages

Slow sand Operational simplicity and reliability NOl suitable for waler with high turbidity
Low cost Maintenance needs of filter surfaces
Ability to achieve greater than 99.9
percent Giardia cyst removal

Diatomaceous earth Compact size Most suitable lor raw water with low
bacterial counts and low turbidity
(less than 10 NTU)

Simplicity of operation Requires coagulanl and Iiller aids


for effective virus removal

Exceflent cyst and turbidity removal Potential difficulty in maintaining


complete and uniform thickness of
diatomaceous earth on filter septum

Membrane Extremely compact Little informaoon available to


establish design criteria or
operating parameters

Automated Most suitable for raw water with


less than 1 NTU; usually must be
preceded by high levels 01
prelTeatment

Easily clogged with colloids and


algae

Shon filter runs

Concerns about membrane failure

Complex repairs of automaled


controls

High percenl 01 water losl in


backllushing

Cartridge Easy to operate and maintain utt1e Information available to


establish design crilena and
operating paramelers

Can be quickly clogged by algae and


colloids

Requires low turbidity Influent

Can require relatively large


operating budget
236 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Table 4-2. Removal Capacities of Four Filter Options

Achievable
Glardl. Cyst Achievable
Levels Virus
Filtration Options (percent Removal) Levels

Slow sand 99.99 99.9999


Diatomaceous earth 99.99" 99.95b
Membrane 100 Very low
Cartridge >99 Little data available

"Aided by coagulation.
twith filter aid.

Source: U.S. Environmental Protection Agency, Office of Drinking Water and Center for Environmental Research Information,
Technologies lor Upgrading Existing or Designing New Drinking Water Treatment Facilities. March 1990. EPA 625/4-89-023.

chemical ones. A sticky mat of Water with high turbidity can quickly Because of the absence of coagulation.
suspended biological matter. called a clog the fine sand in these filters. slow sand filtration is limited to certain
"schmutzdecke: forms on the sand Water is generally applied to slow types of raw water quality. Slow sand fil-
surface, where particles are trapped sand filters without any pretreatment ters do not provide very good removal of
and organic matter is biologically when it has turbidity levels lower than organic chemicals, dissolved inorganic
degraded. Water applied to slow sand 10 NTU (nephelometric turbidity units). substances such as heavy metals. and
filters is usually not prechlorinated, The upper turbidity limit for slow sand trihalomethane precursors (chemical
since the chlorine destroys the filters is between 20 and 50 NTU. compounds, formed when natural or-
organisms in the schmutzdecke that When slow sand filters are used with ganic substances dissolve in water, that
help remove microbiological. organic. surface waters having widely varying might form THMs when mixed with
and other contaminants, (Sometimes turbidity levels. they can be preceded chlorine). Also, waters with very fine
water is prechlorinated and then by infiltration galleries or roughing fil- clays are not easily treated using sand
dechlorinated before slow sand ters, such as upflow gravel filters, to Mers. High algae blooms will result in
filtration.) reduce turbidity. short filter runs.

A slow sand filter must be cleaned


when the fine sand becomes clogged
(as measured by the head loss). The
length of time between cleanings can
range from several weeks to a year,
RAW
depending on the raw water quality.
WATER VENTILATION Cleaning is performed by scraping off
INLET the top layer of the filter bed. A "ripen-
c ing period" of 1 to 2 days is required
TO DRAIN -
for scraped sand to produce a function-
FILTERBOX .. ing biological filter. The filtered water
FILTERED quality is poor during this time. and the
WATER
OUTlET filtered water must be wasted. Ex-
tended cleaning periods require redun-
I dant or standby systems. In some
TO DRAIN TO DRAIN
small slow sand filters, geotextile filter
material is placed in layers over the
Figure 4-3. Slow sand filter. (Source: International Reference Centre for surface. A layer of tilter cloth can be
Community Water Supply and Sanitation. Technology of Small Water Supply removed periodically so that the upper
Systems in Developing Countries, WHO Collaborating Centre. The Hague, sand layer requires less frequent re-
The Netherlands, 1982.) placement.
Filtration Technologies for Small Systems 237

In climates subject to below-freezing


temperatures, slow sand systems
usually must be housed. Unhoused fil-
ters in harsh climates develop an ice Filtrate

layer that prevents cleaning. Thus.


uncovered slow sand filters will
operate effectively only ~ turbidity
levels of the influent (water flowing in)
are low enough for the tiner to operate
through the winter months w~hout
P'l!(Ool\
cleaning. In warm climates, a cover dum
hne
over the slow sand filter may be
needed to reduce algae growth within
the filter. Figure 4-3 shows a typical '---,--'-_ _ r0 Or J III

slow sand system.

In add~ion to maintenance, slow sand


fi~ers require:
Figure 4-4. Typical pressure diatomaceous earth filtration system.
• Daily inspection
DE fi~ers (Figure 4-4) use a very thin
• Control valve adjustment • Periodic backwashing, every 1 to 4
layer of diatomaceous eanh as a filter
days, depending on raw water
• Daily turbidity monitoring material (3.2 to 6.4 mm [1/8 to 1/4 in.])
quality.
which is coated on a porous septum or
Slow sand fi~ers can achieve 91 to liner element. An appropriate grade of • Disposing of spent filter cake.
99.99 percent removal of viruses and diatomaceous eanh should be used.
• Periodically inspecting the filter sep-
greater than 99.9 percent removal of (Grades vary from fine to coarse, with
tum for cleanliness and damage.
Giardia cysts. line grades removing smaller panicle
sizes but producing shoner filter runs). • Verifying the effluent quality.
Package slow sand filters, constructed The septum is placed in a pressure • Maintaining pumps, mixers,
from lightweight materials and vessel or operated under a vacuum in feeders, valves, and piping needed
transponed for local installation. have an open vessel. Add~ional diatoma- for precoat and body feed opera-
been used successfully in small rural ceous eanh (-body feed") is also tions.
communities in developing added to the influent water during the
countries. 12 These might be ap- fi~ration process to prolong the fi~er DE filters can effectively remove
propriate where community size is run. Higher body feed doses are Giardia cysts, algae, and asbestos,
less than 1.000 people and conven- needed for higher concentrations of and the fine grades of diatomaceous
tional construction of a slow sand fi~er suspended solids in the raw water. eanh can remove bacteria. These fil-
would be too slow or inconvenient. When the fi~er becomes plugged. it is ters require, however, that the water
backwashed and agitated so that the be pretreated with coagulating chemi-
Diatomaceous Earth Filtration diatomaceous eanh falls off the sep- cals and special filter aids to effective-
Diatomaceous eanh (DE) filtration, tum and is flushed from the filter tank. ly remove viruses.
widely used for filtering swimming pool
waters. has also been used success- Operation and maintenance of Plain diatomaceous eanh treatment
fully to remove turbidity and Giardia diatomaceous eanh filters require: (without the use of a coagulant) does
cysts 'rom drinking water. Advantages not provide good removal of very fine
of DE tiners include compact size. • Preparing slurries of fi~er body feed panicles. DE filters also are not
simplic~y of operation. and excellent and precoat diatomaceous eanh. capable of removing dissolved sub-
turbid~y removal. They are most stances, including color-causing
• Adjusting body feed dosages for ef-
su~ed for water systems with low tur- materials. Excessive suspended mat-
fective turbid~y removal.
bidity (less than 10 NTU) and low bac- ter and algae in the raw water can
terial counts. cause short fi~er runs.

'2 B.J. Uoyd. M. Pardon, D. Wheeler. Rural Water Treatment Package Plant: Final Report lor rhe U.K. Overseas Development
Administration. July 1986 DelAgua. P.O. Box 92. Guildford. GU2STO. England.
238 Upgrading Existing or Designing New Drinking Water Treatment Facilities

• Adsorption clarlflar package


plants. These use a contact "bed"
with plastic bead media (an adsorp-
tion clarifier) to replacQ the floccula-
tion and sedimQntation basin.
thereby combining these two steps
into one. A mixed media filter (a fil·
ter with a coarse-to·fine gradation
of filter media or several types of fil-
ter media) completes the treatment.

Package plants can effectively remove


turbidity and bacteria from surface
water of fairly consistent quality.
provided that they are run by com-
petent operators and are properly
maintained. Package plants also can
be designed to remove dissolved sub-
stances from the raw water. including
Figure 4-5. A package filtration aystem In Meredith, New Hampshire.
color-causing substances and
trihalomethane precursors. However.
when the turbidity of the raw water
equipment manufacturer can often per- varies a great deal. these plants re-
Package Plants form these tests. quire a high level of operational skill
and operator attention.
Package plants (Figure 4-5) are treat-
Package plants can be (and usually
ment units that are assembled in a fac-
are) designed to minimize the amount Membrane Filtration (Ultrafiltration)
tory, skid mounted, and transported to
the treatment site or that are trans- of day-to-day attention required to Membrane filtration, also known as
ported as component units to the site operate the equipment. Their opera· ultrafiltration. uses hollow fiber
tion and maintenance are simplified by membranes to remove solids from
and then assembled. They are most
widely used to treat surface water automated devices such as effluent tur· water. It can be an attractive option for
supplies for removal ot turbidity, color, bidimeters connected to chemical feed small systems because of its small
and coliform organisms with filtration controls and other operating para- size and automated operation. and it
meters, such as backwashing. Chemi· does not require coagulation as a
processes. but package plants that
can remove inorganic and/or organic cal feed controls are aspecially pretreatment step.
contaminants are also available. Pack· important for plants without full·time
age plants are often used to treat operators or with variable influent char- Many membrane systems are
acteristics. Even with these automated designed as skid-mounted unrts.
small community water supplies. as
well as supplies in recreational areas, devices, however, the operator needs Figure 4-7 shows an example of this
state parks. construction sites. ski to be properly trained and well ac· type of membrane system.
areas. military installations, and other quainted with the process and control
system. Membrane filtration systems can
areas not selVed by municipal supplies.
remove bacteria, Giardia. and some
Package plants can vary widely in Figure 4-6 depicts a package plant. viruses. They are most surtable for
their design cnteria and operating and The three basic types of package polishing water that has already been
maintenance requirements. The most water treatment systems are: treated by other methods. or for drink-
important factor to consider in select- ing water supplies with turbidity of less
• Convantlonal package plants. .than 1 NTU. Fouling of the fibers is the
ing a package plant is the nature of
These contain the conventional major problem preventing widespread
the influent, including characteristics
processes of coagulation, floccula· application of this technology.
such as temperature, turbidity, and
tion, sedimentation, and filtration.
color levels. Pilot tests (tests that
Traditional membrane filters work by
avaluate treatment processes and • Tube-type clarifier package
feeding water to the inside of the fiber
operations on a small scale to obtain plants. These use tube settlers to
membrane, with the filtrate (filtered
performance criteria) might be neces- reduce settling detention time (the
water) emerging on the outside of the
sary before a final system can be average length of tima water
membrane. State-of-the-art membrane
selected. The package treatment remains in the tank or chamber).
Filtration Technologies for Small Systems 239

~1:;.:ct"S'l~~POE -~
f'l.OCCV\.ATOAS
STAGE (ACW TAAIN

~~~[T~~c~tE~ST"~~RD
EACH
T\IIIl()

T1U.... - . . . . "
\
r SERViCE ACC(SS pl.-Afro"..
"ITH ,.. .. MOAAll AND
TYPICAL STANDARD
.~ ItICll.P\.. .. T(

""UNDER llIIl1)l AOJUSTAIl.(


W[I"S [ACH flUIN

.,r------ WA$HwATEIII COLLECTIOH T"ooG'"


WITH ADJUSTAllE W[lltS
EACH T""'IN

~ - - - SuR'ACE AGITarOA INI.E1'


CONNECTION (ACH T"aIN
SLUDGE: OAAW OF,. TO
WASTE (ACH TRAIN

' - - - - WASMW.al(A OuTLET TO WAST(


(ACH T"AIN

,"ILTEA(D ...... TEA OuT\.ET AND


SETTlEO SOLIDS TAOuGMS !lACk WASH wAH" INlET
«IT ... CQ.LEC TORS ANO
WANlfOlDS

' - - - - - - fllT[R Ml:DIA


[ACH TRAIN
ROTARY SUR'ACE
IloGIUTOl'l (ACH ftLT(A

NO"'-CQlIllllOOlNG FILT£1Il -
UNOERD""'''' $Y$T[M

Figure 4-0. Package plant system for surface water treatment. (Courtesy of Smith and Loveless, Inc.)

finers pass influent to either the inside They use a physical process for Cartridge filters are easy to operate
or outside of the membrane. The hol- finratioo-straining the water through and maintain, making them suitable
low fiber membranes are contained in porous media. Cartridge filtration sys- for small systems w~h low turbidity in-
a pressure vessel or cartridge. The tems require raw water w~h low fluent. Skilled personnel are not
contaminants collect on the end of the turbidity. needed; personnel are needed only
hollow fiber and are discharged to for daily operation and general main-
waste by a reversal of water flow. tenance (cleaning and cartridge re-
Unrafinration membranes exclude par-
ticles larger than 0.2 microns.

The membrane filter system must be


cleaned to clear the hollow fibers. This
is done by backflushing and chemical
cleaning or by air pressure. Some
manufacturers have developed sen-
cleaning systems to extend the time
between chemical cleanings.
. --
--
.-~~- Con1l'Ol.IOf~1
Pul... CIP System,.

....
InMQI"Ify ChectI

One major concern about membrane


finers is the potential for membrane
failure. The failure of a membrane
should trigger an operational shut-
down or an alarm to the operator.

A diagram of a sample membrane sys-


tem is shown in Figure 4-8.

Cartridge Filtration
Cartridge liners consist of ceramic or
polypropylene finer elements that are
pad<ed into pressurized housings. Figure 4-7. Typical skid-mounted membrlne filtration ....mbly.
240 Upgrading EXisting or Designing New Drinking Water Treatment Facilities

placement). Ceramic filters may be


cleaned and used tor repealed filter
FIL TERED
cycles. Polypropylene cartridges be- ? WATER
come louled relatively quickly and
musl be replaced wilh new units. Al-
though these filter systems are opera- _ HOLLOW FIBER
tionally simple, they are not automated MEMBRANES
and can require relatively large operat-
ing budgets.
MEMBRANE
Cartridge filtration systems somelimes CARTRIDGE
use "roughing filters" as pretreatment
to remove large solids. Prechlorination
is recommended to prevent the growth ~ UNFILTERED
RECYCLE
of microorganisms on the filters.
(However, this should be avoided if the
raw water contains organic substan-
ces that can contribute to formation of CHEMICAL , '--- /
trihalomethanes.) Except for a disinfec-
~----~
COAGULANT
MEMBRANE CLEANING
tant, no other chemicals need to be ~ SOLUTION TO SEWER
added.
CLARIFIED AIR INLET FOR
Little information is available concern- RECYCLE ---.i BACK WASHING
ing the effectiveness of cartridge filters !
BACKFLUSH
for virus removal. WASTEWATER

Innovative Filtration Technologies


- BACK FLUSH CLARIFIER
Several other simple low-eost filtration
methods might be appropriate for
some small systems. For example, a
system developed by 3M Company
using disposable filter bags made of
polypropylene fibers (Figure 4-9) can
remove Giardia cysts from drinking Figure 4-8. Flow sheet of membrane filtration system.
water supplies. Small systems in
several states have successfully used
these filters with disinfection for treat-
ment of water tram suriace sources.

Figure 4-9. Simple filter bag system removes particles ranging from 1
to 4 microns. (Courtesy of 3M Filtration Products)
5. Disinfection

Disinfection is the treatment process supply tex1txloks. (See Chapler 8,


used to destroy disease-causing or- Resources.)
ganisms in a water supply. Prim;uy
disinfection refers to the part of the Disinfection Terminology
treatment process that provides the When chlorine is fed into water, it
necessary inactivation of Giarriia reacts with any substances that exert
cysts, bacteria, and viruses in source a "chlorine demand." Chlorine
water. Secondary disinfection refers to demand is a measure of the amount of
maintenance of a disinfectant residual chlorine that will combine with im-
which prevents the regrowth of purities and therefore will not be avail-
microorganisms in the water distribu- able to act as a disinfectant. Impurities
tion system. Systems must disinfect that increase chlorine demand include
surface water according to the require- natural organic malerials, sulfides. fer-
ments of the Surface Water Treat- rous iron, and nitrites.
ment Rule (see Chapler 2).
Chlorine can also combine with am-
Chlorination (the addition of chlorine) monia or other nitrogen compounds to
is the most oommon method of disin- form chlorine compounds that have
fecting drinking water. Other disinfec- some disinfectant properties. These
tants that small systems might want to compounds are called combined avail·
consider are ozone and u~raviolet able chlorine residual. ("Available"
(UV) radiation. Table 5-1 summarizes means available \0 act as a disinfec-
the advantages and disadvantages of tant.)
these three disinfectants. The
preferred application point for each dis- The uncombined chlorine that remains
infectant is shown in Table 5-2. in the water after any combined
residual is formed is called free avail-
Chlorination able chlorine residual. Free chlorine is
When chlorination is perform&<! proper- a much more effective disinlectant
ly, it is a safe, effective, and practical than combined chlorine.
way to destroy disease-eausing or-
ganisms. ~ also provides a stable Free chlorine is not available for disin-
residual (disinfectant remaining in the fection unless the chlorine demand of
water) to prevent regrowth in the dis- the raw water is satisfied. When
tribution system. However. under cer- chlorine dosage exceeds the "break-
tain oonditions, chlorine can combine point"-the point at which chlorine
with remaining organic materials in the demand is satisfied-additional
water to produce potentially harmful chlorine will resun in a free available
by-products such as trihalomethanes. chlorine residual. The chlorine dosage
(See Disinfection By-Products and needed to produce a free r!'sidual
Strategies for Their Control below.) varies with the quality of the water
source.
Complex chemical reactions occur
when chlorine is added to water, but Factors Affecting ChlorInation
EfficIency
these reactions are not always ob-
vious. For example. a chlorine taste or Five factors are important to success-
odIor in finished water is sometimes ful chlorination: concentration of free
the resu~ of too little chlorine rather chlorine. contact time, temperature,
than too much. ~ is important for pH, and turbidity levels.
operators to understand basic chlorina-
tion chemistry and the factors affecting The effectiveness of chlorination is
chlorination efficiency. These topics directly related to the concentration of
are cover&<! thoroughly in many water free available chlorine and the contact
time. Contact time is the length of time

241
242 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Table 5-1. Advantages and Disadvantages of Three Disinfectants

Disinfectant AdvantagllS Disadvantages

Chlorine Very effective; has a proven history Potential for harmful


of protection against waterborne by-produclS under
disease. Widely used. Variety of certain conditio'ns
possible application poinlS.
Inexpensive. ApproRriate as both
primary and secondary disinfectant.
Operators can easily test for chlorine
residual throughout the water system.

Ozone Very effective. Minimal harmful Relatively high cost. More complex
by-products identified to date. operations because it must be
generaled on site.
Requires a secondary disinfectant.

Ultraviolet Very effective for viruses and Inappropriate lor surtace water
radiation bacteria. Readily available. Requires a secondary disinfectant.
No known harmful residuals.
Simple operation and maintenance
for high quality waters.

Source: Adapted from U.S. Environmental Protection Agency. Office ot Drinking Waler and Center for Environmental Research
Information, T9Chnologies for Upgrading Existing or Designing New Drinking Water Treatment Facilities. March 1990.
EPA 625/4-89-023.

the organisms are in physical contact and filtration, or other treatment


• Availability of electricity at the treat-
with the chlorine. If the chlorine con- methods.
ment site
centration is decreased. then the con-
tact time must be increased. Chlorination Chemicals • Operator skills
Chlorine is available as a liquid
The lower the pH, the more effective • Safety considerations
(sodium hypochlorrte), a solid (calcium
the disinfection. The pH also affects
hypochlorrte). or a gas. Small systems Disinfection with SodIum
corrosivity and formation of disinfec-
most commonly use sodium Hypochlorlt. Solution
tion by-products. The effects of pH
hypochlorrte or calcium hypochlorite,
should be considered along with disin- Sodium hypochlorrte (chlorine in liquid
because they are simpler to use and
fection effectiveness. form) is available through chemical
have less extensive safety require-
and swimming pool equipment sup-
ments than gaseous chlorine. The
The higher the temperature, the faster pliers, usually in concentrations of 5 to
choice of a chlorination system~iq­
the disinfection rate. The treatment 15 percent chlorine. It is easier to
uid, solid, or gas--depends on a num-
system operator usually cannot confrol handle than gaseous chlorine or cal-
ber of site-spec~ic factors, including:
the temperature, but then must in- cium hypochlorite. Sodium
crease fhe confact time or dose at hypochlorite is very corrosive,
• Availabilrty and cost of the chlorine
lower temperatures. however, and should be handled and
source chemical
stored with care and kept away from
Chlorine (or any disinfectant) is effec- • Caprtal cost of the chlorination sys- equipment that can be damaged by
tive only ~ rt comes into contact wrth tem corrosion.
the organisms to be killed. High tur-
bidity levels can prevent good contact • Operation and maintenance costs A basic liquid chlorination system or
and protect the organisms. Turbidity of the equipment hypochlorinator (Figure 5-1) includes
should be reduced where necessary • Location of the facilrty two metering pumps (one serving as a
through coagulation, sedimentation standby). a solution tank, a diffuser (to
inject the solution into water), and
Disinfection 243

a
Table 5-2. DesIred Points at Disinfectant Applicallon

Disinfectant Point of Application

Chlorine Towards the end of the water treatment process so that water is as clarified
(organic free) as possible, thereby minimizing THM Ionnation and providng
secondary disinfection.

Ozone Prior to the rapid mixing slop in all treatment processes. In addition,
sufficient time for biodegradation of the oxidation products of the ozonation
of organic compounds is recommended prior to secondary disinfection.

Ultraviolet radiation Towards the end at the walllr treatment process tIC tn4nimize the presence of other
contatn4nanta that inl9rlere with this disinfedant and tIC tn4nimize op8fllIlinll
problems.

"In general. disinfectant dosages will be lessened by placing the point of application towards the end of the water treatment
process because of the lower levels of contaminallls there to interfere with efficient disinfec;lion. However. water plants with
short detention times in clear wells and with nearby ~rst aJstomers might be required to move their point of disinfection
upstream to attain the appropriate CT value (see page 44) under the Surface Water Treatment Rule.

Source: U.S. Environmental Protection Agency. Office of Drinking Water and Center for Environmental Researcl1 Infonnation.
Technologi9s lor Upgrading Existing or D9signing N_ Drinking Wat9f Tr9atment Facilili6s, March 1990. EPA 6251HI9-<l23.

tubing. Hypochlorinators, used with (2-,5-,8-, and 35-pound) cans and is hygroscopic (readily absorbs mois-
chlorine in erther liquid or solid form. 360-kg (800-pound) drums. lure) and reacts slowly with moisturll
are discussed in more detail under in the air to form chlorine gas. There-
Hypochlorination Equipment below. When packaged, calcium hypochlorrte fore, shipping containers must be
is very stable, so that a year's supply emptied completely or carefully
Sodium hypochlorite solutions lose can be bought at one time. However, rt resealed.
their disinfec1ing power during
storage, and should be stored in a
cool, dry, dark area. No more than a
I-month supply should be purchased COUTurNfAD
alone timo, to prevent loss of avail- ODSlttG OIVICI

able chlorine. I
CHfMICAL
SOLUTIO. COIII.lcno. TO
Sodium hypochlorile solution is more MUl·UIt
TAU
costly per pound of available chlorine
than chlorine gas. It also does not con-
tain the high concentration of chlorine
available from chlorine gas. However.
the handling and storage costs are
lower than for chlorine in rts gaseous
form.

Dfsfnf9ction with Solid Calcium


Hypochlorite
Calcium hypochlorite is a white solid
that can be purchased in granular,
powdered, or tablet form. It contains Figure 5-1. Simple liquid chlorination disinfection system for ground·_ter
65 percent available chlorine and is supplies. Source water Is pumped to a servlca reservoir Into which a
easily dissolved in water. The chemi- chlorine solution Is dosed. (Sourcll: International Reference Centre for Com-
cal is available in 1-, 2-, 4-, and 16-kg munity Water Supply and Sanrtation, T9Chnology of Small Water Supply Systems
in D9ve/oping Countries, WHO Collaborating Centre, The Hague, The Nether-
lands, 1981.)
244 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Calcium hypochlorite is dissolved in especially desirable in systems where terrupted when the chlorine solution is
water in a mixing tank. The resu~ing water pressure is low and fluctuating. exhausted.
solution is stored in and fed from a
stock solution vessel made of Aspirator feeders. The aspirator Tablet hypochlorlnators. The tablet
corrosion-resistant materials, such as feeder operates on a simple hydraulic hypochlorinating unit consists of a spe-
plastic, ceramic, glass, or rubber-lined principle that uses the vacuum created cial pot feeder containing calcium
steel. when water flows either through a ven- hypochlorite tablets. Accurately con-
turi tube or perpendicular to a nozzle. trolled by means of a flow meter. small
The equipment used to mix the solu- The vacuum created draws the jets of feed water are injected inlo the
tion and inject rt into the water is the chlorine solution from a container into lower portion of the tablet bed. The
same as that for liquid chlorine. Solu- the chlorinator unit where it is mixed slow dissolution of the tablets prOVides
tions of 1 or 2 percent available with water passing through the unit, a continuous source of fresh hypo-
chlorine can be delivered by a and the solution is then injected into the chlorite solution. The hypochlorinating
diaphragm-type. chemical feed/meter- water system. In most cases. the water unit controls the chlorine solution. This
ing pump. inlet line to the chlorinator is connected type of chlorinator is often used when
to receive water from the discharge side electricity is not available, but requires
Calcium hypochlorrte is a corrosive of the water pump. wrth the chlorine solu- adequate maintenance for eHicient
material with a strong odor. and re- tion being injected back into the suction operation. It can operate where the
quires proper handling. ~ must be kept side of the same pump. The chlorinator water pressure is low.
away from organic materials such as operates only when the pump is operat-
wood, Cloth, and petroleum products. ing. Solution flow rate is regulated by Disinfection with Chlorine Gas
Reactions between calcium hypo- means of a control valve, though pres- Chlorine is a toxic, yellow-green gas at
chlorrte and organic maierial can sure variations may cause changes in standard temperatures and pressures.
generate enough heat to cause a fire the feed rate. It is supplied as a liquid in high-
or explosion.
strength. high-pressure steel cylin-
Suction feeders. One type of suction
ders, and immediately vaporizes when
Hypochlorlnatlon Equipment 13 feeder consists of a single line that
released. Small water systems can
Hypochlorinators. used with chlorine in runs from the chlorine solution con-
purchase the quantities they need
either liquid or solid form, pump or in- lainer through the Chlorinator unit and
from chemical or swimming pool sup-
ject a chlorine solution into the water. connects to lhe suction side of the pliers.
When they are properly maintained, pump. The chlorine solution is pulled
hypochlorinators provide a reliable from the container through suchon Gas chlorinators used in small sys-
method for applying chlorine to disin- created by the operating water pump. lems are often cylinder-mounted or
fect water. wall-mounted systems. Figure 5-2
Another type of suction feeder operates
shows a gas chlorinator. Daily opera·
Types of hypochlorinators include posi- on the Siphon principle, with the chlorine
tion of a gas chlorinator consists of
tive displacement feeders. aspirator solution being introduced direct~ into a regulating the feed rate, starting and
feeds, suction feeders. and tablet well. This type also consists of a single
stopping Ihe chlorinator. and changing
hypochlorinators. line, but the line terminates in the well
the chlorine cylinders.
below the water surlace instead of the in-
Positive displacement feeders. A fluent side of the water pump. When the Chlorine gas. if accidentally released
common type of positive displacement pump is operating. the chlorinator is ac- into the air. irritates the eyes, nasal
hypochlorinator uses a piston or tivated so that a valve is opened and the membranes. and respiratory tract. It IS
diaphragm pump to inject the solution. chlorine solution is passed into the well. lethal at concentrations as low as 0.1
This type of equipment. which is adjus- percent air by volume. Therefore.
table during operation. can be In each of these units. the solution
systems using chlOrine gas musl have
designed to give reliable and accurate flow rale is regulated by means of a
several major pieces 01 safety
feed rates. When electricity is avail- control valve and the chlorinators
equipment:
able. the stopping and starting of the operate only when the pump is operat·
ing. The pump circuit should be con- • Chlorine gas delectors to provide
hypochlorinator can be synchronized
nected to a liquid level control so that early warning of leaks
wrth the pumping unit. A hypo-
chlorinator of this kind can be used the water supply pump operation is in-
• Self-contained breathing apparalus
with any water system: however, rt is for the operator

13 From U.S. Environmenlal ProtectIon Agency. Office of Drinking Wale<. Manual 01 Individual Warer Supply Sysrems. October 1982.
EPA-570~821004.
Disinfection 245

the world, but is relatively new to the fectant. requiring shorter contact time
• A power ventilation system tor
United States as a drinking water disin· than Chlorine for disinfection.
rooms in which chlorine is housed
fectant. A toxic gas formed when air
• Emergency repair kits containing oxygen flows ootween two Ozone gas is unstable and must be
electrodes. ozone is a powerful disin- generated on site. In addition, ~ has a
Chlorination Monitoring
Whenever chlorine is used for disinfec-
tion. the chlorine residual should be
monitored at least daily. Samples
should be taken at various locations
throughout the water distribution sys-
tem. including the farthest points of
the system. Most small systems use a
quick and simple test called the DPD
colorimetric test, available as a kit
from companies specializing in water-
testing equipment and materials
(Figure 5-3). Table 5-3 lists some com-
panies that supply chlorine residual
test kits. Appendix 0 describes how to
take a sample for chlorine residual
analysis.

Ozonation
Ozone (03) is widely used as a
primary diSinfectant in other parts of

Two-stage ozone system. (Courtesy of Carus Chemical Company)

TID
./

Figure 5-2. Typical deep well gas chlorination system. (Courtosy of Fischer & Porter, Inc.)
246 Upgrading EXisting or Designing New Drinking Water Treatment Facilities

Table 5·3. Some Suppliers of Chlorine Residual Test Kits produced and to prevent fouling and
corrosion of equipment.

Capital Controls Co. Box 211, Colmar, PA 18195 Ozone used lor water treatment is
(215) 822-2901 (BOO) 523-2553 usually generated using a corona dis-
charge cell consisting of two
Fischer and Porter Co., County Line Rd., Warminster, PA 18974 eleclrodes separated by a discharge
(215) 674-6000 (800) 421-3411 gap and a dielectric plate (Figure 5-4).
The dried air (or pure oxygen) flows
Hach Co., Box 389, Loveland, CO 80537 between the electrodes and is con-
(303) 669-3050 (800) 227-4224 verted to ozone. Several types of
ozone generators are commercially
Hydro Instruments, Inc., Box 615, buakertown, PA available: horizontal tube, vertical
(215) 538-1367 tube, and plate generators. These
systems are available with varying
Wallace & Tiernan, 25 Main St., Belleville, NJ 07109 operating frequencies and voltages.
(201) 759-8000 An ozone contactor is used to dissolve
the ozone in water. Ozone can be
generated under posilive or negative
pressure, depending on the needs of
the contactor to be used.

As w~h chlorine. the ozone demand 01


the water must be satisfied before an
ozone residual is available lor disinfec-
tion. This can be aocomplished by
using two ozone contacting chambers
(Figure 5·5). The ozone delivered 10
the first chamber satislies the ozone
demand of the water, and the second
chamber maintains the disinfecting
residual.

Because ozone is toxic, the ozone in


exhaust gases from the contactor
must be recycled or removed and
destroyed before venting. Figure 5-6
depicts a complete treatment system
that includes ozone disinfection.

Figure 5-3. Test kit analyzes Iree and total chlorine. (Courtesy 01 Hach The capital costs of ozonation sys-
Company) tems are relatively high and operation
and maintenance are relatively com-
plex. Electricity is a major part of
low solubility in water, so efficient con- oxygen to produce the ozone are avail- operating costs, representing 26 to 43
tact with the water is essential. able lor small systems. percent of total operating and main-
A secondary disinfectant. usually tenance costs for small plants. Opera-
chlof/ne. is required because ozone Air feed systems used lor ozonation tion and maintenance for ozonation
does not maintain an adequate are classified by low, medium, or high systems include periodic repair and
residual in water. operating pressure. (High pressure replacement of equipment parts, peri-
systems typically are used in small-to odic generator cleaning, annual main-
Pure oxygen or ambient (freely circulat- medium-sized applications.) These tenance 01 the contacting chambers.
ing) air can be used in ozone produc- systems vary in their maintenance re- maintenance of the air preparation
tion. Pure oxygen delivers higher quirements, capital costs, and operat- system, and day-to-day operation of
concentrations of ozone. Packaged ing costs. The air feed systems are the generating equipment (averaging
ozone generator systems using necessary to dry the air (lower its dew 1/2 hour per day).
point) to increase the amount of ozone
Disinfection 247

Monitoring the Ozonatlon System


Operation
Proper monitors should be supplied
with the ozonation system, including:

• Gas pressure and temperature


monitors in the air preparation ~- ELECTRODE
system . . . . . . ._ _. . . . . . - - DIELECTRIC
• Continuous monitors to determine
moisture content 01 the dried gas
----.03
fed to the ozone generator
- ELECTRODE
• Generator coolant monitors

• Flow rate, temperature, and pres-


sure monitors, and ozone con-
centration monitor lor the gas
discharged from the ozone gener-
ator to determine the ozone produc- Figura 5-4. Typical ozona generating configuration for a corona discharge
tion rate cell.
• Power input monitor lor the ozone
generator

• Ozone residual monitor


CONTACT
C>W\B[~
The ozone residual should be OH-C"..AS
uNOlaNATE-,;O .;I.~,..-__,
measured at a minimum of two points

IT.
in the contactor(s). Ozone residual
monitoring can be performed using a
manual chemical analyzer (by a
trained laboratory technician) or an in-
line instrument that continuously
: .• • • •
J
.
'~ Ii: i!

;:B!:.k~L
samples the water.
•. l I: !::
...- ::: I' :::
.1.

:'i ,-:::~i':
Ultraviolet Radiation (UV) . • rJ.,I.:!.
Ultraviolet radiation effectively kills
bacteria and viruses. As with ozone, a
V~~/f?./ 1::~{u
secondary disinfectant must be used J'lOW P1(T(A (TYPICAL)
in addition to ultraviolet radiation to
prevent regrowth of microorganisms in
the water distribution system. UV
radialion can be anractive as a
primary disinfectant for a small system
because: Figure 5-5. l'wo-<:ompartment ozone contactor with porous diffusers.

• It is readily available.
those organisms. Therelore, it is ganic maner. These materials can
• It produces no known toxic
recommended only for ground water react with or absorb the UV radiation,
residuals.
not directly influenced by surface reducing the disinfection performance.
• Required contact times are short. water, in which there is no risk of
Giardia cyst contamination. (Future UV radiation is generated by a special
• The equipment is easy to operate ground-water disinfection rules will es- lamp (Figure 5-7). When ultraviolet
and maintain. radiation penetrates the cell wall of an
tablish whether and how UV may be
used.) UV radiation is unsuitable for organism, it destroys the cell's genetic
Ultraviolet radiation, however, does material and the cell dies.
water with high levels 01 suspended
not inactivate Giardia cysts, and can- solids, turbidity, color, or soluble or-
not be used to treat water containing
248 Upgrading Existing or Designing New Drinking Water Treatment Facilities

The effectiveness of UV radiation disin- The most important operating factor and the contact time. T (in minutes).
fection depends on the energy dose for u~raviolet radiation disinfection is EPA's Guidance Manual for Com-
absorbed by the organism, measured the cleanliness at surfaces through pliance with the Filtration and Disinfec-
as the product of the lamp's intensity which the radiation must travel. tion Requirements for Public Water
(the rate at which photons are Surface fouling can resu« in inade- Systems Using Surface Water Sour-
delivered to the target) and the time of quate performance, so a strict main- ces provides CT values for achieving
exposure. H the energy dosage is not tenance schedule should be followed. various levels of inactivation of Giardia
high enough. the organism's genetic Another important operating factor is and viruses. Appendix E presents CT
material might only be damaged in- the timely replacement of the UV values for chlorine and ozone at
stead of destroyed. To provide a safety lamps. because they lose their output several water temperatures and water
factor. the dosage should be higher intensity and this loss is not readily ap- pH levels for inactivating Giardia and
than needed to meet disinfection re- parent. A sensor should be used at all viruses, and Appendix F provides an
quirements. For example. if disinfec- times to ensure the desired dose. example of a CT calculation for a
tion criteria require a 99.99 percent small system.
reduction of viruses. the UV system Obtaining Effective Disinfection:
should be designed to provide a CT Values Di~!nfectlon By-Products and

99.999 percent reduction. To ensure proper disinfection. the disin- Strategies for Their Control
fectant must be in contact with the tar- Adding a disinfectant to water might
Substances in the raw water exert a get organisms for a sufficient amount result in the production of harmful by-
UV demand similar to chlorine of time. CT values describe the products.
demand. The UV demand of the water degree of disinfection that can be ob-
affects the exposure time and intensity tained as a product of the disinfectant Chlorine, for example, can mix with
of the radiation needed for proper residual concentration, C. (in mgil) the natural organic compounds in
disinfection. water to form trihalomethanes (THMs).

AIR PREPARATION ox IDATlON FILTRATION SECONDARY DISINFECTION STORAGE


FLOCCULATION

~f ,lIer Degassmg

"
"
DiD·
!!
= ""
T'''..lra "
W,\lel "

"
"
"
"
"
B.lClo. ... .Il!oh Back ....utl Pump
W,.If"
O"lle.
FIOCCUII"'1

Ol"n
Olln'laUTION

R.llWWIlef
Pump

«d WUL

Flgur. 5-6. Sch.matlc of .yst.m that provide. primary disinfection with ozon., filtration, .econdary disinfection,
and lice.. ozone g. . destruction. (Courtesy of Carus Chemical Company)
Disinfection 249

used alone. For these reasons, it is im-


portant to know what compounds are
in the raw water before choosing
ozone as a disinfectant. Researchers
are continuing to study ozonation by-
products and their potential health
eHects.

Ultraviolet radiation might produce


some by-products from organic com-
pounds. but by-products 01 UV radia-
tion have not yet been identified.

Figure 5-7. Ultraviolet disinfection unit. (Courtesy of Atlantic Ultraviolet


Corporation)

One THM-chloroform-is a
• Changing the point In trealment
suspecled carcinogen. Other common
where chlorine Is added. If
Irihalomethanes are similar to
chlorine is presently added before
chloroform and may cause cancer. treatment (chemical feed, coagula-
tion, sedimentation, and filtration),
The formation of chlorination by-
it can instead be added after filtra-
products depends on several factors,
tion, or just before filtration and
including:
after chem,caltreatment.
• Temperature and pH of the water • Using allernatlve diSinfection
• Chlorine dosage methods. A system with a high con-
centration 01 chlorination by-
• Concentration and types of organic products in the treated water might
male rials in the waler consider alternative disinfection
methods, However, ozonation and
• Contact time tor free chlorine
ultraviolet radiation, the alternative
Several strategies for minimizing harm- methods most practical for small
ful chlorination by·products can be systems, cannot be used as disin-
used by small systems: fectants by themselves. Both re-
quire a secondary disinfectant
• Reducing the concentration 01 or- (usually chlorine) to maintain a
ganic materials belore adding residual,n the distribution system.
chlorine. Common water clarifica-
tion techniques, such as coagula- Ozonallon might also result in the for-
tion, sedimentation, and filtration, mation of some harmful by-products.
can eHectively remove many or- Ozone can produce toxic by-products
ganic malenals. Activated carbon from a few synthetiC organic com-
(described in Chapler 6) m'ght be pounds. such as Ihe pesticide hep-
needed 10 remove organic tachlor, If ozone is added to water
materials at higher concentrations containing bromide ions, it can form
or those not removed by other bromlnated organic compounds such
techniques. as bromine~ontaining trihalo-
methanes, Also, studies have shown
• Reevaluating the amount 01 that the addition 01 ozone followed by
chlorine uSGd. The same degree chlorine or chloramines can result in
of disinlection might be possible higher levels of certain by-products
with lower chlorine dosages, than when these disinfectants are
6. Treating Organic Contaminants in Drinking Water

Some small drinking water systems di1fused aeration. Information about


face contamination of raw water by organics removal effectiveness is not
natural or synthetic organic substan· yet available for the other technologies
ces. Sources of the.. substances described in this chapter.
include leaking underground
gasoline/storage tanks, runoff of Granular Activated Carbon (GAC)
herbicides or pesticides, or improperly Granular activated carbon (GAC)
disposed of chemical wastes. Natural removes many organic contaminants
organic materials might also be from water supplies. Congress has
present in water. designated GAC as the Best Available
Technology (BAn for synthetic organic
The technologies most suitable for or- chemical removal.
ganic contaminant removal in small
systems are granular activated carbon Activated carbon is carbon that has
(GAC) and aeration. Several emerg- been exposed to very high tempera-
ing technologies using aeration may ture, creating a vast network of inter-
also be suitable for small systems. nal pores (Figure 6-1). It removes
contaminants through adsorption, a
Table 6-1 presents operational condi- process in which dissolved con-
tions for the organics treatment tech- taminants adhere to the porous sur-
nologies most suitable for small face of the carbon particles. Because
systems. Table 6-2 presents removal activated carbon is very porous, it has
effectiveness data for organic con- a large internal surface area; 1 gram
taminants by granular activated car- of activated carbon has a surface area
bon, packed column aeration, and equivalent to a football field. One

Table 6-1. Operational Conditions for Organic Treatments

level of
Operational level of
Skill Maintenance Energy
Technology Required Required Requirements

Granular aclivated Medium Low Low


carbon (GAC)

Packed column Low Low Varies


aeration (PCA)

Diffused aeration Low Low Varies

Multiple tray Low Low Low


aeration

Mechanical aeration Low Low Low

Catenary grid Low Low High

Higee aeration Low Medium High

Source: U.S. Environmental Prolection Agency, OIlice of Drinking Water and


Center for Environmental Research Information. Technologies for Upgrading
Existing or Designing New Drinking Waler Tmatmenl Facilities, March 1990.
EPA 62514-89-023.

250
Treating Organic Contaminants in Drinking Water 251

Table 6·2. Treatment Technology Removal Effectiveness Reported for Organic Contaminants (Percent)a

Granular Activated Packlld Column Diffused


Contaminant Carbon (GAC) Allration (peA) Allration

Acrylamide NA 1-29 NA
A1achlor 0-49 70-100 NA
Aldicarb NA 0-29 NA
Benzene 70-100 70-100 NA
Cartlofuran 70-100 0-29 11-20
Carbon tetrachloride 70-100 70-100 NA
Chlordane 70-100 0-29 NA
Chlorobenzene 70-100 70-100 NA
2,4-D 70-100 70-100 NA
l,2-Dichloroethane 70-100 70-100 42-77
1.2-Dichloropropane 70-100 70-100 12-79
Dibromochloropropane 70-100 3O-<i9 NA
Did1lorobenzene 70-100 NA NA
o-Dichlorobenzene 70-100 70-100 14-72
p-Dichlorobenzene 70-100 70-100 NA
1.1-Dichloroethylene 70-100 70-100 97
cis-l,2-Dichloroethylene 70-100 70-100 32-85
tran.r l,2-Dichloroethylene 70-100 70-100 37-96
Epichlorohydrin NA 0-29 NA
Elhylbenzene 70-100 70-100 24-89
Ethylene dibromide 70-100 70-100 NA
Heptachlor 70-100 70-100 NA
Heptachlor epoxide NA NA NA
High molecular weight W NA NA
hydrocarbons (gasoline,
dyes. amines, humics)
Lindane 70-100 0-29 NA
Methoxychlor 70-100 NA NA
Monod1lorobenzene NA NA 14-85
Natural organic material P NA NA
PCBs 70-100 70-100 NA
Phenol and d1lorophenols W NA NA
Pentachlorophenol 70-100 0 NA
Styrene NA NA NA
Tetrachloroethylene 70-100 NA 73-95
Trid1loroethylene 70-100 70-100 53-95
Trid101oroethane 70-100 NA NA
1,1,1-Trid1loroethane 70-100 70-100 58-90
Toluene 70-100 70-100 22-89
2.4,5-TP 70-100 NA NA
Toxaphene 70-100 70-100 NA
Vinyl d1loride 70-100 70-100 NA
Xylenes 70-100 70-100 18-89

"Additional treatment Information is available in EPA Office of Drinking Water Health Advisories for specific contaminants.

W z waH removed. P z poorly removed. NA z nol available.

Note: Linle or no specific perlonnance data were available for:


1. Multiple Tray Aeration 3. Higee Aeration
2. CateNry Aerabon 4. MechaOlcal Aeration

Source: U.S. Environmental Protection Agency, Office of Drinking Waler and Center for Environmental Research
In/onnalion , Technologies (or Upgraolng Existing or Designing New DrinkIng Water Treatmenr Facili~es, March 1990.
EPA 62514-89-023.
252 Upgrading Existing or Designing New Drinking Water Treatment Facilities

pound of activated carbon can adsorb A significant drop in the contaminant


over 1f2 pound of carbon tetrachloride. level in influent water will cause a
GAC filter to desorb, or slough off, ad-
GAC has an affinity for high molecular sorbed contaminants, because GAC is
weight compounds. tt is not effective an equilibrium process. As a resutt,
in removing vinyl chloride, a highly raw water with frequently changing
volatile substance, from water. Table contaminant levels can result in
6-3 lists organics that are readily or treated water of unpredictable quality.
poorly adsorbed by activated carbon.
Bacterial growth on the carbon is
GAC can be used as a replacement another potential problem. Excessive
for eXisting medld (such as sand) In a bacterial growth may cause clogging
conventional filter or it can be used in and higher bacterial counts in the
a separate contac1or (a vertical steel treated water. This means that bac-
pressure vessel used to hold the ac- teriallevels in the treated water must
tivated carbon bed). be closely monitored and the final dis-
infec1ion process must be carefully
GAC contac1ors require monitoring to Figure 6-1. Representation of controlled_
ensure that they work eHectively. A Internal carbon structure. (Reprinted
GAC monitoring system should from Introduction to Water Treatmenr. GAC is available in different grades of
include: Vol. 2, by permission. Copyright 1984, eHectiveness. Low-cost carbon re-
American Water Works Association.) quires a lower initial capital outlay, but
• Laboratory analysis of treated must be replaced more often, resulting
water to ensure that the system is in higher operating costs.
removing organic contaminants which GAC has a greater affinity.
Therefore, the presence of other con-
• Monitoring of head loss (the amount taminants might interfere with the
of energy used by water in moving removal of the contaminants of
from one point to another) through concern.
the contactors to ensure that back-
flushing (reversing the flow to
remove trapped material) is per- Table 6·3. Readily and Poorly Adsorbed Organics
formcd al approprlatc times

Readily Adsorbed Organics


• Bacteria monitoring 01 the
contactor's effluent (since bacteria • AromaliC solvents (benzene. toluene, nilrobonzenos)
can grow rapidly within the ac- • Chlorinated aromatics (pCBs, chlorobenzenes, chloronaphlhalene)
tivated carbon bed)
• Phenol and chiorophenols
• Turbidity monitoring of the • Polynuclear aromatics (acenapthene, benzopyrenes)
contactor's effluent (\0 determine If • PestiCides and herbicides (DDT. aldrin. chlordane, heptachlor)
suspended material is passing • ChlOrinated aliphatlcs (carbon tetrachloride, chloroalkyl ethers)
through GAC bed)
• HI9h molecular weight hydrocarbons (dyes, gasoline, aminos, humics)
After a period of a few months or a
few years, depending on the con- Poorly Adsorbed Organics
centration ot contaminants, the sur-
• Alcohols
face of the pores in the GAC can no
• Low molecular weight ketones, aCids, and aldehydes
longer adsorb contaminants. The
carbon must then be replaced. • Sugars and starches
• Very high molecular weight or collOidal organics
Several operational and maintenance • Low molecular weight allphallcs
factors affec1the performance of GAC.
Contaminants in the water can occupy
GAC adsorption sites, whether they Source' US. Environmental Protection Agency, Office 01 Drinktng Water and
are targeted lor removal or not. Also, Center for EnVIronmental Rosearch Information, Technologies for Upgrading
Emling or Designing New Drtnking Wafer Treatment Faciltries, March 1990.
adsorbed contaminants can be
EPA 625/4-89-023.
replaced by other contaminants with
Treating Organic Contaminants in Drinking Water 253

packed column aeration, diffused


aeration, and multiple tray aeration.
Emerging technologies that use aera·
tion for organics removal include
mechanical aeration, catenary grid,
and Higee aeration.

Packed Column Aeration

Packed column aeration (PCA) or


packed tower aeration (PTA) is a
waterfall aeration process which drops
water over a medium within a tower to
mix the water with air. The medium is
designed to break the water into tiny

1
droplets. and maximize ~s contact with
tiny air bubbles for removal of the con·
taminan!. Air is also blown in from un-
derneath the medium to enhance this
process. Figure 6-2 shows a PCA
system.

Systems using PCA may need


pretreatment to remove iron, solids,
Figure 6-2. Packed tower aeration system. and biological growth to prevent clog-
ging of the packing material. Posttreat-
ment (such as the use of a corrosion
inhibitor) may also be needed to
CQMPI=lESSED AIR PIPING reduce corrosive properties in water
due to increased dissolved oxygen
AERATION BASIN
from the aeration process.
11I~'"""'~---t+-- OQWNCOMEA PIPE

Packed columns usually operate auto-


MIXING PATTERNS--t'<,-----~~ DIFFUSERS
matically, and need only daily visits to
,JL...I-I-_MANIFOlD ensure that the equipment is running
satisfactorily. Maintenance require-
ments include servicing pump and
blower motors and replacing air filters
on the blower, ij necessary.
Figure 6-3. Diffuser aeration system. (Reprinted from Introduction to Water
Treatment. Vol. 2, by permission. Copyright 1984, American Water Works PCA exhaust gas may require treat-
Association.) ment to meet air emissions regula-
tions, which can significantly
increase the costs of this technology.
Aeration specially designed aerator system.
Examples of simple aerators include:
Aeration, also known as air stripping, Diffused Aeration
mixes air with water to volatilize • A system that cascades the
In a diffused aeration system, a dif-
contaminants (turn them to vapor). water or passes it through a
fuser bubbles air through a contact
The volatilized contaminants are either slotted container
chamber for aeration (Figure 6-3).
released directly to the atmosphere or The main advantage of diffused aera-
are treated and then released. Aera- • A system that runs water over a
corrugated surface tion systems is that they can be
tion is used to remove volatile organic created from existing structures, such
chemicals and can also remove radon as storage tanks. However, they are
• An airlift pump that introduces
(see Chapter 7). less effective than packed column
oxygen as water is drawn from
a well aeration, and usually are used only in
A small system might be able to use a
systems w~h adaptable existing
simple aerator constructed from rela- Other aeration systems that might be
structures.
tively common materials instead of a su~able for small systems include
254 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Mu/llple Tray Aeration pounds eftectively. They have higher


energy requirements than PCA sys-
Multiple tray aeration directs water
tems, but their more compact design
through a series of trays made of
slats, perforations, or wire mesh
lowers their caprtaJ cost relative to
-
...
Ovhlf
PCA.
(Figure 6-4). A blower introduces air
from underneath the trays. Higee Ae,.tlon n.uot·ID -
._. -- _.. __
:-,,:~.:.:::-.::':."
._-~

..~_-
~
-_.
..~-.:.~

Multiple tray aeration un~s have less


surface area than do PCA unrts. This
Higee aeration is another variation of
the PCA process. These systems
... ,.w - ~~~~~~~~~~~:

..~,~;
type of aeration is not as effective as pump water into the center of a spin-
PCA, and can experience clogging ning disc of packing material, where
from iron and manganese, biological the water mixes with air (Figure 6-7).
growth, and corrosion problems.
Mu"iple tray aeration unrts are readily Higee unrts require less packing
available from package plant manufac- material than PCA units to achieve the ....,oU -- w:'''l •• -

turers. same removal efticiencies. Because


of their compact size, they can be
Emerging Technologies for used in limited spaces and heights. Figur. 6·4. Sch.matlc of a red·
wood .Iat tray ••r.tor
Organics Removal Current Higee systems are best suited
for temporary applications of less than
Mechanical Aeration
, year wrth capacities up to 380 Irters
(' 00 gallons) per minute.
Mechanical aeration uses mechanical
stirring mechanisms to mix air with the
water (Figure 6-5). These systems
can eftectively remove volatile organic
chemicals (VOCs).

Mechanical aeration units need large


amounts of space because they
demand long detention times for eftec-
tive treatment. As a result, they often
require open-air designs, which can
IiWcft.6llol.cal J",'_" ,... &tOI'
freeze in cold climates. These units
also can have high energy require-
ments. However. mechanical aeration
systems are easy to operate. and are
less susceptible to clogging from
biological growth than PCA systems. ~ II
J .....

j ~
C."enllry Grid
Catenary grid systems are a variation !
u,
ot the packed column aeration
process. The catenary grid directs
,I ':"",rbL".
h U 1Cf
--=~::J

water through a series of wire screens


mounted within the column. The
screens mIx the air and water in the Flgur. 6-5. Schematic of mechanical a.ratlon process_
same way as packing materials in
.PCA systems. Figure 6-6 shows a
catenary grid unit.

These systems can effectively remove


volatile organic chemicals. There is I~­
tie information available about the ef-
fectiveness of catenary grid systems
for other organic compounds, but they
probably would not remove these com-
Treating Organic Contaminants in Drinking Water 255

••ww .. w

~.
;liotl..... ~ ...

~] \ ~::::O:·~:O
"' •• I~W.I'"
$ ....._1'.0

1" •• 11'1:1'1<1 .....


100<1'"

Figure 6-6. Catenary grid system.

<) EXHAUST AIR

AIR IN

PROOUCT
Fil TER
~R
PUMP

Figure 6-7. Schematic of Hlge8 system.


7. Control and Removal of Inorganic Contaminants

Water systems control or remove inor- Table 7·1 shows typical customer com-
ganic contaminants using two different plaints and their causes.
strategies:
Controlling Lead Levels In
1. Preventing Inorganic contamina- Drinking Water
tion of finished water. Corrosion Because it is widespread and highly
controls prevent or minimize the toxic, lead is the corrosion product of
presence of corrosion products greatest concern. Table 7-2 shows the
(such as lead and copper) at the risk factors that can indicate potential-
consumer's tap. ly high lead levels at the tap. lead
levels in drinking water are managed
2. Removing Inorganic indirectly through corrosion controls.
contaminants from raw water.
lead is not typically found in source
Removal technologies treat source water, but rather at the consumer's tap
water that is contaminated with me- as a resuit of the corrosion of the
tals or radioactive substances plumbing or distribution system. The
(radionuclides). 1986 amendments to the Safe Drink-
ing Water Act ban the use of lead
Inorganic contaminants presently
solders, fluxes, and pipes in the instal-
regulated under the Safe Drinking
lation or repair 01 any public water sys-
Water Act (SDWA) include lead,
tem or in any plumbing system
radium, nitrate, arsenic, selenium,
providing water for human consump-
barium, fluoride, cadmium, chromium,
tion. In the past, solder used in plumb-
mercury, and silver.
ing has been 50 percent tin and 50
This chapter describes several tech- percent lead. Using lead-free solders,
nologies for inorganic contaminant such as silver·tin and antimony-tin
removal (reverse osmosis, ion ex- solders is a key factor in lead cor-
change, activated alumina, aeration, rosion control. Replacement of lead
and granular activated carbon). Con- pipes can also be an effective strategy
ventional treatment (coagulation! for reducing lead in drinking water.
fiitration) can also remove inorganic
The current Maximum Contaminant
contaminants and is discussed in this
level (MCl) for lead applies to water
chapter.
delivered by the supplier. New lead
Corrosion regulations might include an MCl for
water at the consumer's tap.
Corrosion is the deterioration of a sub-
stance by chemical action. lead, cad- ff tests for corrosion by-products find
mium, zinc, copper, and iron might be unacceptably high levels of lead, im·
found in water when metals in water mediate steps should be taken to mini-
distribution systems corrode. Drinking mize consumers' exposure until a
water contaminated with certain me- long-term corrosion control plan is im·
tals (such as lead and cadmium) can plemented. Some short-term
harm human heaith. measures the consumer can take
include:
Corrosion also reduces the useful life
of water distribution systems, and can • Running the water for about 1
promote microorganism growth, result· minute before each use
ing in disagreeable tastes, odors,
slimes, and further oorrosion. Otten a • Using home treatment processes in
customer complaint is the first indica- extreme cases
tion of a corrosion problem, and at this • Using bonled water
stage corrosion may be extensive.

256
Control and Removal of Inorganic Contaminants 257

Table 7-1. Typical Customer Water Quality Complaints That Might Be Due to Corrosion

Customer Complaint Posslblll Cause

Red water or reddish-brown staining Corrosion of iron pipes or presence of natural iron in raw water
01 fixtures and laundry

Bluish stains on fixtures Corrosion of copper lines

Black water Sulfide corrosion of copper or iron lines or precipitations of natural


manganese

Foul taste and/or odors By-produClS from miaobial activity

Loss of pressure Excessive scaling, tubercle buildup from pitting corrosion, leak in system
from pitting or other type of corrosion

Lack of hot water Buildup of mineral deposilS in hot water system (can be reduced by selting
therrnostals to under 6O'C (140'FIl

Short service life of household Rapid deterioration 01 pipes from


plumbing pitting or other types of corrosion

Source: U.S. Environmental Protection Agency, Office of Drinking Water, Corrosion Manual for Internal Corrosion of Water Dis-
rribution Systems, Apn11984. EPA 57019-84-001.

Table 7·2. Risk Factors Indicating Potentially High Lead Levels at the Tap

Componenls of the water distribution system or structure's plumbing are made of lead, or lead-containing
matenal such as brass. and/or the structure's plumbing has solder containing lead.

And one or more of the following apply:

The slTucture IS less than 5 years old.

• The tap water is soft and/or acidic.

The water stays ,n the plumbing for 6 or more hours.

The structure's electrical system is grounded to the plumbing system.

Source: Adapted from U.S. Environmental Protection Agency, Office 01 Drinking Water and Center for Environmental Research
Information. Technologies for Upgrading Existtng or Designing New Drinking Water Trearmenr FaCIlities. March 1990 EPA
625/4-89-023.

TechnlquBs for Controfflng homeowner's plumbing can be related Modifying water quality. Table 7-3
Corrosion to conditions in the home (such as shows some of the characteristics of
Solutions to corrosion problems in- pipes that are too small or the use of water that affect its corrosivity. All
clude modifying the water quality dissimilar metals in joining pipes and water is corrosive to some degree, but
(especially pH and alkalinity), and add- valves) rather than to the quality of water that is acidic (less than 7.0 pH)
ing. corrosion inhibitors to form protec- water from the distribution system. is likely to corrode metal more quickly.
tive coatings over metal. " should also
be noted that corrosion in a
258 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Table 7·3. Chemical Factors Influencing Corrosion and Corrosion Control

Factor Effect

pH Low pH may increase corrosion rate: high pH may protect pipes


and decrease corrosion rates

Alkalinity May help form protective calcium carbonate (CaCO)) coating,


helps control pH changes, reduces corrosion

Dissolved oxygen (DO) Increases rate of many corrosion reacnons

Chlorine residual Increases metallic corrosion

Tolal dissolved solids High TDS increases conductivity and corrosion rate
(TDS)

Hardness (Ca and Mg) Ca may precipitate as CaCOJ and thus provide protection
and reduce corrosion rates

Chloride, sullate High levels increase corrosion of iron. copper, and galvanized steel

Hydrogen sullide Increases corrosion rates

Silicate, phosphates May form protec!lve films

Natural color, organic maner May decrease corroSion

Iron, zinc. or manganese May react With compounds on Intenor of asbestos-cement pipe
to form protectille coating

Source: U.S Environmental Protecuon Agency, Office 01 Drinking Water, Corrosion Manual lor Inrernal CorrosIon 01 Water Dis-
triburlcm Systems, April 1984. EPA 570/9·84·001

Adjusting pH and alkalinity IS the most both disinfection and corrosion control Treatment Technologies for
common corrosion control method, be- are effective. Removing Inorganic Contaminants
cause it is simple and inexpensive.
(pH is a measure of the concentration Avoiding high dissolved oxygen levels Inorganic contamination of raw water
of hydrogen ions present in water: decreases water's corrosive activity. supplies can come from a wide variety
alkalinity is a measure of water's Removing oxygen from water IS not of sources. Naturally occurring Inor-
ability to neutralize acids.) Generally. practical because of cost. However, ganics, such as fluoride, arsenic,
an increase in pH and alkalinity can treatment systems might be able to selenium, and radium, are commonly
decrease corrosion rates and help minimize the dissolved oxygen levels found in ground-water sources. Syn-
form a protective layer of scale on cor- by minimizing all/water contact. thetic contaminants are usually found
rodible pipe material. Chemicals com- in surface water supplies. Nitrates and
monly used for pH and alkalinity Corrosion Inhibitors. Corrosion in- nitrites are a problem in agricultural
adjustment are lime, caustic soda. hibitors reduce corrosion by forming areas and areas without sanitary
soda ash, and sodium bicarbonate. Ad- protective coatings on pipes, The most sewer systems, and have been found
justing pH to control corrosion, common corrosion inhibitors are inor- at relatively high levels in both surface
however, might conflict with ideal pH ganic phosphates. sodium silicates, water and ground water.
conditions for disinfection and control and mixtures 01 phosphates and sili-
of disinfection by-products, Drinking cates. These chemicals have proven No single treatment is perfectly suited
water suppliers should carefully successful in reducing corrOSIOn in for all inorganic contaminants, Table
choose the treatment methods so that many systems. 7·4 shows the removal effectiveness
of six inorganic treatment processes
Control and Removal of Inorganic Contaminants 259

Table 7-4. Most Probable Applications of Water Treatment Processes for Inorganic Contaminant Removal

Inorganic Contaminant
Treatment Capability
Effectlvenessa
Treatment Principal Application for Most Probable Application
Process Water Treatment High Moderate Low for Inorganic Removal

Conventional Clarification of surface waters Cd As III Sa Removal of Cd, Cr. As, Ag, or
coagulation Crill SeIV F Pb from surface waters
CrVI Hg (0) N03
AsV Hg (I) Ra
Ag SeVI
Pb

Cation exchange Removal of hardness Sa As Removal 01 Sa or Ra from


from ground water Ra Se ground water
Cd N03
Pb F
Crill CrVI

Anion excl1ange Removal of nitrate from N03 Sa Removal of NO) from


ground waler CrVI Ra ground water
Se Cd
Pb
Crill

.Aotlvated alumina Removal of fluoride from F Sa Removal of F, As, or Se from


ground waler As Ra ground waler
Sa Cd

Granular aCliVated Removal of taste, odors, Hg(l) Cd Ag Removal of Hg from surface or


carbon and organiCS Hg(O) Sa ground water
Ra
Crill
F.
N03

Reverse osmOSIS and Desailing of sea water As V NO) Removal of all inorganics
electrodialysis or brackish ground water Sa As III from ground water
Cr
Pb
Cd
Se
Ag
F
Ra
Hg

'High~reater than 80 percenl: moderate-20 to 80 percent: low-less than 20 percent.

Ag Silver Hg Mercury
As Arsenic N03 Nltrale
Sa Barium Pb Lead
Cd Cadmium Ra Radium
Cr Chromium Se Selenium
F Fluonde

Source Adapted from Thomas J Sorg and Gary S. Logsdon, "Treatment Technology to Meet the Interim Primary
Dnnklng Waler Regulations for Inorganics, Part 5: Journal 01 the American Warer Works Association, July 19ao,
260 Upgrading Existing or Designing New Drinking Water Treatment Facilities

and their most probable application for Table 7-5. Advantages and Disadvantages of Inorganic
inorganic removal. Processes su~able Contaminant Removal Processes
for small systems include:

• Coagulationlfiltration CoagulatlonlFlltratlon

• Membranes (reverse osmosis and Adv.nt.gfttl


electrodialysis) • Low cost lor high volume
• Reliable process well suited to aUlomatic control
• Ion exchange

• Activated alumina Dludv.nt.g..


• Not readily applied to small or intermittenlllows
Table 7-5 presents advantages and High-water-contanl sludge disposal
disadvantages of these processes. • Very low contaminant levels may require two-stage precipitation
Coagulationlfi~rationis generally too Requires highly trained operalors
operationally complex for very small
systems (serving 500 people or Membranes (Reverse Osmosis and Electrodialysis)
fewer). Reverse osmosis, ion ex- Adv.nt.ges
change. and activated alumina are
Removes nearly all contaminant ions and most dissolved non-Ions
simpler operations and may be better
Relatively insensitive 10 flow and total dissolved solids level
a~ernatives for these systems. unless
• Low effluent concentration possible
the contaminant of concern cannot be • In reverse osmosis. bacleria and particles are also removed
removed by these processes. or un- • Automation allows for less operator attention
less the costs or maintenance require-
Dludv.nt.ges
ments are prohibitive. Each of these
processes is described below. Aera- • High capital and operating costs
tion and granular activated carbon, High level 01 prelreatment required in some cases
commonly used to remove radon from Membranes are prone to louling (RO). Electrodes require replace-
menl (ED).
waler supplies. are also discussed
below. Ion Exchange
Adv.nt.ge"
Coagulation/Filtration
• Relatively insensitive to flow variations
Coagulation (described in Chapter 4) Essentially zero level 01 effluent contamination poSSible
is traditionally used to control turbidity. Large variety 01 specllic resins available
hardness. taste. and odors. but is also
D/s.dv.nt.gu
effective in removing some inorganic
contaminants. POlentiallor unacceplable levels (peaks) 01 contamination in eHluent
Wasle requlles careful disposal
Coagulation using aluminum or iron Usually notleasible al high levels 01 total dissolved solids
salts is effective in removing most • Pretreatment required lor most surlace waters
metal ions or colloidally dispersed Activated Alumina
compounds (finely divided substances
Adv.nt.ge"
that do not settle out of water for a
very long period of time). ~ is ineffec- Insensitive to flow and total dissolved solids background
tive in removing nitrate. nitrite. radium. Low effluent oontaminantlevel possible
and barium. Table 7-6 presents poten- Highly selective for fluande and arsenic
tial removal efficiencies using D/s.dv.ntsges
aluminum and iron salts as coag-
Strong acid and base are required for regeneration
ulants. Coagulation to remove inor-
Medium tends to dissolve. producing fine particles
ganics is more expensive than Adsorption is slow
coagulalion to remove turbidity be- Waste requires careful disposal
cause higher dosages of coagulant
are needed. Source: U.S. EnVilonmentai Protecnon Agency. Office of Drinking Water and
Center lor Environmental Research Information. Technologies lor Upgrading
Exisnng or Designing New DrinkIng Warer Treatmenr Faciliries. March 1990.
EPA 62514-89-023.
Control and Removal of Inorganic Contaminants 261

Table 7·6. Removal Efficiency Potential 01 Alum Versus Ferric Reverse osmosis systems are com-
Chloride pact, simple to operate, anc have mini-
mallabcr requirements, making them
suitable for small systems. They are
Removal Elllclency also suitable for systems with a high
degree of seasonal fluctuation in water
Inorganic Alum Iron demand.
Contaminant Coagulant Coagulant
One disadvantage of reverse osmosis
Ag (pH < 8.0) 90% is rts high capital and operating costs.
For systems of less than 3.8 million
Ag (pH = 8.0) 70%
liters per day (1 million gallons per day
AsV 90% [MGD]), operating costs range from $3
to $6 per 3,800 liters (thousand gal-
As V (pH < 7.5) 900/. lons) of treated waler. Capital costs
range from $1 to $2 per 3.8hlers (gal-
As V (pH = 75) 90%
Ion) of capacity, depending on the
Cd (pH ~8.0) level of pretreatment required. Manag-
ing the wastewater (brine solution) is
Cd (pH ~ 85) 70% also a potential problem for systems
Gr(llI) 90% using reverse osmosis,

Cr III (pH = 1(5) 90% Electrodialysis is a process that also


Cr VI (using Fe II) 90% uses membranes. In this process,
however, direct electrical current is
Hg 70"1.
used to allrad ions to one side of the
Pb 900/. treatment chamber. Eledrodialysis sys-
tems include a source of pressurized
Ag Silver water, a direct current power supply,
As Arsenic and a pair of selective membranes.
Cd Cadmium Multistage units, in which membrane
Cr Chromium pairs are "stacked" in the treatment
Hg Mercury
vessel, can increase the removal ef-
Pb Lead
ficiency. Electrodialysis is very eHec-
Source: U.S. Environmental Protection Agency, Office of Drinking Waler tive in removing fluoride and nitrate,
and Cenler lor Environmenlal Research Information. Technologies for and can also remove barium, cad-
Upgrading Existing or Designing New Drinl<ing Warer Treatment Facilities, mium, and selenium.
March 1990. EPA 62514-89~23.
Ion Exchange
Ion exchange units (Figure 7-2) can be
used to remove any ionic (charged)
Reverse Osmosis and Reverse osmosis can eHectively
Electrodialysis substance from water, but are usually
remove nearly all inorganic con-
used to remove hardness and nitrate
Reverse osmosis removes con- taminants from water. ~ removes over
from ground water. Inorganic substan-
taminants from water using a semiper- 70 percent of arsenic(III), arsenic(lV),
ces are removed by adsorption onto
meable membrane that permits only barium, cadmium, chromium(III),
an exchange medium, usually a syn-
water, and not dissolved ions (such as chromium(VI), fluoride, lead, mercury,
thetic resin. One ion is exchanged for
sodium and chloride), to pass through nitrite, selenium(IV), selenium(VI), and
another on the surlace of the medium,
its pores (Figure 7-1). Contaminated silver. Properly operated units will at-
which is regenerated with the ex-
water is subjected to a high pressure tain 96 percent removal rates.
changeable ion before treatment
that forces pure water through the Reverse osmosis can also eHedively
operations. Ion exchange waste is
membrane, leaving contaminants be- remove radium, natural organic sub-
highly concentrated and requires
hind in a brine solution. Membranes stances, pesticides, and microbiologi-
careful disposal.
are avaIlable With a variety of pore cal contaminants.
sizes and characteristics. The ion exchange process, like
reverse osmosis, can be used with
lIuctuating flow rates. Pretreatment
262 Upgrading Existing or Designing New Drinking Water Treatment Facilities

w~h filtration might be needed ~ the in-


fluent has a high level of suspended
solids. Ion exchange units are also
sens~ive to the presence of competing
ions. For example, influent w~h high
levels of hardness will compete with
other cations (pos~ive ions) for space
on the exchange medium, and the ex-
change medium must be regenerated
more frequently.

Ion exchangers often use sodium


chloride to regenerate the exchange
medium because of the chemical's low
cost. However, this might res un in a
high sodium residual in the finished
water. High sodium residual might be
unacceptable for individuals with salt-
restricted diets. This problem can be
avoided by using other regenerant
materials, such as potassium chloride.

Ion exchange ef1ectively removes


more than 90 percent of barium, cad·
mium, chromium (III), silver, radium,
n~r~es, selenium, arsenic (V),
chromium (VI), and nitrate. Ion ex-
change is usually the best choice
for small systems to remove
radio nuclides.

Act/V.fed Alumina Figure 7·1. Two typee of reverae oamoala membranes.


Activated alumina systems are il-
lustrated in Figure 7-3. Activated
alumina is a commercially available
ion exchange medium, primarily used
to remove fluoride from ground water.
The activated alumina medium is
regenerated using a strong sodium
hydroxide solution. Because this in-
creases the pH level of the water, sul-
furic acid must be added to the water
after it leaves the exchange unit.

Activated alumina removes over 90


percent of arsenic (V), fluoride, and
selenium (IV), and 70 percent of
selenium (VI). n also ef1ectively
removes iron. n is not effective in
removing barium, cadmium, and
radium.

While activated alumina effectively


removes several contaminants, it can
be hazardous because of the strongly
acidic and basic solutions used. Figure 7·2. Ion exchange treatment Iyatlm.
Another disadvantage 01 activated
Control and Removal of Inorganic Contaminants 263

Maine. This process is applied in a


box-shaped, low profile vessel made
of high density polyethylene. It is
designed to remove volatile organic
chemicals as well as radon. Packed
tower aeration also is commonly used
to remove radon. The size of the pack-
ed lower required for radon removal
generally is much less than for or-
ganics removal.
r, ••unenl UI'ItI

Granular activated carbon (GAG) can


also effectively remove radon from
water. There are, however, concerns
TRE .... rME·IT ).NO e.-.CKIVI>SH aND
OO"""""LOW l=l;11'<SC vP.... OW RINSe. about worker safety and disposal of
carbon that is contaminated with
radon. 's

Aeration and GAC are discussed in


greater detail in Chapler 6, Treating
Organic Contaminants in Drinking
Water.

UPFLCW OQwNFLCw
RE.:teNe':"ATIOM I1EGE"'<fRATIOII

Figura 7·3. Activated alumina systems: Operating mode flow schematics.

alumina is the long contact time reo water. (Radon is a naturally occurring
quired (5 minutes, compared to 2 to 3 radioactive gas that contaminates
minutes for ion exchange). Finally, ac- ground water in some geographical
tivated alumina's costs are higher than areas.) These techniques include
those lor ion exchange. Waste open air storage with no mixing, a flow-
management might also increase through reservoir system with influent
costs because of high concentrations control devices, and a flow-through
of aluminum and other contaminants reservoir with bubble aeration. Initial
in the waste stream, as well as high studies found that minimal aeration
pH. applied during 30 hours of storage can
achieve more than 95 percent radon
Technologies (or Radon Removal: removal."
Aeration and Granular Activated
Carbon Another relatively low·cost aeration
Several 10w-eostl1ow-technology aera· technique is a multistaged diffused
tion techniques can effectively lower bubble aeration system manufactured
the concentration of radon in drinking by Lowry Engineering, Inc. of Unity,

" N.E. Kinner, C.E. Lessard. G.S. Schell, and K.R. Fox. "Low·CosVLow·Technology Aeration Techniques tor Removing Radon from Drink-
ing Waler.· Environmental Research Brief. U.S. Environmental Protection Agency. Office ot Research and Development, September 1987.
EPA/600/M-871031.

15 N.E. Kinner. C.E. Lessard, and G.S. Schell, Radon Removal/rom Small Community Warer Supplies Using Granular Activated Carbon
and Low Technology/Low Cost Techniques. U.S. Environmental Protection Agency Cooperative Research Agreement CR-81-2602-{)1-{l.
8. Resources

Safe Drinking Water Hotline capacity of rural community ollicials to


solve local water problems. It provides
1-800-426-4791
onsite technical assistance, training,
1-202-382-5533
and publications, and works to im-
This hotline, run by the U.S. Environ- prove federal and state government
mental Prot9Clion Aldency, provides responsiveness to the needs of rural
information on drinking water regula- communities. Table 8·3 lists the six
tions, polici9S, and documents to the RCAP regional agencies.
public, state and IocaJ government,
public wate, systems, and consu~ants. Farmers Home Administration
The Safe Drinking Water Hotline's (FmHA)
hours are 8:30 a.m. to 4:30 p.m. East- The Farmers Home Administration
ern Standard Time, Monday through provides grants and loans for rural
Friday excluding holidays. water systems and communities with
populations less than 25,000. Contact
U.S. Environmental Protection FmHA at USDAlFmHA, 14th and Inde-
Agency Reglonel Offices pendence Avenue SW, Washington.
Regional ollic9S of the U.S. Environ- DC 20250 (202-447-4323).
mental Prot9Clion Aldency are listed in
Table 8-1. Publications
General
Stet. Drinking Wet.r Agencies
American Water Works Association.
State agenci9S responsible for public
Basic Management Principles for
water supervision are listed in Table 8-2.
Small Water Systems. Denv'Jr, Co.
1982.
Orgenlzatlons Assisting Small
Systems American Water Works Association.
American Wet.r Worle. Assocletlon Design and Construction of Small
(AWWA) Small Systems Progrem Water Systems-A Guide for
Managers. 1984.
This program provides information,
training, and technical assistance to
American Water Works Association.
small systems. in coordination with
Introduction to Water Treatment. Den·
state regulatory agencies and other
ver. CO. 1984.
organizations assisting small systems.
Contact the AWWA at 6666 W. Quincy
Concern. Inc. Drinking Water: A Com·
Avenue, Denver, CO 80235 (303·794-
munity Action Guide. Washington. DC
7711) for the name of a contact for the
(1794 Columbia Road. NW. Washing·
small systems program in your area.
ton, DC 20009), December 1986.

Nellonel Runl Weter AssoclBtlon National Rural Water Association.


(NRWA)
Water System Decision Makers: An
This organization provides training Introduction to Water System Opera·
and technical assistance to small sys· tion and Maintenance. Duncan. OK.
tems. Contact the NRWA ollice at P.O. 1988.
Box 1428. Duncan. OK 73534 (405-
252-0629) for national information and Opflow. A monthly publication of the
the name of your local contact. American Water Works Association
focusing on the "nuls and bolts- can·
Rural Community Assistance cerns of treatment plant operators.
Program (ReAP)
This program consists of six regional
agencies formed to develop the

264
Resources 265

Table 8·1. EPA Regional Offices

EPA Headquarters
401 M Street. SW
Washing1on, DC 20460
202-382-5043

EPA Region 1 EPA Region 6


JFK Federal Building 1445 Ross Avenue
Boston, MA 02203 Dallas. TX 75202
61 7-565-3424 214·655·2200
Connecticut, Massachusetts. Arkansas. Louisiana. New Mexico,
Maine. New Hampshire. Oklahoma, Texas
Rhode Island. Vermont
EPA Region 7
EPA Region 2
726 Minnesota Avenue
26 Federal Plaza Kansas City, KS 66101
New York. NY 10278 913-236-2803
21 2-264-2515
Iowa. Kansas, Missouri, Nebraska
New Jersey, New York.
Puerto Rico. Virgin Islands
EPA Region 8
One Denver Place
EPA Region 3
999 18th Street, Suite 1300
841 Chestnut Sireet Denver, CO 80202
Philadelphia. PA 19107 303-293-1692
215-597-9370
Colorado, Montana, North Dakota.
Delaware, Maryland, Pennsylvania.
South Dakota, Utah. Wyoming
Virginia, West Virginia.
District of Columbia
EPA Region 9
EPA Region 4 215 Fremont Street
San Francisco, CA 94105
345 Courtland Street, NE
415-974-8083
Atlanta. GA 30365
404-34 7-3004 Anzona. California, Hawaii. Nevada.
American Samoa. Guam,
Alabama. Flonda. Georgia,
Trust Territories of the Pacific
Kentucky, MississippI.
North Carolina, South Carolina,
Tennessee EPA Region 10
1200 Sixth Avenue
EPA Region 5 Seanle, WA 98101
206-442-1465
230 South Dearborn Street
Chicago, IL 60604 Alaska. Idaho, Oregon, Washington
312-353-2000

Illinois. Indiana. Ohio. Michigan


Minnesota, Wisconsin
266 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Table 8·2. Slale Drinking Water Agencies

Region I
Bureau of Public Water Supply Protection
Connecticut Department of Health Services
New York State Department of Heahh
Water Supplies Section
2 University Place
150 Washington Street
Hartford, CT 06106 Western Avenue, Room 406
Albany, NY 12203-3313
203-566-1251
518-458-6731
Division of Water Supply
Department of Environmental Protection Water Supply Supervision Program
One Winter Street, 9th Floor Puerto Rico Department of Health
Boston, MA 02108 P.O. Box 70184
617-292-5529 San Juan, PR 00936
809-766-1616
Drinking Water Program
Division of Health Engineering Planning and Natural Resources
Maine Department of Human Services Government of Virgin Islands
State House (STA 10) Nifky Center, Suite 231
Augusta, ME 04333 SI. Thomas, Virgin Islands 00802
207-289-3826
Region (((
Water Supply Engineering Bureau Office of Sanitary Engineering
Department of Environmental Services Delaware Division of Public Heahh
P.O. Box 95, Hazen Drive Cooper Building
Concord, NH 03302-0095 P. O. Box 637
603-271-3503 Dover, DE 19903
302-736-4731
Division of Drinking Water Quality
Rhode Island Department of Health Water Supply Program
75 Davis Street, Cannon Building Maryland Department of the Environment
Providence, RI 02908 Point Breeze Building 40, Room 8L
401-277-6867 2500 Broening Highway
Dundalk, MD 27224
Water Supply Program 3Ot-631-3702
Vermont Department of Health
60 Main Street Water Hygiene Branch
P.O. Box 70 Department of Consumer and Regulatory Affairs
Burlington, VT 05402 5010 Overkiok Avenue, SW
802·863-7220 Washington, DC 20032
202-767-7370
Region II
Bureau of Safe Drinking Water Division of Water Supplies
Division of Water Resources Pennsylvania Department of Environmental
New Jersey Department of Resources
Environmental Protection P.O. Box 2357
P.O. Box CN-029 Harrisburg, PA 17105-2357
Trenton, NJ 06825 717-787-9035
609·984-7945

(continued on next page)


Resources 267

Table 8·2. Slale Drinking Waler Agencies (continued)

Evironmental Engineering Division Public Water Supply Section


Office of Environmental Health Services Division ot Environmental Health
State uepartment of Health Department of Environment, Health and
Caprtal Complex Building 3, Room 550 Natural Resources
1900 Kanawha Blvd., East P.O. Box 27687
Charleston, WV 25305 Raleigh, NC 27611-7687
304-348-2981 919-733-2321

Division of Water Supply Engineering Bureau of Drinking Water Protect


Virginia Department of Health Department of Health and
James Madison Builping Environmental Control
109 Governor Street 2600 Bull Street
Richmond. VA 23219 Columbia, SC 29201
804-786-1 766 803-734-5310

Region IV Division of Water Supply


Tennessee Department of Health
Water Supply Branch
and Environment
Department of Environmental Management
150 Ninth Avenue, North
1751 Congressional W. L. Dickinson Drive
Terra Building, 1st Floor
Montgomery, AL 36130
Nashville. TN 37219-5404
205-271-7773
615-741·6636
Drinking Water Section
Region V
Department of Environmental Regulation
Twin Towers Office Building Division of Public Water Supplies
2600 Blair Stone Road Illinois Environmental Protection Agency
Tallahassee, FL 32399-2400 2200 Churchill Road
904-48 7-'779 P.O. Box 19276
Springfield. IL 62794·9276
Drinking Water Program 217-785-8653
Georgia Environmental Protection DiviSion
Floyd Towers East, Room 1066 Public Water Supply Section
205 Butler Street. S.E. Office of Water Management
Atlanta. GA 30334 Indiana Department of Environmental Manage·
404·656-5660 ment
105 South Meridian
Drinking Water Branch P.O. Box 6015
Division of Water Indianapolis, IN 46206
Depanment of Environmental Protection 317-633-0174
18 Reilly Road, Frankfort Office Park
Frankfort, KY 40601 Division of Waler Supply
502-564-3410 Michigan Department of Public Heallh
P.O. Box 30195
Division of Water Supply LanSing. MI 48909
Stale Board of Health 517 -335-8318
PO. Box 1700
Jackson, MS 39215-1700 Minnesota Department of Health
601-354-6616/490-4211 Section of Waler Supply and Well Management
DiviSion of Environmental Health
925 S.E. Delaware Street
P.O. Box 59040
Minneapolis, MN 55459-0040
612-627-5170

(continued on next page)


268 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Table 8-2. State Drinking Water Agencies (continued)

Division of Public Drinking Water 515-281-8998


Ohio Environmental Protection Agency Public Water Supply Section
1800 WaterMark Drive Bureau of Water
P.O. Box 1049 Kansas Department of Health and Environment
Columbus,OH 43266-0149 Forbes Field, Building 740
614-644-2752 Topeka. KS 66620
913-296-1500
Bureau of Water Supply
Department of Natural Resources Public Drinking Water Program
P.O. Box 7921 Division of Environmental Quality
Madison. WI 53707 Missouri Department of Natural Resources
608·267·7651 P.O. Box 176
Jefferson City, MO 65102
Region VI 314·751·5331
Division of Engineering
Division of Drinking Water and Environmental
Arkansas Department of Health
Sanitation
4815 West Markham Street - Mail Slot 37
Nebraska Department of Health
Linle Rock, AR 72205-3867
301 Sentenial Mall South
501·661-2000
P.O. Box 95007, 3rd Floor
Lincoln. NE 68509
Office of Public Health
402-471·2541
Louisiana Department of Health and Hospitals
P.O. Box 60630
Region VIII
New Orleans. LA 70160
504-568-5105 Drinking Water Program
Colorado Department of Health
Drinking Water Section 4210 East 11th Avenue
New Mexico Health and Environment Department Denver. CO 80220
1190 St. Francis Drive 303-320-8333
Room South 2058
Santa Fe, NM 87503 Water Quality Bureau
505-827-2778 Department of Health and Environmental Scien·
ces
Water Quality Service Cogswell Building. Room A206
Oklahoma State Department of Health Helena. MT 59620
P.O. Box 53551 406-444-2406
Oklahoma City. OK 73152
405-271-5204 Division of Water Supply and Pollution Control
ND State Department of Health and Consolidated
Bureau of Environmental Health Laboratories
Texas Department of Health 1200 Missouri Avenue
1100 W. 49th Street P. O. Box 5520
Austin. TX 78756-3199 Bismark. ND 58502-5520
512-458-7533 702-224-2370

Region VII Office of Drinking Water


Department of Water and Natural Resources
Surface and Groundwater Protection Bureau
Joe Foss Building
EnVIronmental Protection Division
523 East Capital Avenue
Iowa Department of Natural Resources
Pierre. SO 57501
Wallace Slate Office Building
605-773-3151
900 East Grand Street
Des Moines. IA 50319

(continued on next page)


Resources 269

Table 8-2. Slate Drinking Water Agencies (continued)

Bureau of Drinking WaterlSanitation Marshall Islands Environmental Protection


Utah Department of Health Authority
P.O. Box 16690 P.O. Box 1322
Sak lake City, UT 84116-0690 Majuro, Marshall Islands 96960
801-538-6159 Via Honolulu

DEQ - Water Quality Government of the Federated States of Micronesia


Herschler Building, 4 West Department of Human Resources
122 West 25th Street Kolonia, Pohnpei 96941
Cheyenne, VVY 82002
307-777-7781 Palau Environmental Quality Protec1ion Board
Hosp~al
Region IX Koror, Palau 96940
Field Services Section
Region X
Office of Waler Quality
2655 East Magnolia Street Alaska Drinking Water Program
Phoenix, AR 85034 Wastewater and Water Treatment Section
602-257-2305 Department of Environmental Conservation
P.O. Box 0
Office of Drinking Water Juneau, AK 99811-1800
California Department of Heakh Services 907-465-2653
714 P Street, Room 692
Sacramento, CA 95814 Bureau of Water Quality
916-323-6111 Division of Environmental Quality
Idaho Department of Health
Safe Drinking Water Branch and Wenare
Environmental Management Division Statehouse Mail
P.O. Box 3378 Boise, ID 83720
Honolulu, HI 96801-9984 208-334-5867
808-548-4682
Drinking Water Program
Public Health Engineering Department of Human Resources
Nevada Department of Human Resources Heakh Division
Consumer Heakh Protection Services 1400 S.w. 5th Avenue, Room 608
505 East King Street, Room 103 Portland, OR 97201
Carson City, NV 89710 503-229-6310
702-885-4750
Drinking Water Section
Guam Environmental Protec1ion Agency Department of Heakh
Government of Guam Mail Stop lD-11, Building 3
Harmon Plaza Complex Un~ D-107 Airdustrial Park
130 Rojas Street Olympia, WA 98504
Harmon, Guam 96911 206-753-5954

Division of Environmental Quality


Commonweakh of the Northern Mariana Islands
P.O. Box 1304
Saipan, CM 96950
670-322-9355
270 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Table 8·3. Rural Community Assistance Program (RCAP) Agencies Flnratlon


American Water Works Association
Research Foundation. Manual of
Community Resources Group•. Ioc. Design for Slow Sand Filtration. (To be
2705 Chapman published Fall 1990.)
Springdale. AR 72764
501-756-2900 Huisman. L. and Wood. w.E. Slow
Sand Filtration. World Hea~h Or-
Great Lakes Rural Network ganization. Geneva. 1974.
109 South Front Street
Freemont. OH 43420 Slezak. L.A. and Sims. R.C. "The Ap-
~9-334-8911 plication and Effectiveness of Slow
Sand Fi~ration in the United States."
Midwest Assistance Program. Inc. JoumaIAWWA, 76:1238-43.1984.
P.O. Box 81
New Prague. MN 56071 Visscher. J.T.• Paramasivam. R.•
612-758-4334 Raman. A.. and Heijnen. HA Slow
Sand Filtration for Community Water
Rural Community Assistance Corporation Supply. Technical Paper 24. Interna-
2125 19th Street. Suite 203 tional Reference Centre for Com-
Sacramento. CA 95818 munity Water Supply and Sanitation.
916-447-2854 The Hague. The Netherlands. 1987.

Rural Housing Improvement. Inc. Dlsfnfectlon


218 Central Street. Box 429 American Water Works Association.
Winchendon. MA 01475-0429 Water Chlorination Principles and
617-297-1376 Practices (M20). 1973.

Virginia Water Project. Inc. SMC Martin. Inc. Microorganism


Southeastern Rural Community Removal for Smafl Water Systems.
Assistance Program
EPA 570/9-83-012. Valley Forge. PA.
702 Shenandoah Avenue. NW
June 1983.
P.O. Box 2868
Roanoke. VA 24001 Corrosion Control
703·345-6781
U.S. Environmental Protection Agen·
cy. Office of Drinking Water. Corrosion
Manual for Intemaf Corrosion of Water
Distribution Systems. EPA 57019-84.
Publications (continued) Surlace Water Sources. EPA 570/9-89-
DOt. April t 984.
Schautz. Jane W. The Self-Help 018. October 1989.
Handbook. This manual gives specific Economic and Engineering Services.
U.S. Environmental Protection
guidelines and techniques for estab- Lead Control Strategies. American
lishing seij-help projects (projects Agency. Office of Drinking Water.
Waler Works Association Research
where the community does some of Manuaf of Individuaf Water Supply
Foundation. Denver. CO. 1989.
the work itseH to save money). Focus Systems. EPA 57019-82-004. October
is on improving or creating water and 1982.
Radlonuclfde Removal
wastewater systems in small rural
communities. For ordering information.
Samplfng Kinner. N.E.. Lessar. C.E .• Schell.
contact: Rensselaerville Institute. G.S.• and Fox. K.R.. "Low Costllow-
U.S. Environmental Protection Agen-
Rensselaerville. NY 12147 (518-797- Technology Aeration Techniques for
cy. Office of Research and Develop-
3783). Removing Radon from Drinking
ment. Handbook for Sampling and
Water." EPA/600/M-87-031. U.S.
Sample Preservation of Water and
U.S. Environmental Protection Environmental Protection Agency.
Wastewater. EPA 600/4-82-029. Sep-
Agency. Office of Drinking Water. tember 1982. Office of Research and Development.
Guidance Manual for Compliance with September t987.
the Filtration and Disinfection Require-
ments for Public Water Systems Using
Resources 271

SMC Martin. Inc. Radionuclide Consultants


Removal for Small Public Water Sys- Directory-Professional Engineers in
tems. EPA 57019-83-01 O. Valley
Private Practice. Published by the
Forge, PA. June 1983. National Society of Professional
Engineers. Contact SPE Order Depart-
Wellhead Protsctlon
ment, 1420 King Street, Alexandria.
U.S. Environmental Protection Agency, VA22314.
Office of Ground-Water Protection.
Wellhead ProtBetion: A DBeision Who's Who in Environmental Engineer-
Maker's Guide. 1987. ing. Published by the American
Academy of Environmental Engineers.
U.S. Environmental Protection Agency, Contact the American Academy of
Office of Ground-Water Protection. Environmental Engineers, 132 Holiday
Developing a State Wellhead Protec- Court. Suite 206. Annapolis. MD
tion Program: A User's Guide to Assist 21401.
State Agencies Under the Safe Drink-
ing Water Act. 1988. The Federal Register
The Federal Register is published
U.S. Environmental Protection Agency,
daily to make available to the public
Office of Ground-Water Protection.
regulations and legal notices issued
Ground-Water Protection Document
by federal agencies. It is distributed by
Request Form.
the U.S. Government Frinting Office.
Washington. DC 20402. To order
Costs/Financial Management
copies. call 1-202- 783·3238.
American Water Works Association.
Water Utility Capital Financing (M 29).
1988.
A wide variety of publications on
Gumerman, R.C., Burris, B.E.. and specific topics of concern to water sys-
Hansen, S.P. Estimation of Small Sys- lems is available from the American
tem Water Treatment Costs. Final Water Works Association and the Na·
Report. CulpiWesner/Culp. Municipal tional Rural Water Association.
Environmental Research Lab, Cincin-
nati, OH. 1984.

U.S. Environmental Protection Agency,


Office of Water. A Water and Waste-
water Manager's Guide for Staying
Financially Healthy. EPA 430109-89-
004. July 1989.

U.S. Environmental Protection Agency,


Office of Ground-Water Protection.
Local Financing for Wellhead Protec-
tion.1989.
Appendix A-How to Take Bacteriological Samples

Routine and special bacteriological 5. Without adjusting the flow, fill the
samples must be taken in acoordanee sample container, leaving about 20
with established procedures to prevent percent air space at the top.
accidental contamination, and Replace the cap immediately. ~ the
analyzed by an EPA· or state-certijied sample is taken incorrectly, take
laboratory. The laboratory will usually another sample container --<lo not
provide specially prepared sampling reuse the original bonle.
containers, properly sterilized and con·
taining sodium thiosulfate to destroy 6. Take a second sample and
any remaining chlorine. The following measure the concentration of the
steps should be followed in colijorm disinfectant and record relevant in-
sampling: formation (date, time, concentra·
tion, place, sampler, etc.).
1. Use only containers that are
provided by the bacteriological 7. Package the bacteriological sample
laboratory and that have been for delivery to the laboratory.
prepared for col~orm sampling. Fol· Record all pertinent field informa-
low all instructions for sample con· tion on a form and on the sample
tainer handling and storage. container label.

The containers are sterile. Do 8. Samples must be cool during ship·


not open them before u.. and do ment to the laboratory. Use insu-
not rln.. them. lated boxes tor shipping containers
ij needed, or refrigerate during
2. Take samples at the consumer's transit.
faucet, but avoid:
Do not allow more than 30 hours be·
• Faucets with aerators (unless
tween sampling and test times.
removed) or swivel spouts

• Taps inside homes served by Be sure the laboratory can process


home waler treatment units the samples immediately upon receipt.
such as water softeners

• Locations where the water


enters separate storage tanks

• Leaking faucets that permit


water to run over the outside of
the faucet

3. Always allow the water to flow


moderately from a faucet 2 or 3
minutes before taking the sample.

4. Hold the sample container at the


base, keeping hands away from
the container neck. Be sure the in·
side of the container cap is
protected and does not touch
anything.

SoUrce: SMC Martin. Inc., Miaootganism Removal for Small Warer Systems. June 1983.
EPA 57~3-010.

272
Appendix B-Checklist: Some Factors Affecting
Water Treatment System Performance

Admlnlltlltlon 5. Are equipment relerence sources


1. Does the manager have lirst-hand available (such as operation and
knowledge 01 plant needs through maintenance manuals. equipment
plant visits and discussions with catalogs. etc.)?
operators?
6. Does the plant staff have neces-
2. Are there long-range plans lor sary expenise to keep equipment
lacility replacement. altemalive operating and to make equipment
source walers. emergency repairs when necessary?
response, etc.?
7. Are technical resources (such as
3. Is there an adequate number 01 per- equipment suppliers or contract ser-
vice) available to provide guidance
sonnelto accomplish necessary
operational activities? tor repairing. maintaining, or install·
ing equipment?
4. Are staff adequately trained and
able to make proper operation and 8. Are old or outdated pieces of equip-
maintenance decisions? ment replaced as necessary to
prevent excessive equipment
5. Are adequate lunds available for downtime <If ineNicient process per-
spare pans, improvements or re- formance/reliability?
placement 01 equipment. required
chemicals. etc.?
Design
6. Are the plant unit processes ade- 1. Is the plant design adequate for
quate to meet the demand for raw water quality (e.g., turbidity,
finished water? temperature. seasonal variation.
etc.)?
7. Is the staff aware of the potential
sources of contamination that might 2. Do facilities exist to control raw
affect the drinking water supply and water quality entering the plant
the available management (9.g., can intake levels be varied,
methods? can chemicals be added to control
aquatic growth. do watershed
management practices adequately
Maintenance protect raw water quality)?
1. Is there an effective scheduling and
recording procedure to prevent 3. Are the size of filters and type,
equipment failures. excessive depth. and eNective size of filtration
downtime. etc. resu~ing in plant per- media adequate? lire the surface
formance or reliability problems? wash and backwash facilities ade-
quate to maintain a clear filter bed?
2. Is the spare pans inventory ade-
quate to prevent long delays in 4. Are design features of the disinfec-
equipment repairs? tion system adequate (proper
mixing, detention time, feed rates.
3. Are procedures available to initiate proportional feed, etc.)?
maintenance activities on equip-
ment operating irregularities? Are 5. Are sludge facilities and size of the
emergency response procedures in sludge disposal area adequate?
place to protect process needs ~
critical equipment breaks down? 6. Do process control features provide
adequate measurement of.plant
4. Are good housekeeping proce- flow rate. backwash flow rate, filtra-
dures followed? tion rate. and flocculation mixing in-

273
274 Upgrading Existing or Designing New Drinking Water Treatment Facilities

puts? Do chemical feed facilities


provide adjustable feed ranges that
are easily set for operation at all re-
quired dosages? Are chemical
feed rates easily measured?

7. Are automatic monitoring or control


devices used where needed to
avoid excessive operator time for
process control and monitoring?

8. Are standby units for key equip-


ment available to maintain process
perlormance during breakdown or
during preventive maintenance ac-
tivities?

Operation
,. Are plant and distribution monitor-
ing tests representative ot parlor-
mance?

2. Is the proper process controltest-


ing perlormed to support opera-
tional control decisions?

3. Does the plant staff have sufficient


understanding of water treatment
process control testing and plant
needs to make proper process con-
trol adjustments?

4. Has the plant staff received ap-


propriate operational information
from technical resources (e.g ..
deSign engineer. equipment repre-
sentative. state trainer or inspector)
to enable them to make proper
operational decisions?

5. Does the operalion and main·


tenance manual/procedure provide
appropriate gUidance tor operation-
al decisions? Do operators utilize
the manual?

6. Are distribution system operating


procedures adequate to proteC1the
integrity of finished water (e.g ..
flushing, reservoir management)?

Source: Adapted from U.S. Environmental Protection Agency, Office 01 Research and Development and Center for Environmental Re-
search Information. Summary Repor!.· Optimizing Water Treatment Plant Performance with the Composire Correction Program, March
1990. EPA 62518-90/017
Appendix C-Selecting a Consulting Engineer

Selec1ing the right consultant involves


• The person in your town to con-
the following steps: tact for additional information

1. ldentlfylng potential engineering • Standards by which the


firma. Start by drawing up a list of proposals will be judged
at least five firms that might be able
to maet your needs. Sources of • The place and time the proposal
names include your own past eX- must be submitted
perience or that of neighboring
3. Intarvlawlng candldata engineer.
towns, lists maintained by your
Ing firms. When you receive the
state drinking water agency, and
proposals, check to see ~ they
suggestions received from the local
maet your judging standards, and
Rural Community Assistance
are within an acceptabl. cost
Program. Local professional lIn-
range. From those that meet the
gineering societies may be able to
standards, select three or four and
provide lists of members who spe-
interview each firm individually.
cialize in drinking water treatment
war!<. The National Society of
The following criteria may be help-
Professional Engineers and the
ful in evaluating engineering firms:
American Academy of Environmen-
tal Engineers have lists of their • Small town experience. Does
members available (5ge Chapter 8, the firm have experience with
Resources). communilies like yours? Which
towns have they wor!<ed with in
2. Issuing a Raquest for Proposals.
the recent past?
Notify engineering firms that you
are interested in their services. One • System design experience.
good way to do this is by preparing Does the firm have experience
a Request for Proposals (RFP). In in designing systems for small
your RFP, briefly describe your communities? What types of
'"lown's water treatment problem systems has the firm actually
and request proposals from consult- recommended. designed, and in-
anls on how thay would solve it. stalled? When were they in-
stalled? How are these systems
Depending on your community's wor!<ing? What were the es-
size and the nature of your timated costs? Whal are the
problem, the RFP may be a leiter to present operation and main-
Ihe engineering firms on your list or tenance needs and costs of
it may be a longer, more formal these systems? What systems
document. You may wish to adver- has the firm recommended for
tise your RFP. In any case, it communities that are most like
should include at least the following: your own? Ask for the cost per
dwelling serviced, the up-front
• A brief description of the problem assessments, and monthly char-
• A statement telling what it is you ges for the last few projects of a
want the consulting firm to do size and technology comparable
to your situation.
• The deadline by which your
town must receive the proposal

Sources: Adapted from U.S. Environmental Protection Agency. Office of Municipal Pollution
Control, Irs YOLK ChoioB: A Guidebook tor Local Officials on Small Community Wastewater
Management Options, September 1987. EPA 43019-87-006: SMC Martin. Inc., Microor·
ganism Removal tor Small Walllr Systems, June 1983. EPA 57019-83-012.

275
276 Upgrading Existing or Designing New Drinking Water Treatment Facilities

• Experience with financial In~ firm propose for completing your representatives of the drinking
stltutlons and funding agen- work? Does the firm use sub- water treatment facility, and be-
cies. What experience has the contractors for certain work? fI come familiar with the uniquefea-
firm had in helping communities so, which firms and for what tures and requirements of the
get financing from commercial work? utility, as well as the responsive-
sources (banks, bond sales)? ness of regional chemical suppliers
• Cost oillngineering work. Be
What experience has the firm and equipment vendors.
prepared to pay for good en-
had in dealing with state grant
gineering work. Do not choose
or loan programs or Farmers • Conduct sufficient laboratory and
your engineer only on the basis pilot plant studies and observations
Home Administration grant and
of cost. It is well worth spending
loan programs? What ex- of the source waters to fully charac-
a linle extra to get an engineer terize them. New facilities should
perience has the firm had in
who will design a system that be adequate to handle the full
working with lending institutions
will provide service at lower cost range of expected water condi-
or financial consultants?
for years to come. Ask the en-
tions, including foreseeable water
• Experience with state and gineer to briefly explain the
quality deterioration.
county agencies. What ex- firm's estimated fee. Make sure
perience does the firm have in you understand exaCtly what ser- • Initiate the design process with a
working wrth the state and county vices will be provided. Is there a thorough review of all possible non-
environmental agencies, the distinc1ion between basic ser- treatment or minimal treatment ap-
health department, etc.? vices and additional services? proaches. Consider potential ease
What circumstances could sig- of maintenance, adequate space
• Willingness to work with the nificanlly change the estimate? and light, and simplicity in the
community. If your community
design and equipment. Avoid over·
came up with a range of accept- 4. Checking references. Be sure to ly elaborate control systems. and in-
able user costs, would the en- check references for the firms you clude appropriate redundance (i.e ..
gineer be willing to use these thought were best. Talk to repre- never only one chlorinator).
estimates as guidelines to sentatives from communrties the
design a drinking water treat- firm has recently worked for. Ask • Avoid dead ends in the distribution
ment system? How does the about the overall experience, system. Provide equipment for
firm plan to handle public par- problems or special situations that flushing and sampling. lor storage,
ticipation in this project? arose, delays, etc. and for emergency chlorination of
• Willingness to work for the the distribution system.
5. Selecting a firm and contracting
community. Does the firm have
for Its services. The final selection • Allow operating personnel to par-
any experience in using tech-
of a firm involves evaluating all the ticipate in design decisions and ob-
nologies and maintenance
information you have gathered. serve construction progress.
programs that are differenl from
Once you have selected a firm, you
what the state and county agen·
must negotiate and sign a contract • Prepare operation and main-
cies have traditionally ac-
for their services. The form of this tenance manuals. which include
cepted? Does the firm have the
contract and the payment may be the following information:
willingness and capability to util·
governed by the method your town
ize innovative or alternative tech-
will use to finance this part of your the original design concepts
nology where appropriale?
project. Be sure to consider this
(Some engineers have dealt - description and drawings of
aspect in your evaluation. When
only WIth large centralized treal· the facility as constructed
that is done. you are ready to begin
menl systems and might not be
working with the engineer to normal operational procedures
familiar or experienced with
evaluate and solve your town's
other alternatives.) emergency operational proce-
drinking water treatment problems.
dures
• Stall capabilities and
workload. What projects is the The consultant should do the following organized collection of
firm now working on and what to achieve the best system design and vendors'literature
new ones may be coming soon? to simplify the operator's job:
safety considerations and re-
Which people on their stafl will quirements
be devoted to your project? • Establish a high level of com·
What time schedule does the munication wrth the community and
Appendix C-Selecting a Consulting Engineer 277

schematics with all valves num·


bered to correspond to
detailed operational procedures
maintenance procedures

• Provide startup assistance and


training and followup engineering
services.
Appendix D-Chlorine Residual Monitoring

Chlorine in a most active ~ Mlilntalnlng a Re.ldual Entering


"lr. . residual chlorine"-iI stable only the System
in the absence of agitation, sunlight, The SWTR requires that a r9Sidual of
and certain organic and inorganic 0.2 mglL be maintained in the water
materials with which it can react. entering the distribution system at all
times. Continuous monitoring at the
Reactions of free residual chlorine entry point(s) to the distribution sys-
with chlorine demanding substances tem is required to ensure that a detec-
continue over long periods of time.
table residual is maintained. Any time
Therefore, the sample taken for disin-
the residual drops below 0.2 mg/L, the
fectant analysis should be analyzed system must notify the Primacy Agen-
immediately. Specially prepared sam- cye prior to the end of the next busi-
pling containers, property cleansed,
ness day. The system is in violation of
sterilized, and not containing sodium a treatment technique if the residual
thiosunate should be used. level is not restored to 0.2 mgll within
4 hours and fi~ration must be installed.
In general, the same sampling precau-
(n the Primacy Agency finds that the
tions described in Appendix A for
exceedence was caused by an un-
taking comorm samples should be ob-
usual and unpredictable circumstance,
served, but in addition:
it may choose not to require filtration.)
1. Draw the sample gently, avoiding In cases where the continuous
monitoring equipment fails, grab
agitation.
samples every 4 hours may be used
2. Analyze immediately in the shade for a period of 5 working days while
or subdued light. Do not store the the equipment is restored to operable
sample. conditions.

3. Do not use a bacteriological sam- The system must record, each day of
pling container, which may contain the month, the lowest disinfectant
a chemical to counteract or destroy residual entering the system and this
the disinfecting agent. a residual must not be less than 0.2
mgJL Systems serving less than or
Demonstration of Maintaining a equal to 3,300 pecple may take grab
b
Resldual samples in lieu of continuous monitor-
The Surtace Water Treatment Rule ing at the frequencies shown in the
(SWTR) establishes two requirements box below:
pertaining to the maintenance of a
residual. The first requirement is to
System Population Samples/day"
maintain a minimum residual of 0.2
<500 1
mgllentering the distribution system. 501-1.000 2
Also, a detectable residual must be 1,001-2,500 3
maintained throughout the distribution >2,501-3,000 4
system. These requirements are fur-
·Samples must be taken at dispersed
ther explained in the following sections. time intervals as approved by the
Primacy Agency.

• From SMC Martin. Inc., Microbiological Removal lor Small Water Systems. June 1983.
EPA 57019-83~10.
b Adapted from U.S. Environmental Protection Agency, Guidance for Compliance with the
Filtration and Disintectioo Requirements for Public Water Systems Using Surface Water
Sources, October 1989.
C The Primary Agency is a state with primary enforcement responsibility for public water sup-

plies. or EPA in the case 01 a state that has not obtained primacy.

278
Appendix D-Chlorine Residual Monitoring 279

~ the residual concentration falls APHA, AWWA, and WPCF. Wash- that may be exerting a disinfectant
below 0.2 mgil, another sample must ington, DC, 1985. d demand
be taken within 4 hours and sampling
continued at least every 4 hours until The SWTR requires that a detectable • Flushing and disinfection of the por-
the disinfectant residual is a minimum disinfectant residual be present in 95 tions of the distribution system in
of 0.2 mgil. percent or more of the monthly distribu- which a residual is not maintained
tion system samples. In systems that
Maintaining a Residual within the do not filter, a violation of this require- • Installation of satellite disinfection
System ment for 2 consecutive months caused feed facilities with booster
The SWTR also requires that a detec- by a deficiency in treating the source chlorinators within the distribution
table disinfectant residual be main- water will trigger a requirement for system
tained throughout the distribution filtration to be installed. Therefore, a
system that does not maintain a For systems unable to maintain a
system, with measurements taken at a
residual in 95 percent of the samples residual, the Primacy Agency may
minimum frequency equal to that re-
for 1 month because of treatment determine that it is not feasible for the
quired by the Total Coliform Rule (54
deficiencies, bU1 is maintaining a system to monitor HPCs and judge
FR 27543-27568). The same sampling
residual in 95 percent of the samples that disinfection is adequate based on
locations as required for the coliform
for the following month, will meet this site-specific conditions.
regulation must be used for taking the
disinfectant residual or HPC (hetero- requirement.
Additional information on maintaining
trophic plate count) samples. How-
The absence of a detectable disinfec- a residual in the system is available in
ever, for systems with both ground-
tant residual in the distribution system the American Water Works Assoc-
water and surlace water sources (or
may be due to a number of factors, iation's Manual of Water Supply Prac-
ground water under the direct in-
including: tices and Water Chlorination Principles
fluence of surlace water) entering the
and Practices.
distribution system, residuals may be
measured at points other than coliform • Insufficient chlorine applied at the
sampling points if these points are treatment plant
more representative of the disinfected
surlace water and allowed by the • Interruption of chlorination
Primacy Agency. An HPC level of less
than 500/mL is considered equivalent • A change in chlorine demand in
to a detectable residual for the pur- either the source water or the dis-
pose of determining compliance with tribution system
this requirement, since the absence 01
a disinfectant residual does not neces- • Long standing times and/or long
transmission distances
sarily indicate microbiological con-
tamination.
Available options for systems to cor-
rect the problem of low disinfectant
Disinfectant residual can be measured
residuals within their distnbution sys-
as total chlorine, free chlorine, com-
bined chlorine, or chlorine dioxide (or tem Include:
HPC level). The SWTR lists the ap-
proved analytical methods for these • Routine flushing
analyses. For example, several fest
methods can be used to test for • Increasing diSinfectant doses at the
chlorine residual in the water, including plant
amperometric titration. DPD
colorimetric method, DPD ferrous • Cleaning of the pipes (either
titrimetric method, and iodometric mechanicaily by pigging or by the
method, as described in the 16th Edi- addition of chemicals to dissolve
tion of Standard Methods for the Ex- the deposits) in the distnbution sys-
amination of Water and Wastewater, tem fo remove accumulated debris

dAlso, portable test kits are available that can be used in the field to detect residual upon approval of the Primacy Agency. These kits may
employ btrabon or colonmetnc lesl methods The colonmetTlc kits employ either a visual detection of a residual through the use 01 a color
wheel, or the detection of the residual through the use 01 a hand held spectrophotometer.
Appendix E-CT Values

CT Values for Achieving Inactivation of Viruses at pH 6 through 9


(In mg/L-m)

Log Temperatur.
lnaet!-
vaton o.S·C S'C 10'C 1S'C 2O'C 25'C

Free 2 6 4 3 2 1 1
chlorine- 3 9 6 4 3 2 1
4 12 8 6 4 3 2

Ozone 2 0.9 0.6 0.5 0.3 0.25 0.15


3 1.4 0.9 0.8 0.5 0.4 0.25
4 1.8 1.2 1.0 0.6 0.5 0.3

acT values include a safety fador of 3.

CT Values for Achieving 99.9 Percent Inactivation of


a
Giardia Lamblia

Temperatura

Disinfectant pH O.5"C S'C 10'C 1S'C 20'C 25'C

Free 6 165 116 87 58 44 29


chlorine b 7 236 165 124 83 62 41
8 346 243 182 122 91 61
9 500 353 265 177 132 88

Ozone 6-9 2.9 1.9 1.4 0.95 0.72 0.48

-rhese CT values for free chlorine, chlorine dioxide, and ozone will guaran-
tee greater than 99.99 percent inadivation of enteric viruses.

b CT values will vary depending on concentration of free chlorine. Values


indicated are for 2.0 mgJl of free chlorine. CT values for different free
chlorine concenlrations are spec~ied in tables in the EPA Guidance Manual
for Compliance with the Filtration and Disinf9Ction Requirements for Public
Water Systems Using Surface Wat9r Sourc9S.

280
Appendix E-CT Values 281

CT Values for Achieving 90 Percent Inactivation of Giardia Lamblia


. (in mg/L·m)

Temperlture

Disinfectant pH O.S'C S'C 10'C 1S'C 2O'C 25'C

Free 6 55 39 29 19 15 10
chlorine" 7 79 55 41 28 21 14
(2 mgil) 8 115 81 61 41 30 20
9 167 118 88 59 44 29

Ozone 6·9 0.97 0.63 0.48 0.32 0.24 0.16

aCT values will vary depending on concentration of free chlorine. Values indi-
cated are tor 2.0 mg/L of free chlorine. CT values tor different tree chlorine con-
centrations are specified In tables in the EPA Guidancti Manual for Compliance
wIth the FtI'ration and Disinfection Requirements for Public Water Systems
Using Surface Water Sources.
Appendix F-Sample of CT Calculation for Achieving
1-log Giardia, 2-log Virus Inactivation with Chlorine Disinfection

A SO,OOO GPD slow sand filtration The contact time multiplied by the con-
plant supplies a community of SOC centration (mgIL) of residual chlorine
people with drinking water Irom a in the water is the calculated CT value
reservoir in a protected watershed. for the system. Prover1 inactivation 01
The raw water supply has the lollow· Giardia and viruses are correlated to
ing characteristics: calculated CT values in EPA's
Guidance Manual for Compliance with
• Turbidity: S to 10 NlU the Filtration and Disinfection Require-
ments for Public Water Systems Using
• Total estimated Giardia cyst level: Surfacs Water Sources. (Appendix E
less than 1 per 100 mL contains excerpts from the CT tables
in the manual.)
• pH: 6.S to 7.S
The T,o lor the clearwell basin can be
• Temperature: S' to 1S'C determined by tracer studies. (Tracer
study procedures are described in
An overall removallinactivation of 3 EPA's Guidance Manual.)
logs lor Giardia and 4 logs lor viruses
is sufficient for this system. The On the day represented in this ex-
Primacy Agency credits t~e slow sand ample, the tracer study showed that
filter, which produces water with t\;r· the T,o lor the c1earwell was 40
bidity ranging Irom 0.6 to O.B NTU, minutes at the peak hourly flow rate.
with a 2-1og Giardia and virus removal. At this llow rate, water travels through
Disinfection must achieve an addition- the transmission main at 211 feet per
all-log Giardia and 2-1'Jg virus minute. The distance between the
removallinactivation to meet overall plant and the lirst customer is 1,000
treatment objectives. feel. Thus, the T,o for the distribution
main is 4.7 minutes (1,000 leet divided
To begin the calculations lor determin- by 211 leet per minute).
ing the adequacy of the inactivations
achieved by the disinfection system, Other data required for the calculation
the total contact time must be deter- are:
mined.
• Measured chlorine residual: 2.0
In this plant, chlorine for disinlection is mgil for the c1earwell basin and 1.2
added prior to the clearwell, which has mgil for the distribution main
a 2,000-gallon capacity. The distance
from the plant to the first customer is • Water temperature: S'C
bridged by a l,OOO·foot 2-inch trans-
mission main. The contact time • Water pH: 7.S
provided in both the clearwell basin
and the distribution pipe up to the first CT values required to achieve various
customer comprises the total contact levels 01 inactivation of Giardia and
time lor disinfection. viruses depending on the water
temperature, pH, and chlorine residual
In the calculations, contact time is rep- are provided in the Guidance Manual.
resented by T,o -the time needed lor The calculated CT values (CTcalc)
10 percent of the water to pass based on actual system data are com-
through the basin. In other words, T,o pared to the CT values in the
describes the time, in minutes, that 90 Guidance Manual (CT999 in the case
percent of the water remains in the 01 a I-log inactivation) to determine
basin. (For the distribution pipe, con- whether the inactivations achieved are
tact time is tOO percent of the time adequate.
that water remains in the pipe.)

282
Appendix F-Sample CT Calculation 283

Since, with free chlorine, a 1-log


Giardia inactivation provides greater
than a 4·log virus inactivation, inactiva-
tion of Giardia is the controlling factor
for determining overall reductions.

The calculation of CT and comparison


to CT values for 1-log inactivation of
Giardia provided in the EPA Guidance
Manual is shown in the box below.

For the basin:


CTcarc • Chlorine residual x contact time or
• 2.0 mg/L x 40 minutes. 80 mglL-min

From the EPA Guidance Manual, CT99.9 (3-log inactivation) is


200 mglL·min at S'C, 2 mg/L chlorine residual, and 7.5 pH.

CTcaldCT999. 80 mgll-min • 0.4


200 mg/L-min

For the distribution system:


CT calc • 1.2 mg/L x 4.7 minutes. 5.64 mgll-min

From the EPA Guidance Manual, CT999 is 183 mgll-min at S'C, 1.2 mgll
chlorine residual, and 7.5 pH.

CTcaldCT99 9. 5.64 mgiL-min .0.03


183 mgll-min

Summing CTcaldCT99.9 for both the basin and the main results in 0.43. This is
equivalent \0 a 1.29·109 Giardia inactivation determined by:

3 X CTcaldCT99.9
• 3 x 0.43 • 1.29 log

(This calculation is based on a 3-Iog inactivation; therefore, the ratio is


multiplied by 3.)

Thus, the 1.29-109 inactivation


achieved by diSinfection in this system
exceeds the 1-log additional inactiva-
tion required to meet overall treatment
objectives.
Part III

The Composite Correction Program for


Optimizing Water Treatment Plant Performance

The information in Part III is from Summary Report-


Optimizing Water Treatment Plant Performance with the
Composite Correction Program, prepared by Robert C.
Renner and Bob A. Hegg of Process Applications, Inc.
and Jon H. Bender of the U.S. Environmental Protection
Agency, Technical Support Division for the U.S. En-
vironmental Protection Agency. March 1990.

285
Acknowledgments

Although many individuals contributed to the preparation and review of this document, the
assistance of the individuals listed below is especially acknowledged. Those who assisted by
participating in the actual field studies described in the document are noted below as "Onsite
Field Participants."

Major Authors:
Robert C. Renner and Bob A. Hegg, Process Applications, Inc., Ft. Collins. Colorado
Jon H. Bender, EPA Technical Support Division (TSD). Cincinnati. Ohio

Technical Writing and Editing:


Heidi Schultz and Susan Richmond, Eastern Research Group. Inc., Arlington. Massachusetts

Project Managers:
Jon H. Bender, EPA TSD. Cincinnati, Ohio
James E. Smith. EPA Center for Environmental Research Information (CERI).
Cincinnati, Ohio

Reviewers:
Eric Bissonette, EPA TSD. Cincinnati. Ohio
Robert Blanco, EPA Office of Drinking Water (ODW), Washington. D.C.
Peter Cook. EPA ODW, Washington, D.C.
Dan Fraser, Montana Department of Health and Environmental Sciences (DHES).
Helena. Montana
Denis J. Lussier, EPA CERI. Cincinnati. Ohio
James Westrick. EPA TSD, Cincinnati, Ohio

Onalte Aeld Participants:

EPA
Jon H. Bender, EPA TSD, Cincinnati, Ohio
Eric Bissonette, EPA TSD. Cincinnati. Ohio
Dean Chaussee, EPA Region VIII, Helena, Montana
Ben W. Lykins. EPA, Office of Research and Development, Cincinnati, Ohio
James E. Smith. EPA CERI, Cincinnati, Ohio
James Westrick. EPA TSD, Cincinnati, Ohio

State of Montana
Dave Aune. Montana DHES, Helena. Montana
Jerry Bums. Montana DHES. Billings. Montana
Dan Fraser. Montana. DHES. Helena. Montana
Donna Howell. Montana DHES, Helena, Montana
Denise Ingman. Montana DHES, Helena. Montana
James Melsted, Montana DHES, Helena. Montana
Kate Miller. Montana DHES. Billings, Montana
Rick Rosa. Montana DHES, Helena. Montana
Mark Smith. Montana DHES, Helena. Montana
Roy Wells. Montana DHES, Helena. Montana

State of Ohio
Gary Cutler, Ohio EPA. Columbus. Ohio
Steve Severyn, Ohio EPA. Columbus, Ohio

State of Kentucky
Fred Cooper, Kentucky Department of Environmental Protection (DEP), Morehead, Kentucky
George Schureck. Kentucky DEP, Frankfort. Kentucky
Tom Stern, Kentucky DEP, Frankfort, Kentucky
Damon White. Kentucky DEP. Hazard. Kentucky

286
1. Introduction

Purpose Since 88 percent of the 60,000 community drinking


water systems in the United States are small systems
This document summarizes the results of an ongoing serving fewer than 3,300 individuals. the opportunity
project to evaluate the utility of the Composite for widespread impacts are large. These small
Correction Program (CCP) approach to improving the systems account for approximately 92 percent of the
performance of drinking water treatment facilities. The SDWA compliance problems reported each year. In
CCP approach, which has already proven successful 1987, more than 80 percent of the community drinking
when applied to wastewater treatment plants, is water systems experiencing significant compliance
described and the results of evaluating it at 13 difficulties were small systems.
drinking water plants to date are summarized.
Small systems frequently can neither readily identify
The 13 "case studies" focus on the potential for the and address the factors that cause their compliance
CCP approach to improve the performance of small problems nor easily finance the upgrading of their
drinking water systems in meeting the turbidity facilities. The staff may be inadequate in numbers,
removal requirements of the Surface Water Treatment experience and training to effectively solve the
Rule (SWTR). problems. Successful application of the CCP
approach can identify cost-effective measures that
The CCP approach is still under development. The can be taken to improve plant performance and
end product of this project will be a publication that comply with drinking water requirements.
describes the refined CCP approach and allows it to
be applied by others. The CCP approach is another tool that federal, state,
or local regulators. technical personnel, and
Background consultants familiar with the procedure can use to
identify and correct factors that limit a plant's
Many communities are now considering either performance, Results to date suggest that it is both
construction of new faCilities or major modifications to highly successful and cost effective.
existing ones to meet drinking water regulations, An
approach that allows communities to meet regulatory Content
requirements by implementing changes in their
operation, maintenance, and administration Section 2 of this document details the CCP approach,
procedures Instead of major capital improvements has including facility review, performance analysis, and
obVIOUS advantages. By maximizing the operational implementation of corrective measures. Section 3
efficiency of their facilities, local administrators can summarizes the results of the case studies.
both improve the ability of the facility to meet Safe highlighting specific instances where the CCP
Drinking Water Act (SDWA) requirements and approach revealed problems that were not previously
minimize the financial impact to the community obvious to drinking water treatment plant operators.
associated with major upgrades to the plant. Also highlighted are instances where the CCP
approach saved the facilities money that otherwise
Recognizing that the CCP approach had been would have been spent in plant modification, Section
successfully developed and applied to small 4 includes expanded information on the 13 case
wastewater treatment plants to accomplish the same studies that have been conducted to date should the
objectives, the State of Montana decided to evaluate reader desire additional information.
the potential of modifying it for use at small drinking
water plants. Based on the initial success of this
evaluation, U.S. EPA decided to further develop and
demonstrate the approach to ensure its applicability to
other parts of the country.

287
2. The Composite Correction Program Approach

The CCP approach consists of a Comprehensive maintenance records, and reviewing budgets.
Performance Evaluation (CPE) and a Composite Standardized forms are used to collect the data on
Correction Program (CCP). The CPE is a systematic raw and treated water quality. design and operating
step-by-step evaluation of an existing treatment plant conditions for individual plant processes. plant
resulting in a comprehensive assessment of the unit operator coverage. user fees for water treatment.
treatment process capabilities and the impact of the maintenance scheduling. and operating budgets.
operation. maintenance and administrative practices While the data collection efforts focus on the current
on performance of the plant. status of the plant. the review also includes past
records to account for factors such as seasonal
It is conducted by a team of individuals with variations In raw water quality and peak demand. and
knowledge of drinking water treatment and results in to establish an accurate record of plant performance.
the identification of a unique combinatIOn of factors
limiting plant performance. This team reviews and In addition to gathering existing data. the CPE may
analyzes the plant's physical capacity as well as Its involve collecting new data by conducting speCial
operational capability and associated maintenance and studies. For example. the CPE team usually develops
administration. Based on this analysIs. the team a turbidity vs. time profile on a plant's filters before
projects the capabilities of the major unit processes and alter back washing to determine whether the filters
within the plant. and identifies and prioritizes those were performing adequately (see Figure 2-2). At
factors affecting plant performance. nearly all plants. such a profile revealed that a
Significant breakthrough of turbidity occurred alter the
If the CPE indicates that optimization of existing major backwash. When the CPE team sampled the clearwell
unit processes can result in desired finished water at one end. they discovered turbidity values of 6.3
quality. the CCP phase IS implemented. The CCP NTU. which clearly exceeded the regulatory cnteria.
systematically addresses those factors identified and Other special studies conducted as part of the CPE
prioritized in the CPE phase. olten reveal similar performance problems that may
not be obvious to the plant staff.
Figure 2- t graphically illustrates the CPE CCP
approach. The CPE team usually is composed of two Design Components
individuals experienced in the design and operation of The ePE team determines a plant's capacity by
drinking water treatment faCilities and In trouble reviewing plant drawings and specifications, making
shooting their operation. Teams composed of up to field measurements. and reviewing information
seven individuals were employed for each of the 13 provided by the plant staff. In addition. the team
case studies described in this document. although It is applies ItS experience based on evaluations performed
anticipated that teams this large will not be required to at other plants. The ePE evaluators then determine
apply the finalized CPE:CCP approach. These larger the projected capacity at which plant major unit
teams were used to help evaluate and further refine processes (flocculation, sedimentation. filtration, and
the procedure as well as familiarize regulatory diSinfection) can provide acceptable treated water
personnel with It. quality. Projected values are compared with peak
Instantaneous operating capacity and current plant
The Comprehensive Performance production. The comparison results are summarized
using a performance potential graph (see Figure 2-3).
Evaluation which illustrates the strengths and weaknesses of the
The CPE begins with a plant tour and collection of plant's unit processes.
information from plant records. Data are obtained by
intervieWing plant stall and key administrative Operation and Maintenance
personnel (for example. the mayor and other city Operational factors are assessed by evaluating
administrators). reviewing the plant's physical procedures that the plant uses for process control
capacity, examining the plant's operation and adjustments and by determining if steps the plant

288
The Composite Correction Program Approach 289

Figure 2-1 CPElCCP schematic 01 activities.

AdminIstrators
Recognize Need To
Improve Plant Pertormance

CPE Evaluation
of
Major Unit Processes

Major Unit Processes Malor Unit Processes


Are Are
Adequate Inadequate

Implement CCP to Implement CCP to Do Not Implement CCP


AchIeve DesIred OptimIze EXIsting FaclliIJes Evaluate Options For
Pertormance Before Initiating FacIlity Modifications
From EXIsting FaClh!l8S FacIlity Mod,lJcations

Desired Pertormance AchIeved

takes to modify operations are based on proper factors that limit this performance. They use a
application of water treatment concepts and methods. checklist containing more than 65 performance-
The CPE team discusses process control measures in limiting factors (see Table 2-1) and define each factor
detail with plant operators. This enables them to according to its specific cause of poor plant
accurately assess the plant" s operation and to avoid performance. Once the factors have been identified.
any misunderstandings related to terminology. they are prioritized according to the magntlude of their
Maintenance capabilities are evaluated by reviewing adverse eHects on plant performance. This IS the
maintenance schedules and records. observing spare major output from a CPE: a priorttized list of
parts InventOries, observing the condition of plant performance limiting factors.
equipment. and discussing maintenance activities with
plant personnel. Reporting
The CPE team conducts an eXit meeting with
Administration administrative and operations personnel to
The CPE evaluators interview plant operators and communicate the results of the CPE directly to all
administrative personnel (for example, city managers. concerned. ThiS is followed up with a brief written
town clerks, water board oHiclals, etc.) to consider report. The purpose of the report is to summarize the
administrative factors such as staffing (including results of the CPE and list the priOritized factors
training, motivation, and morale), budgets. and rate limiting plant performance. A typical CPE report is 8 to
structures. 12 pages in length and addresses the following topics:

Evaluating the Factors that Limit Performance • Facility description


After Critically studying the plant design, performance. • Major untl process evaluation
maintenance, administration and operation, the CPE • Performance assessment
team assesses the performance of the plant and • Performance-limiting factors
conducts an In-depth analysis to identify the speCific • Projected impact of a CCP
290 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Figure 2-2. Flher effluent turbidity profile.

Turtlodity. NTU
40

30

20

10

Present ReqU"ements

o 10 15 20 25 30

Mtnutes

Figure 2·3. Sample performance potential graph.

Flow. mgd
Unit Process
02 04 06 08 10

Flocculation'
HOT. m,n 113 57 38 28 : 23 I
Sedimentation'
SOA. gpdlsq ~ 150 302 452 603; I
Flhratlon J
HLR. gpmJsq ~ 0.4 0.7 1.1 15 ; 1.8 2.2 2.5 2+

Olslnfe<:lion 4
Contact time. mIn 180 90 6<1

Peak Instantaneous Operattng Flow


Rate. One Pump • 300 gpm

I Raled al 20 m,n . assumes vallable soeed Orove woul<l be 3OOed.


, Rated at 750 gpdlsq ~ . 12.s·~ Oeplh c,scourages hrghe< raung.
J RatBO al 3 gpmlsq ~ . COnirol System conslderBO I,m,ung.
4 Rated at CT • 127 w,th 2.4 mgtl CI 2 aose. which requrres a 53·min HOT; CT base<l on 4 log required reduction· 2.5 log ,n plant 1.5
log c,s,nfect,on. pH • 8. temperature = S·C. Assumes 10'1'0 01 usable clearweli vOlume lor contact ume..
The Composite Correction Program Approach 291

Table 2·1. Performance-LI",ltlng Faetorl hble 2-1. Perfonnance-Umltlnll Faelors (continued)

ADMINISTRA TION DESIGN


Plant Administrators Raw Water
· Policies - THM precursors
· F am~lanty WIth plant needs • Turbidity
· SupetVlslon - Seasonal vanaiion
· Planning - Watershed/Reservo" management
Plant Staff Un,t DeSign Adequacy
· ManpOWer - Pretreatment
· number · intake structure
· plant coverage · pre-sedimentation basin
· war!< load distribution · pre-chlonnation
· personnel turnover · Low service pumping
· Morale - Flash mIX
· motlValKln · Flocculation
· pay - Sedimentation
· war!< enVIronment - Filtration
· Staff qualificalIOns - DISinfection
· aptitude - Sludge treatment
· level of education - Ultimate sludge disposal
• certification • Fluondatlon
· ProductiVIty Miscellaneous
Financial • Process fteXlb,lity
· InsuffiCient funding · Process controllability
· Unnecessary expenditures · Process automation
· Bond Indebtedness · Lack 01 standby units for key eqUipment
Waler Demand · Flow proportioning units
· Alarm systems
MAINTENANCE - Alternate power source
Prevenbve - Laboratory space and equipment
· Lack of program • Sample laps
· Spare par1S Inventory · Plant Inoperablltty due to weather
Corrective · Return process streams
· Procedures
· Cntical pans procurement OPERATION
General Testing
· Housekeeping • Perlormance monttonng
• ReferencltS aVailable - Process control testing
- Staff expertise Process Control Adiustments
• Technical gUidance · Waler veatment understandIng
· EqUipment age · Application of concepts and leStlng to process control
· Technical gUidance (operallOns)
• Training
• InsuffiCient time on Job
0&"1 ManuallProcedure
A CPE report does not recommend specific actions to - Adequacy
be taken to correct individual performance-limiting · Use
factors, since this could lead to a piecemeal rather D,stnbullOn System
than an integrated approach to corrective actions. MISCELLANEOUS
Corrective actions should be undertaken in the next
phase - the CCP - with the help of the CPE team or
similarly experienced individuals.

The Composite Correction Program program. A factor in anyone of the performance-


limiting areas (design, maintenance. administration.
The objective of this phase is to improve the and operation) can contribute to poor performance. It
performance of a drinking water treatment plant by is unlikely. however, that a single factor limits
implementing the findings of the CPE when it performance: rather it is usually a unique combination
indicates that the plant is likely to meet treatment of factors that causes poor water quality. Plant
requirements with the existing major unit processes. operators and administrators must understand the
The CCP focuses on systematically addressing the relationship among these areas and water treatment
factors that limit the plant in achieving the desired plant product water quality. It is the operation of the
finished water quality. plant that enables a phySically capable plant to
produce adequately treated water.
Implementing the Composite Correction Program
To successfully implement the CCP and achieve Maintaining Long-Term Involvement
improved performance, facilities must utilize the CPE One of the keys. as already noted, to successfully
results and implement a long-term process control implementing a CCP program is long-term effort
292 Upgrading Existing or Designing New Drinking Water Treatment Facilities

(typically involving several months to a year). Long- • Necessary changes in staff attitude may mean
term involvement is cntical for several reasons: personnel changes are needed. If the staff do not
support the CCP approach, the CCP will require
• Repeat training is more effective than one-time only additional effort and perhaps personnel changes to
training. Training should be conducted under a be successful.
variety of operating and administrative conditions
(for example, when seasonal water quality or • Time is required to identify and eliminate any
demand changes) in order for staff to develop additional performance-limiting factors found dunng
confidence in new techniques or procedures. the CCP.

• Time is required to make the necessary physical Since the goal of implementing the CCP is to correct
and procedural changes. This is especially true for performance-limiting factors until the desired water
any changes that require administrative approval or quality IS achieved, the details of the implementation
funding appropriations. often will be site-specific and, therefore, should be left
to the individuals implementing the CCP.
3. Results of Case Studies

Thirteen CPEs were conducted in 1988 and 1989: 11 interviews with the plant staff. While the operating
at conventional drinking water treatment facilities and records provided only a hint of a problem, the CPE
2 at facilities using direct filtration. Of the 13. 9 were team was able to pinpoint the problem by posing
completed in Montana. and 2 each were completed in directed questions to the staff.
Ohio and Kentucky. The plants ranged in size from
3.8 to 202 Ls (86.000-10,000.000 gpd). Table 3-1 • At Plant 3. a direct discharge of backwash water to
summanzes the design capacity and type of plants a stream was identified. This practice violated the
evaluated. Conventional plants are defined as using State's discharge regulations.
fl~~h mix, flocculation, sedimentation, filtratIon, and
c,'nfection un,t processes pnmarily for turbidity • At Plant 12, the CPE team learned that the
reo ·,.'val and disinfectIOn. dilapidated condition of the plant prvvented it from
providing acceptable finished water to the
Table 3·1. Summary of Plants Where CPEs Have Been community. While it originally appeared that the
Conducted
plant would not be able to afford the necessary
Plant No. DesIgn Capacity Process Type repairs, the CPE team's review of the plant's
7 mgd
operating budget and available resources led tile
Lime Sonenlng'/Convenllonal
community to believe that sufficient funds were in
2 3 mgd Conventional
fact available to repair the plant and to redirect
3 5 mgd Conventional priorities.
4 60 gpm Conven Iional
5 3 mgd Direct FiltratIon The case studies also clearly indicate that the
6 4 mgd Lime sonenlng'/Conven/lonaJ
involvement of community administrators is a critical
part of the CPE and, ultimately, to improving a plant's
10 mgd Conventional
performance. Administrators frequently had not been
8 250 gpm LIme sonen,ng'/Conven/lonaJ informed of previous inspection results and potential
9 650 gpm Direct F,lIrat,on or existing problems and, therefore, had not
to 350 gpm ConventIonal implemented remedial actions. In a CPE, the
300 gpm Conventional administrators are involved from the outset and
"12 500 gpm Conventional
informed of the evaluatIOn results during the exit
meeting. Informing administrators of performance
13 1.5 mgd Conventional
problems during this meeting often led to their
• EQuIpped with reactor clantiers combining flocculation and
decision to change priorities regarding water treatment
sedimentatIon In one basin. improvements and policies at the plant. Without the
CPE results, existing plant staff frequently were
unable to enlist the support of administrators or to set
CPE Findings priOrities for remedial actions.

Nearly all 13 case studies revealed significant • At Plant 2. the CPE team discovered that the plant
,nformat,on about each plant's condition, operated at its peak rate 24 hr/day during maximum
administration, and operation, including findings that demand seasons. When the team reviewed data on
had not been identified in previous inspections. the service population, they learned that per capita
water use was excessive. By lowering peak
• At Plant 6, the CPE team discovered that plant staff demands to more typical rates. the plant could
bypassed the reactor clarifier during winter months operate within acceptable loading rates to achieve
and proceeded to operate uSing direct filtration compliance with applicable standards. When the
without any chemical coagulant aids. This practice CPE team informed administrators of this fact
was discovered by thoroughly examining plant during the exit meeting. they decided to change the
operating records and conducting follow up water rate structure and penalize high consumption.

293
294 Upgrading Existing or Designing New Drinking Water Treatment Facilities

At the same time. administrators initiated a leak delay the backwash resulted in a significant
detection survey and identified a major leak into an increase in filtered water turbidity lust prior to
old. abandoned oak stave pipeline. Together. these iMiating the backwash cycle, Similarly. when the
administrative actIOns substantially reduced water CPE team reviewed operating data for a l-yr period
demand and enabled the plant to achieve at Plant 12, they learned that finished water
acceptable treated water quality without malor turbidities were very consistent and rarely
expenditures. exceeded 1.0 NTU (see Figure 3-2). However,
when the team measured turbidities during the
• At Plant 12. the team identified severe finished CPE. they discovered clearwell turbidities in excess
water quality problems (very high finished water of 6.0 NTU. The data reported to the State must be
turbidity levels) that previously went undetected or representative of actual operating conditions.
unreported. Mechanical equipment was in a state of
disrepair. thereby adding to the plant's performance These results indicate that data from daily grab
problems. In addition. plant administrators had samples may not reflect true performance and that
scheduled several major extensions to the data collected over shorter time periods (such as
distribution system; however, when informed by the hours or minutes) is necessary. This suggests that
CPE team of the performance problems, the facilities should perform either in-line continuous
administrators intended to redirect their resources turbidity moMoring and recording on each filter, or
from the distribution extensions to upgrading the manual moMoring of each filter effluent on an hourly
water treatment plant facilities. basis. Less frequent monitoring would likely miss
turbidity spikes.
• Town administrators for Plant 13 had signed long-
term agreements to supply water to a new industry The case studies indicated that plant operators and
and another water district. Some aspects of the administrators generally did not recognize the serious
agreement were considered malor concessions to public health Impacts of short-term digressions in
attract the industry, whIch would employ 500 treated water quality. For example. at Plants 2. 4, and
people. First, it was estimated that when these 6. plant operators and administrators did not take
users came on line they would represent one third immediate action to correct short-term breakdowns
of the plant's current capacity. possibly even when they were aware of performance problems.
necessitating facility modifications to provide
additional plant capacity. In addition. the agreement A key finding of the studies is that. because most
also required the town to supply the water at a small water treatment facilities are only operated for 8
lower cost than that currently paid by the town's or 12 hr'day, they tend to have excess capacity. The
own drinking water customers. When the CPE team excess capacity results from being able to operate at
presented town administrators With thiS informatIOn, a lower flow rate for longer periods of time, enabling
they indicated that they would consider Initialing a many small plants to address uM process limitations.
rate study and examine the need to renegotiate For example. Plant 8 operated at its 16-L's (250-gpm)
these agreements to supply water. capacity for only several hours each day even when
turbIdity levels in the surface water exceeded the
The case studies showed that in all plants but three, plant's treatment capability. The CPE projected that,
plant performance was much worse than preViously by redUCIng the plant flow to 8 Ls (125 gpm) and
reported data had indicated. In two cases. finished operating for up to 12 hr. the plant could treat
water quality was so poor that the state threatened to turbidIties of any anticipated level. Likewise, at Plant
institute a boil order unless the facilities Immediately 5. reducing the plant flow from 132 Us (2.100 gpm) to
made improvements. The CPE teams discovered 69 us (1,100 gpm) relieved a severe air binding
these performance problems despite the fact that problem and enabled the plant to operate
monthly operating reports usually showed that finished successfully. At Plants 8 and 5, water demands were
water quality met drinking water standards. These met even with reduced plant flows; however. this may
findings indicate that the present requirement to not always be the case.
sample turbidity from the clearwell on a daily baSIS
does not accurately reflect actual finished water The case studies revealed that proper control of the
quality at many plants. The CPE team iMlated special filtration process is key to improving plant
studies that included developing turbidity vs. time performance.
profiles on filtered water.
• At Plant 2, filter rate controllers malfunctioned
• At Plant 2. 12 months of data previously submitted causing the filter effluent valves to open and close
to the State revealed no violations. However. when every few seconds. ThFl filter flow rate changed
the CPE team measured turbidity before and after from 0 to 63 L's (1.000 gpm). The filter effluent
the filter backwash and plotted the data (see Figure turbidilles also varied. indicating that particles
3-1). they discovered a turbidity breakthrough of previously filtered were washed through the filter to
5.8 NTU. Figure 3-1 also reveals that a decision to the clearwell. While plant staff knew that the valve
Results of Case Studies 295

FlQure 3- 1. Plant 2 tumidity prollie.

TUrllldity, NTU

5 FILTER BACKWASH-

o 100 200 300 400


MINUTES AFTER START OF SPECIAL STUDY

Figure 3·2. Plant 12 finished water tumidity prollle.

(TJ-j-+-++-~--t--+--+----t"""-+--+---+--I--+--+---+

(\J

...z
:;)

lJ")

if
~
;;
... l>

Present
Requirement

lJ") Future
Requirement

lSl-j-+--o+--+--+-....-+--t+-+--!--+--f--+--+---tf--+
SEP88 OCT NOV DEC JRN89 FEB MRR RPR MRY JUN JUL RUG SEP OCT
296 Upgrading Existing or Designing New Drinking Water Treatment Facilities

"jumped around," they did not realize that It produce excellent water quality despite high raw
aHected filter performance. The studies clearly water turbidities. The CCP also revealed that
indicate that filter rate controllers must be properly accurate coagulant doses could be selected by
maintained to allow the filters to operate correctly. using the jar tesVfilter paper procedure.

• At Plant 2, when plant staH removed one filter for • At Plant 1. the CCP dramatically improved plant
washing. the entire plant flow was directed to the performance. Turbidity removal in the reactor
remaining filter. This caused severe turbidity clarifiers was improved and stabilized. and chemical
breakthrough. requirements were minimized. The improvement
resulted from a combination of process control and
• At Plant 8, filters were "started dirty," causing a monitoring. as well as several major process
serious detrimental eHect on filtered water quality. adjustments.

• At Plant 2. operators changed the flow rate without To achieve the desired results. CCPs should be
adjusting chemical feed rates. This resulted in implemented over a period of at least 6 months, since
improper feed of coagulant chemicals and time is necessary to implement process control
subsequent degradation In finished water quality. programs. purchase equipment. provide training, and
document stable finished water quality for vanable raw
The case studies revealed that all 13 plants water conditions. The case studies demonstrated that
implemented only limited process control eHorts. Little process control programs Improved the performance
testing or data interpretation, both of which are of individual unit processes at the two plants, thereby
Imperative to making ,nformed operating deciSions, leading to improved finished water quality.
were conducted. As a result. improper operating
practices, such as bumping filters or wailing too long Overall Factors limiting Performance
to backwash filters, were widespread. For example. at
Plant 12. water was allowed to drop from the troughs A CPE team evaluated 65 performance-limiting factors
onto the filter media, which clearly violates basIc at each of the 13 plants: the top 10 performance-
principles of filter operation. limiting factors are presented in Table 3-2. It IS
important to remember that no one factor was
CCP Findings responSible for limillng plant performance. but rather a
unique combination of factors contflbuted to
CCPs were Implemented at 2 of the 13 plants. The performance problems.
objective of the CCP studies was to determine if the
approach could Improve plant performance and enable Table 3-2. Top Ranking Performance-Limiting Factors
the plants to comply with the SWTR without major Identified at 13 Facilities
capital Improvements. The specific findings of the Rank Factor NO. Plants Category
CCPs, which were conducted at Plants 1 and 5, are
Operator ApphcaUon of 13 Operallons
also presented In Section 4. Concepts and TeSling to
Process Control
Implementation of these CCPs enabled both plants to 2 Process Control Testing 11 Operalions
meet the future finished water turbidity requirements
3 Process Conlrollal>lilyl 13 Design
of the SWTR by implementing process control Flexlollily
programs and providing operator training. The
4 DISinfection 9 DeSIgn
approach demonstrated the potential for drinking water
5 Sed,menlat,on 10 Design
treatment facilities to meet regulatory requirements
through improved operation, maintenance, and 6 Staff Number Administration
administration rather than major capital improvements. Filtration Design
8 PoliCIes AdministratIon
• At Plant 5, city administrators originally had planned 9 Flocculation 6 Design
to spend approximately $1 million to construct
10 Maintenance Maintenance
sedimentation basin facilities and related
improvements. They felt the major capital
improvements were necessary to ensure that the
plant could achieve compliance with the • The highest ranking performance limiting factors fell
forthcoming SWTR turbidity requirements. Af1er the in the operations category, and were related to the
CCP was conducted, however, construction of the inability of plant staH to respond to water quality
improvements was delayed until such time that changes with appropriate chemicals in appropriate
water demands required that the plant operate at doses. In addition. plant staff frequently made
higher rates, As a result of the CCP, plant staH improper operating decisions because they lacked
developed increased confidence that, by understanding of unit processes and associated
implementing process controls. the plant could controls. Compounding these problems was a lack
Results of Case Studies 297

of process control testing programs at all but two of • Filters presented problems at 7 of the facilities. The
the plants. CPE team identified this factor because of air
binding (2 plants), backwash limitations (2 plants),
• Design factors represented 5 of the performance and possible filter underdrain or support gravel
limiting factors in the top 10 list. Process problems (three plants). The team felt that the air
controllability and flexibility was the highest ranked binding and backwash limitations could be
design limitation. It was cited most frequently minimized or overcome by improved operational
because of. limitations In type and location of practices, and that the underdrain or support gravel
chemical feed options, and in control aspects such damage could have been avoided if the operations
as filter- regulating valves or plant flow control personnel had beller understood the filtration
valves. The CPE team noted that minor process. This damage appeared to be caused by
modifications could address these performance introduction of air or by excessive instantaneous
limiting factors at the facilities. hydraulic load at the beginning of a backwash

• Disinfection facilities were identified as a • Flocculation capability was Identified as


performance limiting factor at 9 plants because of performance limiting at 6 plants because of limited
inadequate detention time in clearwells or basin volume and lack of staging. The CPE team
transmission lines. The SWTR will require a plant to concluded that improved operaliOr,s could minimize
provide a certain CT value, which is obtained by the impact of this factor (for example, lowenng
multiplying the disinfectant concentration by the hydraulic loadings, Installing baffles, modifying
actual contact lime. Most of the plants relied on coagulants).
unbaffled clearwells to provide most of the reqUired
detention time. These clearwells were projected to • Administrative factors (including staff number and
provide inadequate CT because of expected severe administrative policies) also were included In the
short circuiting. However, modificaltOns to the top , 0 list. An inadequate number of staff to
clearwells, such as installing baffles. may allow properly run the facilities was noted at 7 plants.
these plants to meet the CT reqUIrements of the This deficiency was critical considering the need to
SWTA. Findings of disinfection inadequacy were add a process control program and assocIated
based on the CPE team's estimates of the responsibilities at these plants. Frequently,
allowable contact time at each plant. No thorough admonistrators were unaware of operating
hydraulic analyses, as required by the SWTR, requirements, or had set water rates too low to
could be conducted within the scope of this project. maIntain adequate treatment or establish a self-
The identification of disinfection inadequacy was sustaining utility. Few administratIve personnel
tentative and is meant as a signal that current understood the seventy of short-term excursions
operation might not be adequate to meet the from high quality treated water.
disinfection requirements to be established by each
State. • Maintenance factors were identified as impaCllnq 7
of the plants. Operators who lacked understanding
• Sedimentation basin design was identified as a of process operations abandoned many of the
performance-limiting factor at 10 plants. The impact automatic and/or manual control systems at the
was periodic and seasonal during high turbidity or plants. The CPE team Identified several facilities
high- demand episodes. The CPE team projected where maintenance activllies were completely
that improved operallon could minimize the impact neglected, sometimes due to administrative
that the margInal basins had on plant performance indifference. The team concluded that Improved
(for example, longer run times at lower flow rates or understanding oi operations and maintenance,
improved coagulallOn control). coupled with an improved administrative attitude,
could lead to improved plant performance.
4. Case Studies

The following case histories provide a detailed summary of the results from each of the 13 CPEs on which
this report is based. Each case history consists of a facility description. results of the Major Unit Process
Evaluation and Performance Assessment. and a discussion of the factors found limiting the plant's
performance. The applicability of a CCP IS also discussed for each plant as are the results of the CCPs
completed at two plants. These CPEs were completed as part of the project to develop and formalize these
procedures for water treatment plants. Some aspects of the procedures were refined as more CPEs were
completed. As the procedures evolved through these refinements. some of the ways in which the results are
presented have changed. Some inconsistencies between the presentation of the results of the different case
histories. therefore. may be observed.

Plant 1 • Fluoridation (sodium silica lIuoride)

Facility Description • Sludge removal and thickening


Constructed in 1974. Plant 1 is owned and operated
by the city and serves approximately 10.000 persons. • Sludge drying beds
with no Significant industrial water users. It consists of
a pre-sedimentatIOn basin followed by conventional • Three high-service pumps: 100. 113. and 157 L's
treatment and IS used as a softening facility dunng (1.600. 1.800 and 2.500 gpm)
winter months. Average daily flow for a 12- month
penod was 66 L s (1.5 mgd). with an average daily The three raw water pumps transfer water from the
flow during the peak month of 131 Us (3 mgd). The nearby river through a 61-cm (24-in) line to the pre-
plant includes the following unit processes (see Figure sedimentation basin. The three low-service pumps lift
4-1): the water from the pre·sedimentation basin
approximately 18 cm (7 in). so that it can flow by
• Three constant-speed. raw water pumps: two 25- gravity through the plant's unit processes.
hp. 126-Us (2.000-gpm) and one 15-hp. 91-Us
(1,450-gpm) Powdered activated carbon is added to the water in
the pre-sedimentation basin for taste and odor control
• 8.7 million·L (2.3 mil-gal) earthen pre-sedimentation and a cationic polymer is injected following the raw
basin water pumps. Alum and lime are added in the
flocculation area of the two 22-Us (500.000-gpd)
• Three 15-hp. 113-L's (1.800-gpm) constant-speed. solids contact clarifiers.
low·service pumps
The clarifiers are square with circular sludge removal
• Chemical addition (alum. lime. and Dycafloc 587-C) mechanisms. Prior to the clanfiers. water from a 14-
L's (220- gpm) "soda well" is pumped into the raw
• Two 17.7-m x 17.7-m (58-ft x 58-ft) upflow water stream. Hydraulic conditions cause an uneven
clarifiers. 6.4 m (21 ft) and 6.0 m (19.8 ft) deep split of this softer water between the two clarifiers.

• Recarbonation With liquid carbon dioxide Water flows through the clarifiers to a recarbonation
basin. where the pH is lowered. and onto four dual
• Four 7.6-m x 7.6-m (25-ft x 25-ft) dual media filters media filters. From the filters. the finished water flows
into the clearwell. Chlorine and fluoride are added to
• 1.1 million·L (300.000-gal) clearwell the water as It enters the clearwell. The three high-
service pumps deliver finished water to the distribution
• Disinfection system.

298
..,
~

•!
c:

River 50lymer
Bypass Line
...
ii
. /fPlSo<h a
.i<' LJ Well

Presedimentation Bastn
li
Q

f ~

~
1

\
\
\
Coniaci
Basin

¥--, Contact
Basin
2

~----'I _ _---.J

I
\
+-- . --~
A Carbon
Dioxide
\ I
Recarbonalion

0
\ I I
Basin
KEY I I

Nonnal Flow Palh ~


\
\
I
II
'~
Thickener G _LP
Fluoride

,~
\
Ahemale Flow Palh _ _ - \ ----- P
\
Chemical Feed _ _ \ I I P
~ -

/ ~
\
Sludge Flow - - - - - -
I

:f/~""M
\
Chemical Feed -6. \
'-(.- Filter>
Pump 0 \ 2P
Backwash
Yahe ~ I , Waler
i Basin ()
Meier _
o '"
V>

EJ
Ctl

learwell ,...
U'l
c
a.
Drying Beds Backwash ~- ~ ~ )( 0 e;;'
V>
To
High City
To Landfill
O\,orflows 1 Service i'V
CO
CO
300 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Major Unit Process Evaluation process control to be integrated into the daily
Figure 4-2 illustrates the assessed capacity and routine of the plant staH. This involves collecting
projected performance of each of the plant's major additional process control data and interpreting
unit processes in a performance potential graph. The the data to direct process control that optimizes
vertical broken lines indicate the annual average flow plant performance.
of 70 Us (1.6 mgd), the peak monthly flow of 131 Us
(3 mgd), and the design capacity of 307 Us (7 mgd). 2. Process Control Testing - Operation: The lack of
process control testing has resulted in incomplete
As Figure 4-2 shows, the raw water pumps. low- data being collected to determine the level of
service pumps, filters, and high- service pumps are plant performance. In addition. it is probably
rated at the 307-Us (7-mgd) plant design flow. masking periods of production of poor quality
Potential capacities of the pre-sedimentation basin finished water. Items of particular concern are lack
and the clarifiers are rated at less than design. of turbidity testing of raw water and water from the
clarifiers. and continuous monitoring of individual
The pre-sedimentation basin was derated because of filter eHluent. Plant data indicate that filter eHluent
short circuiting through the basin and no capability to "spikes" were occurring, but it was not possible to
add coagulant aids. Also, return of backwash water to determine their seventy or duration.
the eHluent end of the basin results in excessive
turbidity levels in the raw water. The basin was rated 3. Process Automation - Design: There is a need for
above the peak monthly flow of the plant. continuous monitoring and recording turbidimeters
on the raw water. each filter, and the clearwell.
The clarifier/flocculator was rated at 136 Us (3.1 mgd)
with one unit in service and 272 Us (6.2 mgd) with The factors identified as having either a minimal effect
both in service. The corner sweeps on the sludge on a long-term repetitive basis or a major eHect on a
mechanisms have failed allowing excessive amounts periodic basis were priOritized and are summarized
of sludge to build up in the basin corners. Sloughing below.
of the sludge coupled with inconsistent weir elevations
has resulted in periodic solids loss. The clan- 1. Process FleXibility - Design: More flexibility is
f1occulators were derated because of these necessary in types of chemicals added and points
conditions. of chemical addition. At the time of the evaluation.
alum and lime were added to the flocculation
Performance Assessment portion of the clarifier and polymer was added
Plant 1 is currently required to produce finished water after the low-service pumps. There was no
with turbidity levels less than 1.0 NTU and with free process available to feed a filter aid or flocculent
chlorine at levels that will ensure less than 0.2 mg/L at aid. Flexibility to move the alum feed to a point
all points in the distribution system. Fluonde is added with greater mixing could reduce the alum feed
to achieve a 0.9-1.1 mg/L residual in the finished rate and may also allow the flocculator speed to
water. A comparison of plant monitoring data with be reduced to better optimize flocculation. The
state requirements indicated the plant was operating ability to add a filter aid would improve filter
in compliance with applicable regulations. A review of performance.
operating records, however, indicated numerous
excursions of filter effluent turbidities above 2. Lack of Standby Units - Design: There is no
acceptable levels. standby backwash pump. If the existing pump
falls, the plant IS out of operation until it can be
The SWTR will require plants to demonstrate by repaired or replaced.
regular turbidity monitoring or constant recording
turbidimeters, turbidity at less than 0.5 NTU greater 3. Preliminary Treatment - Design: Obvious problems
than 95 percent of the time. Additionally. theoretical 3- exist with the pre-sedimentation basin, including
log removal and/or inactivation of Giardia cysts and 4- no chemical feed (except carbon) to the basin,
log removal and/or Inactivation of enteric viruses must short cirCUiting,. backwash water fouling of the
be demonstrated. intake area of the low-service pumps. and
diHiculty cleaning the pre-sedimentation basin.
Performance-limiting Factors
The factors identified as having a major eHect on 4. StaH Number - Administration: The current major
performance on a long-term repetitive basis are emphasis is on maintenance; more time must be
summarized below in order of priority. spent on process control to Improve performance.
In addition. the superintendent is working extra
1. Operator Application of Concepts and Testing to shifts to keep the plant operating. At least two
Process Control - Operation: Operation of the additional plant operators are necessary for both
plant is maintenance rather than process control operations and maintenance activities to be
driven. Priorities need to be established that allow adequately addressed.
Case Studies 301

Figure 4-2. Plant 1 performance potential graph.

FlOw, mgd
Unit Process 2 3 4 6 8 9 10

Raw Water Pumps' 1

Pre-sedlmentation 2 I
Low-service PumpsJ 1

Clari-flocculato'"
SOR. gpd/SQ ~ (' untl) 329 656 98t
SOR. gpd/sQ ~ (2 un,ls) 154 329 492 656 820 98 4 1

Filters 5
HLR. gpm/sQ ~ (' li~ers) 03 0.6 0;8 '.1 14 1.7 ~O
HLR. gpmJSQ ~ (2 lilters) 0.6 1.2 ',5 1

High-service Pumps6 I
Annual Avg. Peak Month DeSIgn
6/86·6/87 5/86·5/87 Capacity

, 7 8 mgd with Individual pumps· assumes deSign capacIty with contInued operation.
2 Assumes run 2 li~ers al 2 gpm/sQ ~ and backwash 2 "iters at end of day.
J 7.8 mgd with IndIVIdual pumps· assumes deSign capacity with continued operation.
• Rated at 800 gpdlsQ ~ lor turb,d'ty removal· use 1.000 gpd/sq ~ lor so~enlng capacity (we" Imbalance).
5 Rated at 2 gpm/sq It
62.5 mgd with IndIVIdual pumps assumes deSign capacIty with continued operation

5. Sedimentation - Design: The corners of the evaluatIOn as having a minor effect on perlormance.
clanfiers need to be grouted to prevent buildup of Action taken to address these factors may not
sludge ,n the baSins and the flow through the noticeably improve plant performance. but may
weirs needs to be balanced so that resulting Improve the efficiency of plant operation:
hydraulic gradients do not impact the quality of
the water from the clarifiers. The combination of • The pre-sedimentation baSin is not designed to
un-level weirs and sludge accumulation In the handle high turbidities caused by runoff and Ice
baSin corners has led to periodic solrds loss In the jams.
effluent. which severally degrades plant
perlormance. • Flash mixing of chemicals appears Inadequate

6. DiSinfectIOn - DeSign: Short circuiling of finished • The abilrty to sample the sludge return and the
water through the c1earweli results in inadequate sludge concentration within the clarifiers (top to
contact time for proper disinfectIOn. This problem bottom) is inadequate.
will be amplified With colder water and higher pHs.
Projected Impact ot a CCP
7. Return Process Streams: Return of the backwash The CPE indicated that operator training through a
water to the pre-sedimentation baSin near the low- CCP would be beneficial for improving process
service pumps· suction negatively impacts raw stability and finished water.
water qualrty.
CCP Results
In addllJOn to the above major factors limiting A CCP was Iniliated. Monitonng of the two reactor
perlormance. other factors were noted during the clarifiers during the CPE phase had revealed problems
302 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Figure 4-3. SettllKl w81er turbidities from the reactor clarifiers at Plant 1.

Turbidity, NTU
12 ; - - - - - - - - - - - - - - - - - - - - - - - - - - - ,

11

10

/I


J

0~~_=-...,..-__:_:__:'1___::':""""-._-':':"'"_:r._=-___r-__::::__:"~:::
1.-Moy-1l9 ll-Jun-89 08-Au9-89
28--nay-/l9 2~-Jun-89

o BASIN #2 + BASIN"

with c1anfier solids control. Thus. the CCP efforts and a gradual Increase in turbidities from the pre-
began by expanding process control in the clanfiers. sedimentation basin during the last month of the CCP.
Each clarifier was taken out of service so that several
feet of anaerobic lime sludge that had accumulated in The CCP resulted in dramatic improvement of plant
the basins could be removed. performance without major capital improvements.
Process control and monitoring activities. coupled with
Figure 4-3 shows the finished water turbidity from the several major process adjustments. improved and
two clarifiers from the time the cleaning operation was stabilized turbidity removal in the reactor clarifiers.
completed In May until the CCP was concluded In
August. The basins' settled water turbidities gradually
Improved and stabilized at 1 to 2 NTU; both basins
exhibited equal performance. Activities that
contributed to this consistent performance Included
controlled flow splitting. equalized chemical doses to
each baSin. and shutting off of a well that was
contnbuting a disproportionate amount of flow to basin
2. The reactor clanfiers achieved thiS performance
despite van able influent turbtdiltes to the baSins from
the pre-sedimentation pond. as shown in Figure 4-4.

Most importantly. the improved reactor clarifier


performance "carried over" to improve the
conSistency of turbidity removal by the filters. Figure
4-5 shows the overall plant finished water turbidity.
which stabilized at less than 0.2 NTU since the end of
June 1989. COinCiding with stable performance from
the contact clanfiers. The improved performance was
achieved despite an Increase in treated water volume
Case Studies 303

Figure 4-4. Effluent turbidity lrom the presedlmentatlon pond lor Plant 1.

Turbidity, NTU
100 ,--.----;r---,,-----------------..,
90

70

·60

50

20

10

O~::r__=_-__r-~~......-:7""'"-.__-_:_:_"""T""
_ _-___r---,._...J
, <-Moy-69 11-Jun-69 09-Jul-69 0~-"u9-89
Z8-Moy-89 Z~-Jun-89 2J-Jul-89

Figure 4·5. Finished water turbidity prollle Irom Plant 1.

Turbidity, NTU
1.8 . - - - - - - - - - - - - - - - - - - - - - - - - - - - - - ,
1.7
1.6
1.5
1..
1.3
1.2
1.1

0.9
0.8
0.7
0.6 Future Requirement
0.51-----+-+---;::;------------------....,
0 .•
0.3
0.2
0.1
OL..---,-----.----.-------.-----.----~--~----'
"-MQy-89 1 l-Jun-89
28-Nay-89 25-Jun-89
304 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Plant 2 the influent flow varies because raw water pumps are
initiated and terminated several times a day. After
Facility Description chemIcal addition. the water flows to two parallel
Plant 2. constructed in 1931 and expanded and flocculation basins with two parallel- operating variable
upgraded in 1976. has approximately 1.000 servIce speed turbine f1occulators. Subsequently. water flows
connections and no significant industrial water users. to two sedimentation basins equipped with tube
It consists of a conventional treatment process selliers. During the evaluation. flow did not appear
including flash mix, flocculatIOn, sedimentation. and evenly split between the two
filtration, and no pre-sedimentation. Source water is flocculation/sedimentation treatment trains, and the
provided by the nearby river. clarifier weirs were uneven.

Plant records for a 12-month period show daily water Sludge is manually removed from the sedimentation
production to be 1.1-6.1 million L (0.3-1,6 mil gal), basins approximately once a year. FollowIng
Flow records are obtained from the plant finished sedimentation. the water flows through a weir to three
water meter and do not include water used for filter mixed media filters with Leopold underdrains.
backwash. During the CPE. water was produced over Powdered activated carbon is added once a day to
an 8-hr day at an effluent flow rate of about 89 Us (2 the water prior to filtration. Filtration rates are
mgd); peak effluent flow was 131 Us (3 mgd). Plant automatically adjusted by filter water level floats that
influent flow IS not measured because the raw water control the filter effluent valves. At the time of the
meter is inoperable. Plant 2 includes the following unit evaluation, one filter was out of service because of an
processes (see Figure 4-6): inoperable effluent control valve, Another filter effluent
control valve was malfunctioning and was observed to
• Two vertical turbine raw water pumps rated at 76 readjust the flow rate by over 69 L s (1,100 gpm)
and 69 Us (1,200 and 1,100 gpm). and one 28-L s repeatedly WIthin a few minutes.
(450-gpm) engine driven raw water pump
The filters are normally backwashed once a day; the
• Chemical addition of alum, polymer. and lime with process IS Initiated manually but at automatically timed
in-line mechanical flash mix for the alum intervals and includes surface washing, The backwash
rate can be set as high as 189 Us (3,000 gpm).
• Two 206.430-L (54.540-gal) parallel flocculation
basins WIth two variable speed turbine mixtures in Filtered water is disinfected with chlorine and is stored
parallel in the clearwell. The three high-service pumps deliver
the water to the distribution system.
• Two 666.160-L (176,000-gal) sedimentation basins
with tube settlers over half their surface area Spent filter backwash water flows by gravity to a
sedimentation basin. which overflows to the raw water
• Three 3.4-m x 4.9-m (ll-ft x 16-ft) mIxed media wet well. Each backwash sedimentation basin has the
filters with Leopold underdrains capacity for one backwash before excessive solids
overflow the effluent weirs and return to the raw water
• Gas chlorination system wet well. The backwash sedimentation basins are
normally cleaned once a year.
• 204.390-L (50.400-gal) clearwell
Sludge from the flocculation. sedimentation basins.
• Two centrifugal hIgh-service pumps rated at 72 and and backwash water sedimentation basins are
94 Us (1,150 and 1,500 gpm), and one standby pumped to two sand drying beds. The dried sludge is
natural gas driven vertical turbIne pump rated at 27 disposed of at the landfill.
Us (425 gpm)
Major Unit Process Evaluation
Water flows by gravity to a wet well through either a The performance potential graph is shown in Figure 4-
shallow culvert near the bank of the river or through a 7. The flows listed across the top of the graph are the
second pipe extending toward the center of the river maximum at which the plant can operate while
at an unknown distance and depth. The vertical remaining in compliance with applicable regulations.
turbine pumps deliver water to the plant from the wet N8Ither the raw water pumps nor the high-service
well. The turbine pump supplies water from the pumps were rated because the condition of the
surface intake. A float control in the c1earwell initiates impellers and the actual pump output were not known.
the raw water pumps.
The flash mix was rated at 131 Us (3 mgd). where it
Alum IS added to the raw water prior to an in-line can produce a G value of approximately 3.000 sec-'.
mechanical rapid mixer; lime and polymer are added The in-line mechanical mixer would probably be
downstream of the mixer prior to the flocculation limited by water velocity in the pipeline rather than
basins. ChemIcal addition is not flow. paced. although miXIng capability.
!!
'"c:

to
!"

Flocculalors
...
;;
:!.
'"
~
:•
Surface Inlake
~
0.

~L-Me-"~T.
Pump Alum In Line Polymer M
Mullarllcal
'"~
~

2 2
\ Parallel Parallel
I Flocculation Sedimentation Basins
Basins wi'"
Tube Senlers

o
Raw Wale,
0 Raw Wale, Pumps
We,Weli 4 Mud Valves
Wei:r
OverOaw

\
OverOow
to
Filters

\
Spent Backwash
Water and Sludge
Sedimentation Basin

Cl2 ()

'"'"Cll
Clearwell 3 Mixed Media Filters
en
"c
a.
Cll
'"
w
o
Ul
306 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Filiur. 4-7. Plant 2 performance potential liraph.

Flow. mQd
Unit Process 2 3

Raw-water Pumpe Not Rated


: :

Flaah Mix
:
Flocculation 1 :
Detention time, min 78 ';'i.'<:H
Sedlmentatlon 2
SOR. gpdlSQ " 1,287
I
Filtration]
HLR. gpmlSQ " 1.3 Z6 40

Disinfection
Chlonnabon

Contact Tim" I
HllIh-servlce Pumpe Not Rated

Present Plant Design


Max. Flow Flow

1 Rated at 45 min because of single stage.


2 Rated at 2,000 gpdlSQ " . may be able 10 use process control to Increase capacity. AI50 rated on summer water Quality. but may be
able to direct filter In winter.
3 Rated al 4 gpmlSQ ft . media and underdraln Integnty. need to be verified to Justify this rao"",.
• Based on 2·hr detention I,me

The flocculation basins were rated below the plant precise process control. the filters could operate
design flow at 77 Us (1.75 mgd) for a detention time successfully at up to 293 m3/m 2 /d (5 gpm/sq It);
of 45 minutes. because the flocculation basins are however. the media and underdrain integrity must be
single-stage units. Altering the basins to provide verified to justify either rate. If further evaluation
multiple-stage flocculation would better control floc indicates damage to the filter underdrain or support
formation and justify increasing the basin capacity to gravel. the filters would be a major performance-
153 us (3.5 mgd). limiting factor.

The sedimentation basins were rated at 66 Us (1.5 The disinfection system was rated as two processes:
mgd). which results in a surface overflow rate of 81 chlorination capacity and contact time. The
m 3/m 2/d (2.000 gpd/sq It) (based on tube settler area). chlorination capacity rating of 131 Us (3 mgd)
However. improved flocculation and process control indicates that the capacity of the feed unit is sufficient.
could increase thiS rating. Direct filtration might be an The contact time, however. was rated at only 26 Us
optIOn during winter months. which would decrease (0.6 mgd), the maximum flow through the plant that
reliance on the sedimentation basins for solids would provide the recommended 2-hr detention time.
settling. The limited detention time provided at normal plant
flows compounds the importance of effective
The mixed media filters were rated at 131 Us (3 mgd) performance 01 the other treatment processes for the
for a loading rate of 234 m 3/m 2/d (4 gpm/sq tt). With removal of pathogens.
Case Studies 307

In summary. the performance potenllal graph indicates produced. evidenced by the filtered water turbidity
the plant should be operated at less than 66 L's (1.5 of 46 NTU from Filter 3 when Filter 1 was
mgd) during periods of high raw water lurbidity. Flow backwashed (figues 4-8 and 4-9). Plant operations
rates above 66 Us (1.5 mgd) may be possible without staff need to vary coagulant and f10cculant
adversely affecting finished water quality: however. dosages and to change plant water flow rates
filter run times will probably be significantly reduced when backwashing filters in response to raw water
because of excessive solids loading. thereby reducing of variable quality.
lotal plan! capacIty 10 66 L s (1.5 mgd).
2. Process Control Testing - Operation: Testing to
During winter months. if raw water turbidities allow monitor the treatment process is inadequate to
effective direct or In-line direct filtration. plant capacity detect problem areas and indicate necessary
adjustments. This lack of testing allows periods of
may be able to Increase to between 66 and 131 Lis
extremely poor water to go undetected and
(1.5 and 3.0 mgd). depending on whether or not
uncorrected as shown 10 Figure 4·12. At a
flocculatIOn is required for successful operation.
minimum. additional Jar testing and turbidity
measurements of the raw water, sedimentation
Performance Assessment
basin effluent, and filter effluent will be reqwed to
A review of the finished water quality monitoring data
indicate appropriate plant chemical dosages and
Indicated the plant has been operating in compliance
flow rate adjustments.
with the current turbidity Maximum Contaminant level
(MCl) of less than 10 NTU on a monthly average. 3. Maintenance: Preventive maintenance IS lacking at
The plant has. however. had periodic excursions the plant. The plant equipment is maintained on a
above 1.0 NTU. The mOnJloring data also Indicate the crisis basis and plant performance is directly
plant may have difficulty meeting the SWTR turbidity compromIsed. Major treatment components were
maximum of 0.5 NTU tor 95 percent of the time (as out of service dUring the evaluation and have
measured every 4 hr of water production). Figures 4- evidently not been repaired for up to several
8. 4·9. and 4-11 show effluent turbidittes obtained years. Examples of eqUipment in need of repair or
th~ough special studies. Figures 4-10 and 4-12 show out of service include the raw water meter. filter
turbiditIes trom plant data. effluent control valves. raw water pump. and alum
feeder f1occulator paddles.
Investigation of the filter media in the out-of'service
4. Staff Number - Administration: The present
filter bed revealed that the media was clean with no
staffing level does not allow the water plant to be
evidence of mudball formatIOn. However. the
adequately operated or maintained. A minimum of
evaluatIOn team discovered numerous depressions of
two additional staff members are needed to
up to 10 cm (4 in) In the surface of the media. and by
sufficiently cover the utility needs. With adequate
probing the filters found that support gravel had
staff. one operator could focus on plant process
migrated and mounded. The operators mentioned that
control and other utility employees could
sand. anthracite. and garnet had been removed from
speCialize in either water or wastewater treatment.
the clearwell dUring c1eanlOg. Measurement of the
media pIle that had been removed from the clearwell 5. Familiarity with Plant Needs - Administration: Plant
indicated that approximately 0.2 m3 (7 cu ft) of filter administration needs to become more familiar with
media had passed through the support gravel. The the requirements of the plant. Better
migration of the support gravel. depressions in the understanding of the plant's reqUirements would
surface of the filter bed. and the passing of 0.2 m3 (7 help garner the administrative support necessary
cu It) of media Ihrough Ihe filter all indicate serious to operate and maintain the plant properly.
damage to the filter support gravel and media.
TYPIcally. this type of damage occurs when air IS 6. Filtrallon - Design: A limited evaluation of the
introduced Inlo Ihe backwash water. filters revealed potentially serious problems in the
support gravel. Depending on the outcome of a
Performance-Limiting Factors subsequent detailed evalualton of filter integrity.
The factors identified as having a major effect on the filters may be found adequate to 131 Lis (3
performance on a long-term repetitive basis are mgd). However, if the filters are found deficient,
summarized below in order of priority. they would probably have to be repaired before
the plant could produce consistent high quality
Operator Application of Concepts - Operation: finished water on a continuous basis.
Process control is needed so that operators
respond directly to raw water quality changes. At Factors identified as having either a minimal effect on
the lime of the evaluation, finished water quality a routine basis, or a major effect on a periodic basis
was fluctuating drastically with periods of poor are summarized below.
finished water production. Figures 4-8 through 4-
1 1 show actual plant data indicating the variability 1. ~ - Administration: The extremely low pay scale
In water quality and the extremely poor water and lack of employee incentives will make it
308 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Figure 4-8. Filter 3 turbidity prollle, June 8, 1988 - Plant 2.

Turbidity. NTU
460
6
Filter _1
Backwash
5

4
Clearwell

3

~~~e~~~~~~~~_~~~~u~_~~~~!~ _

_____ ~~~_~~~~~~~_~~~~u~_~~~~!~_~-~-~-~-;-----~-~-~-~-~-~_~_~ _

Noon 1'00 2:00 3:00 4:00 5:00 6:00


Time of Day

Figure 4-9. Filter 1 turbidity profile. June 8, 1988 - Plant 2.

Filler -1
Backwash
6

Clearwell

Curren, Regulatory Maximum· 10 NTU

o
11:00 12:00 1:00 2:00 3:00 4:00 5:00 6:00
Time of Day
Case Studies 309

Figure 4-10. Plant 2 performance.

2.0
Turbidity. NTU

1.2

1.0

0.8

0.6
Fulure Regulatory Maximum = 0.5 NTU

0.4

0.2
Desired Operating • 0.1 NTU

o 10 15 20 25 30 35
Days

Figure 4·11. Filter effluent turbidities profile, June 8,1988 - Plant 2.

Turbidity. NTU
Fulure Regulatory Maximum = 0.5 NTU
0.5 -----------------------------------------------------~------------~
I" \ I I
, \ I
, \ I
Filter #3 ., ......,
\
,
I',\ I
"
\
0.4 r' \
, \ I
, \ I
\ , , I
, ' I
, ,, 'II

,,,
\
\

03
\
, ,
N, ,,
\
FItter
\
,
,
\
\

02
", , I
I

--------------------------.----- ..------~-----------------------------
" DeSired Operaong • 0.1 NTU
0.1

o 4 6 8 10

Minutes
310 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Figure 4- 12. Plant 2 performance - May 19BB.

Turbidity. NTU 1.6 1.6


1.2

1.0

0.8

0.6

0.4

0.2
Desired Operallng = 0.1 NTU

o 5 10 15 20 25 30
Days

difficult to retain present employees and to attract allow plant flow rates to be adJusted more
additional qualified help. gradually and set at various rates.

2. Turbidity/Pre-sedimentation - Design: Excessive 5. Disinfection - Design: The lack of a standby


turbidity levels during portions of the year and chlOrinator. mixing. proportional feed capability.
fluctuations during high demand periods have contact lime. and automatic sWltchover could
degraded effluent quality. Pre-sedimentation would result In inadequate disinfection on a periodic
minimize turbidity fluctuations and result In a more basis.
consistent raw water quality. ContinUing
operallons Without a pre-sedimentallon basin may 6. Process Automation Design: Effluent
reqUire reducing plant flow rates dUring high turbidimeters with recorders on each filter effluent
turbidity periods. would be beneficial to monitor water quality.
Without such continuous monitoring. an operator
3. Sedimentation - Design: Surface overflow rates at would have to take frequent measurements (i.e ..
flows above 66 lis (1.5 mgd) may not allow hourly) to monitor plant performance.
adequate settling of the sludge. Poor settling can
cause excessive solids loading to the filters and 7. Chemical Feed - Design: The carbon feeder
subsequently degrade filter efficiency. Uneven should be returned to service to replace manual
overflow weirs also cause poor distribution of addition. The capability to add two polymers would
water within the sediment altOn basin, further be desirable, with additional flexibility In chemical
impairing settling. Manual sludge removal twice a feed pOints. A backup alum feeder is also needed.
year may be inadequate to prevent solids
carryover from the sed,mentallon basins. ThiS 8. Flocculation Basins - Design: The Single-stage
practice is also operator intensive. flocculation basin makes control of proper floc
formatton difficult. The retention time is adequate
4. Hydraulic Loading - Design: Fluctuations in plant and minor modifications may allow two-stage
flows due to cycling of constant-speed, raw water operation with variable energy input in each stage.
pumps during high demand periods reqUire
additional operator attention to maintain finished 9. Process Controllability - Design: Chemical feeders
water quality. An inlluent flow control valve would should be flow paced or manually adjusted to
Case Studies 311

complement and control raw water quality Projected Impact of a CCP


changes. Results of the CPE indicate performance to be
severely limited by a number of administrative. design.
10. Standby Units - Design: No standby units are maintenance. and operations factors. Every major unit
available for critical process components. process was identified as a performance-limiting
including backwash pump. alum feeder. and factor. The evaluation team reached a consensus that
chlorinator. significant improvements in water quality could likely
be achieved with a CCP. but that major capital
11. Working Conditions - Administration: Conditions at expenditures may also be required for the plant to
the water plant discourage staff from spending meet the proposed finished water quality criteria.
time at the plant and encourage neglect. Provision
of a comfortable climate controlled working area Many of the unit process limitations described in the
would improve operator morale. performance potential graph could be eliminated if the
plant were to be operated at a lower capacity than the
In addition to the above major factors limiting present summer water demand. which appears
performance. other factors were noted during the excessive. Water conservation measures and lowered
evaluation as having a minor effect. Action taken to water demands together with a CCP were
address these factors may not noticeably improve recommended.
plant performance. but may improve efficiency in plant
operation:

• Preliminary treatment: Grit in the raw water wet well


and the lack of screens on the intake piping
produce operational problems because of silt and
debris accumulation.

• Flow proportioning to units: Raw water flow to the


sedimentation basins was not evenly proportioned.
Operators can control the distribution of flow to the
units by frequently adjusting valves located in the
sedimentation basin influent piping.
312 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Plant 3 vertical mixing unit that supplies up to a G value of 70


sec". Following flocculation. the water flows by
Facility Description gravity into two parallel sedimentation basins equipped
Plant 3 is owned and operated by the city. with tube settlers. During the evaluation. the flow was
Constructed in 1950 and expanded and upgraded in not evenly split between the two sedimentation basins.
1975 and 1976. it currently serves approximately
5.000 people with no significant industrial water users. Following sedimentation. the water flows to three
The plant uses a conventional treatment process mixed media filters with Leopold underdrains. The
consisting of flash mix. flocculation. sedimentation. filters appeared to be in good condition. but some
and filtration. chemical residue had accumulated on the anthracite.
These filters are typically washed at about 189 Us
Plant records for a 12·month period indicate that the (3.800 gpm). which corresponds to 972 m 3 /m 2:d (16.6
average amount of water treated dally was 41 L's gpm/sq ft). Washing typically occurs at the end of the
(0.94 mgd). with a minimum of 22 Lis (0.5 mgd) and a day so that the filters start clean the following
maximum of 114 Us (2.6 mgd). These daily flows morning.
were pumped through the plant In less than 24 hr and
therefore do not indicate the operational capacity of The filtered water is disinfected with chlorine. then
the plant. It is typically operated at three standard flows into the clearwell. where the four high-service
rates - 69. 101, or 202 Lis (1.100. 1.600. or 3.200 pumps are available.
gpm) - for less than 24 hr. Plant 3 includes the
following unit processes (see Figure 4-13): Sludge from the sedimentation basins and backwash
water from the filters are directed to two earthen
• Three vertical turbine raw water pumps: two 25-hp. sludge settling ponds. At the time of the evaluation.
110-Us (1.750-gpm) and one 15-hp. 63-los (1.000- the plant was discharging overflow from the ponds to
gpm) the creek without a National Pollutant Discharge
Elimination System (NPDES) permit.
• Chemical addition (alum and polymer) with an in-
line mixer Major Unit Process Evaluation
The performance potential graph IS shown in Figure 4·
• 302.800-L (80.000-gal) flocculation basin with a 14. As Figure 4-14 shows. the potential capacities of
variable speed vertical mixing unit the raw water pumps. high-service pumps. and filters
were rated at the 219-Us (5-mgd) plant design flow.
• Two sedimentation basins (each 4.1 m x 18.3 m. The flocculation basin. sedimentation basins. and the
3.7-m deep [13.5 ft x 60 ft. 12-ft deep!) disinfection system were rated at less than plant
design flow. The single-stage flocculation basin was
• Three mixed media rapid sand filters with Leopold derated because control of floc formation is more
underdrains diHicult with a single-stage than with a multiple-stage
flocculation system. The sedimentation basins are
• Gas chlorination system limited by a high surface overflow rate. which can
allow solids to be carried over to the filters. The
• 246.000-L (65.000-gal) clearwell disinfection system was not considered adequate at
flow rates above 166 Us (3.8 mgd). because the
• Four vertical turbine finished-water pumps: two clearwell and transmission lines provide inadequate
200-hp. 94·L's (1.500-gpm). one 100-hp. 38-Ls detention times at these rates.
(600-gpm). and one 50-hp. 22-L's (350 gpm)
The sludge settling ponds were not rated but were
Raw water flows by gravity through three 46-cm (18- determined inadequate at current flows. unless the
in) diameter perforated pipes located beneath the plant obtains an NPDES permit to allow discharge to
source creek into a raw water wet well. The three raw the creek. Without operational changes such as more
water pumps deliver water from the wet well to the frequent cleaning of the ponds. the eHluent quality in
flocculaliOn basin. the ponds may not meet typical permit requirements
(i.e.. 30 mg/L total suspended solids and 1.0 mgtL
Alum and polymer are added to the flow prior to the total dissolved aluminum).
flash mix. The plant's in-line mechanical flash mixer
was not in use at the time of the site viSit because of Figure 4-14 indicates that the plant should be
maintenance problems; therefore. the only chemical operated at less than 101 Us (2.3 mgd). if possible.
mixing was caused by turbulence in the line at nearby Short periods of increased flow may be possible
elbows. Without adversely affecting finished water quality;
however. filter run times will probably be reduced
After chemical addition. the water flows into the because of excessive solids loading. The plant may
Single-stage flocculation baSin with a variable speed be operated more hours at the 101 Lis (1.600 gpm)
Case Studies 313

Figure 4-13. Proce.. !low diagram 01 Plant 3.

Creck

3.J8.>~
pcrro;.~~·1 ~
Pipes

G,avilY
Flow Raw Water
WelWell
1--0
1--0 3 Raw Waler Pumps

1--0
POI[mCr
Alum
j
• FILSh Mi, Unil (Nol in Service)

reb
l "~,,,. Sludce

Oackwish W &.Ic:r

L-,-_....J
f' M',oo Moo', "'""

~3g
~ 4 High Service Pumps
SJudce
Se'tlirlc
Ponds
314 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Figure 4-14. Plant 3 performance potential graph.

FlOw. mod
Unit Process 2 3 4 5

Raw-water Pumpa'

Flash 1.41_ 2

Flocculation)
Detention tIme. min 58
115 I
Sedimentation'
SOR. gpdlsq It 926 1.852 I

FiltrationS
HLR. gpm/sq It ; 2 3 4

Dlslnfe<:1lon 6
Contact time. min ~ 36 24 I

High-service Pumps I
Sludge 5enlln9 Ponda Not Rated· Inadequate at present how

Current Annual Present Dally7


Avg. Flow Max. Flow

, Peak Ilow tIlat plant can treat at worst water Quality.


2 Out 01 service.
J Based on detention lime 01 45 min and Single stage.
• Based on 2.000 gpdlsq It.
S Based on 5 gpm/sq It.
a Based on allowing 2 m,n of contact lime In clwarwell and 18 min of contact bme In 2,~OO ft Of 18-,n ttansml",;n line.
7 Not based on 24 hr/day. SO actual nows are higher

rate to overcome the limitation of the NTU more than 95 percent of the time. as measured
flocculation/sedimentation basins. Also. when raw by regular daily monitoring or constant recording
water turbidities are low the plant may be operated in turbidimeters. Additionally. the plant needs to
a direct filtration mode. which eliminates the need for demonstrate theoretical 3-log removal and/or
sedimentation. inactivation of Giardia cysts and 4-log removal andlor
inactivation of enteric viruses. In order to meet these
Performance Assessment regulations. surlace water treatment plants need to
The city is currently required to produce finished optimize process controls to minimize or eliminate
water with turbidity levels less than 1.0 NTU on a "spikes" of turbidity in the finished water at critical
monthly average and w,th free chlorine at levels that times. such as immediately after backwash.
will ensure a chlorine residual in e_cess of 0.2 mg/L at
all POints in the distribution system. A review of Performance-Limiting Factors
monitOring data indicated that the plant was operating The following factor was identified as having a major
In compliance with the applicable regulations. eHect on perlormance on a long-term repetitive basis.

In the SWTR, the minimum requirements for finished 1. Sludge Treatment - Unit Design Adequacy: The
water turbidity are much more stringent. Plants need sludge holding ponds are currently discharging to
to produce finished water with a turbidity less than 0.5 the creek, but the plant has no permit to allow this
Case Studies 315

discharge. A letter from the Permits Section of the indicate whether or not chemical addition has
Water Quality Bureau states that the plant must been optimized.
apply for a permit, which will require certain
effluent limitations. Permit limitations may require The use of the flash mix unit, especially during the
major modifications to the plant. Failure to obtain a times of the year when direct filtration can be
permit and to meet permit limitations could result utilized. would probably reduce chemical usage.
in sizeable fines being levied against the town. Also. additional experimentation wIth polymer
products could result in the selection of more
Factors identified as having either a minimal effect on effective coagulantiflocculent aids.
a routine basis, or a major effect on a periodic basis
were prioritized and are summarized below. 5. Policies - Administration: Administrative policy
limits the frequency with which the raw water
1. Sedimentation - Unit Design Adequacy: The intake can be backwashed. As a result. the intake
surface overflow rate and weirs limit the pipes can accumulate a signIficant amount of silt
performance of the sedimentation basin. If the before backwashing. thus reducing the plant's
plant were operated at a rate in excess of 101 Us intake capability. More frequent backwashing
(2.3 mgd), the resulting high surface overflow rate would eliminate these periodic IimitalIOns in raw
would not allow for adequate settling. Excessive water supply.
solids would be carned over to the filters, thereby
degrading filtration performance. 6. Chemical Feed FaCilities - UnIt Design Adequacy:
Inability to feed a filter aid and. or flocculant aid
2. Flocculation - Unit Design Adequacy: The single- could result in poor plant performance dunng
stage flocculation basin makes control of proper periods of variable raw water quality.
floc formalIOn difficult. Controlling plant flows at
rates below 101 Ls (1.600 gpm) may allow 7. Alternate Power - Unit Design Adequacy: There is
additional detention time to compensate for the no standby power capability at the plant.
lack of a multiple-stage unit. Therefore. water would not be supplied to the
distribution system dunng a power outage.
3. Lack of Standby Units - Unit Design Adequacy:
There are no standby units for adding chlorine or 8. Hydraulic - Unit Design Adequacy: Low stream
alum. Failure of either of these chemical feeders flows and upstream water use have resulted in
would result in unacceptable finished water penods when no raw water IS available to be
quality. which may require the plant to shut down. pumped into the plant. Studies are presently
In addition, there is no spare backwash pump. underway to incorporate in-stream or off- stream
although. in an emergency. the distnbution system storage to alleviate this problem.
could provide limited backwash.
In addition to the major factors limiting performance
4. Application of Concepts and Testing to Process discussed above. other minor performance-limiting
Control - Operation: The plant performance could factors were noted during the evaluation. AclIOn taken
be improved dunng periods of variable raw water to address these factors may not noticeably Improve
quality by application of a thorough process plant performance, but may improve efficiency In plant
control program. For example. more frequent jar operation:
testing would provide data on which to base
chemical feed pOints. By monitOring turbIdity from • The lack of adequate disinfection could be a
the sedimentation basins several times each day, problem when operallng the plant above 166 Us
chemical doses could be adjusted to optimize (3.8 mgd) because of the potential for short
sedimentation basin performance. circuiting of the clearwell and the limited detention
time provided in the transmission mains.
It would also be good practice to monitor the
turbidity of water from each of the filters. At the • The lack of automatic continuous turbidity
present time. a dally turbidity value is being monitoring and recording on the raw and finished
recorded for water from the clearwell. This water from each filter requires the operations staff
measurement may mask higher turbidities coming to obtain this information manually on a periodic
out of the filters. Significant breakthrough may be baSIS. Not only does this require an additional time
occurnng that would not be detected by the commitment from the operators. but penodic
present monitoring practice. With continuous information is not as effective as a continuouS
monitoring and recording of the turbidity of each record.
filter effluent, the Increase in turbidity following
backwash could be observed along with the length • The pay of the chief operator/superintendent is
of time the elevated turbidity occurs. ThiS would approximately the same as the shift operators.
316 Upgrading Existing or Designing New Drinking Water Treatment Facilities

This pay differential does not recognize the chief Projected Impact of a CCP
operator's additional responsibility. Results of the CPE indicated that plant performance.
based on daily measurements of turbidity from the
• It is very difficult to sample the sludge discharge clearwell and a filter turbidity profile conducted during
lines from the sedimentation basins. the evaluation. was in compliance with applicable
drinking water regulations. The plant monitoring data
• Additional process control testing should be done also showed very consistent plant performance for the
to provide more of a basis for process control 12-month period evaluated. As a result. a CCP was
decisions. Examples of further testing would not recommended at Plant 3.
include more frequent analysis of raw water
turbidity and alkalinity. along with measurements
of turbidities of the water leaving each
sedimentation basin and filter.
Case Stud ies 317

Plant 4 • 1.S-hp, 3.8-Us (60-gpm)filtered water pump (to


clearwell)
Facility Description
Plant 4 is owned and operated by the county water • S-hp, 14-Us (220-gpm) centrifugal backwash pump
and sewer district. It was constructed in 1970 and
serves approximately 81 connections, including the • Chlorination system consisting of a calcium
school. It is a packaged, conventional plant and its hypochlorinator; solutions are made as needed in
processes include pre-sedimentation, flocculation, 189-L (SO-gal) batches
sedimentation, and filtration.
• 4S,420-L (12,000-gal) clearwell
A stream fed largely by return flows from the local
irrigation district. supplies the plant. Historically, the • Two high-service centrifugal pumps: one 7.S-hp, S-
creek flowed only intermittently, but the importation of Us (80-gpm), and one 3-hp, capacity unknown,
irrigation water with subsequent water losses to creek backup
drainage have significantly increased stream flows.
These artificially increased stream flows, coupled with Raw water is pumped through a 10-cm (4-in) line from
naturally erosive soils. have caused a severe the creek to the pre- sedimentation basin. The 1.S-hp
sedimentation and turbidity problem in the creek. turbine pump delivers water from the pre-
sedimentation basin through as-em (2-in) line to the
Plant flow records indicate an average daily water plant. This pump is started automatically by a float
production of 0.4 Us (10,000 gpd) in the winter and control in the clearwell and the flow rate is regulated
2.6 Us (60,000 gpd) in the summer. Flow records, by a float-controlled valve in the chemical mix
obtained from the plant effluent meter, do not include chamber.
water used for filter backwash.
Alum and polymer are added in line to the raw water
The CPE did not determine the accuracy of the prior to the chemical mix chamber. The manufacturer
effluent flow meter, but the inconsistency of readings originally designed this basin for the addition of calcite
taken over the day indicated a problem exists. The to stabilize the raw water. Some mixing, but no flash
influent flow rate is measured by a rectangular weir mixing, is provided both in the line and through the
located just prior to the flocculation basin. chamber. After chemical addition, the water flows over
a rectangular weir and into a single-stage flocculation
At the time of the site visit, the influent flow rate was basin with a detention time of 10 minutes at both
3.8 Lis (60 gpm). The plant operates at this rate for design and operating flows. The f1occulator is a
various hours per day depending on demand. Plant 4 constant-speed (12 rpm), vertical paddle wheel type.
(see Figure 4-1S) includes the following unit
processes: Following flocculation, the water flows into the settling
chamber equipped with 6-degree horizontal tube
• Manually-operated, 2-hp centrifugal raw water pump settlers, Settled water flows over a weir onto a mixed
media filter equipped with a perforated pipe
• 22.7 mlilion-L (6 mil gal) earthen pre-sedimentation underdrain. The filter operates on a constant
basin rate/variable head basIs; a float-controlled effluent
valve regulates flow rate. Design filtration rate is 293
• 1.S-hp, 3.8-Us (60-gpm) turbine-type settled water m 3 /m 2 'd (S gpm/sq ft); the operating filtration rate was
pump determined to be 281 m 3 /m 2/d (4.8 gpm/sq tt). Water
is pumped out of the filter to the clearwell.
• Chemical addition of alum and polymer without
flash mixing; both alum and polymer are mixed in Backwash is initiated manually or automatically by
batches in 189-L and 114-L (50-gal and 30-gal) headloss across the filter: the filter IS not equipped
tanks, respectively with a surface wash. The backwash flow rate of 13.9
Us (220 gpm) corresponds to a wash rate of 978
• 1.2-m x 1.2-m (3.8-ft x 4-ft), 2,290-L (60S-gal) m 3 /m 2!d (16.7 gpm/sq ft). Filter backwash water flows
flocculation chamber with a constant-speed, vertical by gravity to an earthen storage pond located
paddle wheel northwest of the plant, adjacent to the stream. Sludge
from the tube settlers is removed with the backwash
• 0.9-m x 1.6-m (3-ft x S.4-ftl. 2,290-L (60S-gal) cycle,
sedimentallOn chamber with 6-degree lube settlers
Filtered water is disinfected with a calcium
• 0.9-m x 1.2-m (3.1-ft x 4.1-ft) mixed media filter hypochlorite solution within the clearwell and pumped
with perforated pipe underdrain system approximately 2.6 km (1.6 mil to the town reservoir
318 Upgrading EXisting or Designing New Drinking Water Treatment Facilities

Flgur.4-'5. Plant 4 procua IIow diagram.

~
~
II
I!.
~
e
0
·3
5

-;
• u j
~5
~ ~
a.., hi j ~
:i:

I!.
~
:;;: . .lli
~~
~
.!i
;,
.•
j

~ - d
~'M
'~ ] ftlnt""~lf
u, e
c
~. .2
d

~uO
" e
]~
~u

~
i
.~

CD
[fl

~o.

'"
Case Studies 319

Figure 4-IB. Planl 4 performance potential graph.

Flow. Qom
Unit Process 20 40 60 80 100

Raw-water Pumps Not Rated

Flash Mix> None

Flocculation'
Detention time. min II 2+

Se<llmentalion'
SOR. gpmtsQ " 1.4
I
Filtration
HLR, gpm/SQ " 1.7 3.3

Disinfection
Contact time. min 600 300 2<iO 150 '~O

High-service Pumps I

Operating and
DeSign Flow

, 4S·mln HOT for sweep flOC: 22.S·mln HDT for direct Iiltrat,on.
2 Based on SOR • 2 gpmtSQ ".

(capacity of 50.000 gal [189.250 L) through a 10-cm The sedimentation basin was rated at 1.8 lis (28
(4-in) line. gpm) or 117 m 3 /m 2 1d (2.0 gpm/sq It) surface overflow
rate (SOA) based on the capability of the tube settler
Major Unit Process Evaluation to produce a clear water with the existing raw water
The performance potential graph is shown in Figure 4- conditions. The design surface overflow rate of 251
16. The slashed vertical line on the graph represents m 3 /m 2 1d (4.28 gpm/sq It) is too high given the raw
both the design and operating flow rate of 3.8 Us (60 water conditions and has resulted in excessive
gpm). carryover of solids to the filter.

As seen In Figure 4- I 6, the raw water pumps were not The mixed media filter was rated at the design flow of
rated because actual pump output was not known. 3.8 Us (60 gpm), which corresponds to a loading rate
The high-service pumps were rated at 4.4 Us (70 of 293 m 3/m 21 d (5 gpm/sq It). The backwash rate was
gpm), slightly above the design flow. The flocculation also rated at design, for up to " I 71 m 3/ m 2/d (20
basin was rated at 0.8 Us (13 gpm) when operated in gpm/sq It). The condition of the media must be
the sweep mode of coagulation and at 1.6 Us (26 restored (mudballs eliminated) and the support gravel
gpm) when operated in the direct filtration mode. The and underdrain integrity must be verified to justify this
rated flow was doubled for direct filtratIOn because rating. The backwash cycle must be extended and
shorter flocculation times are acceptable in this surface wash facilities should be added or the filter
operational mode. The flocculation basin was derated will again become a major limiting factor at this plant.
because it is a single-stage unit and only provides 10
minutes of detention time at a 3.8 Us (60 gpm) flow Disinfection contact time was rated at 2.0 hr at flows
rate. Single-stage flocculation and short detention of up to 6.2 Us (98 gpm). This rating was justified
times make formation of an adequate floc more because of the long transmission line to town and the
difficult. 189,250·L (50,OOO-gal) storage reservoir.
320 Upgrading Existing or Designing New Drinking Water Treatment Facilities

The performance potential graph indicates that the 2. Process Control Testing - Operation: Testing to
plant capacity is limited by the flocculation and monitor the effectiveness of the treatment process
sedimentation unit processes. To achieve acceptable is inadequate. The available data clearly
finished water quality, the plant capacity may have to substantiates turbidity MCl violations. (See
be limited to about 1.9 Us (30 gpm) and operated for Figures 4-17 through 4-20.) During certain times,
longer periods of time to produce the daily amount unacceptable water is supplied to the public,
required. exposing users to an unacceptable risk of
contracting waterborne diseases. At a minimum,
Performance Assessment turbidities should be monitored in the plant
A review of the finished water quality monitoring data influent, sedimentation basin effluent, and filter
submitted to the State indicate that the plant is effluent several times each day. Jar testing should
operating at the current MCl for turbidity, 1.0 NTU, on be done daily or at least when raw water
a monthly average. However, much of this data was conditions change, to fine-tune chemical dosages.
collected from the storage reservoir located in town, Continuous·recording, in-line turbidimeters would
2.6 km (1.6 mil from the plant. Sampling at this point be very beneficial in providing information for
violates the regulations and intent of the law. optimizing process control.
Unreported records of turbidity data taken directly
from the filter reveal that numerous excursions above 3. Plant Coverage - Administration: The operator
the 1.0 NTU occurred on a regular basis for several makes one brief visit to the plant each day, which
months. These data indicate the plant will have is not adequate to perform proper process control
difficulty meeting the SWTR turbidity MCl of 0.5 NTU testing, experimentation. and adjustments.
for 95 percent of the time. Figures 4-17. 4-18 and 4- Addition of contJnuous-recording turbldimeters and
20 depict turbidity data from plant records, while appropriate alarms could reduce the time spent at
Figurs 4-19 shows turbidity data generated from the plant, but a minimum of 2 hr each day will still
special CPE studies. be required for process control testing.

InspectIOn of the filter 'evealed a heavy layer of 4. Flocculation - Design: The plant design only
chemical floc on the media surface and numerous allows for single-stage flocculation and the
mudballs with,n the anthracite. The depths of media detention time is too short to allow optimum floc
were found to be adequate: however, probing of the formation before water flows to the sedimentation
media indicated the support gravel varied by as much chamber. The effect of flocculation may be
as a couple of inches In some areas of the fitter. In reduced if the plant could be run in the direct
the past 18 years of operation. the filter media has filtration mode. However, the plant flow rate may
"solidified" to a concrete-like state tWice, the last time need to be reduced to overcome flocculation
about 5 years ago. deficiency.

Media removal necessitated the use of a Jackhammer. Factors identified as having either a minimal effect on
Recently, the c1earwell and In-town reservoir were a routine basIs or a major effect on a periodic basis
both cleaned for the first time. Four 189-l (50-gal) were prioritized and are summarized below.
barrels of media and sediment were removed from the
clearwell and a conSiderable amount of sediment was ,. laboratory Space and Equipment - Design: The
removed from the finished water storage reservOir in operator has no equipment other than a
town. turbidimeter to perform the necessary tests to
determine raw and finished water quality. The
Inspection of the automatic backwash cycle indicated accuracy of the meter cannot even be verified,
that the backwash time was Inadequate to clean the because of a lack of equipment. There is no jar
filter media. This, coupled with the lack of a surface testing or other equipment and supplies for
wash, has allowed the filt~' In :>ccumulate mudballs. process control testing.
The Inadequate backwash not only results ,n costly
replacement of the media, hut also makes production 2. Operator Pay - Administration: The operator pay is
of an acceptable finished water quality Impossible. too low to compensate the operator for the
number of hours necessary to run the plant
Performance-limiting Factors properly. ThiS pay does not offer an incentive for
The factors Identified as having a major effect on keeping qualified operational personnel or for
performance on a long-term repetitive basIs are prOViding adequate plant coverage.
summarized below in order of priority:
3. Alarm Systems - Design: Because of a lack of an
1. Water Treatment Understanding - Operation: The alarm system, particularly on the finished water
operator is newly hired and, while enthusiastic and turbidity, unacceptable water has been pumped
willing to learn, has had no training or background into the distribution system on many occasions.
in the science of water treatment. Without alarm systems and with minimal plant
Case Studies 321

Figure 4-17. Plant 4 turllidlty proflle.

Turtltdity, NTU

3.0
Unreported

2.5
, Maximum
,
"
1\
1\
"
, \
,\

I ' , \
I \ I \
I' ,\
2.0 ,, \
\ ,
I \
\ , ~\
I \ J' I \
, \ I \ I ,
, \ I ' I \
, \ I \ ' ,

1.5 I ' I ' . ., / \

: ~,' " ,",


"'.JIr"
... o4 \
\

. .
I \ I \ I
" \ I ',,,'" ,

: " \ " w... \


1.0 ----------r------~-~--------~-r----------------------- -----------------
Regulatory MaJumum • 1.0 NTU \
,
,
I "
\
\I
,: \ I
\
\

0.5
:
, Reported
\
I
: Average •
• ' D.?~I~~ _O~r.?~n_g_:_O~,_~~U _

6/87 8187 10i87 12/8 2/88 4/88 6/88

Date

Figure 4-18, Plant 4 turllidlly profile, August 1987. Figure 4-19. Plant 4 turtlidlty proflle before and immediately
following backwaah.
Turtltdily, NTU
Turbldily. NTU
3.0 1.0

Reported

0.8

2.0

0.6

0.4
10

0.2

Prior to Backwash

17 20 24 28 31
o 10 20 30 40
Day 01 Month Minutes A«er Backwash
322 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Figun! 4·20. Plant 4 tumidity prolll•. January. 1988.


Turbidity. NTU 20+ 20 +
20 + 20+
20
UnrePOrted
MaxImum

15

10

Regulatory Maximum • 1.0 NTU


~-~.=.=--------~-----

5 10 15 20 25 30
Day 01 Month

coverage. the operator is not aware of the water • Lack of a preventive maintenance program may
quality being supplied to the town. result in excessive equipment downtime. which
could be significant since there are no backup
4. Staff Number . Administration: The district needs systems.
to have a backup operator. so that the plant
manager can leave town for business or personal • Lack of filter surface wash may be contributing to
reasons without leaving the plant unmanned. the inefficient washing of the filter.

5. Flash Mix - Design: Lack of a flash mix unit will • Lack of standby units for key equipment could
limit coagulation effectiveness and increase the cause periods of plant downtime.
chemical requirement. particularly ,f the plant IS
run in the direct filtration mode. Projected Impact of a CCP
Results of the CPE indicated performance was limited
6. Sedimentation - Design: The surface loading rate by a number of factors in operation. administration.
within the sedimentation chamber is too high to and design. The evaluation team judged that a CCP
allow floc to settle prior to flowing to the filter. could help the plant make significant improvements in
Better settling would improve or lengthen filter finished water quality. However. design limitations may
runs. Operation in the direct filtration mode would require the plant to operate at a reduced rate to
eliminate the sedimentation baSin as a potential produce an acceptable finished water quality. In
problem since the floc that is produced is addition. since many of the limiting factors are in the
filterable. but not settleable. areas of administration and design. some minor capital
improvements must be made and greater
In addition to the above major factors limiting administrative support to the plant (i.e .. higher
performance. other factors identified as having a operator salary) must be provided to significantly
minor effect were noted during the evaluation. Action improve plant performance.
taken to address these factors may not noticeably
improve plant performance. but may improve the
efficiency of plant operation:
Case Studies 323

Plant 5 one of them stops. The pumping rate through the


plant was measured at approximately 136 uS (3.1
Facility Description mgd). at the time of the visit. This is the "normal"
Plant 5 is a direct filtration plant that was installed in plant flow rate and remains constant.
1978 to replace an older pressure filter. At the time of
the CPE. the system had a total of 1.122 connections. The influent flow meter was found to measure almost
Water for the plant is supplied from the southeast end 13 percent less flow than was calculated to be
of a multiple use lake located about 29 km (18 mil to entering the clearwell. Meters typically measure less
the northwest. flow with age. so the meter may need to be
recalibrated.
The plant's water treatment processes include
coagulant chemical feed (alum and polymer). A valve controls the flow to the plant during startup of
flocculation in a reaction basin. optional pre' the raw water pumps. An onfice plate is located
chlorination and non· ionic polymer filter aid feed. downstream of the valve to regulate pressure for
filtration through four dual media filters. post- optimum operation.
chlorination. and gravity flow from the plant to storage
and distribution. The plant was designed for a flow Alum and a cationic coagulant polymer are added to
rate of 131 us (3 mgd). Plant flow records for a 12- the influent alter the orifice plate. Typical feed rates
month period indicated an average daily flow of 26 us were 5-10 mg/L for alum and 0.1 mgtL for the cationic
(0.6 mgd). Average mon1hly flows during the same polymer. No flash mixing is provided; however.
period were 12-50 L's (0.28 and 1.16 mgd). Plant 5 moving the chemical feed points to the onfice plate
(see Figure 4-21) Includes the following unit would probably result in a hydraulic flash mix. At low
processes: alum feed rates in a direct filtration plant. some type
of flash mixer must be provided.
• Two 100-hp. 125-L's (1.980-gpm), vertical turbine
raw water pumps The three compartment reaction basin allows some
flocculation to occur before filtration. Flow of the water
• 36-cm (14-in) propeller influent flow meter at the through the baffled compartments provides hydraulic
plant "agitation" of the water to promote floc formation. No
outside energy is input to the water. The reaction
• Chemical addition of alum and cationic polymer basis has apparently been designed to decrease
without flash mixing turbulence as flow proceeds through the
compartments. Approximately 8 minutes of detention
• Three-compartment "reaction" basin that allows time are provided in the reaction basin at the normal
mixing and detention time for flocculation. 6.1 m flow rate.
(20 It) in diameter, with a water depth of 2.3 m (7.6
It) Chlorine and a non-IoniC polymer can be added just
ahead of the filters. but chlorine normally is added
• Four dual media filters. each 3.5 m (11.5 It) square only alter filtration. The non-ionic polymer. Serapan, is
and 2.4 m (7.75 It) deep, with heaterrlateral·type added ahead of the filters during some periods of the
underdralns year, for example, in winter when turbidities are low.

• Two 60·hp, 167-Us (2,650-gpm) vertical turbine Water then enters an open influent channel where it is
backwash pumps distnbuted to the dual media filters. Flow through the
filters is regulated by effluent valves. As the headloss
• 10-hp, 7-lIs (112-gpm) vertical turbine surface across the filter builds up. the effluent valve gradually
wash pump opens to counteract this increase. Since the influent
flow to each filter cannot be equally split, the flow rate
• Two 378.500-L (1 OO.OOO-gal) hypalon-lined through each filter is unknown. 61 cm (24 in) of 0.90-
backwash water storage basins mm effective size anthracite and 15 cm (6 in) of 0.45-
to 0.55-mm effective size silica sand lie above the 53
• Two submersible backwash water recycle pumps cm (21 in) of layered support gravel. A header/lateral
piping underdrain system is located at the bottoms of
• Gas chlorination system the aluminum filter boxes. The filtration rate under
normal plant flow conditions is 234 m3tm 2 /d (4 gpm/sq
• Two 3.8 million-L (1 mil-gal) steel ground level It).
reservoirs
One interesting feature of the filters is the design of
Water is pumped to the plant through a pair of 30-em the discharge header. The header profile rises in
(12-in) transmiSSion mains. One pump is used at a elevation just before the water is discharged into the
time and the pumps automatically alternate each time clearwell. This discharge pipe then terminates just
324 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Figure 4-21. P..nt 5 ptOCft8 flow dWQrllm.

~ 0---"
I
I
.p I
.,,10--.1 I
u'£ I
I .5
:l
Q:l
I
I
s--'
li
I '""
I
I L -lll
L
L
_
I,

I_- ~- - -- - -'
I
I
I o
I
I
o
l!
·c
o
:c
u
Case Studies 325

above the floor of the clearwell. creating a negative provide 10 minutes. the plant flow rate would have to
head in the discharge header from the filters. The be reduced to 114 Us (2.6 mgd). Further evaluation
intensity of the negative head depends on the amount would be required to determine the actual detention
of headloss across the filters. the plant flow rate, and times necessary for successful flocculation.
the depth of water in the clearwel!. A plug can be
removed from the top of the tee fitting. thus The filters are the weakest major process. The most
eliminating the negative head condition. It is not significant factor affecting filter performance is air
known 'what effect this negative head condition has on binding. Air binding (accumulations of air within the
the filter operation, but its effects should be evaluated. filters) has been a problem since the plant was
constructed. It is believed that the air coming into the
Backwash is initiated by adjustable headloss controls. plant from the lake is very high in dissolved gases. As
The two 60-hp pumps each provide 167 Us (2.650 water flows into the plant from the transmission pipe.
gpm) of backwash water from the clearwell at 21 m the pressure on the water is relieved. allOWing the
(70 tt) of head. A rotating arm surface wash system dissolved gases to escape. The air accumulations are
helps break up sediment at the surface of the filters so serious that the filters have to be allowed to "rest"
during backwash. The 10-hp surface washwater pump for several minutes prior to backwashing to allow the
can provide 7 lis (112 gpm) at 79 m (260 tt) of head. gases to escape, so that they will not disrupt the
support gravel and filter media during backwash.
The spent plant backwash water is stored in the two Based upon the uneven surface of the top of the
hypalon-lined basins adjacent to the plant. After media in one filter, it is possible that accumulations
settling overnight. the decant water is recycled back have disrupted the support gravel in the past. If
into the plant by two submersible pumps. The valve disruption of the support gravel has occurred, it would
described previously holds the flow rate through the need to be removed from the filter(s) and replaced.
plant constant during recycle of the backwash water. Further investigation IS needed to evaluate thiS
potential problem before taking such drastic action.
The production of the backwash pumps was restricted
by a butterfly valve to about 118 Ls (1.870 gpm) each During air binding. the effective filtration rate through
to prevent excessive media loss and disruption during the filters IS increased because the water must flow
backwash. ThiS represents a backwash rate of only around the air bubbles. The net area of the filters is
about 820 m 3 /m 2 d (14 gpm/sq ft): the backwash therefore reduced. resulting in an increase In effective
pumps are capable of pumping as much as 167 Us filtration rate. The only way to reduce the filtration rate
(2,650 gpm), or 1.171 m 3/ m 2 id (20 gpm1sq ft). Dual is to either eliminate the air binding problem. or to
media filters are typically backwashed at 878-1,171 reduce the plant flow rate. The plant is now operating
m 3t m 2 !d (15-20 gpm/sq ft). at. or slightly above. the deSign filtration rate of 234
m 3 /m 2 /d (4 gpmlsq ft) or 136 lis (3.09 mgd). Because
Major Unit Process Evaluation about 20 percent of the finished water must be used
The performance potential graph is shown in Figure 4- for backwash water, the effective filtration rate is
22. The raw water pumps appear to be capable of reduced to 187 m 3im 2 /d (3.2 gpm/sq ft) or 107 Us
pumping at the plant design capacity of 131 Us (3 (2.44 mgd). After observing the amount of air released
mgd). During the plant evaluation. the plant flow rate from the filters. the evaluatIOn team estimated that the
was measured by drawing down the clearwell and effective filtration rate of the filters may have to be
measUring the average rate at which it filled with time. reduced to about 117 m 3 /m 2,'d (2 gpm/sq ft) or 67 Us
ThiS rate was measured at 136 L's (3.1 mgd)). (1.52 mgd).

The reaction basin was not a typical design and. The automatic filter backwash cycles (headloss
therefore, was rated in terms of detention time initiated) did not appear to be long enough to properly
provided. A detention time in flocculation basins of at clean the filters. The water was still dirty as the
least 20 minutes normally is desirable to permit backwash cycle ended. Improper cleaning of the filters
adequate floc formation In cold water conditions. (The can lead to poor treatment and short filter runs.
reaction rates of the coagulants are slowed
considerably In cold water.) To provide a 20-minute Performance Assessment
detention time. the plant flow rate would have to be Finished water quality data from the past year
decreased to 57 L s ('.3 mgd). Achievement of this indicated that the plant was operating within the 1.0-
rate would reqUire the raw water pumps to be NTU standard for finished water turbidity, although
restricted in their output dUring the winter months. incoming turbidities are extremely variable. Winter
turbidities are often less than 5 NTU. and usually less
During warm water conditions. the flow rate than 10 NTU. while spring and summer turbidities can
theoretically can be Increased because reaction rales vary Widely, even within a given day. Prevailing
of the coagulant chemicals are higher. The typical westerly winds often stir up sediment in the relatively
minimum flocculation time for warm water. which shallow lake. resulting in raw water turbidities of 50-
appears In the design standards. is 10 minutes. To 280 NTU. Monthly average turbidities are reported
326 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Figure 4-22. Plant 5 performance potential graph.

Flow. mod
Unit Process 3

Raw-water Pumps I
W.nler . 20 mIn Summer' 10 min
Reaction Tank (1.3 mgd) (2.6 mgd)
HOT. min 26 I 13 I
Filters' (1.5 mgd) (2.4 mgd)

HlR. gpmlSQ " I I


Chlorination
Cobntact time. min 360 180 '2

1 Flow reduced 10 2.44 mgd 10 account for 20 percent of Ihe production being used for backwash waler.
The flow may have 10 oe reduced even further (to approXimately 1.5 mgd) because of 3Ir binding.

typically at 0.5 NTU or less. but turbidities of 0.5-0.6 Over the past several years. the Slate has completed
NTU range were reported for a few months. microscopic particulate examinallOns (MPEs) of the
finished water to determine the effectiveness of the
Finished water Quality data is taken from the plant filtration process. These have shown that the finIshed
potable water system. which pumps water from the water contains a large amount of particles (algae.
clearwell. Readings are typically taken at about 9 a.m. insect eggs. elc.). which indicates that. at times. the
alter the plant has been shut down overnight. Figure filters are not operating well.
4-23 shows. under current operating condillons. the
number of days each month that Plant 5 would be in Performance-limiting Factors
Violation of the SWTR effluent turbIdity standard. The factors Identified as having a major effect on
which will require that finoshed water turbidity be 0.5 performance on a long-term repetitive baSIS are
NTU or less at least 95 percent of the tome. summanzed below ,n order of priority:

Inspection of Filter 2 atter dewatering did not detect 1. Operator Application of Concepts - Operation:
the presence of any mud balls; there was. however. Varying raw water quality requires changes In
deep penetrallOn of the floc into the anthracite. This In chemical feed rates and plant flow rates to
Itself is not necessarily indicative of a problem. but a maIntain acceptable finished water quality. The
vanallon of about 5 cm (2 In) in the elevation of the plant had no organized process control program to
top of the anthracite media indicates a potentially provide Information to base operational decisions
serious problem with the condition of the support upon. Although the operators had a good
gravel. understanding of water treatment. they were not
applying that knowledge fully.
The presence of filter sand and anthracite in the
clearwell also indicates a potential problem. This 2. Process Control Tesllng . Operation: Proper
material was cleaned out of the clearwell previously operation of a direct filtration plant requires regular
but has SInce accumulated. If the support gravel were process control testing so that chemical doses
in proper condition, the filter sand and anthracite can be optimized. Jar te~.' 'J. follo,',ed by filtration
layers would not break through to the underdrains. through Whatman #40 paper filters. is a
Readings laken after backwash of Filter 2 during the reasonable simulation of the direct filtration
site visit Indicated that the filter effluent turbidity process. ThiS regular testing was not being done
exceeded 1.0 NTU. This IS of concern both for lhe at the plant.
present standard of 1.0 NTU and the proposed 0.5-
NTU standard. These data are presented in a graph in 3. Filtration· Design: Turbidity measurements taken
Figure 4-24. at the time of the evaluation and MPE testing
Case Studies 327

Figure 4-23. Plant 5 potential lor SWTR compliance.

"0
~ 15
"0

~
CIl
>.
~
:e
.=
"0
1;l
8. 10
a
0:
'0
c:
."a
]'"
>
'"
V>
>.
5
a'"
.....
a
~ ( Days Allowed
E
:> Above 0.5 NTU
z
0
---
Dec Jan Feb Mar Apr May Jun Jut
87 88 88 88 88 88 88 88

done by the State demonstrate that the filters do water. Nearly constant backwashing may also be
have a performance problem. The presence of necessary because of the solids load to the IiIter.
filter media in the clearwell indicates that the which will reduce the eHective plant capacity.
support gravel may have been damaged in the
past. If the support gravel is in fact damaged. 2. Plant Coverage - Administration: The plant is not
replacement to the media will be necessary. Filter attended on weekends and the operators are
capacity and finished water quality is being often away from the plant during weekdays
aHected by severe air binding. Periods of high performing other duties. As a result, periods of
turbidity also require frequent backwashing. which poor finished water quality could go undetected.
further reduces plant capacity. If the support
gravel is in good condition. then the filters could 3. Lack of Standby Units . Design: There are no
be rated as a lower priority factor and the filtration standby alum and polymer feed units. Failure of
rate decreased because of the air binding problem one of the units would severely affect plant
and backwash water requirements. performance.

Factors identified as having either a minimal eHect on 4. Automated Process Monitoring - Design: A lack of
a routine basis. or a major eHect on a periodic basis continuous monitoring turbidimeters on the raw
were prioritized and are summarized below. water quality at the lake pumping station and on
the eHluent line from each filter makes plant
1. Turbidity - Design: The turbidity of the raw water operation more diHicult. Turbidity monitoring on
often exceeds that normally recommended for the the raw water would allow increases in raw water
direct filtration treatment mode. During periods of turbidity to be anticipated. so that treatment could
high turbidity. it may be necessary to reduce plant likely be improved. Monitoring of the filter eHluent
flow rates to produce an acceptable finished quality would provide information necessary to
328 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Figure 4-24. Finishe<l water turbidity profile after backwash 01 Iliter 2 - Plant 5.

Backwash
Recycle
Pump On

CurrcntMCL
1.0

Proposed MCL

20 40 60 80 100 120

Minutes

adjust chemical feed rates and. therefore, • Better communication of priorities to the plant staH
optimize filter performance. and better teamwork among staff members could
improve plant performance.
5. Reactor Basin . Design: The reactor basin may
provide inadequate detention time to allow • Return of spent backwash water to the influent can
chemicals to react and flocculate during cold result In increased raw water turbidities and a
water conditions. As a result. the plant flow may change in raw water chemical characteristics.
have to be reduced substantially dunng tile winter.
• The alum and cationic polymer feed points provide
6. Inoperabliity Due to Weather - Design: As no flash mix of the chemicals. Movement of the
discovered during the summer of 1985. drought chemical feed points to just prior to the orifice plate
can severely impact the availability of water from would provide better mixing.
the lake. The raw water intake is located in a
shallow corner of the lake and considerable Projected Impact of a CCP
attention has been given to relocating the Intake It was projected that a CCP would help Plant 5
to a deeper portion of the lake. An engineering achieve better performance. Results of the CPE
study has been completed that evaluates the indicated that the plant was limited by a number of
alternatives for another intake location. factors, primarily in operation and design. Because of
the design limitations, the plant would need to reduce
In addition to the above major factors limiting its operating rate to produce an acceptable finished
performance. other minor factors were noted during water quality. Some capital investment could be
the evaluallOn. Action taken to address these factors necessary, depending upon the condition of the filters.
may not noticeably improve plant performance, but
may improve the eHiciency of plant operallOn: It was recommended that the city continue its eHorts
to construct a new water intake on the lake. The new
• Funding the operation of the wastewater system intake would appear to improve the quality of the raw
through water revenues is not a good practice. water, as well as the reliability of the water source.
Each utility should be self suHicient.
Case Studies 329

CCP Results to be treatable with direct filtration (e.g., less than 50


A CCP was initiated, at which time the plant flow rate NTU), Another indication of Improved performance
was reduced so that design-related limitations could was that filter effluent turbidity following a backwash
be addressed. During the initial site visit. the CCP did not exceed 0,3 NTU and returned to 0.15 NTU
team developed a daily data sheet and implemented a Within minutes after the wash.
procedure describing process control testing. In
addition, procedures were developed for calibrating The CCP proved that the plant could achieve
chemical feeders and calculating chemical dosages so compliance with SWTR turbidity requirements without
that chemical feed rates could be accurately applied. major capital improvements. City administrators had
Special studies were implemented to determine the planned on spending an estimated $1,000,000 on
effect of operating the plant at a reduced flow rate and construction of sedimentation basin facilities and
operating the filters without a negative pressure. related improvements. After the CCP, they decided to
delay any construction until water demands required
At the conclusion of the first visit. the plant was the plant to be operated at higher rates. The plant
operated at 69 L's (1,100 gpm) rather than at 132 Us staff developed increased confidence that excellent
(2,100 gpm) and a plug was removed from the filter quality water could be produced despite high raw
effluent header to release the negative pressure from water turbidities, and they developed a level of pride
the filter. Chemical feed rates were not changed. The that did not allow them to accept marginal finished
CCP team developed an action list and assigned tasks water quality, In addition, the jar test/filter paper
to the operating staff and administrators with due procedure proved to be a valuable process control
dates to ensure activity continued until the next site tool that allowed accurate selection of coagulant
visit. doses.

During an additional two site visits and weekly phone


consultation sessions, the CCP team explained the
conduct and interpretation of the jar tesVfilter paper
procedure to the operating staff. This, coupled with
activities from the first site visit. launched full
implementation of the process control program,
including evaluating raw water quality and making a
determination of the correct coagulant and filter aid
feed rates.

The only physical change to the plant was the


relocation of the feed pOints for alum and cationic
polymer addition to take advantage of a hydraulic flash
mix at an orifice plate located in the influent piping.
City administrators were also convinced to allow time
for the operating staff to remain at the plant so that
they could conduct process control testing and make
plant adjustments.

Figure 4-25 shows the significant improvement in


plant performance achieved by Plant 5 dUring the
conduct of the CCP. Plant operation improved after
reducing the plant flow rate and eliminating the
negative pressure on the filter bottoms In April, but
performance remained erratic until process control,
including chemical adjustments, was implemented in
July. After July, plant finished water turbidities
remained very consistent at about 0,1 to 0.2 NTU
through the duration of the project, even though raw
water turbidities varied widely (Figure 4-26). Plant
finished water quality remained below 0.3 NTU even
when the raw water turbidities reached 70 NTU,
because the operating staff consistently monitored
varying raw water quality and responded by changing
chemical feed rates,

Plant performance was especially impressive since


influent turbidities frequently exceeded values thought
330 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Figure 4-25. Flnlehed water turbldltl.. during CCP lor Plant 5.

1.4
1.3
1.2
1.1

0.9
:J
I- 0.11
Z
>. 0.7
:2 0.11
-e
::J 0.5
Fulura Requirement
I-
0.4
0.3
0.2
0.1

O+--~:"""'-....-,..-lil...-r---r--r-.....-.----,-......,...-.....---l
02-Apr-89 12-l'1oy-89 21-J",,-89 31-JuI-89 09-Sep-B9 19-0e1-89 28-Noy-89

Figure 4-26. Raw water turbldltl.. during CCP lor Plant 5.

BO

70

60

:\0
:J

il
I-
Z
>. 40
:2.0
:s

i~
I- 30

,,£;:j '~
20

10

Ii!I
0
02-Apr 12-Moy 21-Jun 31-Jul 09-Sep 19-001 2B-Nov
Case Studies 331

Plant 6 Raw water is pumped Irom the river to the pre-


sedimentation basin. Two of the raw water pumps are
Facility Description located in the basement of the plant and two are
Plant 6 was built in 1916, and underwent major located in a pumping station north of the plant. A
modifications in 1923 and 1960. To date the plant hand-cleaned basket strainer provides preliminary
serves approximately 5,000 people, with no significant screening for the two raw water pumps located in the
industrial users. The plant's water source is the plant basement. The five separate chambers in the
nearby river. The plant is a 175-Us (4-mgd) lime pre-sedimentation baSin are normally run in series. but
softening facility with pre-sedimentation, rapid mix and may be run in series, parallel. or any combination.
flocculation (in a solids contact unit), filtration, and Sludge is removed from the basin manually. and when
disinfection. The maximum quantity of water produced the pre-sedimentation baSin is cleaned the sludge IS
over a 12-month period was 1,182 Us (4.15 mgd) and discharged directly to the river. From the pre-
the lowest was 17 Us (0.39 mgd). Plant flow is sedimentation basin, water IS pumped by the three
measured by a raw water meter ahead of the solids intermediate pumps to the solids contact unit. ChlOrine
contact unit, and by a master meter following the may be added ahead of the solids contact unit. if
high-service pumps. necessary.

The city is currently under a compliance schedule In the solids contact unit. alum. polymer. lime. soda
from the State to curtail the discharge of untreated water. and activated carbon are added for softening.
sludge into the river. The city has applied for funds for turbidity removal, and taste and odor control.
the constructIOn of sludge handling facilities and for Flocculation and rapid mixing both take place In the
other plant modifications. Plant 6 consists of the contact unit. From the solids contact unit. water flows
following unit processes (see Figure 4- 27): by gravity to the recarbonatlon chamber. where
carbon diOXide generated onsite by burning natural
• Four split-case centrifugal raw water pumps: two 69 gas IS added to reduce pH after softening.
lis (1.800 gpm) and two 50 Ls (800 gpm) Polyphosphates are added prior to filtration.

• Pre-sedimentation basin divided into five sectIOns, The design filtration rate is 117 m 3 'm 2,d (2 gpm,sq ftl.
total volume 01 about 3.8 million L (1 mil gal) Filtration rates are controlled by an effluent regulator
on the basis of flow and headloss. The filters are
• Three non-clog centrifugal pumps (one 63-lIs back washed With finished water from the clearwell
[1.000-gpm] and two 94-L's [1 ,500-gpm]) as pumped by one 189-L s (3.000-gpm) backwash pump.
intermediate pumps for transferring water from the Backwash water flow is measured by a Venturi meter
pre-sedimentation basin to the solids contact unit and totalized at the main instrument control panel.

• Chemical addition of liquid cationiC coagulant, alum, Filtered water is disinfected with chlorine dosed at a
filter ald. polyphosphate, lime. soda water. carbon feed rate of 1-2 ppm. Disinfected water then flows to
dioxide. activated carbon, and gaseous chlorine the c1earwell. The clearwell detention time is 46
minutes at a plant flow of 175 L·s (4 mgd).
• Solids contact unit With a total surface area of 144
m 2 (1.555 sq It) and a 3.1·m (10-ft) depth. A The three high-service pumps deliver the water to the
mechanical 10·hp rapid mix unit IS located In the distribution system.
solids contact unit.
There are no facilities available for the disposal of
• Recarbonatlon chamber to reduce pH after sludge. Sludge generated from cleaning the pre-
soltening sedimentation basins. solids contact unit, and the
spent filter backwash water is piped untreated to the
• Seven Single media sand filters With a total surface (Iver.
area of 167 m 2 (1,800 sQ It): three in the plant
addition With surface area of 33 m 2 (360 SQ It) Major Unit Process Evaluation
each. and four in the old plant with surface area of The performance potential graph IS shown In Figure 4-
17m 2 (180 sQ It) 28. The raw water pumps were rated at 208 uS (4.75
mgd), based on Information from a prevIous
Two chlOrinators engineering study. The pre-sedimentation basin was
rated at 175 Ls (4 mgd) assuming that a coagulant
• 487, 130-L (128, 700-gal) clearwell aid could be added prior to the basin. At t 75 Ls (4
mgd), the baSin has a detention time greater than the
• Four vertical turbine high-service pumps: three 88- 3-hr minimum recommended by the Ten State
Us (1,400 gpm) and one 63-lIs (1,000 gpm) Standards.
ACliV3Led W
w
Carbon ."
ii I I'.)
Chlorine
Addilion I Polyphosphate

"
.,~ C
Soda
Waler
I :-' "0
~
OJ

;a'"
Alum and OJ Q:
Polymer Fillers in New ::l
Building
PresedimenlJltion Basin
I, '"
'C
(Cl

1-----
Lime

I, I,
Chlorine
Addition ....~ ~.
;::::l .
~ (Cl

I ~ 0,

Raw water I I I. Carbon


0.
OJ
I
Dioxide .-.- 4_ To
Distribution ~
OJ
0
'"
V>
1- - -

I I I System ?I <0'
::l

F}
::l
I
I §=
(Cl

1---
I
I leuwen
Z
'"~
I I Q-
I I I .5
:3 High Service
,0
:J
I
I I al Pumps "
: :l
"o
ii
(Cl

...g;:E
I
I " OJ
L __
I
I
i'" --i
;;;
I Drain 10
Inlermediale
Pumps ...3
OJ

4 River
....,'"
::l

OJ
Drain 10 Q.
Pre-sed bypass River
=.
'"
V>

Sotids ConlJlcl Unit Bypass


Wastewater
(spent backwash)
Filter 10
River
Pre-sed and Solids Contacl Unil bypass
Case Studies 333

Figure 4-28. Plant 8 performance potential graph.

Flow. mod
Unit Process 2 4 6

Raw-water Pumpa' I
Pr.aecUmentatlon Ponela'
HOT. min

Intennedllte Pumps' I
Contact Unlt 3
SOR. gPdlSQ II 643 1.286 I
Flltratlon'
Loading. gpmlSQ II
I
Disinfections
Contact lIme. min
=:J
High-service Pumpa I

Sludge Handling InadeQuate

Raled DeSign
Capaclly

1 Based on prevIous master plan study.


2 Based on additIOn of coagulanl alds . detention greater than 3 hr mmimum ,n Ten Stale Standards.
) Based on 1.440 gPdlSQ II.
• Based on 2 gpmlSQ II and assumes filler aJd wllh labor'lntenSlve backwash.
S Based on 120·mln contact lime and [olal clearwell volume.

The intermediate pumps were rated at 167 Us (3.8 1.0 NTU. the chlorine contact time must be adequate
mgd). slightly less than the 175-Us (4-mgd) design to ensure disinfection of any pathogenic organisms
capacity. based on a previous study. Based on a that may have passed through the plant's previous
surface overflow rate of 58 m 3 /m 2/d (1 gpm/sq tt). the treatment steps.
solids contact unit was rated at 96 Us (2.2 mgd). well
below the plant design capacity. The high-service pumps are rated at 241 Lis (5.5
mgd).Jn summary. the solids contact unit limits plant
The filters were rated at 228 Us (5.2 mgd) based on a capacity to 96 Lis (2.2 mgd) during periods of high
filtration rate of 117 m3 /m 2/d (2 gpm/sq tt). This rating turbidity. An operational rate greater than 96 Lis (2.2
assumes successful use of a filter aid and hand-raking mgd) may result in increased solids carryover from
of the filters during backwash. the solids contact unit to the filters. with a subsequent
decrease in filter performance and overall plant
Because the clearwell has such a small volume in performance. Also. if bypass of the solids contact unit
relation to total plant production, the disinfection becomes necessary, provisions must be made to add
contact lime was rated at only 8.8 Us (0.2 mgd). coagulants prior to filtration.
Current design standards require a minimum
disinfection contact time of 2 hr, but the clearwell Performance Assessment
provides only 46 minutes at 175 Us (4 mgd). When Finished water quality monitoring data indicated that
the plant produces water with a turbidity greater than the plant had been operating in compliance with the
334 Upgrading Existing or Designing New Drinking Water Treatment Facilities

current turbidity MCl of less than 1.0 NTU on a Figure 4-29. Plant 6 direct IIItratlon results. February 1988.
monthly average. However. the MCl had been
exceeded on days within certain months. The highest
finished water turbidities were noted at a time when
2.5
the solids contact unit was being bypassed for
cleaning. Plant records indicated that no coagulant
aids were added at times when the solids contact unit
was bypassed. an unacceptable practice.
2.0
Figures 4-29 through 4-31 show selected plant
operating data. The data shown in these figures
suggest that the plant will experience difficulty in
1.5
complying with the SWTR. which establishes a MCl
for turbidity of 0.5 NTU for 95 percent of the time.

Performance-limiting Factors
The factors identified as having a major effect on 1.0
performance on a long-term repetitive basis are
summarized below and listed in order of prionty.
Proposed Finished Water MCL
1. Operator Applications of Concepts and Testing to 0.5
Process Control - Operation: The operators had a
good understanding of water treatment but were
not applying that knowledge fully. For example.
direct filtration without chemical pretreatment can
result in a significant health nsk to the consumers. 10 15
The Filter 6 effluent turbidity plot presented in
Day of Month
Figure 4- 30 is from a specIal CPE study and
Indicates that the coagulation process has not
been optimized.

2. Disinfection - Design: The limited volume of the Figure 4-30. Plant 6 e"luent tumidity after backwash,
September 21, 1988.
clearwell results in inadequate chlorine contact
tIme. which in turn limits the time the chlorine has TUrbidity. NTU
to act on pathogenic organisms that may have
passed through the previous treatment process. 1.0
There was also an inadequate free chlorine
residual in the finished water leaving the plant.

3. Sludge Disposal - Design: No sludge disposal 0.8


facilities eXist to treat the sludge generated In the
water treatment process. The CIty is in vIolation of
the State Clean Water Act because of the
discharge of sludge to the river and has hired a 06
consultant to evaluate the problem and design a
remedy.
Proposed Finished Water MCL
Factors Identified as having either a minimal effect on 0.4
a routine basis. or a major effect on a periodic basis
were priontlzed and are summarized below.

1. Process Flexibility . Design: There are no 0.2


provisions for chemIcal feed to the pre-
sedimentation baSin or to the filter influent.
Addition of these chemical feed points would allow
operators to bet1er control finished water quality.
o 20 40 60 90
2. Process Control Testing - OperaliOn: The absence
Time. min
of or the wrong type of process control testing
results in improper operational control decisions.
The solids contact unit should be monitored more
Case Studies 335

Figur. 4-31. Plant 6 turtlldlty.

12

II

10
:::l
!-
9
Z
.,., 8
0
"8 7
"8
8><
I.1J 6
'">-
...
Cl 5
4

o Jun Jul
88 88
Date

thoroughly and more frequently. Sample taps and there is a heavy focus on maintenance at the expense
continuous monitoring turbidimeters should be of adequate process control. Action taken to address
installed on each filter eHluent so that discrete these factors may not noticeably improve plant
changes in filter performance may be observed performance. but may improve the eHiciency of plant
before substandard quality water is produced. operation:
Process control activities should be integrated into
the daily routine of operation. • The screens on the raw water pumps plug with
moss at certain times of the year. Cleaning the
3. Solids Contact Unit - Design: The solids contact screens is very labor intensive. Frequent plugging
unit severely limits flow during peak periods when of the screens aHects water quantity more than
the surface overflow rate becomes excessive. and finished water quality.
results in a high solids loading to the filters.
• If filter aids are used to improve filter performance.
4. Planning and Guidance - AdministratIOn: More it may be necessary to agitate the tilters by hand
emphasis should be placed on ongoing capital during backwash to facilitate cleaning. This practice
improvement and replacement of equipment. will be rather labor intensive.
Long-term reliability of the plant has been
jeopardized by a reluctance to make necessary Projected Impact of a CCP
expenditures. Admmistration should develop an The evaluation team believed that a CCP that
integrated approach to selting goals. not only for addressed factors related to plant operation. such as
maintenance and process control. but also for operator application of concepts. process control
meeting minimum finished water quality goals and testing. and flexibility could improve plant performance
sludge discharge limitations. for the majority of the year. However. continuous
compliance with proposed regulations for finished
In addition to the above major factors limiting water turbidity and disinfection. and with existing
performance. other factors identified as having a NPDES permit limitations on sludge discharge to the
minor eHect were noted dunng the evaluation. Overall. river. would require design modifications as well.
336 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Plant 7 River water is drawn from the concrete intake


structure located in the middle of the river through
Facility Description 41-, 36-. and 51-cm (16-, 14-, and 20-in) lines. The
Plant 7 is a conventional plant that treats water from a intake structure is modified during winter months to
river for industrial and domestic use by the city. The include a perforated culvert intake. which lessens the
peak day demand for a 12·month period was 202 Us impact of pack ice. A portion of the raw water from
(4.6 mgd) based on plant records. Plant 7 includes the the intake structure is pumped directly to a local
following unit processes (see Figure 4-32). refinery.

• Concrete river intake structure that houses Raw water pumps located in a pump station adjacent
manually cleaned bar screens, 1.9-cm (3/4-in) to the plant deliver the desired volume of water for
spacing. treatment. The operators indicated that the turbine
pump is unable to achieve rated capacity output. The
• Three manually operated raw water pumps: one pumps discharge to a Parshall flume. At the time of
turbine - 126 Us (2.000 gpm). and two centrifugal - the CPE, sodium silica fluoride and liquid alum were
126 Us (2,000 gpm) and 88 Us (1,400 gpm) being added just upstream of the throat of the Parshall
flume. The superintendent had only recently switched
• Three turbine pumps used to pump raw water from dry to liquid alum. Historically, polymer had also
directly to a large refinery; one 76 Us (1,200 gpm) been added at the flume location.
and two 126 Us (2,000 gpm)
After the flume. the water flows to the "mud basins."
• 46-cm (18-in) Parshall flume used to monitor raw a series of six tanks used for flocculation. The basins
water influent flow appeared to be functioning well even though the only
energy impact was that generated hydraulically. The
• Liquid alum feed pump with a backup dry alum feed operators Indicated that observing the floc formation in
system these basins was critical in adjusting the chemical
dosages to the raw water. These basins are not used
• Polymer feed pump in the winter because of icing problems. The basins
have noticeably deteriorating concrete, and a
• Fluoride feed system for sodium silica fluoride consultant had been retained to evaluate their
structural integrity.
• Six manually-cleaned, uncovered "hydraulic"
flocculation basins ("mud basins"), each with a From the mud basins, water flows in series through
volume of approximately 81,400 L (21,500 gal) two manually cleaned sedimentation basins. The first
basin is uncovered and IS used only during warm
• Two manually cleaned sedimentation basins. one weather months. During winter operation, the covered
covered and one uncuvered. Each basin has a basin is used for both flocculation and sedimentation.
volume of approximately 2.8 million L (750,000 gal). The basins are cleaned manually approximately six
Basin eHluent is transferred to the lilters through a times each year, and the sludge is pumped to the
S1-em (24-io) pipe earth sludge holding pond; 3-6 hr are required to
clean the basins. and no routine sampling is done on
• Two 8.5-m x 7.6-m (28-ft x 25-ft) mixed media them.
filters
EHluent from the covered sedimentation basin is
• Gas chlorination system collected in a 61 -cm (24-in) pipe and delivered to the
mixed-media filters. The filtration rate is controlled by
• Two filtered water "transfer" pumps controlled by a transfer pump located at the discharge from each
an altitude valve that "transfer" filtered water to a filter. The transfer pump discharges filtered water to
standpipe or to the high-service pumps the distribution headers of the high-service pumps and
to an on-site standpipe. The transfer pumping rate is
• On-site standpipe with a capacity of approximately controlled by demand and by an altitude valve which
94.625 L (25.000 gal) receives a Signal from a pressure sensor on the on-
site standpipe. This filter control has been set to
• Five high-service centrifugal pumps: two 157 Us respond relatively slowly so as to minimize rapid
(2.500 gpm). one 126 Us (2,000 gpm). one 63 Us fluctuations in filtration rate.
(1,000 gpm). and 38 lis (600 gpm)
Backwash consists of a surface wash and gravity
• Sludge holding lagoon of approximately 3.4 million backwash using the water stored in the standpipe. At
L (900,000 gal) capacity the lime of the CPE, filter runs were approximately 10-
20 days, and a dirty filter was started and stopped on
• On-site storage of sludge from the holding lagoon a daily basis. Backwashing was intended to be
Case Stud ies 337

Figure 4-32. Plant 7 proceu flow diagram.

Sludge Holding
Lagoon

Flocculation
I I
Basins
[I I I

~MSMll
Sedimentation
Basin

(uncovered)
)
~ume

Liquid
I Alum--
j
- - - Fluoride
Sedimentation I
Basin

(covered)
~
U
-r\
-LOI 1----1-1

~ \ Intake

Filter Filter Plant Raw


Water Pumps

( )
r r ~ High
Service

Transfer
~ I~ () Pumps

Pumps '-...J '---.../


()
( ) River

CI2
To Distribution Systcm
1
338 Upgrading Existing or Designing New Drinking Water Treatment Facilities

initiated when the headloss exceeded 2.4-2.7 m (8-9 sedimentation basins was assessed as 263 Us (6
ft) or the effluent turbidity exceeded 0.5 NTU. mgd).
Backwash water was directed to the sludge holding
pond. The mixed media filters were assessed based on a
filtration rate of 293 m3 /m 2/d (5 gpm/sq ft). At this
After filtration, the plant effluent is chlorinated and rate, a potential capacity of 438 Us (10 mgd) was
pumped by the high- service pumps to the distribution projected for the existing filters.
system. The plant does not have a clearwell to allow
contact time of the chlorine with the treated water, Disinfection capability was severely limited based on
and the first consumer is located approximately 275 m current state criteria of a 2-hr detention time after
(300 yd) from the plant site. The target chlorine chlorination. A capacity of 11 Lis (0.25 mgd) was
residual is 0.7-0.8 mg/L. projected based on the 2-hr standard and the
approximation that there are 274 m (300 yd) of 36-cm
Operators attempt to match water demand with high- and 41-cm (14- and 16-in) pipes prior to the first
service pumping. However, caution IS used in turning system user. Despite revisions of existing regulations
on high-service pumps so as to avoid high pressure in that may allow lower detention times, the absence of
the distribution system. Excessive system pressures any clearwell or contact basin will remain a unit
have resulted in broken lines because of the existence process limitation.
of old pipes in the system. The 38-Us (600-gpm)
high-service pump was out of service at the lime of Except for the noted limitation in the disinfection
the CPE. process, unit processes were assessed adequate to
meet the current peak demand of 202 Lis (4.6 mgd).
Sludge from the sedimentation basins and the filter
backwashes is stored in the unlined sludge storage Performance Assessment
lagoon from which it is dredged approximately every 2 In general, the CPE indicated that a high quality water
years and stored In a ditch adjacent to the lagoon. (e.g.. turbidities less than 0.15 NTU) was produced.
Supernatant from the lagoons is pumped back to the However, some difficulty with performance was
head of the plant. indicated during winter operation. Figure 4-34, which
shows a plot of filter turbidity and headloss vs. time,
Dried sludge is removed from the ditch and stockpiled indicates an excessively long recovery time after
on the plant side. Present stockpile volumes have backwash of Filter 1 (e.g., 36 hr until effluent turbidity
practically filled the available space, and alternative stabilized). Ideally, performance stability would be
disposal options will be necessary in the foreseeable achieved in less than 10 minut-es, Also shown is a
future, period of approximately 12 hr in which the turbidity
exceeded the current state criteria of 1 NTU. Cold,
Operation is almost totally manual. Pump settings, low turbidity, low alkalinity water and use of only the
chemical dosage rates, and filter backwashing are all covered sedimentation basin may have contributed in
initiated manually by the plant operators. The plant is part to the noted difficulties. Greater operational
staffed 24 hr every day with four operators rotating control and perhaps additional chemical conditioning
shifts every 28 days. A relief operator is available. may be required to Improve performance.

Major Unit Process Evaluation Another problem indicated on Figure 4-34 is the time
The performance potential graph is shown in Figure 4- of deteriorated performance before a filter backwash
33. The flocculation basins were rated based on a was implemented. Turbidities increased from the 0.1
detention time of 40 minutes, which resulted In a NTU range to the 0.35 NTU range, and approximately
capacity rating of approximately 206 Us (4.7 mgd). 1 full day passed before the operators initiated
The 40-mlnute detention time rating is relatively backwash. Failure to backwash when turbidity begins
conservative but is Justified because mechanical to increase can allow Significant breakthrough of
mixing is not provided. This rating is slightly higher particles, causing a potential health risk for the
than the current peak demand: therefore, the unit community. Figure 4-35 shows March 1989 turbidity
process is deemed satisfactory. data for Filter 2 With results similar to those in Figure
4-34. The performance problem Indicated on Figure 4-
The sedimentatIOn basins were rated on a hydraulic 35. however, is much more severe. Inillal turbidities,
loading rate of 12 m3 /m 2 /d (300 gpd/sq It ) and a after filter startup, exceeded the state requirements
detention time of 5.4 hr because of the relatively poor (1.0 NTU) for over a day. Additionally, backwashing
outlet structure (e.g., a 61-cm [24-inJ pipe) and the was delayed for approximately 3.5 days from the time
absence of mechanical sludge removal. Polymer indicated by plant data. Proper chemical conditioning
addition was assumed in establishing thiS raling. and closer attention to the need to backwash would
Based on these criteria, the capability of the have lessened the potential for health risk
demonstrated by these data analyses.
Case Studies 339

Figure 4·33. Plant 7 performance potential graph.

Flow. mgd
Unit Process 4 6 8 10 '2

Flocculation I
DetenlJon time. min 93 46 I
Sedlmentation 2

*
SOA. gpd/SQ ~ 100 200
HOT, hr 16.2 8.1 51 4

Filtration
HLR. gpm/SQ ~ 1 2 3 4
~

Dislnleclion 3
Contact tIme. min
~
Actual Max. Day
5/88·4/89
I Flocculation basin rated at 4Q·m,n HOT because there IS no mechanlcaJ mixing.
2 Rated at 300 gpd/sQ ~ and 5.4 hr because Ihere IS a poor outlet and no sludge collectIOn equipment Assumes polymer use..
3 Based on current Slate standard of 2·hr HOT. Assumed 300 yd of 14- and 16-ln pipe 10 first user. Standards are being revised and
lower detention tImes may be allowed for eXIsting plan IS.

During the CPE, a special study on filter startup was Factors identified as having either a minimal effect on
conducted under spring runoff water conditions. a routine basis, or a major effect on a periodic basis
During the study. Filter 1 was removed from service were prioritized and are summarized below:
for backwashing and Filter 2 was placed in service.
The results of this analysis showed that turbidiltes 1. Supervision/Staff Morale - Administrative:
never exceeded 0.15 NTU during and after startup of Communication between administrative personnel
Filter 2. The stability of the filter's performance was and the plant staff. and communication among the
also demonstrated when the flow was increased staff is limited and strained. As a result of the
dramatically over a short period of time and no limited communication: data interpretation.
increase in turbidity occurred. An additional initiative. maintenance efforts. proper process
evaluation of the surface of Filter 1 after backwash adjustments. understanding of standard operating
revealed no mudballs. indicating that the backwash procedures. and "cross training" of personnel are
procedure was adequate. limited. The operators seem to function
independently on each shift. This limited
Performance-Limiting Factors interaction is believed to be a major contributing
The following factor was identified as having a major factor to the poor operational decisions that
adverse effect on performance on a long-term resulted In the deteriorated filter performance
repetitive basis. documented in Figures 4-34 and 4-35.
Administrative skills and operator attitude will have
,. Disinfection - Design: A detention time of 2 hr IS to be addressed to eliminate the impact of this
not available at the plant to ensure effective factor.
disinfection of the treated water prior to use. The
new regulations that will be promulgated as a 2. Application of Concepts - Operation: Inability to
result of the SWTR and/or current state criteria apply proper concepts to optimize unit process
may necessitate capital improvements in order to performance was identified for several reasons.
prOVide adequate disinfection capability. This First the supervision/staff morale problem limits
factor was asterisked because the final rule and the capability of current personnel to learn from
direction could not be established until the latter each other and. therefore. limits their capability to
part of 1989. . properly and consistently apply basic concepts to
operational decisions. Also. the practice of
340 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Figure 4-34. Turbidity and headless prollles lor IIIler 1 - Plant 7.

.2 -s;l,----------------------------,

Current Standard

0
:::.
.:::
0.9

<Ii 0.8
'"0
'0
, Headloss
Cll
Ql
0.7 -t----t-------------------::::o...c---\t1
J:
0.6
"c
Cll
:J 0.5
I-
z
i- 0.4-
:g Turbidity
.a 0.3
I-
:s
02

0.\

0
2 3 5 6 7 8
February 1989

Figure 4-35. Turbidity prollle lor filter 2 - Planl 7.

5
.4-
3
2

Current Standard
:J
I- 09
z
i- 08
'6 Recommend
:J5 0.7
:s
I-
Initiate
Backwash
} t<lJ
06
0.5
1 ~ ,"', ~
\ f=~
0.4- ~
"\ ~ 1 ~
0.3
0.2 ~~
~",~F~
"'<:I .r
0.1
riP
'" CI

o
11 12 13 \ 4- \5 \6 \7 18 19 20

March 1989
Case Studies 341

routinely starting up a dirty filter and of occasion- Projected Impact of a CCP


ally "bumping" the filters with large changes in Generally. the plant was producing a high quality
flow rate can and has resulted In turbidity spikes treated water. As such. a CCP would not dramatically
and the associated breakthrough of particles. Improve water quality. However. intangibles such as
Particle breakthrough represents a high potential communication. operational concepts. and a "tight"
health risk. fiscal policy represent deep-seated problems that may
be difficult to address using past practices. From this
3. Policies· Administration: The implementation of point of view. an external facilitator may be necessary
very tight fiscal policies limits the expenditure of to impact the current situation.
funds for necessary items. The lack of in-line
turbidimeters for the two fillers. an additional In lieu of a CCP. it was recommended that the
polymer feed pump. and use of polymer was communtly pursue a correction effort to address the
assessed to be impacting plant performance. The factors identified as limiting plant performance.
CPE did not identify where these policies were recognizing that they Will be difficult and time
originating. It was felt that improved consuming to eliminate.
communication would greatly assist in addressing
this factor.

Chemical feed capability was identified as having a


minor impact on performance. Moving the
alum/coagulant feed pOint from ItS present location to
the throat of the Parshall flume would allow a better
rapid mix to occur and may result In lower chemical
use. As a side benefit. the raw water sample could be
taken in the chemical feed room. although the sample
would contain fluoride.
342 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Plant 8 River water flows to the raw water pump wet well
through a pipe that is located inside the corrugated
Facility Description steel intake structure. The original Intake configuration
Plant 8. operated by the county sewer and water consisted of an infiltration gallery located in the river:
district. treats water from a nearby river for domestic however. due to plugging of the gallery with sediment.
use by the town and a local rural water system. The the system was replaced in 1986 with the current
plant. which began operation in 1981. utilizes a solids intake pipe. The operator and board members
contact clarification/filtration process. Extreme indicated that the original infiltration gallery provided
variations in turbidity exist in the river. During the water of much lower turbidity than the current Intake
winter. turbidity is generally below 30 NTU and can be structure. However. as the gallery became plugged
as low as 10 NTU: however. during spring and With sediment. raw water capacity was reduced to the
summer. turbidity ranges from 40 NTU to as high as extent that the board decided to replace the infiltration
several thousand NTU as the res\'1t of storm events. gallery with the present system.
Plant 8 includes the following unit processes (see
Figure 4·36). The two raw water pumps. located in a wet well
adjacent to the river. each deliver the rated plant
• 1.2-m (4-ft) diameter corrugated steel intake which capacity. Pump operation is based on a level signal
extends below the river bottom. The steel intake from the plant clearwell. and operation is alternated to
pipe IS perforated and packed with gravel in the maintain even run times on both units. The plant
area below the river bottom operator reported that a pump is typically operated at
full capacity: however. a butterfly valve located in the
• Wet well with two 16·Ls (250-gpm) submersible plant can be used to reduce pump capacity. In an
raw water pumps. The pumps are automatically attempt to maintain a sludge blanket in the solids
operated from a level signal in the plant clearwell. contact unit. the operator had throttled pump capacity
to reduce How to the unit. The operator felt that this
• 15-cm (6-ln) diameter propeller meter for measuring experiment did not stabilize the process. and Since
the raw water flow rate that time the plant has operated at 16 Ls (250 gpm).

• 110-4.375 kg (240·9.645 Ib)/d dry alum feeder The raw water flow rate is indicated and totalized
through a propeller meter located in the raw water line
• 3·15 L (0.1-0.5 cu ft)/hr lime feeder inSide the plant building. FollOWing the flow meter. raw
water IS directed into the solids contact unit.
• 3·15 L (0.1-0.5 cu ft)/hr soda ash feeder Components of this unit include an uoflow mixing
column with a turbine mixer. a downflow flocculation
• Polymer feed system with a 170-L (45-gal) dilution cone. and an upflow clarilier with tube settlers
tank and a 3.7-20.8 L (1' 5.5 gal)/hr leed pump throughout the settling area.

• Sleel. 5.5-m (18-ft) diameter. 3.2·m (10.5-ft) deep Water treatment chemicals including alum. lime. soda
solids contact unit with 45· tube settlers ash, and polymer can be added into the miXing area.
DUring the winter. when water hardness increases in
• 1.3-6.8 L (0.34-1.8 gal)/hr polyphosphate feeder the river. the operator can add lime and soda ash to
the solids contact unit for softening. During the
• 11-kg (24-lb)'d carbon dioxide feed system and an remainder of the year. alum and a polymer are
associated in-line static mixer typically used for turbidity removal. The operator
reported that the plant was not operated in the
• Steel 3.2-m (10.5-ft) diameter dual-media filter softening mode last winter and that alum and a
cationic polymer were used for coagulation during this
• 9-kg (20 lb),d gas chlorination system period. At the time of the site visit. the operator
indicated that 80 mg/L of alum and 2 mg/L of cationic
• 56.000-L (14.800-gal) clearwell with baffled polymer were being fed to the mixing tube.
compartments
A variable speed turbine mixer located on top of the
• Two 16·Ls (250-gpm) vertical turbine pumps which mixing tube directs the incoming raw water upward
pump treated water to the distribution system and a through the tube and out into the flocculation cone.
storage tank The mixing tube is open at the bottom. thus allowing
the reCirculation of flocculated water with the incoming
• Two sludge holding lagoons with a total capacity of raw water. The amount of reCirculation depends on
approximately 658.200 L (173.900 gal) the speed of the turbine mixer. According to the
operator. the turbine mixer speed has been operated
• Spent backwash storage lagoon of approximately throughout the entire range. without changing the
2.1 ml'lion·L (556.170-gal) capacity performance characteristics of the solids contact unit
343

FlgUre~4~.~36~'--------------
Plant 8 -------------------
Case Stud'ies

procesl flow dlag ram.

River

/
Inlake
Pipe

Raw
Waler
Pumps

Spent
Backwash
Pond Backwash

I
Conl101
Valve
Backwash
Supply
Line

Flow
Meier

High
Service
Pumps

1
Stalic
Mixer

Alum, Lime
Soda Ash '
Polymer'

To
Dislribution
Syslem
344 Upgrading Existing or Designing New Drinking Water Treatment Facilities

except at high speed. At high speeds. turbulent the elevated storage tank. When operated in
conditions in the unit degraded clarifier performance. automatic mode, the plant may start and stop more
The current practice is to operate the turbine at low to than once daily with each start on a "dirty" filter.
medium speed. and the turbine and sludge scraper According to the operator. the clearwell is taken out of
continuously, even when the plant is not treating service once each year to remove accumulated
water. sediment in the tank bottom.

FollOWing the mixing tube. the conditioned water Operation is usually by the automatic mode (i.e ..
enters an inverted cone where flocculation occurs. At storage tank elevation). During peak water usage
the bottom of the flocculation cone. water moves periods in the summer months. the plant operates 12-
outward into the clarifier and, theoretically. through a 14 hr each day at the design rate 16 Us (250 gpm).
developed sludge blanket. Water then flows upward Less operating time is required during the winter. The
through 45-degree tube settlers before overflowing a plant is staffed about 2-3 hr each day by one operator.
peripheral weir. The operator reported that he has had Board members assist with plant maintenance and
a difficult time controlling the sludge blanket in the repair on an as-needed baSIS. The plant operator is
solids contact unit. It appeared that sludge solids also responsible for a booster pump station and
generated In the process are typically lost over the distribution system serving the rural water members of
weir and deposited on the filter. the district. The operator and board members reported
that maintaining this part of the system can be very
Excess sludge generated in the process can be time consuming because of frequent leaks that occur
directed to two sludge lagoons by opening a sludge in areas of the rural distribution system.
blow-off valve. The operator indicated that sludge has
never been removed from the two sludge lagoons Major Unit Process Evaluation
since startup of the plant. Effluent from the solids The performance potential graph is shown in Figure 4-
contact unit flows onto a dual-media filter. If the plant 37. Mixing. flocculation, and sedimentation processes
is operating in the sottening mode. carbon dioxide can all occur in the solids contact unit. Although
be added for pH adjustment atter the solids contact flocculation IS typically evaluated on the performance
unit. Mixing is accomplished by an in-line static mixer. potential graph, It was not included here because the
Polyphosphale can also be added at this locatIOn to process occurs Within the solids contact unit.
control calcium carbonate bUildup in the filter. Even Flocculation in solids contact units is enhanced by the
though the plant was not operating In the sottening recirculation of flocculated water through the mixing
mode during the site visit. polyphosphate was stili fed tube With the incoming raw water. The sedimentation
to the solids contact unit effluent. component of the solids contact unit was rated at a
surface overflow rate of 58 m 3/m 2 /d (1,421 gpd/sq ttl.
Water level In the filter IS controlled by a modulating which is slightly less than the design rate. This rating
flow control valve. According to the operator and is also based on incoming raw water turbidity levels of
several board members. this control valve has never less than 500 NTU, which is typical for most of the
provided a constant-rate operating condition as year, and a relatively shallow depth of 3 m (10 ttl. For
intended by the original design. Until just recently. occasions when turbidity IS greater than 500 NTU. the
water level in the filter would fluctuate dramatically sedimentation component was rated at a surface
dUring a filter run. causing sudden, high magnitude overflow rate of 32 m 3 /m 2 /d (790 gpd/sq ttl or 8 Us
flow rate changes through the filter. ThiS condition has (125 gpm). At thiS turbidity level. control of the sludge
improved significantly since the plant operator and a blanket might be difficult and solids loss from the unit
board member fabricated an adjustment mechanism to would begin to affect filter performance. The plant
control the travel distance of the valve seat. could overcome this limitation by operating at the
lower flow rate of 8 Us (125 gpm) over a longer
Backwashlng of the filter can be initiated by either period of the day.
filter headloss or filter effluent turbidity. The backwash
supply water IS provided by the pressurized The dual media filter was assessed based on a
distribution system including an elevated storage tank, filtration rate of 176 m 3 /m 2;d (3 gpm/sq ttl. At this
and the backwash rate is controlled by a pressure rate. a potential capacity of 16 Us (250 gpm) was
redUCing valve. Spent backwash water is directed to a projected for the existing filter. In some cases. dual-
storage lagoon located adjacent to the plant. The media filters have been rated over 176 m3/m 2/d (3
operator reported that spent backwash water seeps gpm/sq ft): however. because of the complex
into the ground water and has never accumulated In operations associated with the solids contact unit. a
the lagoon. conservative filtration rate was selected. This rating
also assumes that the existing effluent control valve
Treated water from the filter is chlOrinated and limits extreme variatIOns in water flow rate through the
directed to a baffled clearwell. The two vertical turbine filter. Although the modifications made to the control
pumps deliver treated water to the distribution system. valve appear to have significantly improved filter
Pump operation IS based on a water level Signal from operation. adequate time was not available during the
Case Stud ies 345

Figure 4-37. Plant 8 performance potential graph.

FlOw, gpm
Unit Process 50 100 150 200 250 300

SOlids Contact Unit'


SOR. gpdlsq n 316 632 947 1.263 I

Filtration 2
HLR. gpmlsq n 06 1.2 1.7 2.3 21 9

Disinfection
Contact lime. hr 4.9 25 16 1.2 110

Design
Flow

I Capacity reduced to approXimalely 125 gpm when lurbldlty > 500 NTU. Shallow. 10·n deep clanfler I,m'ls capacity to < 250 gpm
when turbidity> 500 NTU.
2 Assumes adeQuate filler eHluenl control valve.

site visit to thoroughly evaluate the valve's basis over the past year, and have been more severe
effectiveness. during the winter months when low turbidity, cold
water was treated by the plant.
The chlorine contact basin was rated at 16 Us (250
gpm) based on a hydraulic residence time of 1 hr. During the CPE, a special study on filter startup under
Disinfection capability is typically based on current two different conditions was conducted. The first
State criteria of a 2-hr residence time after condition consisted of starting a dirty filter (i.e.. a
chlorination. The 1-hr residence time was allowed in condition in which a backwash had not occurred
this case because of the efficient baffling that exists in before filter startup) and monitoring effluent turbidity.
the contact basin. The operator indicated that thiS condition occurs
rOUlinely at the plant because of the automatic mode
At the present time, the plant operates at the 16 Us of operation. A graph of filter performance under this
(250 gpm) rate for about 12-14 hr/day dunng the peak condition is shown in Figure 4-39. As shown. follOWing
demand period. The performance potential graph filter startup effluent turbidity immediately increased to
indicates that this peak demand can be met when raw over 5 NTU and then gradually decreased to about
turbidity is less than 500 NTU Without significant 1.5 NTU after 20 minutes. Results of this test indicate
changes In hours of operation. To meet the peak that starting a dirty filter results In turbidity levels
demand when turbidity IS greater than 500 NTU. the above the 0.5-NTU limit for an extended period of
plant would have to operate at a reduced rate of time and presents a Significant danger of passing
approximately 8 Us (125 gpm) over a 24-hr period. pathogeniC organisms through the filter.

Performance Assessment The second condition of the special study involved


ThiS plant has hlstoncallY had operaliOnal problems backwashing the filter and monitOring the effluent
associated with the operation of the solids contact turbidity after it was placed In operation. As shown in
unit. During the CPE site Visit. the plant operator Figure 4-40, turbidity after backwash increased
reported that maintaining a sludge blanket in the solids immediately to a peak value of 13.5 NTU. About 25
contact unit has been a problem since startup. Other minutes after the filter startup, the effluent turbidity
operations-related information obtained during the site decreased to the 0.5-NTU level. This condition also
VISit Indicated that performance problems have been indicates the potential for pathogenic organisms to
more common than indicated by plant monitOring pass through the filter. Properly condiliOned filters
reports. Figure 4-38 shows the number of days that typically expenence a turbidity spike of less than 0.2
treated water turbidity exceeded 0.5 NTU (the SWTR NTU for less than 15 minutes.
turbidity standard for treated surface water) for each
month over the past year. This analysis indicated that Several problems associated with the backwash
performance problems have occurred on a frequent contributed to the subsequent poor performance of
346 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Figure 4-38. Plant 8 performance.

12

11

10

9
::J
I- 8
z
'"0 7
A
~
:a 6
~:::>
l-
S
'">.
0'" 4
0
z
3

o
Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May

the filter. When the backwash sequence was initiated, NTU at the time of the site visit, alum was probably
the surface wash valve did not open. According to the being overfed at this dosage rate. Following further
operator, the surface wash normally operates the investigatIOn into this problem, it was determined that
entire length of the backwash period. The filter was the operator had exchanged the alum feeder with the
washed for over 23 minutes, a much longer than lime feeder. ThiS change was implemented so that the
normal duration for this functIOn, without getting clean. feeder with a vibrator could be used for lime addition.
Because of the dirty condition of the filter, the The operator was not aware that the new alum feeder
backwash rate was manually increased by adjusting had a 3.8-cm (1.5-in) feed screw instead of a 1.9-cm
the control valve, but never succeeded in cleaning the (3/4-in) feed screw, thus causing a higher-than-
filter. Measurements of the rise rate In the filter at the expected alum dosage.
beginning of the backwash indicated a backwash rate
of approximately 645 m 3l m 2 :d (11 gpm/sq ftl. less Performance-limiting Factors
than the minimum recommended value of 878 The factors Identified as having a major eHect on
m 3 /m 2 :d (15 gpm/sq ft). Proper adlustment of the performance on a long-term repetitive basis were
backwash control valve would allow an adequate prioritized and are summarized below:
backwash flow rate.
1. Water Treatment Understanding - Operation: Lack
An additIOnal factor that could have contributed to the of operator understar,ding of water treatment has
plant's poor performance during the CPE site visit been a major cause of the performance problems
relates to the alum feed rate to the solids contact unit. experienced at the plant. The plant operator never
The operator indicated that the alum feed rate to this received any formal training on operation of the
unit was approximately 80 mg/l. Upon checking the plant. ThiS situatIOn is compounded by the fact
alum feeder by weighing a sample of dry alum that the solids contact process is complex and
collected over a selected time period, it was requires a high level of process control to achieve
determined that the actual alum feed rate was about good performance. Examples of this lack of
177 mg/L. Since the raw water turbidity was about 80 understanding include incorrect calculations of
Case Studies 347

Figure 4-39. Turbidity prollle • dirty filter stanup - Plant 8. alum and polymer feed dosages, startup of dirty
filters, inadequate process control testing, and an
Effluent TUrbidity, NTU inadequate filter backwashing procedure.
6
2, Process Control Testing· Operation: The current
process control testing in eHect at the plant is
inadequate. The only process control testing
currently done is periodic jar testing to determine
chemical dosages. Jar testing should be
completed whenever raw water quality changes
significantly. Control of the solids contact unit
requires daily monitOring of influent and effluent
turbidity, sludge blanket concentration and
location, and blowoff sludge volume and
3 concentration. The plant does not have adequate
testing eqUipment to perform many of these tests.
Although a jar testing apparatus is available at the
2 plant, it does not accurately simulate the
flocculalton and coagulation processes. Equipment
necessary to monitor the solids contact unit
Includes a blanket finder and a centrifuge.

3. Plant Coverage - AdministraliOn: Paid coverage at


the plant is presently 2- 3 hr,day, 7 daysiWeek. If
unanticipated problems develop at the plant or
o 10 15 20 distribution system that require time beyond the
routine amount. the operator and board members
volunteer time to correct the Situation. The
operator's son currently acts as a backup
operator: however, he has not received any formal
traming and is not certified in water treatment.
~ur. 4-40. Tumidity prollle • IIIter stanup aMer backwash Given the complex nature of the plant and the
Plant 8.
potential for extreme variations in raw water
quality. coverage at the plant needs to be
Nluenl Turbidity. NTU
extended. Ideally, whenever the plant is in
14 operation. an operator should be at the plant
mOnitoring its pertormance. A minimum of 4 hr
each day probably would be required to perform
12 routine process control testing, data analySIS,
reporting, and preventive maintenance activities.
Realistically, It would probably be difficult to staH
10 the plant at all times that It is in operation because
of the small size of the district. A possible
compromise could include increasing plant
8 coverage to a minimum of 4 hr/day, not Including
the distribution system. and adding an alarm,dialer
system at the plant.
6
4. Insufficient Funding/Bonded Indebtedness .
Administration: The district received a 40-year
4
loan from the Farmer's Home Administration to
fund the constructIOn of the plant in 1981.
Repayment of this loan requires approximately 50
percent of the district's revenues at the present
time. Because of the small size of the district.
approximately 46 town customers and 62 rural
o 10 15 20 25 customers. and the large indebtedness. water
rates are moderately high relative to other similar
TIme. min systems. Rates were recently increased to assist
in rebuilding a reserve fund that was depleted
when the new intake structure was installed. Even
348 Upgrading Existing or Designing New Drinking Water Treatment Facilities

with these moderately high rates. revenues are particles to pass through the filter. thus affecting
marginal to cover the cost of additional plant treated water quality. Recently, a throttling
coverage and minor capital improvements. mechanism fabricated by the operator and a board
member has limited these fluctuations in water
Factors identified as having either a minimal effect on level. With this modification completed. this factor
a routine basis. or a major effect on a periodic basis moved to a lower priority; however. replacement
were prioritized and are summarized below. of the valve should stili be considered when
funding is available.
1. Plant Staff. Number - Administration: The plant is
staffed by one operator at the present time. and 5. Performance Monitoring - Operation: During the
he provides coverage 7 days/week. Occasionally. CPE site visit. performance monitoring records
the operator's son performs as a backup operator were reviewed. and some performance problems
when the operator is not available. A trained. were occasionally noted. However. interviews and
backup operator is needed at the plant to routinely special studies conducted during the CPE
relieve the regular operator. Several of the board revealed serious performance problems at the
members recognized this problem and expressed plant. Since records did not accurately reflect
their concern with locating a person who would be actual plant performance. regulatory agency
willing to work on a part-time basis at the pay rate reviews were not able to establish that a
they could afford. The operator's son has performance problem existed. Accurate reporting
expressed some interest in becoming certified in would probably have resulted in pressure from the
water treatment. Given his work experience at the regulatory agency and correction of some of the
plant. this option may be worth pursing by the factors noted in this report.
district. Another option could involve utilizing an
interested board member as the backup operator. 6. Raw Water Turbidity - Design: As noted by the
Once trained and certified. the backup operator performance potential graph. raw water turbidity
could provide plant coverage on alternating above 500 NTU is projected to limit plant capacity.
weekends and during vacations. High turbidity water typically occurs during the
spring through fall. and only occasionally during
2. Alarm System - Design: An alarm and automatic this period as the result of runoff from storm
plant shutdown capability are available when high events. The high turbidity problem can most
turbidity is recorded from the filter. As is the case realistically be handled through operational
with any type of automation, this function has changes and minor expenditures for testing
failed on occasion and high turbidity water was equipment and an alarm system. If the current
directed to the clearwell. This alarm function peak demand remains the same. the plant may be
provides a necessary safeguard against able to treat high turbidity water by reducing the
contaminated water entering the distribution flow rate through the plant and operating for more
system and should be routinely checked and hours during the day. If operational measures are
maintained. Because of the high degree of not successful, there should be added flexibility to
variability in the raw water source. it would be direct the raw water through a pre-sedimentation
advantageous to have a similar high turbidity basin With chemical addition capability. A pre-
alarm and automatic plant shutdown capability for sedimentation pond could be used as a backup
this source. This capability could warn the water supply during short runoff events or used to
operator about a change in raw water turbidity and lower raw water turbidity during longer storm
allow time to adjust chemical feed dosages. events.

3. Chemical Feed Facilities - Design: Because of the 7. Sedimentation/Solids Contact Unit - Design: The
exchange made between the lime and alum solids contact unit's capability to treat water under
feeders. the present alum feeder does not appear a variety of conditions may be limited by its
to have a satisfactory range to feed low dosages relatively shallow 3-m (1 O-ft) depth and high
under certain water quality conditions. Since surface overflow rate. Under these conditions.
overfeeding alum can detrimentally affect plant maintaining a sludge blanket in the unit can be
performance, this problem will have to be accomplished; however. considerable process
corrected. The operator may want to Investigate control testing and pOSSible adjustments to the
changing the feeders back to their original plant flow rate are required. The short hydraulic
functions. If lime is to be fed in the future, a new residence time in the unit may limit the plant's
shaker may be required. capability to treat cold. low-turbidity water. Under
cold water conditions. chemical reactions are
4. Process Controllability - Design: Since startup of slower and longer residence times are required.
the plant. the filter effluent control valve has Longer residence times can be achieved by
caused erratic control of the water level in the operating the plant at a lower rate for long periods
filter. This rapid change In filter water level causes of time.
Case Studies 349

A factor identified as having a minor effect on plant Projected Impact of a CCP


performance is the existing chemical feed As indicated by the performance potential graph and
arrangement. which limits the injection of alum. factors limiting performance. this plant does have
polymer. lime. and soda ash to the mixing tube in the some design deficiencies. However. operatIonal
solids contact unit. Under cold. low turbidity water changes at the plant and administrative support could
conditions. it would be advantageous to add the be used to overcome most of these deficiencies. As
coagulation chemicals ahead of the flow meter. This such. implementation of a CCP could demonstrate
injection location would allow more intense mixing and dramatic improvement in treated water quality. Before
slightly more detention time when treating cold. low a CCP could be implemented. however. the district
turbidity water. would have to commit to providing the additional
staffing and coverage reqUired to operate the plant
and the expenditures necessary to purchase the
necessary testing equipment.
350 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Plant 9 or start the plant. based on the water level in the


finished water reservoir. At the time of the evaluation.
Facility Description the valve was being operated manually to maximize
Water is supplied to the city from four sources: a the depth of treated water in the finished water
direct filtration plant and three ground-water wells that reservoir.
augment the water supply during summer months.
The CPE was limited to the direct filtration plant. Alter passing the control valve. the water flows by
which treats water from an infiltration system for gravity through an in-line static mixer and onto the
domestic and commercial use by the city. The peak automatic backwash filter. The filter was designed to
day demand for a 12-month period was estimated at operate at a filtration rate of 117 m 3 /m 2 /d (2 gpm/sq
74 lis (1.7 mgd). Plant 9 includes the following unit It). but was being operated at approximately 211
processes (see Figure 4-41): m3 /m 2 /d (3.6 gpm/sq It) at the time of the CPE. The
filter has approximately 28 cm (11 in) of sand media in
• Infiltration system consisting of perforated 54 20-cm (8-in) sections.
subsurface laterals that are connected to shallow
caissons The filter sections are separated by fiberglass
dividers. which were warped at the top. Because of
• 8.330-L (2.200-gal) basin that serves as a sand trap the warping. some sections were only 2.5-5 cm (1-2
in) wide at the top. while other sections were over 20
• In-line static mixer with two elements cm (8 in) wide at the top. This variation was caused
by migration of the sand media from one section to
• 30-m 2 (324-sq ft) monomedia (sand) travelling another during backwash.
bridge automatic backwash filter
The filter can be backwashed automatically by
• Two polymer feed pumps: one for feeding neat headloss or by timer. During the evaluation. the filter
polymer and one for feeding diluted polymer was being washed automatically based on headloss.
with little consideration given to filter effluent turbidity.
• Gas chlorination system with two chlorinators. each When the filter is backwashed. a travelling bridge
23-kg (50-lb)/d capacity passes across the filter and washes each section.
One pump on the bridge pumps water back up
• On-site treated water reservoir. 33.5 m (110ft) in through the filter section and another pulls the spent
diameter and 2-m (6.S-It) deep (1.7 million L backwash water from the top of each section through
[462.000 gal]) a shroud to a discharge channel adjacent to the filter.
The backwash water flows by gravity to the backwash
• Propeller meter on the discharge line from the storage basin. Following filtration. the water is
treated water reservoir chlorinated prior to the filter level control weir and
flows into the treated water reservoir. Effluent from the
• On-site 12-m x 12-m x 2-m deep (40-ft x 40-ft x reservoir flows by gravity 8 km (5 miles) to town
6.5-ft) concrete backwash holding basin. which through parallel 20-cm (8·in) and 25·cm (1 O-in)
discharges to an irrigation ditch transmission lines.

Water from one of two creeks is diverted onto a hay Sludge from filter backwashes is stored in the
field adjacent to the plant where the water percolates concrete backwash storage basin. Supernatant from
through several feet of soil to perforated laterals the basin is diSCharged over a weir to an irrigation
buried under the field. Water flows through the laterals ditch. According to the operator. sludge is removed
to concrete caissons. which are fitted with metal from the baSin every 3·4 years by wheelbarrow and
covers. A line then carries the composite flow from front-end loader and is spread on adjacent fields.
the caissons to the direct filtration plant. Microscopic
particulate examination of the infiltration system water Operation is primarily manual except for filter
has shown that It is directly impacted by the surface backwashmg. Plant operation. flow rates. and
water and. therefore. should be considered a surface chemical dosage rates are all initiated manually by the
water source. plant operators. The two plant operators spend
approXimately 0.5-1 hr each day at the plant checking
Raw water flows by gravity from the infiltration system the operalton of the plant. During that time. the plant
to a sand trap basin in the plant. CationiC polymer is is Inspected to ensure equipment is operating
fed at the end of the basin alter which the water flows properly: no significant process control activities are
through a control valve that regulates the amount of conducted. The operators are also responsible for the
water treated in the plant. Any excess water from the wastewater treatment plant. wastewater collection
infiltration system flows over a weir at the influent end system. water distnbution system. streets. parks.
of the sand trap basin to an irrigation return ditch. The airport. swimming pool. and grave digging at the
control valve can operate automatically to shut down cemetery.
Case Stud ies 351

Figure 4-41. Plant 9 process flow diagram.

Infiltration
Laterals

Automatic
Backwash Filter
Slatic
Sandtrap~,-_-'- -,
r- Mixer

Overnow _

Control
Valve
Polymer

Bypass Line /

Chlorine/ Filler
Emuent
Weir

Backwash
Storage
Basin
Overflow

To Town
352 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Figure 4-42. Plant 9 performance potential graph.

Flow. mgd
Unit Process 05 , 0 15 20

Filtration'
gpnvsq h ,., I
Disinfection
COnlact time>. hr 2.9 I
Contaci tlme J , hr 63 3.2 2.1

Estimated Max.
= 1.7 mgd

, Rated at 2 gpnvSQ h because there IS only " ,n of sand medium.


Z Based on current State standard of 2·or HOT to first lap. Assumed '0 percent volume 01 clearwell and 0.75 m, of 8· and 'O'ln pipe to
,,,st user. HOT refers to time It takes water to travel from the plant to the first lap.
Z Based on current Slate standard of 2·or HOT to first lap. Assumed 10 percenl volume 01 clearwell and 2.5 m, of 8· and 'O·,n pipe to
hrst user In town.

DUring the evaluation, peak water use in the City was Because of the limitation In the filtration process, the
approximately five times greater than typical water use plant was assessed as inadequate to meet the
for a community with no large Industrial users. This projected peak demand. However, water use on a per
extensive use required the City to augment the surface capita basis was noted to be extremely high. Normal
water supply with three ground-water wells. water use would result in daily water production of
Determining the cause of the excessive water use and about 39 Us (0.9 mgd), which the plant should be
taking measures to lower it to normal levels would able to handle on a continuous basis.
allow the plant to operate at lower flow rates.
Performance Assessment
Major Unit Process Evaluation A review of the operating records indicated that the
The performance potential graph is shown in Figure 4- raw water was of very good quality with peak
42. Peak day demand for the plant was estimated by turbidities of 0.7 NTU. Treated water was also of good
measuring the depth of flow over the filter effluent quality with turbidities normally about 0.2 NTU and
welf and applying that flow on a 24-hr basis. This flow with a peak of 0.45 NTU. However, turbidity of very
was then compared to prolected plant capabilities. As clear waters, such as the water from the infiltration
Figure 4·42 shows, the monomedia sand filter was system, is often not a good indication of
assessed at a potential capacity of 39 L s (0.9 mgd) bacteriological quality. In fact, previous state
based on a filtration rate of 117 m 3.m 2l d (2 gpm/sq ft). particulate tests revealed that the filter was not
The filter capability was limited because a direct filter removing a signIficant number of particles from the
uSing 28 cm (11 in) of sand cannot be expected to raw water. The operating data revealed that only
provide consistent performance at rates higher than about 50 percent of the raw water turbidity was being
117 m3 'm 2 /d (2 gpm1sq tt). removed. Plant data taken during periods of no
chemical feed and some chemical feed indicated little
D,s,nfection capability was rated based on current difference between raw and treated water turbidity
state criteria of a 2-hr detention time after chlorination. between the two.
As Figure 4-42 shows, two conditions were rated:
detention time to the first tap downstream from the During the ePE, a special study was conducted to
plant and to town. DiSinfection detention time was determine the effect of backwashing the filter on
adequate to town at flow rates up to 77 Us (1.75 treated water quality. The filter was backwashed and
mgd). However, detention time to the first tap only samples were collected as near as possible to the
resulted in a plant capacity of 33 Us (0.75 mgd). filter cell being washed and at the filter effluent weir.
Standards are being revised and different criteria may Figures 4-43 and 4-44 present the results of the
be used to allow lower minimum detention times for study. As shown, at both sample points effluent
existing plants, which would likely rely on effective turbidity increased significantly (to 5 and 7 NTU)
filtraliOn. during backwash and remained above the 0.5 NTU
Case Studies 353

Figure 4-43, Turbidity profile dUring and after backwash at IiIter effluent weir· Plant 9,

TurbIdIty, NTU Completed 151 Pass

Completed Wash

o 10 20 30 40 50 60

Minutes

Figure 4-44, Turbidity profile during and after backwash at ceil effluent - Plant 9.

Turbidity, NTU
Completed 151 Pass
8

6
Completed Wash:

FuMe TurbidIty SlaJldard


------------------------------------------------------ -:---------------~----

o
o 4 8 12 16 20 24
354 Upgrading Existing or Designing New Drinking Water Treatment Facilities

limit for over 60 minutes. The study results indicate monitored following a backwash so that chemical
that a significant amount of the material removed by feed could be optimized to reduce the increase in
the filter was allowed to pass through the filter into the turbidity (turbidity spike) that occurs alter a
treated water reservoir. The significant increase in backwash.
turbidity is especially indicative of poor performance.
since the raw water turbidity was only about 0.6 NTU 4. Disinfection - Design: A detention time of 2 hr is
during the special study. needed to ensure eHective disinfection of the
treated water prior to the first user. The new
The special study results. State microscopic regulations that will be promulgated as a result of
evaluations. and plant daily records indicate that the the SWTR and/or current state criteria may
plant is not eHectively removing particles found in the necessitate capital improvements before the water
raw water. Should a significant number of parasitic system has adequate disinfection capability. An
organisms such as Giardia cysts or Cryptospofldium example of a capital improvement would be the
oocysts occur in the raw water (for example, as a installation of baHle walls in the clearwell to keep
result of cattle feeding on the grass above the the water in the basin longer for disinfection rather
infiltration system), they would likely enter the plant than taking a direct route through the basin from
and pass through the filter. Since some cysts are the influent to the effluent pipe. This factor was
resistant to disinfection by chlorine, they could pose a asterisked because the final rule will not be
significant health hazard to the community. effective until June 29. 1993, following
development of State criteria in 1990.
Performance-limiting Factors
The factors identified as having a major eHect on 5. Filtration . Design: The filter is presently being
performance on a long-term repetitive basis were operated at too high a rate to expect adequate
prioritized and are summarized below: performance on a continuous basis. In addition.
warping of the filter section dividers and the
1. Water Treatment Understanding . Operations: The potential inability of the travelling backwash
plant superintendent/utilities director is a very mechanism to properly wash the filter could
motivated operator; however. he has not received impact filter performance. This factor was
any formal training in water treatment practices. asterisked because it may be possible for the
This is compounded by the complexity of plant to operate at a flow rate consistent with its
operating a direct filtration plant treating cold. low capability, if water use is reduced to normal
alkalinity. low turbidity water. Lack of water levels. Under this condition. the filter dividers and
treatment understanding was identified as the top backwash may prove not to significantly impact
ranking factor because it has led to poor operating performance.
deciSions. such as little or no change in chemical
feed rates. filter backwashing based on headloss Factors identified as having either a minimal eHect on
rather than filtered water quality, and bypass of a routine basis. or a major eHect on a periodic basis
untreated raw water to the treated water storage are summarized below in order of priority.
reservoir. Operation of the plant will require a
process control program and an understanding of 1. Process Controllability - Design: The effluent flow
coagulation chemistry including chemical feed meter does not adequately measure the plant flow
calculations. rate because it is located downstream of the
finished water storage reservoir. The flow rate out
2. Process Control Testing - Operations: There was
of the reservoir is not indicative of the plant flow
no process control program in place at the plant.
rate. Since accurate flow measurement is the
Operation of a surface water plant requires that
basis for chemical feed calculations and filter
testing be conducted and results recorded in a
hydraulic loading rates. actual plant flow needs to
systematic manner so that data is available to
be accurately measured.
make process control decisions. Control of the
direct filtration plant will require daily mOOlloring of
2. Laboratory Space and Eguipment - Design: The
influent turbidity. continuous mOOlloring of filter
plant is not equipped with a jar test apparatus.
effluent turbidity. and jar testing to select
Because of the raw water quality characteristics
appropriate coagulant aids and to determine
(low turbidity. alkalinity, and temperature). special
optimum chemical doses.
studies with various coagulant and f10cculant aids
3. Process Automation . Design: The plant is not will likely be required to optimize plant
equipped with a continuous reading and recording performance. A jar test apparatus will be
turbidimeter. which is necessary to adequately necessary to conduct the special studies as well
monitor plant performance since the staHing levels as to optimize plant chemical feeds.
do not allow turbidity tests to be conducted more
than once each day. A continuous reading 3. Alarm System . Design: There were no alarm
turbidimeter would allow filter performance to be systems in the plant to warn the operator of
Case Studies 355

problems. such as chemical feeder shutdown or Projected Impact 01 a CCP


raw water quality changes. Since the plant is only Plant 9 produces water that poses a significant health
checked once each day for about 1 hr, it is risk to consumers. Conducting a CCP could result In
essential that alarms be provided to warn the an improvement in finished water quality. especially
operator of a change in conditions. If alarms are during and after filter backwashing. However, because
not provided, the plant should be staffed any time peak water demands exceed the rated capacity of the
it is in operation. filter and disinfection system. the plant would have to
be operated at a lower flow rate. In addition. design
4. Watershed Management - Design: Allowing cattle aspects of the filter such as backwashing
to graze on top of the infiltration system provides effectiveness and uneven filter dividers. could limit
an unnecessary public health risk. Cattle are filter performance to the extent that the plant could not
known carriers of Cryptosporid;um. a parasitic meet regulatory requirements on a continuous basis.
cyst that is extremely resistant to chlorine The CCP might discover that limitations in the filtration
disinfection and small enough to easily pass system require major capital improvements to ensure
through a poorly operated filter. A direct filtration continuous compliance with applicable regulations.
plant provides a limited number of barriers to
pathogenic organisms and limited response time
for the operator to react to a change in raw water
quality.

No factors in the administration or maintenance areas


were identified as impacting performance.
356 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Plant 10 wet well where it is picked up by the raw water


pumps.
Facility Description
Plant lOis a conventional water treatment plant Raw water pumps located on top of the wet well move
supplied from a nearby river that provides water for the raw water from the intake structure to the plant.
domestic use. The peak operating flow for the plant Only one of the 22-L's (350-gpm) pumps is used.
was established at 22 lis (0.5 mgd) based on a Though the plant was originally designed for 44 Us
review of flow records for the previous year. The plant (700 gpm), the plant staff feels thaI 22 Us (350 gpm)
is normally operated for approximately 8 hr/day; is the maximum flow that can be handled because of
however. on several days the plant is operated for limitations with the sedimentation basins. Since the
longer than 8 hr to meet demands of peak water use. plant is operated at a constant rate of 22 lis (350
On these days, the treatment processes are still gpm), variations in water demand are met by varying
operated only at the 22-Us (0.5-mgd) flow rate. Plant the length of time the plant IS operated. Raw water
10 consists of the following unit processes, shown flows entering the plant are measured and recorded. A
schematically in Figure 4-45: manual rate-of-flow controller is available though not
normally used.
• Intake structure located on the bank of the river,
consisting of a manhole intake structure and a wet Volumetric feeders discharge dry chemicals into tanks
well from which raw-water pumps deliver the water below the feeders, where water IS added and mixed to
to the plant. A bar screen is provided between the make a slurry. As the raw water flows to the
manhole intake structure and the wet well sedimentation basins, alum and lime are added as
slurries into the pipe. Chemical feed rates can be
• Three manually operated vertical turbine raw water adjusted by varying the amount of chemical added to
pumps: two 22 L,s (350 gpm) and one 16 Us (250 the slurry tanks. These adjustments are made based
gpm) on observations of floc formation in raw water
samples that have been placed on a magnetic stimng
• Raw water flow measurement consisting of an 20- apparatus after chemical feed. No mechanical or static
cm (8-in) orifice meter with a chart recorder. Also, flash mixing is provided. Chemical feed rates are not
a manual rate-of-flow controller rOUllnely adjusted.

• Volumetric feeder each for alum, lime, and After chemical addition, raw water flows to the two
powdered activated carbon sedimentation basins. These units were originally
designed as a type of upflow solids contact clarifier.
• Two 9.1-m (30-ft) diameter sedimentation basins, eliminating the need for separate flash mix and
each with a surface area of 66 m 2 (707 sq ft) and a flocculation processes. but they are no longer
volume of 307.000 L (81,100 gal) operated as designed. Flow enters through a 20-cm
(8-in) pipe at the bottom of the unit and strikes a small
• Two 19.300-L (5,1 OO-gal) recarbonation basins baffle redirecting the flow in the basin. Basin effluent
discharges through peripherally mounted submerged
• Four 2.5-m x 3.0-m (8.7-ft x 10-ft) filters with 51 cm orifice weirs. Each basin originally had a rotating arm
(24 In) of sand media located near the bottom powered by preSSUrized basin
effluent. Basin effluent was to be withdrawn and
• 94-Ls (l,500-gpm) backwash pump pumped back through nozzles in the arm causing it to
rotate and promote flocculation. The rotattng arm has
• Gas chlOrination system been removed from one of the units and is not
operational in the second. Sludge is manually
• Two clearwells: one with a capacity of 567.750 L removed from the basins twice a year. The plant
(150.000 gal) and the second with a capacity of discharges thiS sludge to the sanitary sewer or back
56.775-L (15,OOO-gal) to the river.

• Two 47-lis (750·gpm) vertical turbine high-service Settled water flows by gravity from the sedimentation
pumps basins to the four sand filters and is controlled by float
valves in the recarbonation basins. These valves shut
• Venturi-type flow meter, totalizer, and chart off sedimentation basin effluent flow If the level in the
recorder filters exceeds the maximum. Operators visually
monitor the filter water levels and adjust the flow uSing
Water from the river is pooled behind a low head dam rate-of-flow controllers. Flow meters are available for
across the river downstream of the intake structure. each filter. but are not used for filter flow adjustment.
Several pipes extend out into the river from the
manhole Intake structure allowing water to be taken During the CPE, the standard practice for filter
from different locations. The water then flows to the backwashing was to wash two of the filters each day
."
is
tT° SaIlllafY Sewe,
•.
c:

to
---..... Backwash Waler and SluClge
...... !"

I --
Senled Wale,
-- --
I "Cl
OJ
a
I +
0
To River
I ~.
I I ..
I I <;
~
a.
OJ
Sand Fillers
J I

I .
~

I + Filler 3
I ;J

I I

w I F,niShed
I SedimentatIon Basins F,lle, 1 Filler 2 Waler

I I
L BackwaSh

AawlWatcr
Chlollne
Lime

Alum

High' Servrce
Flowmeler Pumps
Q
V>
Raw Water
IrHake
Flowmeter
'"
CJ)

\
0-
ro'
V>

W
01
-...I
358 Upgrading EXisting or Designing New Drinking Water Treatment Facilities

using backwash water from Clearwell 2. Headloss or The filters were rated at 22 los (0.50 mgd) based on a
turbidity measurements were not typically used to filter loading rate of 58 m 3/ m 2 :d (1 gpm/sq ft). This
initiate backwashing. The backwash pump discharges loading rate is lower than typical values because of
through a rate-of-flow controller. but there are no the air binding observed by the CPE team during filter
valves that can easily be operated to slowly start and backwashing. Air binding results in air pockets in the
stop the backwash flow to the filters. The surface of filter media. which prevents water from passing
the sand is manually raked during the backwashing. through that portion of the filter. effectively lowering
The plant discharges backwash water to the sanitary the surface area available for filtration. The filter
sewer or to the river. During the CPE. the backwash loading rate, therefore. was lowered in the
water was being discharged to the river. assessment to compensate for the loss of filter area
due to the observed air binding.
Water from all four filters is discharged to Clearwell 1,
immediately following injection of chlorine gas into the The disinfection system was rated at 24 L s (0.54
pipe. Chlorine doses are controlled to provide a mgd). Future drinking water regulations for disinfection
residual of 2.5-3.0 mgJL. Finished water normally flows will be based on CT values found to be needed ior
from Clearwell 1 to Clearwell 2: piping is provided to various removals of Giardia cysts and Inactivation of
allow bypassing of Clearwell 1. Clearwell 2 also viruses. CT is the disinfectant concentration multiplied
serves as the suction pipe for the high-service and by the actual time the finished water is in contact with
backwash pumps. the disinfectant. To establish the CT required, it was
assumed that the plant's diSinfection system would
Two high-service pumps supply finished water to the have to provide 2 logs (99 percent) of cyst removal
two in-ground storage tanks that feed the Village water with 2 logs of removal credited for the other treatment
distribution system. These pumps are operated processes. The total of 4 logs of cyst removal
manually based on water levels in the storage required for the plant was based on the CPE team's
reservoirs. estimate of the quality of the raw water.

Major Unit Process Evaluation To achieve the 2 logs of cyst removal, the CPE team
The performance potenlial graph is shown in Figure 4- estimated that the disinfection system would have to
46. Flocculation is a key major unit treatment process. provide a CT of 133. This CT value is for chlOrine at a
As originally designed. the sedimentation basins were 2.0 mg/L dose, pH 7.5, and temperature of 5C. The
to provide both flocculation and sedimentation: contact time was based on the chlorine being added
however. the flocculation aspect of these units has ahead of Clearwell 1 and the flow pasSing through
been removed or is inoperable. The CPE team doubts both clearwells. Only 15 percent of the theoretical
that the units were ever capable of providing detention time in the clearwells was used because the
acceptable flocculation even if operated as designed. clearwells are not baffled and because they are
As such. the flocculation process was given a peak subjected to fill and draw operation. The actual levels
instantaneous rated capacity of 0 LJd. This rating of disinfection required for the plant in the future will
implies that the plant cannot be expected to be determined by the State. The CPE estimates of the
consistently produce the desired water quality of less required total number of log reductions and the
than 0.5 NTU without adding flocculation process allowances for actual contact times in the clearwells
capabilities. may change when the final slate regulations are
developed.
The sedimentation basins were rated at 12 lis (0.28
mgd) based on a surface overflow rate of 8.1 m3 /m 2/d The performance potential graph shows that the lack
(200 gpd/sq It). The surface overflow rate is of flocculaliOn severely limits the capabililieS of the
significantly lower than that for other types of circular treatment processes. Without adequate flocculation,
sedimentation basins with the same depth as Plant the CPE team estimates that there is essenlially no
10. The projected peak instantaneous operating flow where the required performance can be Obtained.
capacity of the sedimentation basins was lowered The sedimentatIOn basins also severely limit the
because of the extremely poor inlet conditions. With capacity of the plant, even If adequate flocculation
the inlet structure located in the bottom of the was prOVided. These processes prevent the plant from
sedimentation baSins. the influent flow disrupts the achieving desired performance at the current peak
settled solids and tends to carry them upwards instantaneous operating flow rate of 22 Us (350 gpm).
towards the effluent. Properly designed sedimentation The filters and disinfection system were projected to
basins introduce the Influent water near the surface be adequate to treat this flow.
through an inlet structure that directs the flow into the
basin. promoting the separation of solids from the Performance Assessment
clarified liquid over the entire surface area. This allows Turbidity data from the plant records for the raw
the separated solids to move by gravity to the bottom water, settled water from the sedimentation basins.
of the basin and the clarified effluent to move to the and finished water over a 1-yr period are plotted in
surface where it is removed. Figures 4-47, 4-48, and 4-49, respectively.
Case Studies 359

Figure 4-46. Plant 10 performance potential graph.

Flow, mgd
Unit Process 01 02 03 04 05 06

Sedimentation 1
SOA. gpdlsq ~ 70 141 I
Filtration'
HLR. gpmtsq ~ 0.2 0.4 0.6 0.8 1

Dllinfectlon 3
Contact time. min 356 178 119 89 7:1 I

Peal< Instantaneous Operating


Flow (1 Pump) • 350 gpm

, Rated at 200 gpdlsq ~ because of poor Inlel condillons. turbulence at the basin bonom. and poor dutface area development.
2 Rated at 1 gpmJsQ ~ because 01 observed alf binding.
J Rated at CT • 133 with 2 mglL chlorone dose. whiCh reQUIres a 67·mln HOT; allowed 15 percent 01 available volume dfar contact
lime. temperature' 5'C. pH • 7.5.4.109 reQUIred reducllon. 2 allowe<l lor plant.

Figure 4-47 shows the fluctuation of raw water experienced an approximately O.3-NTU rise in turbidity
turbidity over the year. Well-designed and operated that did not drop back to the original value even after
treatment processes are expected to produce a water 30 minutes. These results could indicate a problem
With consistent turbidity levels even With wide with the filters or that the water being applied to the
vanations in raw water turbidity. As shown in Figure 4- filter has not been pr";Jerly treated and conditioned in
48, the sedimentation basins produced a settled water the preceding unit processes.
that also had significant variations in turbidity. The
filters reduced the levels of turbidity. as shown in An evaluation of the filter media was also performed
Figure 4-49. but still experience variations as raw dunng the CPE. The evaluation team determined that
water turbidity changes. These results indicate design the filters were being adequately backwashed, since
andror operational problems. Figure 4-50 shows the no significant mudballs were found. They did find
finished water turbidity during a 6·month period when some buildup of chemicals on the surface of the
the plant was treating a highly variable turbidity raw media. but these were not considered to affect
water. The applicable regulation for turbidity is performance. Air binding was also observed during the
currently 1.( NTU. Future regulations will require the special studies.
plant to meet a 0.5-NTU finished water turbidity 95
percent of the time. Figure 4-50 shows that the plant Performance-limiting Factors
generally compiles With the 1.0-NTU regulation, but is The factors identified as having a major effect on
conSistently above the 0.5 NTU required by the future performance on a long-term repetitive basis were
regulations. A probability plot of thiS same data is prioritized and are summarized below:
shown in Figure 4-51, which indicates that under
present conditions the plant would only meet the 0.5- 1. Flash Mix and Flocculation - Design: The plant
NTU standard approximately 50 percent of the time. has no flash mix or flocculation treatment
processes. facilities which are required to properly
During the CPE. a special study was completed to condition the raw water with chemicals prior to the
assess the filter performance after backwashing. With sedimentatIon and filtration treatment processes.
adequate facilities and operation of preceding unit Without these capabilities the plant will have
processes, a properly operated filter should produce a significant problems removing enough turbidity to
finished water turbidity of approximately 0.1 NTU and consistently meet future regulations.
only expenence a 0.2-NTU nse in turbidity in the
finished water for approximately 10 minutes after 2. Sedimentation Basins - Design: Adequate
restart following backwashing. Filters 3 and 4 were sedimentation basins are required to remove the
sampled for a 30-minute period after restart following coagulated turbidity from the raw water. The
backwashing. The results of this special study are sedimentation basins have basic limitations related
shown in Figures 4-52 and 4-53. Both filters to the lack of a proper inlet structure. At current
360 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Figure 4-47. Raw water turbidity profile· Plant to.


lSI
lSI-I+o..............+-++++o..................-++-+-++......-t...................+-++t-..................-t+-............-++......
o
_--t-
o
lSI o
CD o 0
"6
o
o
o o
0g 0

o
o
lSI 0
(\J

RUG88 OCT DEC FEB89 RPR JUN RUG OCT

Figure 4-48. Settled water turbidity profile· Plant 10.

0<11>
o
o a
o
o o 0 00
<II> o o o
::::> (\J
o o 00
I-
Z o

lSIt+-.......+-++-+t+....................- t -............ ....,I--......_ - t -...................-t++-++............-+f-<.............+-++-!-


RUG88 OCT DEC FEB89 RPR JUN RUG OCT
Case Studies 361

Figure 4-49. Finished waler lurbldlty profile· Planl 10.

..,.

II'l
cry 0 0

cry 0

II'l
:::>
l- N 0
Z
0
~
:6 N 0
:0 0 <:>
:;
I- II'l
0
0
0 00
0 0

II'l 0

lSI
RUG88 OCT DEC FEB89 RPR JUN RUG OCT

Figure 4-50. Finished wltlr turbidity profile· Plant 10.

N+-----+-----+-----+----o---!------t-;--.. . .- +
o e

o
o o
o 00
o o o
Present
Requirement

Future
Requirement

ISI+---.. . .-FEB
JRN89 t-;----t-+-----!-----f-----f-----+
MAR RPR MAY JUN JUL
362 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Figure 4·51. Probability plot offlnllhtKl wlter furbldlty • Pllnt 10.

m-l-----t-----+-----+--1f--.......---t-----t----t

CD

r.D

::l
I- lf1
Z
~
:2
.0
..,.
~
I-
M

.I 10 50 90 99 99.9
Probability

Figure 4·52. Flull... 3 8"'uenl turbidity profile alter backwash Fl9unl 4-53. Filter" effluent turbidity profile alter baclrweeh •
• Plenll0. Pllnt 10.

TurbKlily, NTU Turbodily. NTU


0.8 0.8

0.6 0.6

0.4 0.4

02 0.2

o 10 15 20 25 30 o 5 10 15 20 25 30
Minutes Minutea
Case Studies 363

loadings (e.g .. 22 LIs [350 gpm]). the basin 2. Lack of Preventive Maintenance Program -
limitations allow high turbidity water to pass to the Maintenance: The plant has no formalized
filters, thus degrading their performance. preventive maintenance program. Equipment is
repaired as it breaks down. A lack of maintenance
3. Number of Plant Staff· Administration: Currently a of a number of key pieces of equipment was
staff of three persons have responsibility for the considered to have a minimal, but continuous
operation and maintenance of the water plant, impact on performance. This key equipment
wastewater plant. distribution system, and included the alum feeder, filter flow measurement
collection system. CPE interviews with the Board devices and control valves, chlorinator controls,
of Public Affairs revealed that a fourth person may clarifier equipment including the weirs, and the
be hired after construction of the new wastewater raw water pumps. Plant staffing levels appeared to
treatment plant is completed. To respond to the impact the level of preventive maintenance, but
variations in raw water turbidity, increased even with adequate staff a formalized program
coverage of the water plant to make process would be needed to assure availability of key
changes will be needed. Given this requirement equipment.
pius other observed responsibilities. even a staff
of four may be inadequate. 3. Chemical Feed Facilities - Design: The plant
needs additional chemical feed facilities to
4. Application of Concepts and Testing to Process consistently meet required performance. A
Control - Operation: The plant staff appeared to polymer feed system is projected as a
have proper training. understood the basic requirement to optimize filter performance.
concepts of process control, and were very especially during cold weather operation. A
motivated. However, the chemical doses were not backup alum feeder would also be required to
changed based on changes In raw water assure a consistent source of chemical feed.
characteristics, the alum dose measured during
the CPE was excessively high. and the filters 4. Alarm Systems - Design: The plant is operated for
were operated without adequate consideration of significant periods of time without any operations
the turbidity levels in the filter effluent. staff present to make process adjustments in
Additionally. dirty filters were regularly started response to variations In raw water characteristics
without any assessment of the impact of this or correct problems with key processes
practice on filter effluent turbidity levels. Because equipment. During periods of unattended
of these practices, the CPE team assessed that operation process performance could degrade to
the plant staff was not consistently applying a point where it poses a potential health risk to
proper water treatment concepts and process the village. A turbidity monitoring system tied to
control testing to optimize the plant's raw and finished water could be used to alert the
performance. plant staff to process problems before finished
water quality reached undesirable levels.
Factors identified as havmg either a minimal effect on
a routine basIs or a major effect on a periodic basis The age of some of the equipment was Identified as a
are summarized below in order of priority. minor factor limiting performance. Though not a
performance-limiting factor, the current practice of
1. Filtration - Design: The configuration of the disposing of sludge and backwash water into the river
filtration system allows conditions that create air is in violation of State regulations.
binding In the filters. The air binding is caused by
negative pressures being created in the filter Projected Impact of a CCP
media as solids are removed and head loss The CPE identified numerous design problems related
increases. This causes the dissolved air in the to the key unit treatment processes, which must be
water to come out of solution and be retained in corrected before any process optimization through
the filter media. Water cannot pass through the use of a CCP could be successful. A CCP, therefore,
portions of the media where the air is retained, was not recommended.
which effectively reduces the surface area
available for filtration. This condition reduces the
plant's capacity and can significantly affect filter
performance.
364 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Plant 11 Raw water flows to the rapid mix basin where ferric
chloride solution is added via a diaphragm metering
Facility Description pump. A hydrated lime slurry is also fed into the rapid
Plant 11 is a conventional water treatment plant that mix basin via a volumetric feeder. A volumetric feeder
treats water from a nearby river for domestic use by is in place to feed alum. although it IS currently not is
the village. Based on a review of flaw records for the use. Flash mixing was not performed during the CPE
previous year. the peak operating flow for the plant due to a bearing problem with the mixer. However, the
was established at 19 lis (0.43 mgd). Plant 11 CPE team calculated the G value for the rapid mix
consists of the following unit processes shown basin to be adequate at 1,418 sec" if the mixer were
schematically in Figure 4-54. operating. Chemical feed rates can be adjusted
manually by the amount of chemical added to the
• Intake structure slurry tanks. or by adjusting the stroke on the
metering pump, but they were- not routinely changed.
• 49 million-L (13 mil-gal) reservoir
Raw water flows to the center flocculation cone of the
• Three vertical turbine raw water pumps: two t9 Lis solids contact clarifier. After flocculation, water enters
(300 gpm) and one 38 uS (600 gpm) the outer clarifier portion of the unit and is removed
through peripheral v-notched weir troughs. Sludge IS
• 19-Us (300-gpm) submersible raw water pump periodically removed automatically from the clarifier.
through use of a timed blow-down. and discharged to
• Metering pump to feed a ferric chloride solution a sanitary sewer for disposal at the wastewater
treatment plant.
• Two volumetric feeders. one for lime and another
for alum addition Settled water flows by gravity from the reactor
clarifiers to a basin (originally designed as a
• Rapid mix basin with a surface area of 0.46 m 2 (5.0 recarbonation basin), where the flow IS split to the
sq It); and a depth of 0.6 m (2 It) three filters. Rate of flow through the filters is
controlled by float-activated butterfly valves that open
• Solids contact clarifier 9.1 m (30 It) in diameter and or close to maintain a constant water level above the
3.8-4.1 m (12.5-13.6 It) deep; f1occulator volume. filter media.
29.800 L (7.875 gal); clarifier effective surface area.
61.6 m2 (663 sq It), and volume, 205,330 L (54,250 Filter runs are normally 24-27 hr. With the plant
gal) operating a total of 8-9 hr/day. One filter is
backwashed each day, so that filters operate 3 days
• Three dual media filters, two 2.4 m x 2.5 m (8 It x 8 before backwashing. No individual filter headloss or
It), and one 2.1 m x 2.7 m (7 It x 9 It) that contain turbidity monitoring equipment exists. although the
46-51 cm (18-20 in) of sand and 20-25 cm (8- 10 village intends to install headloss gauges. Backwash
in) of anthracite media water is supplied by distribution system pressure from
operation of the high-service pumps. The backwash
• Vacuum-controlled solution feed chlorination rate was determined during the CPE to be
system fed from 68 kg (150-lb) cylinders approximately 1.110 m 3/ m 2 id (19 gpm1sq tt). The
surface of the media is manually raked during
• 151,400-L (40,000-gal) clearwell backwashing. Backwash water is discharged to the
sanitary sewer.
• 321.725-L (85,000-gal) clearwell
Water from each filter flows through separate pipes
• Two 24-Us (375 gpm) vertical turbine high-service into Clearwell 1, where a chlorine solution is injected
pumps to maintain residuals between 2.3 and 2.5 mg/L.
Finished water normally flows from Clearwell 1 to
• 15-cm (6·;n) orifice plate With a totalizer-indicator- Clearwell 2. The high-service pumps take suction from
recorder Clearwell 1.

Water from the river is pooled behind a low dam Two high-service pumps supply finished water to the
across the river downstream of the intake structure. two elevated storage tanks that feed the water
River water IS normally pumped to the reservoir by the distribution system. These pumps are operated
submersible pump. Water from the reservoir flows by automatically based on water levels in the elevated
gravity to the wet well beneath the Intake structure. storage tanks. The plant comes on and off line usually
where it is pumped to the plant by the vertical turbine two to three limes during the day based on water
raw water pumps. These pumps can also pump river levels in the clearwell. Typically. the plant operates 8-
water directly to the plant, bypassing the reservoir. 9 hr/day at a constant rate of 24 Us (375 gpm).
'TI
.E
c
iil
Backwash Water f'

-
To Sanitary Sewer _
---, and Sludge '"
!"
."
ii'
a
::
Solids '0

.~
COniaCI
Unll III

i
D-
o;
It)
iil
;!

I ChlorIne

t Clearwell 2

RawlWaler F,ller 1

- F,ller 2

--
Dual·Med,a FIllers
Clearwell 1

HIgh-Service

lIme
L Backwash Waler --Pumps

-- Flowmeter

Fernc Chloride

Raw Waler Intake

.;_.c,_ .;~ ; a; : ,r.; " .\M,r and Pumps


00 Elevaled Tanks
(')
Ql
Vl
co
(/)
C
0-
co
V>

w
0)
tTl
366 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Variations in demand are met by varying the length of operating flow rate of 19 Us (0.43 mgd). The filtration
time the plant is operated. system. rated at 36 uS (0.82 mgd). was rated
considerably more than adequate to treat this flow.
Major Unit Process Evaluation Lack of baffling in the clearwells and the piping
The performance potential graph is shown in Figure 4- arrangement between the two clearwells limited the
55. The instantaneous peak operating flow of 19 Us projected capacity of the disinfection process.
(0.43 mgd) assumes that only one of the small raw
water pumps will supply the plant. Performance Assessment
Figure 4-56 shows turbidity data from the plant
The flocculation capabilities of the reactor clarifier records. The current applicable regulation for turbidity
were rated at 25 Us (0.57 mgd) based on achieving a is 1.0 NTU. The plant normally produces water with
hydraulic detention time of 20 minutes. This rating turbidities less than 1.0 NTU, but js frequently above
depends on retrofitting the basin with the variable the 0.5 NTU level. which will be required by the
speed drive for the mixers. which was included in the SWTR. as shown in Figure 4-56. A probability plot of
plant's original design. this same data. shown In Figure 4-57. indicates that
the plant would only meet the 0.5-NTU requirement
The sedimentation capabilities of the reactor clarifier approximately 30 percent of the time.
were rated at 22 Us (0.50 mgd) based on a surface
overflow rate of 30 m3 lm 2 /d (750 gpdlsq ft). The During the CPE. a special study was conducted to
projected capacity of the basins was lowered due to assess filter performance after backwashing. With
the constraints of the 3.8- m (12.6-ft) basin depth. adequate facilities and operaliOn of preceding unit
processes. a properly operated filter should
The filters were rated at 36 L's (0.82 mgd) based on a expenence a 0.2-NTU nse in turbidity in the finished
filter loading rate of 176 m 3im 2 ,'d (3 gpm1sq ft). This water for approximately 10 minutes after restart
loading rate was decreased from more typical values following backwashing. Figure 4-58 shows the results
for dual media filters because of the rate control of a study that sampled Filter 3 for a 40-minute period
system. after restart following backwashing. The filter
experienced an approximate 1.0-NTU rise in turbidity
The disinfection system was rated at 15 Us (0.34 that did not drop back to the original value even after
mgd). Future drinking water regulations for disinfection 40 minutes. This delay may be attributed to a problem
will be based on CT values found to be needed for in the filter. or to improper treatment and conditioning
removal of Giardia cysts and inactivation of viruses. of the water prior to filtering.
This evaluation used a CT of 127. which is for
chlorine at a 2.4 mg/L dose. pH 8.0. and temperature After backwash. Filter 3 was drained and 2.5- to 3.8-
of 5OC. It was assumed that the disinfection system cm (1- to 1.5-in) mudballs were observed in the
would have to provide 1.5 logs of cyst removal with media. As mudballs Increase in size. they can settle to
2.5 logs of removal credited for the other treatment the bottom of the media and limit the flow through
processes. The 4 logs of total cyst removal required these portions of the filter. The filter flow IS then
for the plant was based on the CPE team's estimate forced through the remaining media at higher rates.
of the potential for contamination of the raw water. which can impact filter capacity and performance.
The contact time was based on the chlorine being Proper backwashing procedures. such as adequate
added ahead of Clearwell 1 and the flow passing length of backwash. gradual increase in backwash
through both clearwells. Only 10 percent of the flow rates. and sufficient agitaliOn of the media. can
nominal detention time in the clearwells was used minimize the occurrence of mudballs.
because (1) the clearwells are not baffled. (2) the
piping arrangement does not assure that all of the flow Performance-limiting Factors
passes through both c1earwells. and (3) the clearwells The factors identified as having a major effect on
are subject to fill and draw operaliOn. In the future. the performance on a long-term repetitive basis were
actual levels of disinfeclion required for the plant will prioritized and are summarized below:
be determined by the State. The estimates of the
required total number of log reductions and the 1. Application of Concepts and Testing Process
allowances for actual contact times in the clearwells Control - Operation: Several operational practices
may change after the final State regulations are performed by the plant staff. including applying
developed. ferric c'lloride and lime at the same point.
regularly started dirty filters, and an unawareness
As shown in the performance potential graph. the of the condition of the filter media. impaired plant
major unit processes. with the excepl10n of the performance. Lime raises the raw water pH above
disinfection process. have a rated capacity close to or the range necessary to achieve optimum
exceeding the instantaneous peak operating flow. The coagulation and flocculation using ferric chloride.
flocculation and sedimentation process. although Starting dirty filters without monitoring the impact
borderline. were prolected adequate to treat the on finished water turbidity may represent a
Case Studies 367

Figure 4-55. Plant 11 performance potential graph.

Flow. mgd
Unit Process 02 04 06 08 10

Flocculation'
HOT. mIn 57 28 I
SlKIlmentatlon'
SOR. gpdlSQ It 302 603 I
Flitratlon J
HLR. gpmlSQ It 0.7 1.5 2.2 291

Ol.infectlo....
ConlaCt ~me. mIn 90 I
Peak Instantaneous Opera~ng
Flow (1 Pump) • 350 gpm

1 Rated at 20 ·m,n HOT - assumes varIable speed drive would be added.


2 Rated at 750 gpdlSQ It - , 2.5-1t depth dIscourages hIgher rating.
J Rated at 3 gpm/sQ It. Rate control system conSIdered limIting.
• Rated at CT • 127 WIth 2.4 mglL chlOrine dose. whIch reoUlres a 53·min HOT; allowed' 0 percent 01 avaIlable volume for contact
t,me. temperature· 5'C. pH • 8. 4-log reoulred reductIon. 2.5 log ,n plant. 1.5 log dls,nfec~on.

Figure 4-56. Raw water turbidity profile· Plant 11.

o
U1
"-

U1
o 0 o
U1
N
o

:::> Present
I-
Z Requirement

.~
'0 U1
l5 "-
OJ
I-
Future
U1
Requirement

U1
N

CSl4----l--->+f-+----+---....,.....
MARSS MAY JUL SEP NOIf
--o-+---+t+----.
JANS9 MAR
. . . - --+JUL
MAY
368 Upgrading EXisting or Designing New Drinking Water Treatment Facilities

Figure 4-57. Probability plot 01 Ilnl'heel water lurtlldlty - Planl 10.

N+----+------+-----...---+----o---+------+----+

In
N
::l
!Z
~
:g
.t:J
:s In

=_.
~ I'-

In-+-

In
N

1SJ-f----+----+----oJ-_-+_---<_ _-+--l.---+-_-f-
.1 10 50 90 99 99.9
Probability

Figure 4-58. Filter 3 ellluenl lurtlldlty prollJe alter backwash - potential health risk to the village. If the condition
Planll0. of the media is not known (e.g .. mudballs),
operational practices causing these problems
Turo,diry. NTU
cannot be changed. The staff's lack of attention to
20 repairing and using the rapid mixer also
contributed to this factor's high rating.

2. Process Control Testing - Operation: Process


control testing was observed to be minimal.
1.5 Proper operational decisions require a routine and
comprehensive process control testing program.
Jar testing is necessary to evaluate changing raw
water conditions, and determine optimum
1.0
coagulant types and feed rates. Also, the solids
contact unit requires constant monitoring of solids
(e.g.. sludge depth, sludge discharge volume,
etc.) to optimize turbidity removals. Raw water
.... ...
--- ------.------ ... ------.
Clearwell
monitonng is also necessary.
0.5
3. Disinfection - Design: The SWTR will require all
plants to achieve a particular CT value for the
disinfection portion of their plant. Concerns about
the formation of tnhalomethanes, formed as a by-
product of chlorination, will limit the level of
o 10 20 30 40 chlorine concentrations used by a public water
Minutes
system. In lieu of increasing chlorine dosages, a
utility will be compelled to increase the effective
contact time. The lack of baffling in the clearwells
may limit the effective disinfectant contact time (T)
that the utility will need to comply with the SWTR.
Case Studies 369

The State's interpretation of this rule may lead to 5. Plant Staff Morale Pay - Administration: The
different results than indicated by this current pay structure for the staff may discourage
evaluation. Factors identified as having either a more highly qualified people from applying for
minimal effect on a routine basis or a major effect operator positions. The village does not currently
on a periodic basis are summarized below in order offer a pay scale competitive with other faCilities.
of priority.
6. Chemical Feed Facilities· Design: The plant lacks
1. Supervision· Administration: There is no clear the capability to feed chemicals to various points
definition or delegation of responsibilities between in the treatment process. The option to apply
the superintendent and other members of the chemicals (e.g., lime) will enable optimal use of
staff. To maintain continuity in plant operations, chemicals and chemical dosages. Additional
specific tasks should be assigned to each staff chemical feed facilities were projected to be
member. This would alleviate incidents of poor required. A polymer feed system could be used to
communication between the plant staff (e.g., staff optimize filter performance especially during cold
members arbitrarily resetting chemical feed rates weather operation.
after they have been adjusted by other staff
members). Once plant personnel have clear The CPE team identified additional factors that had a
definition of duties, daily planning and priority minor effect on plant performance. Specifically, the
setting, which is presently minimal, can be lack of a preventive maintenance program, the lack of
optimized to increase plant performance. variable speed mixing capabilities during flocculation,
and the minimal depth of the sedimentation baSin.
2. Process Accessibility for Sampling . Design: The Headloss gauges should also be installed on each of
lack of sampling locations to evaluate various the filters to enable optimization of filter runs based on
plant unit processes limits implementation of an headloss and/or effluent turbidity levels. The inability
acceptable process control program. At minimum, to control flow distribution to the filters and control the
the plant should have taps to determine influent rate of flow at each filter will also adversely impact
and effluent turbidity levels for the reservoir (and filter operation.
river, if pumping directly to the plant), solids
contact clarifier, each of the three filters, and both Projected Impact of a CCP
clearwells. Alleviating the identified factors would appreciably
improve the performance of Plant 11. As such,
3. Alarm Systems· Design: The plant is operated for implementation of a CCP, if accepted by the Village
significant periods of time without any operations personnel, represented a viable option for the plant.
staff present to make process adjustments in
response to variations in raw water characteristics
or correct problems with key processes
equipment. As such, finished water quality could
degrade and pose a potential health risk to the
village. A turbidity monitoring system tied to the
clearwell effluent could be used to alert the plant
staff to process problems before finished water
quality reaches undeSirable levels.

4. Number of Plant Staff - Administration: Presently,


a staff of four persons have responsibility for the
operation and maintenance of the water plant,
wastewater plant, distribution system, collection
system, and street maintenance. To properly
respond to the variations in raw water turbidity,
implement a process control program, provide
sample taps, etc., increased coverage of the plant
will be needed. Given these responsibilities and
requirements, a staff of four is inadequate.
370 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Plant 12 • Propeller-type finished water flow meter

Facility Description • 333,080-L (88,000-gal) backwash water and sludge


Plant 12 is a conventional plant that treats water from holding basin
a nearby river to provide water for domestic use by
the city. It consists of two separate sets of treatment Water is taken from the middle of the river through an
process trains that operate in parallel. One of these, intake pipe. Either of two submersible pumps is used
designated the "old" plant, consists of the solids to supply raw water to the plant. The plant is operated
contact clarifier and the two circular dual media filters. usually 6-8 hr'day and meets higher demands by
These were the original treatment processes before operating for longer periods.
the plant was expanded in 1977. The second set,
designated the "new" plant, consists of two package The raw water is split between the "old" and "new"
plants. Based on a review of plant records, the peak plant by separate valves at the flow split. Capabilities
operating flow for the entire plant was 32 L's (0.72 exist to pre-chlorinate the raw water. The raw water
mgd). Plant 12 consists of two separate sets of unit flow meter for the "old" plant had been removed for
processes shown schematically in Figure 4-59. repair and had never been replaced. For the "new"
plant, raw water flows can be measured and
• Raw water intake structure containing two controlled. During the CPE, the flow meter was
submersible pumps: one 25 Ls (400 gpm) and the operational, but had not been calibrated and, thus,
other 3t Us (500 gpm) was not used by the plant staH. The rate-of-f1ow
controller was out of service. Neither the flow meter or
• Propeller-type raw water flow meter controller had been used for a long time.

• Seven volumetric chemical feeders: two for alum, Alum and lime are added to both of the plants using
three for lime, one for fluoride, and one for volumetric dry chemical feeders. On the "new" plant,
KMn04/PAC alum and lime slurries were prepared by adding water
to the dry chemicals in mixing tanks beneath the
• Two 3.4-m (11-ft) diameter flocculation basins, 2.7 feeders. These slurries were conveyed by gravity to
m (8.8 ft) deep, each eqUipped with vertical paddle the pipe carrYing the raw water to the f1occulators. No
flocculators and variable speed drives. Each unit is mechanical or static flash mixing was provided.
divided into two sections by a mid-depth horizontal Chemical feed rates for both plants were adjusted with
perforated baH Ie changes in raw water turbidity based on the operator's
experience. A jar testing apparatus was available at
• Two package plants each with sedimentation and the plant, but not used. Feed rates were not routinely
filtration. The sedimentation section has a surface adjusted.
area of 4.8 m2 (51.8 sq ft) and contains a 1.7-m
(5.7- ft) high module of 7.5-degree tube settlers. The two f1occulators on the "new" plant had been
Each filter has 10.2 m 2 (110 sq ft) of surface area modified by removing the horizontal, perforated
and 76 cm (30 in) of mixed media baHles, These baHles, intended to separate each of
these units into two stages, would have provided
• 6.4-m (21-ft) diameter, 3.0-m (10-ft) deep upflow improved flocculation. The mechanical mixers on
solids contact clarifier. Center flocculation cone has these basins were also not operational.
a volume of 8,515 L (2,250 gal) with a vertical
paddle mixer From the flocculation basins, flow to the "new" plant
enters a tank that has a sedimentation and filtration
• Two 2.7-m (9-ft) diameter dual media filters section. Flow is directed to the bottom of the
containing 69 cm (27 in) of media sedimentation section and then flows up through a
module of 7.5 degree-tube settlers. Settled water
• Two clearwells: one "old", with 246,782-L (65,200- discharges Into a trough that conveys it to the filtration
gal) capacity and the other "new", with 199,470-L section. After passing through the mixed media filter,
(52,700-gal) capacity the finished water flows to the clearwells.

• Two backwash pumps: one "new" 63 Us (1,000 The filtration units were designed for automatic
gpm), and the other "old", est. at 31 Us (500 gpm) operation With electrically actuated valves controlling
the filter flow rate based on level measurements.
• Vacuum-controlled solution feed chlorination Backwashing was also designed to use automatic
system fed from 68-kg (150-lb) cylinders electrically actuated valves. Continuous turbidimeters
for the raw and finished water were provided. During
• Three high-service pumps: two "old", estimated at the CPE, however, none of the automatic valves or
16 Us (250 gpm) and 38 Us (600 gpm) and the turbidimeters were operational. Instead. the plants
third "new" 31 Us (500 gpm) were operated manually, but the plant staH could not
."
New Plant C
c:
--, I --- -
Multl·Medla
.-'----'-. I
t
r- 0;
...
i"
!"
Filters
Dual-MeOla .....
Tulle Selliers TuDe Selliers
+I + Fillers ~

III '"
"~
hi ...•.
oII
Fioccuialors I~s~ Q
ii
'!AI:!
Lime
Alum
III ?

Flowmeler

Lime
FluoflOe
Chlonne
I I I
Valve ICIOse<l1

+t t
I
Raw Water Inlake New Clearwell Old Clearwell
ana Pumps

III
High-ServICe
Pump
_~!L
Backwash Pumps ana Waler
__ High-Service
Pumps

Finished Waler« EJ 1 ~ J J
Flowmeler
(")

BaCkwash Waler ana ~


----l ct>
SIu<lge BaSIn
Supernale 10 River _ _ Solids 10 Lanallil ~
c::
a-
ct>
on

w
....
~
372 Upgrading Existing or Designing New Drinking Water Treatment Facilities

adequately control the flow through the filters. Settled service pumps are operated manually to supply the
water was observed cascading onto the surface of the four storage tanks in the distribution system.
filter media. When a filter is properly operated, the
surface of the filter media is flooded, but under the Major Unit Process Evaluation
observed conditions, flow was passing through only a The performance potential graph is shown in Figure 4-
portion of the media. Both the filters and the tube 60. Three sets of bars are presented for each unit
settlers were backwashed at the same time using process representing the "old" plant, the "new" plant.
water pumped from the clearwells. Backwash water and the total for both plants combined. Disinfection
was discharged to the holding basin. was only evaluated for the combined plants because
the units are interconnected. The shortest bar
In the "old" plant, a volumetric feeder adds dry alum represents the treatment process that limits plant
directly into the center flocculation zone and lime is capacity to achieve the desired performance of less
prepared as a slurry and piped to the flocculation than 0.5 NTU.
zone. A mechanical mixer provides mixing in the
zone. The instantaneous peak operating flow for the plant
was established at 16 Us (0.36 mgd) for each of the
The upflow solids contact clarifier on the "old" plant two plants, or 32 lis (0.72 mgd) for the total plant.
consists of two sections that provide for both This flow is based on a review of flow records for the
flocculation and sedimentation. Flocculation occurs in previous year and the practice of only operating one
a mechanically mixed cone-shaped center section, of the raw water pumps. On days the plant operates
while sedimentation occurs in the outer portion. Raw for longer than 8 hr, the treatment processes are stili
water enters the flocculation section and then flows operated at a maximum flow rate of 32 lis (0.36
downward before proceeding through the bottom of mgd).
this section into the upflow sedimentation section. The
mechanical mixer promotes flocculation and settled The flocculation basins were rated at 48 Lis (1.1) mgd
water discharges over peripheral weirs. Proper for the total plant. Most of the flocculation capabilities.
operation of solids contact clarifiers relies on the however, are provided by the "new" plant. The two
measurement and control of the solids maintained in flocculators were rated at 39 Us (0.90 mgd). under
the unit. At the plant. solids levels and concentrations the assumption that the horizontal perforated baffles
were not measured or controlled. would be replaced and the mechanical mixers made
operational. The flocculation portion of the "old"
The "old" plant filters were circular steel tanks that plant's solids contact clarifier was rated at 8.3 lis
showed significant signs of corrosion. Two troughs (0.19 mgd).
above the media distributed the settled water to the
filters and collected the backwash water. Filter flow For sedimentation, the solids contact clarifier was
rates were controlled by float-actuated valves that rated at 14 Us (0.32 mgd) and the combined package
were intended to maintain a constant water level plants at 17 Us (0.38 mgd) for a total plant capacity of
above the filter. During the CPE, these valves were 31 Us (0.7 mgd). The shallow depth and configuration
not operational. Settled water was observed cascading of the solids contact clarifier were judged to limit its
onto the media surface instead of flooding the filter capacity. Higher surface overflow rates were applied
media. These filters are backwashed with water to the package plants because of the tube settlers.
pumped from the clearwells. Backwash water is then
discharged to the holding basin. Total filtration capacity for the plant was rated at 53
Us (1.2 mgd). A filter loading rate of 117 m 3 /m 2 /d (2
Finished water from both plants combines ahead of gpmlsq ft) was used for the "old" plant, which
the "new" clearwell where chlorine, fluoride. and lime resulted in a rated capacity of 16 Us (0.36 mgd). The
are added. Chlorine doses are adjusted to maintain a filters on the "new" plant were rated at 35 Us (0.79
residual of 2.0 mg/L in the finished water leaving the mgd) based on a loading rate of 293 m 3 /m 2 /d (5
c1earwell. Both pre- and post-chlorination are used. gpm/sq ttl. These ratings assume the rate control
The staff attempts to maintain a 50/50 split between valves on both plants will be operational.
the two addition points, but no proviSions are available
to measure this split. The disinfection system was rated at 18 Us (0.42
mgd). Future drinking water regulations for disinfection
All finished water enters the "new" clearwell, but both will be based on CT values needed for various
clearwells are interconnected so that the flow is removals of Giardia cysts and inactivation of viruses.
distributed between them. Separate high- service To eS1ablish the CT required, it was assumed the
pumps draw from each clearwell, an arrangement that plant's disinfection system would have to provide 1.5
prevents the two clearwells from operating in series, logs of cyst removal with 2.5 logs of removal credited
and thereby optimizing the contact time with the for the other treatment processes. The total of 4 logs
chlorine. During the CPE, one of the "old" high- of cyst removal required was based on the CPE
service pumps had been removed for service. High- team's estimate of the quality of the raw water.
Case Studies 373

Figure 4·60. Plant 12 per1ormar.ce potential graph.

Flow, mgd
Unit Process 02 04 06 08 10 12

Flocculation 1
Old Plant HOT. min 1~
New Planl HOT. min 90 45 30 23 I
Total
I
Sedlmentation'·3
Old Plant SOR, gprnlSQ ft 0.43 I
New Plant SOR, gpm/SQ ft 1.3 .I
TOlal
I
Filtration O
Old Plant HLR, gpm/SQ ft 1.1 I
New Plant HLR, gprnlsQ ft 1.3 25 38 I
Total
I

Disinfection 6
HDT, min 85 4j2

Peak Instantaneous Operating Peak Instantaneous Operallng


Flow ~ 250 gpm Flow ~ 500 gpm

, Rated at 20-mln HOT - assumes baHles reinstalled and functIonal mlcers In new floc basin.
2 Old plant rated at 0.7 gpmlSQ ft . shallow deplh and conflgurallon conSidered limiting.
3 New plant rated at 2.5 gpmlSQ ft . lube senlers allow higher rates.
4 Old plant rated at 2 gpmlsQ ft . Integn!y of tanks was assumed to be adequate.
5 New plant rated at 5 gpm/sQ ft . assumes rate control valves are operational.
6 Rated at CT = 100 wllh 2.5 mglL chlonne dose, which reQUlles a 40·mln HOT; allowed 10 percenl of available volume for contact
lime, temperature' 5'C. pH = 7.5. 4·log reQUlled reduction. 2.5 log allowed for plant " operated well.

To achieve the 1.5 logs of cyst removal. the CPE rated capacity close to or exceeding the peak
team estimated that the diSinfection system would instantaneous operallng flow of 32 L's (0.72 mgd),
have to prOVide a CT of 100, ThiS CT value is for with the exception of the disinfection process,
chlOrine at a 2.5 mg;L dose, pH 7.5, and temperature FlocculaliOn IS adequate up to 48 L s (1,1 mgd) If the
of 5C. The contact time was based on the chlOrine f1occulators on the "new" plant are returned to their
being added ahead of the "new" clearwell and the Original condition. The sedimentation processes are
flow passing through both clearwells. Only 10 percent projected adequate to treat a flow of 31 Us (0.70
of the theoretical detention time in the clearwells was mgd). which is borderline, The filtration system, rated
used because the clearwells are not baffled and at 53 L s (1.2 mgd), was rated considerably more than
because they are subjected to fill and draw operation. adequate to treat the peak instantaneous operating
The piping arrangement. which does not assure that a flow. Lack of baffling in the c1earwells and the piping
flow passes through both clearwells, also contributed arrangement between the two clearwells limited the
to thiS rating. The actual levels of disinfection reqUIred projected capacity of the disinfection process.
for the plant in the future will be determined by the
State, The estimates in thiS CPE of lhe required total On an Individual plant baSIS, the performance potential
number of log reduclJOns and the allowances for graph shows that both plants do not have equal
actual contact limes in the clearwells may change capabilities. For the total plant to have a rallng of 32
when final regulalJOns are developed, lis (0.72 mgd), more than half of the flow will have to
be treated In the "new" plant. The solids contact
The performance potential graph shows that. on a clarifier on the "old" plant limits the flow it can
total plant basis. the major unit processes have a adequately treat to 8.3 L s (0.19 mgd).
374 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Performance Assessment 1. Performance Monitoring - Operation: The practice


Figure 4-61 shows the finished water turbidity as of sampling at optimum times, though allowed by
reported by the plant staH for a 12-month period. The current regulations, has resulted in an inaccurate
current State regulation for turbidity is 1.0 NTU. The assessment of the plant's true performance.
federal SWTR will require the plant to meet a 0.5-NTU Accurate monitoring would have alerted the plant
finished water turbidity 95 percent of the time. The staH to the serious performance problems at the
plant generally complies with the 1.0-NTU regulation, plant and likely would have resulted in regulatory
but is consistently above the 0.5-NTU required by the pressure to correct them. Improperly operating
SWTA. A probability plot of this same data, shown in laboratory instruments used for monitoring also
Figure 4-62, indicates that under present conditions led to an improper interpretation of performance.
this plant would only meet 0.5 NTU less than 10
percent of the time. 2. Plant Administrator's Policies - Administration:
Current and historical actions by the mayor and/or
During the CPE, special studies were conducted to city council were inadequate in recognizing the
assess the performance of both the "old" and "new" significance of poor water quality and
filters alter backwashing. With adequate facilities and inappropriate in that they did not aggreSSively
operation of proceeding unit processes, a properly address the causes of the situation. The existing
operated filter should produce a finished water new plant, constructed in 1977, had been allowed
turbidity of approximately 0.1 NTU and only to deteriorate. Repairs and maintenance to protect
experience a 0.2-NTU rise in turbidity in the finished system integrity had been largely ignored. StaH
water for approximately 10 minutes after being with expertise in water treatment were performing
restarted following backwashlng. For this special numerous other city functions away from the
study, both filters were sampled for a 30-mlnute water plant, and staH with virtually no training in
period after being restarted following backwashlng. water treatment were manning the plant for only
portions of the time it was operating. "Muddy"
Figure 4-63 shows the results for the "old" filter. Prior water was accepted as a way of life. Agreements
to backwashing, this filter was prodUCing 0.22-NTU to provide water to other communities were
water. Alter backwashing, however, the water quality negotiated and perpetuated despite the increased
peaked at 38 NTU after 1 minute and did not drop demand that was placed on a marginally
back to the original value even after 30 minutes. A functioning system. A significant change in past
turbidity of 1.0 NTU was not achieved for almost 20 policies and in emphasis on the water plant will be
minutes. Figure 4-64 shows the results after necessary to reduce the health risk associated
backwashing one of the "new" fillers. Prior to with current water plant performance.
backwashing, the filter was producing a 12.5-NTU
water, significantly above the 1.0-NTU regulation. 3. Maintenance: Years of neglect of all plant
Alter backwashing, the water quality improved. but equipment have degraded a potentially well-
had not achieved adequate performance even after 20 equipped plant to essentially a nonfunctional state,
minutes. ConSiderable expenditures will be required to
make this equipment operational and to keep it
These results Indicate a significant performance maintained.
problem that may be attributed to the filters or to the
fact that the water being applied to the filter has not 4. Water Treatment Understanding - Operation: The
been properly treated and conditioned in the plant staH demonstrated a significant lack of
preceding unit processes. During the backwash of the understanding of even basic concepts of water
"new" filter, large amounts of air were observed treatment, allowing the water to cascade onto the
bubbling up through the media. filter media, starting dirty filters, performing no
process control testing, and essentially providing
During the two special studies, the team also no adjustment of chemical feed rates. There was
collected samples of the finished water from the also a lack of urgency to repair and/or replace
clearwells. These results. shown in Figure 4-65, improperly functioning equipment essential to
Indicated that significantly high levels of turbidity were providing water treatment.
passing into the city water systems: well above the
levels allowed by the State. Such high levels of 5. Process Control Testing - Operation: A process
turbidity pose a significant health risk to the control testing program to optimize unit process
community. performance did not exist at the plant. Process
control testing is essential for water plants served
Performance-Limiting Factors by surface sources because of the frequent and
The factors Identified as having a major eHect on rapid changes in raw water quality. Basic
performance on a tong-term repetitive basIs were equipment was available to conduct this testing,
prioritized and are summarized below: but was not used.
Case Studies 375

Flilure 4-81. Flnllhed wlt.r tUrblcll1y profile· Pllnt 12.

In

C\J

C\J

Present
Requirement

Future
Requirement


CSl-f-+......-o+--+-.....-+o--+o+............-t+--t+--t-.....-+o--t--t+--O+--....,f---+
SEP88 OCT NOV DEC JAN89 FEB MAR APR MAY JUN JUL AUG SEP OCT

Figure 4-82. Protleblll1y plot of flnllhed wlter tUrbkll1y . Plant 10.

C?li----+-----+-------+------+-----f----+

II'l

C'\J

C'\J

~
.....
Z
II'l
.~
:g
-e
::::J
..... ... •

CSl+---'"""'i--.-~'"""'i------t_-----+_----+_--__I_
•1 10 50 90 99 99.9
Probability
376 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Figure 4-63. Olel ptent north filter effluent turtlkllty profile Figure 4-65. ClealWell turtlidlty prolile aftar IIIt.r backwash •
efter blIc:kwe'" • Plent 12. Plant 12.

Turbidity, NTU
Turbidity, NTU
40 8

30 6

20 4

10 2

Present Requirements
Present Requirements

o 5 10 15 20 25 30 o 10 20 30 40 50 60 70

Minutes
Minules

Figure 4-64. N_ plent north Illter eHluent turtlidlty prolile 6. Filtration - Design": This factor has an asterisk
aftar backwaah • Plent 12.
because of the air observed above the filters
during backwash. Air entering the filters during
Turbidlly, NTU backwash may have disturbed the filter media to
12 the point that it will have to be replaced. The
condition of the filter media was not verified during
the CPE. The filter tanks on the "old" plant are
severely corroded to the point that they could fail
entirely.

7. Disinfection· Design': This factor has an asterisk


8 because it was assessed based on the initial
disinfection requirements of the new regulations.
These requirements may change when final
regulations are developed by the Slate. On thiS
basis, however, inadequate contact time is
provided because of a lack of baffling in the two
4 clearwells and a lack of piping to allow them to
operate in series.

Factors identified as having either a minimal effect on


a routine basis, or a major effect on a periodic basis
are summarized below in order of priority.

o 5 10 15 20 1. Staff Number . Administration: Additional staff are


required to provide adequate coverage of the
Minutes
plant, to perform the necessary process control,
and to complete maintenance functions.

2. Staff Qualification . Administration: All of the plant


staff must have high levels of education to make
Case Studies 377

proper operation and maintenance decisions: all The amount of bond indebtedness of the city was
staH should also be certified. considered a minor factor, because it could limit the
ability to properly fund operation and maintenance or
3. Process Flexibility - Design: The capability is needed repairs to the plant. Practices used for
needed to feed polymers and filter aids at diHerent disposal of plant sludges were not considered
locations in the plant to optimize performance. environmentally sound, but had no impact on plant
This will be especially critical if the plant is to performance. A lack of simple taps on all of the filters
consistently meet a required finished water quality prevented proper monitoring of filter performance.
of 0.5 NTU.
Projected Impact of CCP
4. Alarm Systems . Design: The plant experiences Data collected during the CPE indicated severe
rapid variations in raw water quality and has a performance problems. Correcting the identified
limited number of operations staH present to make factors would appreciably improve the plant's
necessary adjustments in response to these performance and allow it to meet both current and
variations (e.g., adjust chemical or chlorine doses, future regulations. As such, implementation of a CCP
or correct problems with key process equipment). represented a viable option for the plant.
On these occasions, process performance could
degrade to a point where it poses a potential
health risk to the city. A turbidity monitoring
system tied to raw and finished water and a
chlorine residual monitoring system could be used
to alert the plant staH to process problems before
finished water quality reached undesirable levels.

5. Flow Proportioning to Units . Design: Flow


measurement and flow control devices are
needed to accurately split flow to ensure that each
plant receives the proper flow rate. This is
especially critical because the flow to the "old"
plant must be limited to achieve desired
performance.

6. Flocculation . Design: In the "new" plant, the


f1occulator's original horizontal perforated baHies
will have to be replaced and the mechanical
mixers made operational. The flow to the "old"
plant must be limited because of the size of the
flocculation section of the solids contact clarifier.

7. Sedimentation - Design: The sedimentation


capabilities of the plant are marginal because of
the shallow depth of the solids contact clarifier
and the limited surface area of the sedimentation
sections of the package plants.
378 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Plant 13 pipes supply water to a wet well. Either of two vertical


turbine pumps is used to supply raw water from the
Facility Description wet well to the plant. The plant is usually operated 17-
Plant 13 is a conventional water treatment plant that 20 hr/day. Higher demands are met by operating the
supplies water to the city for domestic use. Its source plant for longer periods.
is a nearby river. Based on a review of plant records
for the year. the peak flow was 66 Us (1.5 mgd). The An orifice plate measures raw water flow rates just
plant includes the following unit processes and is prior to chemical addition. Flow rates are charted on a
shown schematically in Figure 4-66: totalizer-recorder located on a panel with the reservoir
level alarm system.
• Raw water intake structure and two vertical turbine
60-hp. 55-Us (1 ,040·gpm) pumps Chlorine is injected prior to lime and alum addition at
concentrations high enough to maintain a 1.5-mg/L
• Orifice plate with a totalizer-recorder to measure residual on top of the fillers - 45 kg (100 Ib)/d. The
raw water flows rate is changed only when "muddy" waters are
observed at the plant influent.
• Five volumetric chemical feeders: one each for
alum, lime. soda ash. potassium permanganate. Alum and lime are added· using dry volumetric
and powdered activated carbon feeders. Alum and lime slurries are prepared by
adding dry chemicals to mixing tanks beneath the
• Mechanical 3-hp flash mixer feeders and conveyed by gravity to a trough carrying
raw water to the rapid mixer. Volumetric feeders are
• Two dual-stage flocculation basins. Each stage is also in place to feed potassium permanganate.
4.9-m (16-11) square and 3.7-m (12.2-11) deep. Each powdered activated carbon, and soda ash slurries to
stage contains a vertical paddle flocculator meet seasonal variations in raw water quality.
operated from a central l-hp variable speed drive
Alum feed rates are adjusted based on visual
• Two sedimentation basins 11 m (36 11) long and inspection of the floc particles in the flocculation
6.1 m (20 11) wide. containing 60· tube settlers 11 basins. A Jar testing apparatus is in place, but is used
m x 4.5 m (3611 x 14.711) and 3.7-m (12.2·11) deep. infrequently (10-' 5 times/yr). The lime feeder is
Each basin has a weir length of 43.9 m (144 11) currently operating at maximum output. and is not
adjusted.
• Two mixed media filters 3.4 m x 3.8 m (11 11 x 12.3
11) fitted with rotary surface wash. and containing 84 The chemical slurries and raw water enter a 0.9-m
cm (33 in) of media square (3-11 square) and 2.4- m (7.8-11) deep basin
containing a mechanical flash mixer. The basin has a
• Backwash water and settling basin sludge decant hydraulic detention time of 30 seconds. The CPE
basin team calculated the G value for the mixer to be
adequate at 894 sec-'.
• Four sludge drying beds
Al1er exiting the rapid mix basin the water splits
• Diaphragm metering pump to feed hydrofluosilicic hydraulically and flows through a parallel train of
acid identical flocculation, sedimentation, and filter basins.
prior to entering the two clearwells.
• Vacuum-controlled solution feed pre-/post-
chlorination system fed from 68-kg (150-lb) The coagulation/flocculation process is performed in a
cylinders dual stage system; each stage is fitted with horizontal
paddles. A pair of variable speed motors. one for each
• Two 575,300-L (152.000-gal) clearwells stage. drives a central shal1 which in turn drives the
paddles for each stage of the parallel trains. Each
• 189-Us (3,000-gpm) vertical turbine backwash basin has a detention time of 45 minutes and a peak
pump G of 77 sec".

• Two vertical turbine high·service pumps with a Flocculation basin effluent is directed to the bottom of
capacity of 66 Us (1,050 gpm) the two pairs of sedimenlation basins and flows up
through a set of 60-degree tube settlers. Each of the
• 6.340-L (1.675-gal) wet well for the backwash and two basins has a detention time of 2.1 hr. The basins
high-service pumps have a combined surface overflow rate of 58 m 3 /m 2/d
(1,420 gpd/sq ft). Sludge is manually removed
Water is taken from the river through any of three approximately every 2 months and washed to a
intake pipes located at different depths. The Intake decant basin. Originally, decant from the baSin was
."
~
c
Dual-Stage FI occulaLOrs
~~
Tube Senlers
coedbCU ••

++
o<>o"'ooooouoooooooooooooooo
~
OVOOOOOOOOOOUOOUOOOOOOOOOOOOOOOOOOOOgOOOgOOOO
~
ii
oooooooo,,>oo..>oou~OO\)OOOO
'00000000000000000000000
a
BS9999999999929900g0099'
..,'"

++
~eo606o6odo6o6o6o8o6o6odobobo6o'o6o6odoGobob~
it7fJfrlrfuvfrvf&lfcf.1Jll,?filfl
(I (I au (I 00 000 00000 (I (I (I (I ou Q
99P9P9QQQQQRQQQQ99'
,b6S66666bbb06a6dbd
I
9
°0°0°0°0°0°0 0°0°0°0°0°0°0°0°0°0°0°0°0°0°0°0°
f
a.
i
il
;l

To
SaMaJY Decant BaSIn .--l Sludge and
Landlill Backwash Water

Senled
Waler
Decant Recycled
To Head 01 Plant Multr-Medra

I
Backwash Waler
, Frllers

I I
Clearwell 1
~T
I L
To
Distribution
Syslern
I ("')
Ql

Hryh-Servrcc
L. '"
ct)

Pumps Clcarwcll 2 ~
c:
0-
ct)

'"
CAl
.....,
(0
380 Upgrading Existing or Designing New Drinking Water Treatment Facilities

recycled to the head of the plant, but the recycle Filtration capacity for the plant was rated at 85 Ls
pump is out of service so the decant is discharged to (1.95 mgd) based on the state maximum allowable
the river. Sludge from the decant basin is washed to loading rate of 293 m3 /m 2 /d (5 gpm/sq It).
the drying beds for ultimate discharge to a sanitary
landfill. The disinfection system was rated at 26 Us (0.6
mgd). Future drinking water regulations for disinfection
Settled water then enters a pair of mixed media filters will be based on CT values needed for various
through metal troughs above 12.6 m 2 (136 sq It) of removals of Giardia cysts and inactivation of viruses.
media. Filter flow rates are controlled by raw water To establish the CT required, it was assumed that the
pumping rates, which in turn actuate pneumatic plant's disinfection system would have to provide 1.5
valves, maintaining an approximate loading rate of 222 logs of cyst removal with 2.5 logs of removal credited
m3 /m 2/d (3.8 gpm/sq It). for the other treatment processes. The total of 4 logs
(99.99 percent) of cyst removal required for Plant 13
Backwash frequency is determined by measuring was based on the CPE team's estimate of the quality
headloss across the filters or by observing a rise of of the raw water.
clearwell turbidity. Filter runs are routinely 40-45 hr.
Backwashing routinely consists of 6-7 minutes of To achieve the 1.5 logs of cyst removal the CPE team
surface wash at a rate of 26 m3 /m 2/d (0.45 gpm/sq It). estimates that the disinfection system would have to
The media beds are then washed at a rate of 1,055 provide a CT of 183. This CT value is for chlorine at a
m3 /m 2/d (18 gpm/sq It) for an indiscriminate duration. 2.5 mg/L dose, pH 8.0, and temperature of S·C. The
contact time was based exclusively on the post-
The influent troughs are then used to discharge chlorine dose. Only 10 percent of the theoretical
backwash water to the decant basin. The filters are detention time in the clearwells was used because the
capable of filter-to-waste operation, although this is clearwells are not baffled and thus are subject to
not commonly practiced. hydraulic short circuiting. The actual levels of
disinfection required for the plant in the future Will be
Chlorine (23 kg [50 Ib]/d), hydrofluosilicic acid (91 L determined by the State. The estimates in this CPE of
[24 gall/d), and, occasionally, soda ash are then the required number of log reductions of Giardia cysts
added to filtered water prior to entering the two and viruses and the allowances for actual contact
clearwells. Each c1earwell has a hydraulic detention times in the clearwells may change when final
time of 4.9 hr. Current piping configurations do not disinfection regulations are developed.
permit operating the c1earwells in series.
Raw and finished water pumping capacity was rated at
Finished water then flows into the wet well which 66 Us (1.5 mgd). This rating was based on use of a
supplies water for the two high-service pumps, as well single raw/finished water pump with one pump out of
as the backwash pump. service.

Major Unit Process Evaluation As shown in the performance potential graph. the
The performance potential graph is shown in Figure 4- major unit processes have a rating capacity exceeding
67. The shortest bar represents the treatment process the instantaneous peak operating flow, with the
limiting the plant's capacity to achieve the desired exception of the disinfection process. Again, the lack
performance of less than 0.5 NTU. of baffling in the clearwells limited the projected
capacity of the disinfection process.
The instantaneous peak operating flow. was
established 66 Us (1,050 gpm). This is based on a Performance Assessment
review of flow records for the previous year and the Figure 4-68 shows the settled water turbidity
practice of only operating one of the raw water measured by the plant staff over the previous 12-
pumps. The plant IS normally operated approximately month period. Settled water turbidities were generally
17-20 hr/day. To meet events of peak water use less than 2.0 NTUs, although there were also several
during the summer, the plant is operated 24 hr/day. periods of higher turbidity. These appeared to be
On those days the treatment processes are still related to periods of high raw water turbidity,
operated at a maximum flow rate of 66 Us (1,050 indicating that chemical feed rates were not properly
gpm). adjusted to compensate for the changes In raw water
turbidity.
The flocculation basins were rated at 149 Us (3.4
mgd) based on a 20-minute hydraulic detention time Figure 4-69 shows the finished water turbidity reported
and two-stage flocculation with a variable speed input. over the previous 12 months. Current regulations for
finished water turbidity are 1.0 NTU. Future
Due to the use of tube settlers, the sedimentation regulations will require the plant to meet a O.S-NTU
basins were rated at 101 Us (2.3 mgd) based on a finished water turbidity 95 percent of the time. Except
surface overflow rate of 88 m3 /m 2 /d (1.5 gpm/sq It). for several days in the first 3 months, the plant met
Case Studies 381

Figure 4-67. Plant 13 performance potential graph.

FlOW. mgd
Unit Process 10 20 30 40

Flocculation 1
HOT, min 68 34 23 I

Sedlmentatlon 2
SOR. gpm/SQ It 0.66 13 I
Flltration J
HLR. gpm/SQ It 2.6 I
Disinfection'
Contact ume. min
0
Raw/Finished Pumping 5
694
I
Peak Instantaneous Operaung
Flow (I Pump) = 1.050 gpm

1 Ratec at 20 ·mln HOT· based on 2-stage with vanable energy Input.


2 Ratec at 1.5 gpm/SQ It - based on tube senlers With annual sludge removal.
J Ratec at 5 gprnlSQ It - based on State maximum allowable loading.
• Ratec at CT = 183 With 2.5 mgiL chlortne dose. which reqU"es a 73-mln HOT; allowec 10 percent of avaliable volume lor contact
time. lemperalure = S·C. pH = 8. 4-log reqU"ec reduction. 2.5 log In pianL 1.5 log diSinfection.
5 Assumes firm Capacity at 1.5 mgd with one pump out of service.

the 1.0-NTU turbidity requirements and was regularly Figure 4-71 shows the results of this special study.
below the 0.5-NTU required by the SWTR. Figure 4- Prior to backwashing, the filter was producing 1.0-
70 presents a probability plot of this data and shows NTU water. After backwashing, the turbidity levels
that the plant produces' a finished water turbidity of Increased to 3.6 NTU and did not stabilize at the 0.1-
less than 0.5 NTU approximately 70 percent of the NTU level for 30 minutes. The special study also
time. Comparing Figures 4-68 and 4-69 reveals that found a problem with an inadequate amount of
higher finished water turbidity occurred dUring the backwash water flow at the beginning of the backwash
same periods that the senled water turbidity was cycle. A period of essenlially no flow was followed by
above 2.0 NTUs. providing further evidence that a violent eruption of the filter media as the backwash
chemical feed rates were not properly adjusted when water finally started entering the filter. Further
raw water turbidities changed. investigation of the problem revealed that the valve
that controls the backwash water flow was sticking In
During the CPE. a special study was conducted to a closed position and would finally snap fully open.
assess the performance of the filters after
backwashing. With adequate facilities and operation of These results indicate that a finished water that meets
preceding unit processes, a properly operated filter current and future regulaliOns is usually produced.
should produce a finished water turbidity of During periods of high raw water turbidity. it appears
approximately 0.1 NTU and experience only a 0.2- that the plant staH is not adequately adjusting process
NTU rise in turbidity in the finished water for control to allow the plant to produce a consistently
approximately 10 minutes after being restarted good quality finished water. Turbidity levels following
follOWing backwashing. For this special study, finished backwashing also indicate that bener process control
water from Filter 1 was sampled for a 40-minute could be practiced to limit the passage of high
period after being restarted following backwashing. turbidity water Into the distrtbution system following
filter backwash.
382 Upgrading Existing or Designing New Drinking Water Treatment Facilities

Flgur. 4·68. Senled wal.r turbldily profile· Planl 13.

(TJ,....-+-..-+--+-~'"--+---+--+-..-+--+-~--+--+--.,..

CD ll>

ll>
"- ll>ll>
ll>
to ll>t!>.

ll> ll>

:::l
l-
Z
lf1 ll>ll> .ll>
ll>

>. ll>
..- ll> (;>(;s>ll> ll>ll>
:2.0 I» ll>
:; ~ll>
ll>
l- (T) 0> . . ll> ll> ll> 0>.

~
It>
• ~ 0> • [l$> ll>0> [l$>

C\J
• ~
~. ll>!*>~ ll> Il+.. tp ll>ll> ~
ll> ll> ~~~"ll> ~ ~ll>~~ll>~ ~
t!> ~.. ~~ll>~~
is)-r...-+.......-+--+--+--+....... -t--,....... . . •
-t-+-..-+-'""'"'!--t--o-t-o--t-
DEC JRN89 rEB MAR APR MRY JUN JUL AUG SEP OCT NOV DEC JAN90

Figure 4-69. Finished waler tUrbidlly prolll. - Planl 13.

..--1-+--+---1-+--+---1-+-.......---1-+-.. . . ---1-+-.. . . ---1-+-.. . . . . . . . .-+-0.......-...+

(T)

lf1
:::l
l-

..
C\J
Z
ll>
~
'6 C\J ll>
:0 ll> ll>
:;
l- lf1 ll>
I» ~
ll> ll> ll>ll>
ll> ll>
ll>ll>
Present
IJ> i'1- ll> ll> Requirement

..
ll> ll>
~IJ> ll>ll>ll>l» ll> ll>
lf1
Il>!» .ll> ll> ll> ll> 1J>ll> ll> ll> ll>
Future
~.ll>
ll>
ll> [;&
lit>
~ll>
ll>
ll>~ll> ~ ~~ ~
• ll> ll>ll>~
ll> ~ ~GlEIt . . . . ll>~"ll> ll> ll>ll>ll> ~~
ll> . . ~ 1'• • • Il8e. e-. EUII ll>
Requirement

is) • ll> ll> ~


t!>""
• • IP'f!> -.ll>. Ii>

DEC JAN89 FEB MAR RPR MAY JUN JUL AUG SEP OCT NOV DEC JAN90
Case Studies 383

Figure 4-70. Probability plot ollinished water turbidity· Plant 13.

..-+---+-------l-------+----+--+------1----+

111
::l
t- N
Z
~
:2
D
N

:s
t-

r~
111

..£
lI1+-

(Sl i
-.-, -;;;;=::;jI~

[l>~,..:[l>~Cl>e::!~E~U~!!!!!!!!!!~~~~~~:==_-_L_--+_----___lf----t
10 50 90 99 99.9
Probability

Figure 4· 71. Filter el1luent turbidity prollle atter backwash • Performance-Limiting Factors
Plant 13.
The following factor was identified as having a major
Turold1ty. NTV effect on a long-term repetitive basis:
4
1. Disinfection· Design·: This factor has an asterisk
because it was assessed based on the initial
disinfection requirements of new regulations.
These requirements may change when final
regulations are developed by the State. Using this
3 basis. however. inadequate contact time is
provided because of a lack of baffling in the two
clearwells and a lack of piping to allow them to
operate in series.

2 Factors identified as having a minimal effect on a


routine basis. or a major effect on a periodic basis
were prioritized and are summarized below:

1. Supervision - Administration: The plant has 24-hr


Present ReQUirement coverage. which requires a large staff. This large
staff works without any formal organizational
structure. no lines of authority. no chief operator.
Fulure ReQUirement
and a total lack of leadership. With this absence
of supervision. essentially no communication
occurs between the staff on the different shifts.
o o 10 20 30 40
There are no regular meetings. no operating log.
and no shift overlaps where essential information
MInutes
on the status of the plant can be discussed.
Without this essential supervision. the productivity
of the plant staff is poor, which encourages poor
384 Upgrading Existing or Designing New Drinking Water Treatment Facilities

performance. For example. the alum feeder that 5. Water Demand - Administration': This factor has
needed repair at the end of the first shift was not an asterisk because it is projected that in the
repaired during either of the next two shifts. Lack spring of 1990 the water demands of the new
of supervision and communication were made industry and development in the city will exceed
worse by a total lack of standard operating the raw water pumping capacity of the plant. Plant
procedures. administrators committed the plant to supply this
water with little regard to its impact on the
2. Application of Concepts and Testing to Process capacity or performance of the plant.
Control • Operation: While the plant staff is
certified and has a knowledge of water treatment, Several of the administration's policies are considered
they are not able to apply their knowledge to to have a minor impact on the performance of the
properly control the treatment processes to plant. Current rate structures do not allow the plant to
optimize performance. Filter-to-waste capabilities be self-sustaining and cover all needed operation and
at the plant are not used to minimize the passage maintenance costs. There is a total lack of long-range
of high turbidity water to the c1earwells after filter planning to allow for growth within the community so
backwash. The plant was also operated for a as to minimize impact on plant's capacity and
month without a turbidimeter. Iron and manganese performance. Funding for the plant is also kept low.
levels were high in the finished water on several preventing repair of key equipment. Other factors
occasions. but no process changes were initiated thought to have a minor impact on performance are
by the plant staff. Lime is fed at the same point as the low pay of the plant staff as compared to other
the alum even though the lime raises the pH out plants in the State and the lack of process flexibility.
of the optimum range for alum coagulation.
Projected Impact of a CCP
3. Process Control Testing - Operation: A process Data collected during the CPE indicated that this plant
control testing program to optimize unit process usually performs satisfactorily, with some problems
performance did not eXIst. Process control testing responding to changes in raw water turbidity.
is essential for water plants served by surface Correcting the identified factors would appreciably
sources because of the frequent and rapid improve the consistency of the plant's performance
changes in raw water quality. Basic equipment and allow it to meet both current and future
was available to conduct this testing, but was not regulations. As such, implementation of a CCP
used. represented a viable option for this plant.

4. Preventive Maintenance - Maintenance: The lack


of a maintenance program has resulted in many
key pieces of equipment needed for optimal
operation not operating or near failure. Filter
controls, the influer,t control valve. the finished
water flow meter. and backwash water reclaim
pumps are not operating. New alum and lime
feeders were not installed to replace the margmal
units still in operation. Backwash control valves
were malfunctioning and the drives on the
flocculators were making excessive noises with no
indication of repairs bemg planned.

Anda mungkin juga menyukai