This article was downloaded by: [202.133.55.

169] On: 27 February 2014, At: 11:40 Publisher: Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Journal of the Air Pollution Control Association

Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/uawm16

Mechanical Dust Collectors

Charles A. Gallaer & J. W. Schindeler
a b a b

Chief Engineer , Buell Engineering Co., Inc. , New York , New York , USA

Director of Research and Development, Buell Engineering Co., Inc. , New York , New York , USA Published online: 19 Mar 2012.

To cite this article: Charles A. Gallaer & J. W. Schindeler (1963) Mechanical Dust Collectors, Journal of the Air Pollution Control Association, 13:12, 574-580, DOI: 10.1080/00022470.1963.10468222 To link to this article: http://dx.doi.org/10.1080/00022470.1963.10468222

PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the Content) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions

MECHANICAL DUST COLLECTORS

CHARLES A. GALLAER, Chief Engineer, and J. W. SCHINDELER, Director of Research and Development, Buell Engineering Co., Inc., New York, New York

Downloaded by [202.133.55.169] at 11:40 27 February 2014

or the purposes of this paper, a mechanical dust collector is defined as a device which separates dust in a dry state from gas through the application of inertial and gravitational forces only. The most widely used type of mechanical dust collector is the cyclone, in which centrifugal force plays the major role in effecting the separation of dust from gas. To fully understand the functioning of a cyclone the major flow patterns within a cyclone should be known. There are three major flow patterns: (a) The descending spiral flow which carries the separated dust down the walls of the cyclone to a dust receptacle. (b) The ascending spiral flow which rotates in the same direction as the descending spiral but which carries the cleaned gas from the cyclone and/or dust receptacle to the gas outlet. (c) The radially inward flow which feeds gas from the descending to the ascending spiral. There is inevitably some flow of gas to the dust receptacle where the final separation of dust from gas takes place. This, dependent on cyclone type, is a relatively minor component. However, it does occur and in this region this component of the total gasflowmakes a complete change of direction and is fed to the ascending spiral flow. Cyclone collectors may have either a tangential or axial inlet (Fig. la and lb), and they have been built in almost every conceivable size and geometric configuration. Cyclone collectors which differ only in size but are geometrically similar (that is, in which the same ratio of dimensions is maintained throughout) are known as a type of cyclone. Cyclones which may or may not be of the same diameter, but in which the ratio of dimensions is different, belong to different types. The diameter of the body of the cyclone is a convenient * Presented at the 56th Annual Meeting of APCA, Sheraton-Cadillac Hotel, June 9-13, 1963, Detroit, Michigan.
574

feature by which cyclones of different sizes may be distinguished. Thus, we may speak of a 6 in. or a 9 in. or a 50 in. cyclone. A type of cyclones may be identified by the ratio of two of the important dimensions, alwaj^s considering that all other dimensions of cyclones of that type are geometrical^ similar.

Thus, a particular type of cyclone in which the height of the gas inlet is equal to 50% of the body diameter may be identified as a type 0.5, for example (Fig. 2). Cyclone collectors may be designed for any required efficiency. However, it must be recognized that efficiency is a

CLEAN GAS OUTLET

FINES SHAVE-OFF FINES DUCT NLET FOR DUST LADEN GAS FINES RE-ENTRY OPENING

DUST OUTLET

Fig. 1 a.

Cutaway of involute inlet cyclone. Journal of the Air Pollution Control Association

COLLECTOR A
1

COLLECTOR B

00

\l
Downloaded by [202.133.55.169] at 11:40 27 February 2014

/ O

2 2 0 0 lbs.

EQUAL GAS VOLUMES

Fig. 1 b.

Cutaway of axial inlet cyclone. Fig. 2.

5350 lbs.
and even more difficult to obtain consistent analyses of identical dust samples in different laboratories or when using different analytical techniques. These problems have been recognized by the ASME, the APCA, the IGCI, and other groups endeavoring to codify dust collector performance testing techniques. At the present time, the most satisfactory methods of obtaining consistent data are to obtain a representative dust sample by an isokinetic sampling traverse as specified by ASME PTC-27, and analyze the sample for particle size distribution in the Bahco MicroParticle Analyzer as specified by ASME PTC-28 (not yet issued). The Bahco device must have been calibrated with special calibrating dust samples obtainable through the ASME. The fractional efficiency curve for any cyclone, operating under given conditions, must be empirically determined. At the risk of disagreeing with some noted authorities in the field, the writers would like to state that there
575

function of energy expended or space occupied; thus an acceptable cyclone must be a compromise which results in an acceptable efficiency at moderate pressure drop within reasonable space requirements. This compromise results in a basic type of cyclone which can be modified for special applications. Efficiency can be improved at the expense of pressure drop without altering the space requirements. It can also be improved by going to a totally different type of cyclone having greater space requirements which may or may not also have a greater pressure drop. All cyclone collectors behave to some extent like classifiers in that coarse dust particles are collected with relatively high efficiency, while finer particles are collected with ever decreasing efficiency. If the collection efficiency of a cyclone is plotted as a function of particle diameter, a typical fractional efficiency curve results (Fig. 3). It should be appreciated that the particle diameters shown on a fractional efficiency curve are not intended to be true measured diameters, but are
December 1963 / Volume 13, No. 12

rather the equivalent Stokes diameter ' of a spherical particle having the same terminal settling velocity. Hence, a more rational presentation of the fractional efficiency curve would show efficiency plotted as a function of particle terminal settling velocity. This method of plotting fractional efficiency curves has not been widely used, mainly because it is much easier to visualize and speak of a particle having a diameter of 10 microns than one having a settling velocity of 15 ipm in still air at standard conditions, for example. Regardless of which method of plotting a fractional efficiency curve is used, it is apparent that the curve may be used to accurately predict the total collection efficiency of a cyclone only if the particle size distribution of the dust (or the distribution of settling velocities) is known. This factor is determined in the laboratory by analyzing a representative sample of the dust for particle size distribution. It can be rather difficult to secure a representative sample,

is, at the present time, no satisfactory mathematical method of calculating the fractional efficiency of a cyclone on a completely theoretical basis in the absence of empirical data. There are, however, some useful mathematical concepts in the treatment of fractional efficiencjr data. If the fractional efficiency curve of certain types of cyclones is plotted on semilog paper, a straight line results, showing that the fractional efficiency curve may be represented by the formula:
_ 1 ad

FRACTIONAL EFFICIENCY CURVES

where: EKa) = fractional efficiency of a particle of size "d" cl = particle diameter, microns a = a constant Thus, a family of curves may be constructed by using different values of a. The fractional efficiency of most cyclones in common use will be represented by one of the c curves of this family. The more efficient the cyclone is, the higher will be the value of a (Fig. 4). The particle size distribution of some dusts may be represented by the following formula, which also produces a straight line on semilogarithmic graph paper (Fig. 7): Z(d) = e-t where: %(d) = that fraction, by weight, of the total dust having a larger particle size than "d" d = particle diameter, microns /3 = a constant Given these conditions, it has been shown1 that the over-all, or total, efficiency of a cyclone may be represented by: F a

Downloaded by [202.133.55.169] at 11:40 27 February 2014

10 15 20 25 30 35 PARTICLE SIZE-MICRONS
Fig. 3.

40

45

FRACTIONAL EFFICIENCY CURVES 99.9

99
i >

O 2

u
Li_

95

LL

II

LJ

For the sake of comparison the more conventional presentation of particle size distribution, on logarithmic probability coordinates is shown on Fig. 6. It should be noted that this formula is useful even for those dusts whose particle size distribution does not plot as a straight line on semilogarithmic paper, as long as the value of /3 approximates the slope of the line for those values of d which significantly effect the cyclone efficiency, i.e., the finer fractions. Inspection of the equation shows that, for high efficiencies, a. must be relatively large in comparison with /3. In other words, a high efficiency cyclone must be used with fine dusts (when /3 is large), whereas for coarse dusts (where
576

O 90-

u
LJ

y-

80-

o 60

70

5O III/// 40- w/7/y ; ill/// 20

unz

10

15 20 25 30 35 PARTICLE SIZE-MICRONS
Fig. 4.

40

45

Journal of the Air Pollution Control Association

C M

in

o
<
oc
UJ

M to C O
<

IZE DIS RIBUTI

z o
H

IBU

(V)

REA

<0 O
CO

Fig. 6.

a: \ \ o \
\ \
if)

cz
<

O LJ

TICI

</>

o O)
m

H Z UJ U

o o
\ \ \ \ \ \

u a: \ui U

3ZI
Ui _i

PAR! ICLE

o >

OC

UJ

LJ CO Q.

Downloaded by [202.133.55.169] at 11:40 27 February 2014

a
9

cc

o m o
\o

'99.99

c > m^t

t ro

(VI

o9 o o o o o m 't ro t\i SNOdOIW ' P 3ZIS 3101iaVd

(3 is small) a cyclone of lesser performance characteristics may still attain a satisfactorily high efficiency. Figures 5 and 8 show over-all efficiencies obtained by two types of cyclones handling dust of the same particle

size analysis. For illustrative purposes over-all efficiencies have been calculated using the a-/3 method (Fig. 8) and the more conventional method (Fig. 5). The most important characteristic of cyclone collectors, other than effi-

Overall Efficiency Calculation

Collector "A"

ciency is pressure drop. The pressure drop across any cyclone collector varies as the square of the gas volume throughput and directly as the density of gas. Unless the dust concentration is rather high, its effect on gas density may be neglected. It is often convenient to express the pressure drop as a function of the entering gas velocity as: AP where: AP Jc P v = = = = C3^clone pressure drop, "W.G. a constant gas density, lb. cu. ft. entering gas velocity, ft./sec.

Particle Size Range -Microns

n '? u&

Particle Size Distribution

1n 1U

Fractional Efficiency
1 i yQ

Percent Collected
1 l. O y

2-4 4-6
6-8

8-10 10-15 15-20 20-25 25-30 30+

8 8 7 6 14 10 8 22

45 63 75.5 83.5 91.5 97.0 98.9 99.5 99.75

100 Collector "B"

3.6 5.05 5.65 5.00 12.80 9.70 7.91 6.96 21.95 80.52 Percent Collected 3.90 6.21 7.34 6.78 5.94 13.96 46.95 91.08

Particle Size~* Range-Mi crons

0-2 2-4 4-6 6-8

Particle Size Distribution

10 8 8 14

Fractional Efficiency
39

8-10 10-15 15-20 20-25 25-30 30+

7 6 -| 47

10n .8

77.75 91.8 97.0 98.9 99.8 99.9

7\ 22 J
100
Fig. 5.

I J

Overall efficiency calculation.

Within any one type of cyclone, the pressure drop constant, k, is always the same, regardless of cyclone size. Cyclones of different types may often have a different pressure drop constant. Within any one type of cyclone, the fractional efficiency curve (value of a) will decrease as the size of the cyclone in increased. However, it is quite possible, by changing cyclone types, to produce large diameter cyclones having the same, or even a higher value of a than the cyclones of smaller size. Generally speaking, high efficiency cyclones require more steel per unit of gas handled, or have a higher draft loss coefficient, or both, than do cyclones having lower values of a. Hence, high efficiency cyclones cost more to build 577

December 1963 / Volume 13, No. 12

and/or more to operate. Were this not so, there would be no reason for making cyclones having lower values of a, even for the collection of coarse dust. The value of a is affected by factors other than the geometry of the cyclone. It is decreased by increases in gas viscosity or by decreases in particle density, because of the change in particle terminal settling velocity. It is increased by increases in gas inlet velocity, within limits, and also by increases in dust concentration, again within limits. The limits depend upon the design (type) of the cyclone. Thus far, we have discussed only the characteristics of a single cyclone. When relatively large quantities of gas are to be handled, it is common practice to use groups of cyclones in parallel, rather than one enormous cyclone. There are two reasons for this practice. One is to avoid too drastic a decrease in a, and the other is to produce a dust collector which does not require excessive headroom for installation. When a number of cyclones are installed so as to discharge dust into a common bin or hopper, it is frequently observed that the collection efficiency of the group is less than that which might have been expected from an individual cyclone handling its proportionate share of the total gas volume. In other words, the value of a for the group is less than that for a single cyclone. This phenomena is caused by a difference in the pressure at the various cyclone dust outlets, which causes a circulation of gas within the hopper from one cyclone to another (Fig. 9). The particular cyclone having a lower pressure at the dust outlet will discharge more gas from the hopper than enters through that cyclone. The effect, so far as that particular cyclone is concerned, is that the ascending axial gas velocity is increased without a corresponding increase in radial velocity, resulting in a decrease in efficiency. This effect is quite similar to that experienced by a single cyclone with a faulty hopper seal which allows air in leakage. In a similar manner, the efficiency of the particular cyclone having a higher pressure at the dust outlet will be increased to some extent. However, the net effect of such hopper circulation is always a decrease in total collection efficiency. This decrease in collection efficiency can be minimized by minimizing the difference in pressure between various cyclone dust discharges. This means that all cyclones on one hopper should be dimensionally identical, and each cyclone must handle the same volume of gas having identical dust loadings and temperature. Careful attention to design and fabrication of the collector and associated duct work can approach this ideal condition, but
578

Overall Efficiency Calculation Collector "A"

a = .20 (Figure 3 or 4) 0 = .05 (Figure 6)

Er =
.20 .20

.20 + .05

.25

= .8 X 100 = 80%
Collector "B'

a = .5 (Figure 3 or 4) /3 = .05 (Figure 6)

Er =

5 = S = -9091 X
Fig. 8. Overall efficiency calculation.

Downloaded by [202.133.55.169] at 11:40 27 February 2014

stratification of gas and dust as well as the effects of dust deposits make perfection impossible. Hence, in design, it is well to anticipate this effect, and use a lower a for a group than for a single cyclone. This would, of course, not apply if each cyclone of a group were to have its own individual dust hopper. Cyclones, or groups of cyclones, are sometimes used in series (Fig. 10) when it is desired to obtain a higher efficiency than is possible with one stage of collection alone, or to protect an expensive, high efficiency cyclone from a heavy concentration of abrasive dust. In these cases, the total pressure drop and efficiency may be represented by the following formulae:

AP r =

It should be noted that the value of a\ and at should be corrected for the different dust loading of the first and second stage cyclones. As previously noted, a is affected by dust concentration. Other mechanical dust collectors include settling chambers and inertial types. One specific type is shown in Fig. 11. Generally speaking, these have low pressure drops and poor fractional efficiency curves, as compared with cyclones. In other words, the values of a and k are usually low. They are, therefore, of value in the collection of

< AR>

Fig. 9.

Hopper circulation. Journal of the Air Pollution Control Association

SECONDARY COLLECTOR PRIMARY COLLECTOR

Downloaded by [202.133.55.169] at 11:40 27 February 2014

Fig. 1 1.

Cutaway of L-R collector.

Fig. 10.

Two cyclones in series.

rather coarse dusts, where the low pressure drop is advantageous. An interesting arrangement is that in which the bulk of the dust is concentrated into a portion of the gas by use of an inertial device, with final collection being accomplished by a cyclone. Such collectors are usually designed so that about 10 or 15% of the

main gas flow passes through the shunt circuit and cyclone. The fractional efficiency (ai) of the concentrating section of the collector is a function of the ratio of shuntflowto total gas flow, with higher fractional efficiencies being obtained with larger shunt flows. The total efficiency of the concentrator section alone is,

obviously, 0% when there is no shunt flow, and 100% when all the gas passes through the shunt circuit. The fractional efficiency of the cyclone (a2) is increased by larger gas volume flows, and higher dust loadings as in the ordinary cyclone installation. It is further apparent that the dust caught in the concentrator section must again be collected by the cyclone before it is completely removed from the gas stream. Hence the fractional efficiency of the complete collector must be less than that of either the cyclone or the concentrator. It has been shown by H. van der Kolk1 that the over-all efficiency of the dust collector arrangement may be expressed by the formula:

Er-l

L_.

I.D. FAN INERTIAL "CONCENTRATOR

7
SHUNT CYCLONE &FAN

Fig. 12. December 1963 / Volume 13, No. 12

where: an and ai apply to the primary concentrator and cyclone . collector, respectively. In the typical installation as shown in Fig. 12, the shunt fan, running at a constant speed will maintain a fairly constant volume of gas flowing through the shunt circuit, regardless of any changes in gas quantity entering the primary concentrator. Hence, if the main flow gas quantity is reduced due to a reduction in boiler load, for example, the shunt flow becomes a greater percentage of the main flow, and the value of ai increases, causing an increase in over-all efficiency. If the main flow were to decrease to a point where all of the flue gas passed through the shunt circuit (about 10%-15% of load) the efficiency of the collector would be the same as that of the cyclone [Er = 1 (0:2/0:2 + |3) ], which is the maximum efficiency for the collector. If the main flow were to increase with a constant shuntflow,the efficiency of the collector decreases due to a reduction in ai. The total efficiency of the collector will approach zero as the main flow approaches infinity. Hence, if a family of fractional effi579

FRACTIONAL EFFICIENCY CURVES LDL COLLECTOR-CONSTANT SHUNT FLOW 100

TRI-STATE CONFERENCE FEATURES AIR POLLUTION

What can be done and what is being done to safeguard the health and lives of millions of residents in the tri-state area of New York, New Jersey, and Connecticut was the topic of an all-day conference held on October 28, at Pace College, 41 Park Row, downtown Manhattan. Various causes of air pollutionfrom pigeon "free-loaders" in city parks to poisonous air-borne fumes, from chimneys, industrial smokestacks, and automobile exhausts, as well as from the open burning of waste, etc.were discussed by specialists on the panels and by authorities in the audience. Civic, industrial, and governmental action necessary to combat this growing menace to the area's population and economy thoroughly explored and defined. The meeting will also sum up the impact on the public of national "Cleaner Air Week" proclaimed each year in October.
Senator Was Speaker

90
^80 y 70 GI 60
UJ50

^ -

DECREASED BOILER LOAD

^-NORMAL BOILER LOAD

^ I N C R E A S E D 301 LER LOAD

z o 40 u
30 20 10 0
Downloaded by [202.133.55.169] at 11:40 27 February 2014

o u

10

20

3 0 4 0 50 60 70 80 PARTICLE SIZE-MICRONS
Fig. 13.

90

100

I.D. FAN INERTIAL CONCENTRATOR

7
SHUNT CYCLONE &FAN
Fig. 14.

ciency curves for different boiler loads are drawn for this particular type of collector, we find that the higher fractional efficiency curves are obtained at lower boiler loads, which is a complete reversal of the condition obtained with a normal cyclone installation (Fig. 13). Sometimes, when making an installation on an old boiler, the shunt fan may be placed in parallel with the existing inside diameter fan, so as to increase total fan capacity (Fig. 14). In this case, shunt flow is not a constant due to the different static pressure heads
580

imposed by the main inside diameter fan at various loads. Shunt flow decreases at high loads and increases at lower loads. The effect is to accentuate the condition already described so, that the efficiency increase at reduced boiler loads is more pronounced, as is the efficiency decrease at higher boiler loads.
REFERENCES 1. H. van der Kolk, "Cyclones in Industry," edited by Rietema & Verver, Elsevier Publishing Co., Amsterdam, 77-85, (1961).

United States Senator Edmund S. Muskie from Maine was the luncheon speaker, beginning at 12:15 P.M. Chairman of the Special Senate Subcommittee on Air Pollution Control (Senate Public Works Committee), the Senator outlined the plans of the federal government to "help clean the air we breathe." Public health aspects of air pollution were discussed by the following panelists: Dr. George James, Commissioner, New York City Department of Health; William H. Megonnell, Regional Program Director, Air Pollution Control, United States Department of Health, Education and Welfare; M. J. Plishner, Executive Director, Queensboro Tuberculosis and Health Association; Alexander Rihm, Executive Secretary, Air Pollution Control Board, New York State Department of Health; and Dr. Amos Turk, Associate Professor, Department of Chemistry, The City College of New York. Other panelists discussed the technical aspects of air pollution: Abraham Kussman, Air Pollution Meteorologist, United States Weather Bureau; George L. Minasian, Assistant to Vice President, Consolidated Edison Company of New York and immediate Past Chairman of the Mid-Atlantic States Section of the Air Pollution Control Association; and Major Alexander de Seversky, President, the Electronatom Corporation and trustee of the Tri-State Air Pollution Foundation. The moderator for the conference was Commissioner Arthur J. Benline, Department of Air Pollution Control, City of New York.
Journal of the Air Pollution Control Association