Anda di halaman 1dari 23

15th LDAR Symposium

May 19-21, 2015


New Orleans, LA

Development of
Emission Factors from
API 622/624 Test Data
Buzz Harris and Bronson Pate
Sage Environmental Consulting
Standards
Certification
Education & Training
Publishing
Conferences & Exhibits
Presenter Information

• Buzz Harris holds a BS in Chemical Engineering with 45


years experience and still learning
• Bronson Pate holds a degree in Engineering with 8 years
of full time experience plus 3 years of internships in
LDAR during university studies

2
Overview

• Low Emission (Low E) Technology Introduction


• Primary US tests for Low E packing and valves use
methane and include Method 21 type readings
• Tests are based on accelerated wear that should be
representative of 5 years or more of field operations
• EPA is requiring Low E in new Consent Decrees and
encouraging voluntary application
• Emission Factors (EF) and control efficiencies can be
developed from the test readings, which, if approved,
would provide more incentive to use Low E in any new
construction
• Conclusions

3
Low E Technology

• There have always been packings and valves that were


more or less likely to leak
• Only recently, however, have we had guarantees,
warranties, and, most importantly, test data to prove
manufacturer’s claims of low emissions
• Based on hard data, EPA has begun to require Low E in
Consent Decrees written over the last 4 to 5 years
• Several companies had begun their own Low E testing
and implementation voluntarily before that
• The use of Low E is gradually being recognized as cost
effective and just good business

4
Fugitive Emissions Testing

A number of valve and packing tests have been developed:


• ISO 15848-1 is an EU and British standard that allows
testing by either helium or methane, with most test data
to date done with helium
• API 622 is a packing test using methane
• API 624 is a valve test using methane
• ChevronTexaco test procedure is similar to API 622, but
it tests packing in a valve and includes more wear cycles
• Shell also has its own packing test
• The API test procedures have more industry wide
applicability and focus on methane testing

5
API 622 and 624 Details

Comparison of Service Parameters: API 622 and API 624


Standard identification API STD 622 API STD 624
Type Testing of Rising Stem Valves Equipped
Type Testing of Process Valve Packing for
Title and edition with Graphite Packing for Fugitive Emissions,
Fugitive Emissions, Second Edition
First Edition
Standard by the American Petroleum Standard by the American Petroleum Institute,
Date
Institute, 10/01/2011 02/01/2014
Pass Criteria 500 ppmv maximum after one adjustment 100 ppmv maximum
Equipment Specified fixture simulating a valve Valve being qualified
Packing adjustment One allowed1 None allowed
Media Methane 97% minimum purity Methane 97% minimum purity
Temperature 500°F (260°C) 500°F (260°C)
Pressure 600 psig (41.4bar-g) 600 psig (41.4bar-g)

Number of valve stem actuations 1510 310

Number of thermal cycles 5 3

Leak measurement method Method 21 Method 21

Leak measurement details Done with stem in static state Done with stem in static and dynamic states

Leak measurement frequency Every 50 actuations of the stem Every 50 actuations of the stem

Prepared by and used with permission of Garlock 6


API 622 Test Data

7
PPM Readings in API 622 and 624

• Probe split in two to simultaneously check


both stem and packing
• Tin foil used as a partial shroud to try to
collect leakage from any point around the
stem and packing
• Record ppmv data over a minute and report
the average and maximum with no
background correction
• Some report both static and dynamic (stem
moving) readings
• This paper uses the maximum readings
recorded in either static or dynamic modes
• This should be conservatively higher than
traditional M21 readings 8
Mass Emissions from ppm

• For petroleum industry:


– Default Zero EF (reading zero) = 0.0000078 kg/hr/source
– Correlation equation = 2.29*10^-6*SV^0.746 kg/hr/source

• For chemical industry:


– Default Zero EF (reading zero) = 0.00000066 kg/hr/source
– Correlation equation = 1.87*10^-6*SV^0.873 kg/hr/source

• No need for Pegged EF in Low E testing!


• These emission estimates are for the instant the
measurement occurs
• Still need to estimate emissions over time

9
Accelerated Wear Time

• API 622 includes 1510 mechanical wear cycles and 5


thermal stress cycles
• API 624 includes 310 mechanical wear cycles and 3
thermal stress cycles, but also requires packing to have
passed API 622 before 624 testing
• The tests take place over 3 to 6 days, but represent a
much longer period because of the accelerated use
• The equivalent process life may vary with:
– Continuous vs. batch processes
– Process application of the valve (manual isolation, drain,
sample, motor operated, control, etc. valve)
– Other factors too process specific for consideration here

10
Continuous Process – Cycles to Time

Accelerated Wear Cycles to Operating Time


Continuous Processes Frequency per year % of Total Weighted Frequency

Valve Application Low End High End Valves Low End High End

Manually operated block isolation valves 1 10 78% 0.78 7.8

Block valves isolating pumps 5 24 2% 0.1 0.48

Drain valves 12 120 5% 0.6 6

Sample valves 52 795 2% 1.04 15.9

Motor operated valves 150 1000 2% 3 20

Process control valves 500 5000 11% 55 550

Totals 720 6949 100% 60.52 600.18

Average Annual Operation Cycles 330.35

Years of Operation for 1510 Operating Cycles 4.57

11
Batch Process – Cycles to Time

Accelerated Wear Cycles to Operating Time


Batch Processes Frequency per year % of Total Weighted Frequency

Valve Application Low End High End Valves Low End High End

Manually operated block isolation valves 12 365 75% 9 273.75

Block valves isolating pumps 24 365 2% 0.48 7.3

Drain valves 12 365 5% 0.6 18.25

Sample valves 52 795 2% 1.04 15.9

Motor operated valves 150 1000 5% 7.5 50

Process control valves 500 5000 11% 55 550

Totals 750 7890 100% 73.62 915.2

Average Annual Operation Cycles 494.41

Years of Operation for 1510 Operating Cycles 3.05

12
Continuous Process – Thermal Cycles to Time

Thermal Cycles to Operating Time

Continuous Processes: Scheduled Unsheduled Total

Run Lengths Between T/A, yrs Shutdowns/yr Shutdowns/yr Shutdowns/yr

5 0.20 0.5 0.70

4 0.25 0.4 0.65

3 0.33 0.3 0.63

2 0.50 0.2 0.70

1 1.00 0.1 1.10

Average 0.76

Years Operation for 5 thermal cycles 6.61

13
Batch Process – Thermal Cycles to Time

Thermal Cycles to Operating Time

Batch Processes: Scheduled Unsheduled Total

Campaign Run Lengths, days Shutdowns/yr Shutdowns/yr Shutdowns/yr

30 12.17 0.05 12.22

20 18.25 0.04 18.29

10 36.50 0.03 36.53

5 73.00 0.02 73.02

1 365.00 0.01 365.01

Average 101.01

Years Operation for 5 thermal cycles 0.05

14
Summary – Accelerated Wear Time

• 1510 mechanical wear cycles represent 3.1 to 4.6 years


of operating time for batch/continuous process units
• 5 thermal cycles represent 0.1 to 6.6 years of operating
time for batch/continuous process units
• EPA defines Low E as <100 ppm operations for 5 years,
and accepts API 622 data as satisfying that definition
– The thermal cycles in API 622 falls far short of the potential
thermal cycles in 5 years for a batch process, but
– API 622 accelerated wear cycles average to around 5 years
equivalent operation for continuous operating units and
mechanical wear for batch units
• This paper, therefore, assumes that the API 622/624
readings are spread evenly over a 5 year period
15
API 622 Emissions Over Time Example

• 1510 cycles with a reading every 50 cycles (including


beginning and end readings) results in 42 readings
• 42 readings spread evenly over 5 years would be one
reading every 45.6 days (roughly every 6 weeks)
• Default zero and correlation equations for petroleum and
SOCMI are used to calculate instantaneous emission
estimates
• Emissions from the previous to current readings are
averaged over the 45.6 days between readings
• Emissions are summed over the 42 total readings and
divided by the time in years to create an average
emission factor in kg/yr/source
• Emission Factor for static and dynamic readings are
averaged 16
API 624 Emissions Over Time Example

• 310 cycles with a reading every 50 cycles (including


beginning and end readings) results in 14 readings
• 14 readings spread evenly over 5 years would be one
reading every 140.4 days (roughly semi-annual)
• Default zero and correlation equations for petroleum and
SOCMI are used to calculate instantaneous emission
estimates
• Emissions from the previous to current readings are
averaged over the 140.4 days between readings
• Emissions are summed over the 14 total readings and
divided by the time in years to create an average emission
factor in kg/yr/source
• Emission Factor for static and dynamic readings are
averaged 17
Emission Factors Summary

Summary of API 622 and 624 Testing-Based Emission Factors


Petroleum Chemical
Methane, ppmv
Calculated Calculated
Static Dynamic Emission Emission
Test Packing Valve Average Maximum Average MaximumFactor, kg/hr Factor, kg/hr
Chevron Teadit 2236 Velan 4 300 4 20 6 23 6.9203E-06 7.0746E-06
624 Teadit 2236 Ladish 4 600 5.1 37 5.7 12 8.8547E-06 1.141E-05
622 Garlock 1303 FEP Velan 4 300 1 5 1 8 7.3361E-06 1.8432E-06
622 Garlock 212 ULE Stuffing Box 1 4 1 4 4.5681E-06 2.228E-06
622 Chesterton 1622 Stuffing Box 16 33 16 34 1.8124E-05 2.1304E-05
622 Chesterton 1622 Velan 4 300 10 38 13 46 1.4043E-05 1.5093E-05
622 Nippon Pillar EDP 15 Stuffing Box 10.1 49 NA NA 1.1793E-05 1.1769E-05
624 Nippon Pillar EDP 15 Velan 4 300 3 15 7 20 9.8515E-06 1.0118E-05

API 622 and 624 tests reports were provided by and thanks go to:
Ron Walters of Teadit North America
Todd Haberkost of Ladish Valves
Jim Drago of Garlock
Scott Boyson and Rodney Roth of A.W. Chesterton Company
Josh Erd of Nippon Pillar Corporation of America
18
Low E Emission Factor Considerations

• All of these emission factors calculated from low


emission packing and valves are near the default zero
emission factor for valves
– One approach to encourage voluntary use of Low E would be to
allow use of the default zero factor to predict emissions for
permitting
– Another approach would be to allow an average emission factor
(or control efficiency) for Low E valves and packing to date to be
used to predict emissions for permitting
– Another approach would be to require applicants to include
calculations of an emission factor for the specific Low E
equipment they will use

19
Control Efficiency Approach

• In lieu of developing emission factors for Low E, it would


also be possible to develop a control efficiency for
application of Low E that could be applied to the normal
average emission factor
• The following table shows control efficiency estimates
calculated as:
CE%=(1-(Average EF/Low E EF))X100

Each Average EF above is for valves in a specific service category


in a specific industry
Each Low E EF above is for valves in a specific industry, but all
testing is done on Gas/Vapor service (methane, 600 psi)

20
Low E Control Efficiency

Low E Control Efficiency Estimates


Valve Avg. EF Low E EF Control
Industry
Service kg/hr/source kg/hr/source Efficiency
Refining Gas/Vapor 0.0268 1.01864E-05 97.32%
Refining LL 0.0109 1.01864E-05 98.91%
Refining HL 0.00023 1.01864E-05 99.98%
Refining Average All 98.74%
Chemical Gas/Vapor 0.00597 1.01049E-05 99.40%
Chemical LL 0.00403 1.01049E-05 99.60%
Chemical HL 0.00023 1.01049E-05 99.98%
Chemical Average All 99.66%

Low E Emission Factors are the average of all the API 622/624 tests using either petroleum
or chemical correlations/default zero emission factors.

21
Conclusions

• The measurements done in API 622/624 testing give


methane ppmv data that can be used to estimate
emissions
• API 622/624 are accelerated wear tests that appear to
represent a five year period for the average valve (with
the exception of thermal cycles on a batch process)
• Emission factors (EFs) developed for API 622/624 data
range in value from 4.6E-6 up to 1.8E-5, falling about
one order of magnitude higher than the default zero EFs
(6.6E-7 up to 7.8E-6)
• Control efficiency numbers for the Low E EFs compared
to EPA Protocol Average EFs range from 97.3 to 99.9%

22
Conclusions

• These Low E EFs and control efficiencies could be used


for permitting, where actual components that can be
monitored do not exist yet
• EPA should also consider allowing use of Low E EFs for
non-monitored valves (such as HL service, UTM, etc.)
• Neither EPA nor the states currently accept these Low E
EFs and control efficiencies for permitting
• We ask that EPA review these calculations and/or
replicate their own approach to similar calculations
• EPA-sanctioned Low E EFs and/or control efficiencies
would add another incentive for every new facility or
modification to be done with Low E technology

23

Anda mungkin juga menyukai