EPA-650/2-74-093
OCTOBER 1974
Environmental Protection Technology Series
-------
EPA-650/2-74-093
FINE PARTICLE SCRUBBER
PERFORMANCE TESTS
by
S. Calvert, N. C. Jhaveri, S. Yuny
A. P. T. , Inc.
P.O. Box 71
Riverside, California 92502
Contract No. 68-02-0285
ROAP No. 21ADJ-037
Program Element No. 1AB012
EPA Project Officer: L. E. Sparks
Control Systems Laboratory
National Environmental Research Center
Research Triangle Park, North Carolina 27711
Prepared for
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, D.C. 20460
October 1974
-------
This report has been reviewed by the Environmental Protection Agency
and approved for publication. Approval does not signify that the
contents necessarily reflect the views and policies of the Agency,
nor does mention of trade names or commercial products constitute
endorsement or recommendation for use.
11
-------
PREFACE
This report, "Fine Particle Scrubber Performance
Tests", is the final report submitted to the Control
Systems Laboratory for E.P.A. Contract No. 68-02-0285.
Scrubber performance data relating fine particle
penetration to particle size and operating parameters
are needed to validate and/or develop engineering design
methods so that performance can be predicted with more
confidence. Careful measurements of particle size and
concentration into and out of the scrubber were completed
for 7 types of scrubbers on various pollution sources.
Useful mathematical models were validated for all but
one of the scrubbers. Recommendations for future work
are given.
Dr. Leslie E. Sparks of.the Control Systems Laboratory,
National Environmental Research Center, Environmental
Protection Agency, was the Project Officer for this program.
Dr. Seymour Calvert of A.P.T., Inc. was the Project
Director.
Eight industrial organizations permitted tests to be
performed at their facilities and assisted the program in
many ways.
-------
CONTENTS
Page
Preface
Pi triirp";
3
List of Figures
List of Tables
10
Sections
Introduction 13
Summary, Conclusions and Recommendations 15
Method 21
Computation and Modeling Methods 27
Valve Tray On Urea Prilling Tower 45
(Koch Flexitray)
Vaned Centrifugal On Potash Dryer
(Ducon Multivane Scrubber)
87
Mobile Bed On Coal-Fired Boiler 113
(T.C.A. Scrubber)
Venturi Scrubber On Coal-Fired Boiler 137
(Chemico Venturi)
Wetted Fibrous Filter On Salt Dryer 159
Impingement Plate Test 193
(Impinjet)
Venturi Rod Scrubber On Cupola 223
References 267
Nomenclature 268
-------
FIGURES
No. Page
2-1 Comparison of Actual Data and Values 29
Calculated From Polynomial
2-2 Integrated (overall) Penetration as a 32
Function of Cut Diameter, Particle
Parameters and Collection Characteristics
2-3 Overall Penetration as a Function of Cut 33
Diameter and Particle Parameters for Common
Scrubber Characteristics, B = 2
2-4 Representative Cut Diameter as a Function 37
of Pressure Drops for Several Scrubber Types
2-5 Ratio of Particle Diameter to Cut Diameter 37
as a Function of Collection Efficiency
2-6 Predicted Particle Collection by Diffusion 40
in Plates, Packing, and Venturi Scrubbers
3-1 Assembly of a Scrubbing Element on the 46
Flexi Tray
3-2 Cumulative Mass Concentration Distribution 51
for Run #6
3-3 Urea Water Solution Drop Diameter Versus 53
Original Dry Urea Particle Diameter
3-4 Cumulative Mass Versus Aerodynamic Particle 57
Size for Run #9
3-5 Cumulative Mass Versus Aerodynamic Particle 58
Diameter for Run #10
3-6 Cumulative Mass Versus Aerodynamic Particle 59
Diameter for Run #11
3-7 Cumulative Mass Versus Aerodynamic Particle 60
Diameter for Run #12
3-8 Cumulative Mass Versus Aerodynamic Particle 61
Diameter for Run #13
3-9 Cumulative Mass Versus Aerodynamic Particle 62
Diameter for Run #14
3-10 Fractional Penetration Curves for Data Set "A" 63
3-11 Fractional Penetration Curves for Data Set "B" 64
3-12 Predicted and Experimental Penetrations for 67
Data Set "A"
-------
FIGURES
3-13 Predicted and Experimental Penetrations for 68
Data Set "B"
3-B-l Particle Size Distribution Measured by the 80
U. W. and Andersen Cascade Impactors
3-B-2 Particle Size Distribution Measured by the 81
U. W. and Andersen Cascade Impactors
3-B-3 Dry Particle Size Distribution Obtained with 82
In-Stack and Ex-Stack U. W. Cascade Impactor
3-B-4 Particle Size Distribution for Data Set A 83
3-B-5 Inlet Particle Size Distribution (Data Set B) 84
3-B-6 Outlet Particle Size Distribution (Data Set B) 85
4-1 Ducon Multivane Scrubber 88
4-2 Cumulative Mass Versus Particle Diameter 94
4-3 Cumulative Mass Versus Particle Diameter 95
4-4 Cumulative Mass Versus Particle Diameter 96
4-5 Experimental and Predicted Penetration 97
4-6 Cut Diameter-Pressure Drop Correlations 99
(Calvert, 1974)
4-7 Predicted Particle Diameter-Penetration IQO
Relationship for Inertial Impaction
(Calvert, 1974)
4-B-l Inlet Particle Size Distribution 110
4-B-2 Outlet Particle Size Distribution 111
5-1 Mobile Bed Scrubber 114
5-2 Duct Arrangements 116
5-3 Inlet Cumulative Mass Concentration Size 119
Distribution
5-4 Outlet Cumulative Mass Concentration Size 120
Distribution
5-5 Particle Penetration Versus Aerodynamic 122
Particle Diameter for T.C.A. Scrubber
-------
FIGURES
No. Page
5-B-l Inlet Particle Size Distribution 134
5-B-2 Outlet Particle Size Distribution 135
6-1 Chemico Venturi
6-2 Inlet Cumulative Mass Concentration ^43
Distribution
6-3 Outlet Cumulative Mass Concentration 144
Distribution
6-4 Particle Penetration Versus Aerodynamic 145
Diameter
6-5 Predicted and Experimental Penetrations 149
for Venturi
6-B-l Inlet and Outlet Particle Size Distributions 155
(log -probability)
6-B-2 Inlet Particle Size Distribution ^57
(log -probability)
7-1 Schematic Diagram of Wet Fiber Scrubber
7-2 Penetration Versus Particle Diameter
(Data Set "A")
7-3 Penetration Versus Particle Diameter
(Data Set "B")
7-4 Predicted and Experimental Penetrations for
Fiber Filter Bed (Data Set "A")
7-5 Predicted and Experimental Penetrations for 170
Fiber Filter Bed (Data Set "B")
7-B-l Inlet and Outlet Particle Size Distribution 180
(Data Set "A")
7-B-2 Inlet and Outlet Particle Size Distribution
(Data Set "B")
7-C-l Cumulative Mass Distribution for Run #3 184
7-C-2 Cumulative Mass Distribution for Run #4 185
7-C-3 Cumulative Mass Distribution for Run #5 186
7-C-4 Cumulative Mass Distribution for Run #6
-------
FIGURES
No. Page
7-C-5 Cumulative Mass Distribution for Run #7 188
7-C-6 Cumulative Mass Distribution for Run #8 189
7-C-7 Cumulative Mass Distribution for Run #9 190
7-C-8 Cumulative Mass Distribution for Run #10 191
8-1 Two Stage No. 245 Sly Impingjet Wet Scrubber 194
Shell - 1/4" FRP
8-2 Schematic Diagram of Scrubber System 195
8-3 Penetration Versus Particle Diameter 199
(Data Set "A")
8-4 Penetration Versus Particle Diameter 200
(Data Set "B")
8-5 Predicted and Experimental Penetration 202
(Data Set "A")
8-6 Predicted and Experimental Penetration 203
(Data Set "B")
8-B-l Inlet Particle Size Distribution for 210
Data Set "A"
8-B-2 Outlet Particle Size Distribution for 211
Data Set "A"
8-B-3 Inlet Particle Size Distribution for 212
Data Set "B"
8-B-4 Outlet Particle Size Distribution for 213
Data Set "B"
8-C-l Mass Concentration Distribution for Run #1 216
8-C-2 Mass Concentration Distribution for Run #2 217
8-C-3 Mass Concentration Distribution for Run #3 218
8-C-4 Mass Concentration Distribution for Run #4 219
8-C-5 Mass Concentration Distribution for Run #5 220
8-C-6 Mass Concentration Distribution for Run #7 221
9-1 Schematic Diagram of Scrubber System 224
9-2 Schematic Diagram of Venturi-Rod Bed 225
-------
FIGURES
No. Page
9-3 Diffusion Battery Assembly 230
9-4 Particle Penetration Versus Diameter for 232
Venturi-Rod Scrubber (Data Set "A")
9-5 Penetration Versus Particle Diameter for 233
Venturi-Rod Scrubber (Data Set "C")
9-6 Penetration Versus Particle Diameter for 234
Venturi-Rod Scrubber (Data Set "B")
9-7 Predicted Particle Cut Diameter Versus 259
Pressure Drop for Venturi Scrubber
9-8 Predicted and Experimental Penetration for 240
Venturi-Rod Scrubber (Data Set MA"-Ductile)
9-9 Predicted and Experimental Penetration for 241
Venturi-Rod Scrubber (Data Set "B"-Gray Iron)
9-10 Predicted and Experimental Penetration for 242
Venturi-Rod Scrubber (Data Set "C"-Ductile)
9-11 Predicted Penetration by Brownian Diffusion 243
and Inertial Impaction
9-B-l Inlet and Outlet Size Distribution (Set "A") 254
9-B-2 Inlet and Outlet Size Distribution (Set "B") 255
9-B-3 Inlet and Outlet Size Distribution (Set "C") 256
9-C-l Cumulative Mass Concentration for Run #1 258
9-C-2 Cumulative Mass Concentration for Run #2 259
9-C-3 Cumulative Mass Concentration for Run #3 260
9-C-4 Cumulative Mass Concentration for Run #7 261
9-C-5 Cumulative Mass Concentration for Run #9 262
9-C-6 Cumulative Mass Concentration for Run #10 263
9-C-7 Cumulative Mass Concentration for Run #11 264
9-C-8 Cumulative Mass Concentration for Run #12 265
9-C-9 Cumulative Mass Concentration for Run #13 266
-------
TABLES
No. Page
1-1 Gas Measurements 22
3-1 Impactor Location 48
3-2 Andersen Sampler Calibration at 50
0.028 m3/min (1 CFM)
3-A-l Particle Data for Run #1 72
3-A-2 Particle Data for Run #2 72
3-A-3 Particle Data for Run #3 73
3-A-4 Particle Data for Run #4 73
3-A-5 Particle Data for Run #5 74
3-A-6 Particle Data for Run #6 74
3-A-7 Particle Data for Run #7 75
3-A-8 Particle Data for Run #8... 75
3-A-9 Inlet and Outlet Sample Particle Data 76
For Simultaneous Run #9
3-A-10 Inlet and Outlet Sample Particle Data 76
For Simultaneous Run #10
3-A-ll Inlet and Outlet Sample Particle Data 77
For Simultaneous Run #11
3-A-12 Inlet and Outlet Sample Particle Data 77
For Simultaneous Run #12
3-A-13 Inlet and Outlet Sample Particle Data 78
For Simultaneous Run #13
3-A-14 Inlet and Outlet Sample Particle Data 78
For Simultaneous Run #14
4-1 Impactor Operating Conditions 90
5-A-l Coal Analysis 130
5-A-2 Inlet Sample Particle Data 131
5-A-3 Outlet Sample Particle Data 132
6-A-l Coal Analysis (As Received) 152
6-A-2 Inlet and Outlet Sample Particle Data 153
for Run #1
10
-------
TABLES
6-A-3 Inlet and Outlet Sample Particle Data 154
for Run #2
6-A-4 Inlet and Outlet Sample Particle Data 154
for Run #4
7-A-l Inlet and Outlet Sample Particle Data 174
for Run #3
7-A-2 Inlet and Outlet Sample Particle Data 174
for Run #4
7-A-3 Inlet and Outlet Sample Particle Data 175
for Run #5
7-A-4 Inlet and Outlet Sample Particle Data 175
for Run #6
7-A-5 Inlet and Outlet Sample Particle Data 176
for Run #7
7-A-6 Inlet and Outlet Sample Particle Data 176
for Run #8
7-A-7 Inlet and Outlet Sample Particle Data 177
for Run #9
7-A-8 Inlet and Outlet Sample Particle Data 177
for Run #10
8-A-l Inlet and Outlet Sample Particle Data 206
for Run # 1
8-A-2 Inlet and Outlet Sample Particle Data 206
for Run #2
8-A-3 Inlet and Outlet Sample Particle Data 207
for Run #3
8-A-4 Inlet and Outlet Sample Particle Data 207
for Run #4
8-A-5 Inlet and Outlet Sample Particle Data 208
for Run #5
8-A-6 Inlet and Outlet Sample Particle Data 208
for Run # 7
9-A-l Inlet and Outlet Sample Particle Data 248
for Run #1
11
-------
TABLES
No. Page
9-A-2 Inlet and Outlet Sample Particle Data 249
for Run #2
9-A-3 Inlet and Outlet Sample Particle Data 249
for Run #3
9-A-4 Inlet and Outlet Sample Particle Data 250
for Run #7
9-A-5 Inlet and Outlet Sample Particle Data 250
for Run #9
9-A-6 Inlet and Outlet Sample Particle Data 251
for Run #10
9-A-7 Inlet and Outlet Sample Particle Data 251
for Run #11
9-A-8 Inlet and Outlet Sample Particle Data 252
for Run #12
9-A-9 Inlet and Outlet Sample Particle Data 252
for Run #13
12
-------
INTRODUCTION
The need for more reliable data on the fine particle
collection efficiency of air pollution control scrubbers
has become increasingly apparent as control requirements
have grown more demanding. Major efforts, such as the Wet
Scrubber System Study (Calvert et al., 1972), to augment
our ability to design better scrubbers and to predict their
performance have illuminated this need. Design methods,
including mathematical models, have been developed from
basic theory plus whatever good data were available but
to a large extent they were untested. Thus, one could not
either predict performance for present scrubber designs
and operating conditions or extrapolate into better
combinations of design and performance with a reasonable
degree of confidence.
It is very difficult to compare scrubber performances
in different situations without knowing efficiency as a
function of particle size, commonly called: "grade efficiency."
Even on an empirical basis, there have been so few carefully
and properly done performance tests that the capabilities of
existing systems, were not known. Collection effiency in terms
of overall particle mass was rarely tested because of a pre-
dominant concern for only the outlet particulate loading or
emission rate. The few data which had been published were
generally unsatisfactory for use because of inadequate
methodology, undefined parameters, insufficient quantity, and
similar inadequacies.
The program reported here was initiated in response to
the need for additional reliable performance data on fine
13
-------
particle scrubbers. The objectives were as follows:
1. Obtain data on fine particle collection efficiency
as a function of particle size for scrubbers
operating on representative industrial emission
sources. Fine particles are those smaller than
several microns in diameter. Record pertinent data
on scrubber design and operating conditions.
2. Reconcile the performance data with existing
mathematical models, such as those presented in
the "Scrubber Handbook" by Calvert, et al. (1972).
Where necessary and to the extent possible, develop
better models and/or design approaches.
3. Obtain data on scrubber system costs for invest-
ment, operation, and maintenance.
4. Compile the available information on scrubber
operating characteristics and problems (including
entrainment), maintenance requirements, corrosion
and erosion experience, and similar items regarding
system behavior.
14
-------
SUMMARY, CONCLUSIONS AND RECOMMENDATIONS
SUMMARY
A summary of the performance test program is given in
tabular form below:
Test
# Source
1
2
3
4
5
6
7
Metal Melting
Urea Prilling Tower
KC1 dryer
Coal fired utility
boiler
Coal fired utility
boiler
NaCl dryer
NaCl dryer
Scrubber
Hybrid
Valve Tray
(Koch Flexitray)
Vaned Centrifugal
(Ducon)
Mobile Bed (T.C.A.)
Venturi
(Chemico)
Wetted fiber
Impingement Plate
Approximate
Cut Diameter
-
1 . 2 pmA
1 .2 pmA
0.4 umA
0.7 umA
0.8 ymA
1 .0 umA
(Sly Impinjet)
8 Foundry Cupola Venturi Rod 0.3 ymA
NOTE: Cut diameter is for 501 penetration.
Test No. 1 was started and had to be postponed due to
operating problems. Upon returning to the plant to resume the
test it was found that the scrubber system had been drastically
modified and it was then decided to abandon the test. The
remaining tests were all completed despite the necessity to
interrupt tests No. 2, 3, 4, and 5 because of plant shut-downs.
The experimental and computational methods were modified
as the program proceeded and experience led to the evolution
of better tools. Initially the focus of interest was in the
particle size range from a few tenths to several microns
diameter; or essentially what could be measured by means of a
cascade impactor. Later in the program there developed a
further concern for smaller particles, ranging down to
0.01 micron diameter. It then became necessary to employ
15
-------
some additional technique for particle size measurement.
A diffusion battery system was chosen for this purpose.
The evolution of a useful and convenient diffusion battery
apparatus is still in progress.
All scrubber performance tests present problems in
coping with entrained liquid in the outlet and many tests
also involve a high loading of large particles in the inlet
gas. Both situations necessitate the use of a pre-cutting
device to remove the heavy liquid or solid loading from
the sample gas before it reaches the cascade impactor.
Several approaches were tried before a satisfactory pre-
cutter was designed and proven in practice.
In general, there are a variety of problems depending
on the specific case and causing the test method possibili-
ties to be less than ideal. Most tests require the exercise
of judgement in deciding on the best compromise which will
yield valid data for the purpose at hand.
Analysis of the data for the computation of particle
penetration as a function of particle size ivas satisfact-
orily accomplished by means of a graphical technique. A
digital computation approach proved to be useful for some
cases but not for all. Consequently, both methods were
used where possible and checked against one another. The
combined effect of errors in experimental measurements on
computational procedures causes the uncertainty of pene-
tration determinations to be +^10% or more at a given diameter
Fortunately, the dependence of penetration on particle
diameter is so great that it usually overshadows the effect
of errors and one can obtain meaningful results for, say,
the particle size at 50% penetration.
16
-------
Comparison of the experimental results with mathe-
matical models (or design equations) was successful in
six out of seven cases. The results of this comparison
may be summarized as follows:
1. Valve tray on urea prill tower - A sieve plate
model compared well with the data after accounting
for particle growth due to water vapor condensation.
2. Vaned centrifugal on KC1 dryer - A gas atomized
spray model gave predictions which agreed with the
data if reasonable allowance is made for growth due
to condensation.
3. Mobile bed on coal-fired boiler - No satisfactory
model is available and attempts to find a reason-
able mechanism to account for the high efficiency
were not successful. Particle growth due to
condensation caused by H2S04 adsorption is a
probable contributor to the performance.
4. Venturi on coal-fired boiler - The model for a
venturi in terms of particle cut diameter
correlated with pressure drop agrees well with
the experimental results.
5. Wetted fiber on NaCl dryer - The model for
collection on fibers yields a good prediction
after allowing for reasonable growth due to
condensation.
6. Impingement plate on NaCl dryer - A model based
on impingement from round jets gives good agreement
with the data after allowing for particle growth
due to condensation.
7. Venturi rod on cupola - The venturi model gives a
good prediction for particles larger than about
17
-------
1.0 micron aerodynamic diameter but does not
account for low penetration for the sub-micron
particles. Brownian diffusion can explain the
high efficiency for particles smaller than a
few tenths micron aerodynamic diameter.
Cooperating Organizations
The names of the organizations who cooperated in this
program are not given because of the agreement to report
the results without identification of the source. We are
appreciative of their help in allowing the tests to be
made and in providing facilities, assistance, and
information.
CONCLUSIONS
The program achieved the principle objective of obtain-
ing reliable performance data for the validation of mathe-
matical models for scrubber design. Scrubbers of several
types on a variety of sources, including two very large
power plant boilers, were studied and the results add very
significantly to our engineering ability. It is possible
to predict performance for fine particle collection with
much more confidence than one could prior to these evalua-
tions.
A recently developed relationship between particle cut
(501 efficiency) diameter and scrubber pressure drop has
been tested with the data of this program. In all except
one case, the new correlation gives very good results and
is shown to be a very powerful and convenient design method.
The experimental methods for measuring fine particle
collection efficiency remain more difficult and less
accurate than one would like, despite improvements evolved
18
-------
by our organization and other investigators. The time
required to take the data also contributes to the cost of
the study because the longer the test period, the higher
the probability that the plant will break down and abort
the test. In several cases, the scrubber system reliability
is a problem while the intrinsic capability of the scrubber
(when operating) is satisfactory.
Data on capital investment, operating costs and
maintenance were generally not available in complete or
reliable form. Operating problems related mostly to those
caused by entrainment, solids deposition, and scaling.
Plugging, corrosion, fan unbalancing, and similar problems
stem from the aforementioned causes.
RECOMMENDATIONS
Recommendations stemming from the present study include
items in the nature of methods whose use appears warranted,
additional tests which should be made, experimental methods
to be improved, model development required and additional
research needed in related areas. For brevity, these are
listed below:
A. Recommended methods
1. Particle penetration predictions for scrubbers
other than mobile beds can be made with reasonable
confidence by means of the cut diameter - pressure
drop correlation.
2. Measurement of fine particle penetration in the
inertial impaction regime can be done with useful
accuracy by means of cascade impactors.
19
-------
B. Additional scrubber tests needed
1. Mobile bed scrubbers on a variety of sources with
and without condensation effects
2. Pre-formed spray scrubbers
3. Venturi scrubbers at high pressure drop, on non-
wettable particles, and on large gas flows
4. Plate type scrubbers on systems without condensation
effects and with non-wettable particles
5. Impingement and entrainment type scrubbers
6. Systems with wet fans
C. Experimental method improvement
1. Better impactor catch iveighing
2. More convenient and reliable diffusion battery
system
3. Instantaneous particle size and concentration
analysis
4. Aerosol dilution system for use with particle
counters and diffusion batteries
5. Particle density measurement
6. Opacity measurement for saturated gas streams.
D. Model development needed
1. Performance model for mobile bed
2. Reliable particle growth prediction for soluble
materials in near-saturated gas
E. Additional related research
1. Particle growth by condensation on soluble
materials at relative humidity of 100% and less
2. Effect of adsorbed gases on particle growth
(e.g., H2S04 on fly ash)
3. Particle collection efficiency in well controlled
mobile bed experiments.
20
-------
METHOD
The method of approach to the program objectives involved
a number of experimental determinations to obtain collection
efficiency data, the aquisition of information on system
characteristics and behavior, and computations which
utilized the performance data and mathematical models. Over
the course of the program the methods and apparatus used
were generally improved and were modified to suit each
specific test situation but the main features were similar
and will be described here.
The most important experimental measurements were those
regarding particle size and concentration. In the beginning
of the program the size range of primary interest was
from a few tenths to a few microns diameter, which is
within the measurement range of a cascade impactor. Later
the size range was extended downward by an order of magnitude
and it was necessary to use a diffusion battery in addition
to the cascade impactor. The apparatus and methods used
are outlined below.
1. Gas velocity distribution and parameters had to be
measured at the inlet and outlet of the scrubber
in order to define the following:
a. Conditions for isokinetic sampling.
b. Particle concentration per unit volume of
dry gas, which is a consistent basis for
comparing inlet with outlet in the computation
of efficiency.
c. Gas flow rate.
d. Amount of liquid entrainment in the outlet.
21
-------
The necessary gas parameters were measured as shown in
Table 1-1 below:
Table 1-1- GAS MEASUREMENTS
Parameter
Method
Velocity and
Flow rate
A type "S" pitot tube was calibrated in the
stack with a standard pitot tube and then
used to make a multipoint traverse.
Temperature
A thermocouple or a dial thermometer was
used for traversing.
Pressure
A water or mercury manometer measured
pressure by means of a static pressure tube
inserted in the duct.
Humidity
Wet and dry bulb temperature measurements
were made on a flowing sample withdrawn
from the stack. Outlets are generally
saturated and require some heating in
order to use this technique. Condensate
and adsorption by a drying tube in a
modified E.P.A. sampling train were also
measured and used for confirmation.
Gas density
Depending on the test, dry gas density was
measured by means of a pyenometer technique
or computed from process conditions. Humid
gas density (to be used in pitot tube
computations) was calculated from dry gas
density and humidity.
Liquid
Entrainment
The quantity of liquid entrainment in the
outlet was measured from the liquid
collected in the pre-cutter used upstream
of the cascade impactor in some runs and
by means of a dye-treated paper technique
for drop size determination in some.
22
-------
2. Particle sampling and size analysis data in
all tests were taken by means of cascade impactors.
The early tests were made with a sampling probe
from the stack to an externally mounted cascade
impactor. Later tests were all made with the
impactor in the stack in order to minimize probe
losses. The three types of cascade impactors which
were used are:
a. An Andersen "viable" sampler for about
302,/min (1 CFM) of sample was used ex-stack.
It was calibrated by means of polystyrene
latex particles and a Climet particle counter.
A glass fiber paper filter was used after the
impactor.
b. A Brink cascade impactor for about 6£/min was used
for both in- and ex-stack measurements. It was
also followed by a glass fiber paper filter to
collect particles smaller than the last stage cut
size. This impactor was calibrated as the
Andersen was.
c. An University of Washington (Pilat) Mark III
Cascade Impactor for about 302,/min was used
for in-stack sampling. It contains a filter
holder after the last impaction stage. The
manufacturer's calibration was used for this
impactor.
All of the impactors were operated with inlet nozzles
appropriately sized to give isokinetic sampling. In the
later tests a pre-cutter was used to remove either the
heavy particle loading from inlet samples or the entrained
liquid from outlet samples. A cyclone separator with
about a 3 umA cut diameter was first used but a round jet
impactor with about an 8 ymA cut diameter was found to have
better characteristics and was adopted for use for both
23
-------
inlet and outlet sampling. The impactors were either given
time to reach the duct gas temperature or were heated to
prevent moisture deposition.
Several types of particle collection substrates were used
with the impactors. Generally a pre-weighed, greased
aluminum foil substrate was used. In some <-.o.ses a glass
fiber paper substrate was used. Silicone vacuum grease was
either wiped on the foils or applied as a solution in
an organic solvent.
Impactor substrates and filters were weighed with
an analytical balance to the nearest tenth milligram (10~lf g) .
Tare weights were taken after drying in an oven and desiccator
Sample weights were taken both before and after drying.
Sample Bias
It is important to note that the program objective
is to investigate scrubber performance on fine particles
and, consequently, it is not necessary that the methods
used be accurate for large particles. This makes the
sampling simpler in the following ways:
a. Isokinetic conditions are not important for fine
particles. For example, the error caused by
sampling 4 ymA particles at a velocity 50% higher
or lower than the gas stream velocity would only be
about 2 or 3% of the concentration.
b. The fine particles will be well distributed in the
gas stream, except in cases where streams with
different particle concentrations have not had time
to mix, so single point sampling is generally
sufficient. To illustrate, we may note that the
Stokes stopping distance of a 3 ymA particle with
an initial velocity of 15 m/sec (50 ft/sec) is
about 0.04 cm (0.016") and for a 1 ymA diameter
particle it is one ninth of that. Since the stopping
distance is the maximum a particle can be displaced
24
-------
from a gas stream line by going around a right-
angle turn, it is obvious that fine particle
distribution in the gas stream will be negligibly
affected by flow direction changes.
c. The effect of a pre-cutter on the size resolution
of a cascade impactor is not significant in the
size range of interest, so long as the pre-cutter
has a cut diameter larger than several microns.
Particle size distributions were plotted on log-probability
paper and described in terms of the approximate mass median
aerodynamic diameter and geometric standard deviation which
were obtained from the best straight line through the data.
Collection efficiency was computed by means of a technique
which utilizes a plot of the cumulative particle mass concen-
tration versus particle aerodynamic diameter. We use the
symbol "ymA" for aerodynamic diameter, which is equal to
particle diameter (d ) in pm (microns) times the square root
of the particle density (p ) in g/cm3 times the square root
of the Cunningham slip correction factor (C'). This computa-
tion is described in detail later in this report.
Comparison of the particle collection performance with
the prediction of mathematical models is described in the
separate chapters on the individual performance tests. Other
data and information specific to each test are also presented
in the appropriate chapters.
25
-------
26
-------
COMPUTATION AND MODELING METHODS
The computation of penetration as a function of particle
size has been done in our previous reports by means of a
method vrttich involved some graphical steps. Curve fitting
to the data points and the measurement of curve slopes were
done "by eye" and involved subjective judgement. In an
effort to standardize the method we developed a completely
defined computational procedure and it is presented in the
following section. By using the same mathematical procedure,
different individuals should arrive at the same answer.
Overall penetration is defined as:
(2-1)
where W is the total particle mass and Pt^ is the penetration
(2-2)
for
particle
Pt
diameter,
f (d ..)
i f Cdpi)
d . , and
pi'
outlet
inlet
it is
dW
d (d . )
- pi -
" dW '
d(dpiJ^
*iven by:
outlet
inlet
where I" dW "I is the slope Of the cumulative mass versus
L Pi J
particle diameter curve at d . and equals f (d .).
In order to calculate this quantity, we must fit the
cumulative mass data with a mathematical function.
Several mathematical functions, e.g., hyperbolic tangent
function, exponential function, Rosin-Rammler distribution
function and polynomial function, etc., have been tried.
None of these functions gives a satisfactory fit for the
whole range of particle size. They fit the data very well
in the upper range of particle size. However, valuable in-
formation were lost in the small particle size range since
27
-------
these functions all have the tendency to smooth the data
points.
In order to overcome these difficulties, several
alternative approaches have been tested. One of these is
to fit only the lower range with a curve instead of the
whole range. This is acceptable because we are only
interested in the ability of the scrubber to control small
particle emissions.
Among these functions, only the high degree polynomial
function follows the data points closely. Therefore, it
was used to compute the scrubber collection efficiencies.
A method of least squares technique, which is presented in
appendix A of this chapter, was used to fit the function to
the experimental data.
COMMENTS ON CURVE FITTING BY POLYNOMIAL
The estimation of parameters by least squares causes a
smoothing of a given set of data and eliminates, to some
degree, errors in observation, measurements, recording, trans-
mission and conversion, as well as other types of random error
which may have been introduced in data. Systematic errors or
bias errors will not be eliminated by the least square technique
Figure 2-1 compares the polynomial fit with actual data.
Curve A is a 3rd degree polynomial fit for the first five data
points. Curve B is a 4th degree polynomial fit for the first
6 points and curve C is a 3rd degree polynomial fit for the
first 6 points. From this figure, it clearly shows that
curves A and B fit the experimental data almost exactly at
the lower end of the curve and oscillate at the upper end.
Curve C tends to smooth the data points. Since we are only
interested in small particles, 3rd degree polynomial fit for
the first 5 points is accurate enough for us to calculate the
slope of the curve and particle penetration.
28
-------
to
o
x
p
bo
W
P
=
u
-" t- -* -f- -~,
izri_ ::?r.r:r-
3rd Degree polynomial
fit for first 5 points
4th Degree polynomial
fit for first 6 points
3rd Degree polynomial
fit for first 6 points
Actual value
10
15
dp,
Figure 2-1 - Comparison of actual data and
values calculated from polynomial.
29
-------
CUT DIAMETER METHOD
Difficulty of separation
The "cut diameter" method, first described in the "Scrubber
Handbook" (Calvert et al. 1972) and further discussed by
Calvert et al. (1974), can be .used as a convenient method
for particle collection efficiency prediction. This method
is based on the idea that the most significant single
parameter to define both the difficulty of separating particles
from gas and the performance of a scrubber is the particle
diameter for which collection efficiency is 0.5 (50%).
For inertial impaction, the most common particle
separation process in presently used scrubbers, aerodynamic
diameter defines the particle properties of importance.
d = d (p C')172, common units = pm(g/cm3) 1;2=umA (2-3)
pa p p
When other separation mechanisms are important, other particle
properties may be more significant but this will occur gener-
ally when "d " is less than a micron.
P
When a range of sizes is involved, the overall collection
efficiency will depend on the amount of each size present and
on the efficiency of collection for that size. We can take
these into account if the difficulty of separation is defined
as the aerodynamic diameter at which collection efficiency
(or penetration) must be 50%, in order that the necessary
overall efficiency for the entire size distribution be attained.
This particle size is the required "separation cut diameter",
"dRC" and it is related to the required overall penetration,
Ft, and the size distribution parameters.
The number and weight size distribution data for most
industrial particulate emissions follow the log probability
law. Hence, the two well established parameters of the log-
normal law adequately describe the size distributions of
30
-------
particulate matter. They are the geometric mean weight
diameter "d a" and the geometric standard deviation "a ".
sr C» Qf
Penetration for many types of inertial collection equip-
ment can be expressed as a function of constants "A " and "B":
a
Pt = exp (-Aa dpa B) (2-4)
One may use the simplifying assumption that this
relationship can be based on actual diameter, d . This will
not introduce much error and it will conservatively utilize
too low an efficiency for particles smaller than a micron or
s o. Thus:
Pt = exp (-Adp B) (2-5)
Packed towers, centrifugal scrubbers, and sieve plate
columns follow the above relationship. For the packed tower
and sieve plate column "B" has a value of 2. For centrifugal
scrubbers "B" is about 0.67. Venturi scrubbers also follow
the above relationship and B = 2 when the throat impaction
parameter is between 1 and 10.
The overall (integrated) penetration, Ft, of any device
on a dust of any type of size distribution will be:
w
Pt = _f (|^) Pt (2-6)
o
The right-hand side of the above equation is the inte-
gral of the product of each weight fraction of dust times
the penetration on that fraction. If equation (2-6) is solved
for a log-normal size distribution and collection as given
by equation (2-5), the resulting equation can be solved to
yield Figures 2-2 and 2-3.
Figure 2-2 is a plot of "Ft" vs.(d 5Q/d )B with "B In
(o )" as a parameter. For a required "Pt" one can find the
31
-------
is:
z"
o
DC
UJ
z
UJ
0.
Q
UJ
oc
a
UJ
I-
z
0.001
0.001
0.01
0.
1.0
Figure 2-2. Integrated (overall) penetration as a
function of cut diameter, particle
parameters and collector characteristic
32
-------
1.0
o O.I -
a:
H-
UJ
z
UJ
Q.
g .01
O
.001
0.001
Pt=EXP -Ad
0.01
O.I
1.0
Figure 2-3. Overall penetration as a function of cut
diameter and particle parameters for common
scrubber characteristic, B = 2.
33
-------
value of dRC when "d ", "a ", and "B" are given. For con-
venience, Figure 2-3 is presented as a plot of "Ft" vs.
(d 5Q/d ) with a as the parameter when B = 2.
To illustrate the use of the separation cut diameter,
assume that 2% penetration is "needed for dust with d = 10 ym,
XT o
p = 3g/cm3 and a =3. If a scrubber such as a packed bed,
sieve plate, or venturi is to be used, Figure 2-3 shows the
cut diameter, d 5Q, must be 0.09 x Cdpg) = °-9 Pm- Tne
corresponding aerodynamic diameter is cLp = 1.7 ym (g/cm3) lfi
= 1.7 ymA. Of course if the scrubber is capable of a smaller
cut diameter, that is good; so "dRC" is the maximum cut
diameter acceptable. Some scrubbers, such as Venturis, are
only approximately fitted by relating penetration to exp (d 2)
and more accurate plots can be prepared by using more
representative performance equations. To avoid confusion these
will not be given here, although they are presented in the
"Scrubber Handbook".
Scrubber Performance
Collection efficiencies have been reported in the form
of "grade efficiency" curves, which are plots of particle
collection efficiency versus particle diameter for "typical"
scrubbers. Unfortunately, there can be great variation in
performance, depending on operating conditions and scrubber
geometry so that one would need a grade efficiency curve for
each important set of parameters.
The cut diameter approach proves to be a much more
compact way to characterize scrubber performance. Performance
graphs for a number of the important types of scrubbers are
presented by Calvert et al. (1974). Capability is defined by
"performance cut diameter", "dpc", which is the aerodynamic
particle diameter at which the scrubber gives 50% collection
efficiency.
34
-------
Once a scrubber type, size, and operating conditions
are chosen by matching the "separation" and "performance" cut
diameters, (i.e., dR~ = dp.-,) a more accurate efficiency-
diameter relationship can be developed and a more accurate
computation of overall penetration can be made. The reason
this step is necessary is that the relationship between
overall penetration and separation cut diameter is shown in
Figures 2-2 and 2-3 is only correct for packed beds and
similar devices and is an approximation for others.
SCRUBBER ENERGY
The energy required for particle scrubbing is mainly a
function of the gas pressure drop, except for pre-formed
sprays and mechanically aided scrubbers. Previously we have
been shown that there is an empirical relationship between
particle penetration and power input to the scrubber for a
given scrubber and a specific particle size distribution
(Lapple and Kamack (1955) and Semerau (I960)). However, this
"power law" did not provide a way to predict performance vs.
power input for any size dust, without first determining the
relationship experimentally.
A new relationship between "dpc" and scrubber pressure
drop (S. Calvert, 1974) is presented here. Figure 2-4
is a plot of performance cut diameter, dp.-,, versus gas
pressure drop for sieve plates, venturi (and similar),
impingement plates, and packed columns. Predictions were
made by means of design methods given in the "Scrubber
Handbook"
1. Sieve plate penetration and pressure drop predictions
for one plate are plotted as lines la and Ib for perforation
diameters of 0.5 cm and 0.3 cm, respectively, and F=0.4. Cut
diameters for other froth densities (F) can be computed from
35
-------
the relationship that they are inversely proportional to "F".
Cut diameters for two and three plates in series would be 84%
and 801 of those for one plate at any given pressure drop.
Note that these predictions are for wettable particles and
that both froth density and pressure drop are dependent on
plate design and operation.
2. Venturi penetration and pressure drop data are
given for f = 0.25 and f = 0.5 in lines 2a and 2b, respect-
ively. The predictions are for a liquid to gas ratio,
QT/Q - 1 £/m3, corresponding to about the minimum pressure
Lt \3
drop for a given penetration. Data recently obtained by
A.P.T. for a large coal-fired power plant scrubber fit a
value of f = 0.5.
3. Impingement plate data used for line #3 were
predicted for one plate. Cut diameters for 2 and 3 plates
in series are 88% and 83% of those shown in line #3.
4. Packed column performance as shown by line #4 is
representative of columns from 1 to 3 meters high and
packing of 2.5 cm nominal diameter.
To estimate the penetration for particle diameters
other than the cut size, under a given set of operating
conditions, one can use the approximation of equation (2-5)
with B = 2.0. Alternatively, one could use more precise data
or predictions for a given scrubber. Figure 2-5 is a plot of
the ratio of particle aerodynamic diameter to cut diameter
versus penetration for that size particle (d ) , on log-
probability paper. One line is for equation (2-5) and the
other is based on data for a venturi scrubber.
Performance Limit for Inertial Impaction
The limit of what one can expect of a scrubber utilizing
inertial impaction is clearly indicated by Figure 2-4. If a
36
-------
3 -
o
0.
o
of 2
UJ
UJ
5
< 1.0
D
O
O 0.5
0.4
0.3
O.I
SCRUBBER
la Sieve, F= O.4,dn = O.5cm.
Ib Sieve, F= 0.4, dh= 0.3 cm.
2a Venturi,F = 0.25
2b Venturi,F=0.5
3 Impingement Plate
4 Packed Column, dc- 2.5 cm.
' ' _ i - 1 - 1
<» 5 IO 20 3O 4O 5O IOO
PRESSURE DROP. cm. W.C.
200 30O
Figure 2-4. Representative cut diameter as a
function of pressure drop for
several scrubber types.
10
O.I
For Venturi
For Pt = exp -Ad
pa
I 10 50 90 99
COLLECTION EFFICIENCY FOR dpa (%)
Figure 2-5. Ratio of particle diameter to
cut diameter as a function of
collection efficiency.
37
-------
cut diameter of 1.0 pmA, or smaller is required, the necessary
pressure drop is in the medium to high energy range. High
efficiency on particles smaller than 0.5 ymA diameter would
require extremely high pressure drop if inertial impaction
were the only mechanism activ.e.
High efficiency scrubbing of sub-micron particles at
moderate pressure drop is possible, but it required either
the application of some particle separation force which is
not dependent on gas velocity or the growth of particles so
that they can be collected easily. Particle separation
phenomena which offer promise and have been proven to some
extent are the "flux forces" due to diffusiophoresis, thermo-
phoresis, and electrophoresis . Brownian diffusion is also
useful when particles are smaller than about 0.1 urn diameter.
Particle growth can be accomplished through:
1. Coagulation (agglomeration)
2. Chemical reaction
3. Condensation on particles
4. Ultrasonic vibrations
5. Electrostatic attraction
Diffusional Collection
Particle collection by Brownian diffusion can be de-
scribed by relationships for mass transfer and it is possible
to outline the magnitude of efficiency which can be attained
with typical scrubbers. The general relationship which
describes particle deposition in any control device in which
turbulent mixing eliminates any concentration gradient normal
to the flow outside the boundary layer and in which the dep-
osition velocity is constant is:
Pt . exp - U (2-7)
where, upD = particle deposition velocity
A, - total outside surface area of drops in scrubber
38
-------
Q = gas volumetric flow rate.
u
The particle deposition velocity for Brownian diffusion,
uRf), can be estimated from penetration theory as:
= 1.13
D \V*
~ (2-8)
For packed columns the penetration time, 9, can be taken
as the time required for the gas to travel one packing diameter,
For plate scrubbers which involve bubbles rising through
liquid, the penetration time for a circulating bubble is about
that for the bubble to rise one diameter, as shown by Taheri
and Calvert (1968). For spray scrubbers the penetration
time is that for the gas to travel one drop diameter.
Predictions of particle penetration due to Brownian
diffusion only were made by means of equations (2-7) and (2-8)
for a typical sieve plate and packed columns. A prediction
for a venturi scrubber was made by means of "Scrubber Hand-
book" equation (5.2.6-17), for gas phase controlled mass
transfer.
The results are plotted on Figure 2-6 as collection
efficiency vs. particle diameter. It can be seen that high
efficiency collection of 0.01 ym diameter particles is
readily attainable with a three plate scrubber, typical of
a. moderately effective device for mass transfer. Collection
efficiency for particles a few tenths micron diameter is
poor, as is well known.
Particle separation by flux force mechanisms is not
amenable to such simple treatment as Brownian diffusion
because of the variation of deposition velocity with heat
and mass transfer rates within the scrubber.
39
-------
too
o
O.OI
PARTICLE DIAMETER, fim
Figure 2-6. Predicted particle
collection by diffusion
in plates, packing, and
venturi scrubbers.
40
-------
APPENDIX 2-A
CURVE FITTING TECHNIQUE
41
-------
CURVE FITTING TECHNIQUE
The principle of least squares is employed to derive
information about the functional relation between particle
diameter and cumulative mass, assuming such a relation
exists, from a set of data pairs (d ., W.) (i = o, n) .
The technique is to fit a function of the form
Ffd ) = a f (d .) + a. f. (d .) + + a f (d .) (7 . -n
* p' oo*- pi' J J pi m m v pi' l^-A-lj
to a set of data pairs.
Where f. (d .) are some arbitrary functions and a ,
J pi 7 o'
a. a are independent parameters to be determined.
These parameters may either be linear or non-linear. In
the present study, we only considered linear parameters.
The difference between the approximating function
value, F(d .)> and the corresponding data value, VL, is
called residual, r., and is defined by the relation
ri = F(dpi) " Wi Ci = o, n) (2-A-2)
The function that best approximates the given set of data
in a least-squares sense is that the function produces
the minimum value of the sum Q of the squared residuals
Q = E [F(dpi) - W^2
= E [a f (d .) + a, f, (d .) +---+ a f
o o pi l i pi ni m
i
(2-A-3)
42
-------
A minimum is obtained when m + 1 partials of Q (aQ> ---
with respect to parameters a., (j = o, m) , simultaneously
vanish, i.e., when
fJL.E 2 I [F(dp.)
W.]
3 F(d
or s 2 I [a0 fQ (d ) * a, f, (d t
i
f(dp.)
fj
In matrix form, this becomes
= 0, --- m
(2-A-4).
o
i1 "i
(2-A-5)
43
-------
The solution (a , a. --- , a^) of equation (2-A-5) can be computt
by inverting the matrix of coefficients in that equation, and
multiplying the right-hand column matrix by this inverse
matrix.
In the case of curve fitting by polynomial,
f. (d .)
j pi
(d
pi
j = 0, 1, --- m
J
and the coefficient matrix is
+ 1
Q. .
pi
z dp. -
I (dpi)* ....
* Cdpi)m+
Once the functional relationship between particle
diameter and cumulative mass is obtained, the calculation
of penetration is straightforward.
44
-------
VALVE TRAY ON UREA PRILLING TOWER
(Koch Flexitray)
SOURCE AND SCRUBBER
An 85 Am3/min (3,000 ACFM) valve tray (Multi-Venturi
Flexitray) scrubber for urea prilling tower exhaust was
chosen for the second performance test. The scrubber, which
was designed and built by Koch Engineering Company utilizes
two trays in series. The bottom tray contains 27 openings
and the top tray contains 70 openings. Each of the openings
is surmounted by a spider cage holding a floating cap, or
"valve" (Figure 3-1) . In addition, each tray is equipped
with downcomers and weir flow baffles that control the
scrubbing liquid as it flows across the tray and then to
the tray below.
The scrubber was a pilot plant installed to determine
its effectiveness in scrubbing particles from the urea prill
tower exhaust gas which was brought down to ground level
from the tox\rer top, about 46 m (150 ft) above ground. The
gas enters the bottom of the scrubber and flows upivard
through the trays. At low gas velocities, the lightweight
caps (located in every other row) rise first, whereas the
heavy weight caps (in the alternate rows) remain in the
closed position. All the caps are fully opened as the vapor
flow attains the design conditions.
The liquid flows across the tray deck and is kept in
a constant froth by the gas which flows from the caps. There
is always a head of froth maintained by the weir. After
passing through the trays the gas passes through a mist
eliminator. The scrubbed gas then flows from the top of the
scrubber to the induced draft fan and a short stack.
45
-------
LIQUID LEVEL
\' tfrr&*-4 1g£
LOW PRESSURE
DROPVENTURI -
ENTRY
V
FLEXITRAY
Figure 3-1 - Assembly of a scrubbing
element on the Flexi Tray
46
-------
Because this scrubber was a pilot plant, there were no
data available on operating problems, maintenance, economics,
etc.
TEST METHOD
The performance of the scrubber was determined by
analyzing the particle size distribution and mass loading
of the inlet and outlet gas samples. As in all of the per-
formance tests, each sample was taken at one point in the
duct. A modified E.P.A. sampling train equipped with a
cascade impactor was used for particle size sampling. Two
types of impactors,an Andersen Sampler and University of
Washington Mark III Cascade Impactor (or Pilat) , were used
for this purpose.
An in-stack cyclone pre-cutter was attached to the
sampling probe. The cyclone collects particles larger than
about 3 ymA diameter and leaves the fine particles to be
collected by the cascade impactor/back-up filter assembly.
The effect of condensation on particle size was
studied during this performance test. Therefore, in some
of the test runs the impactor was kept in-stack and allowed
to heat up to stack temperature before sampling. In other
runs the impactor was ex-stack and with electric heating
tape wrapped around the outside jacket of the impactor. The
temperature of the impactor was controlled by a variac on
the heating circuit. Details on each run's impactor location
were listed in Table 3-1.
Scrubber inlet and outlet gas temperatures were measured
by mercury filled glass bulb thermometers. Gas temperature
at the sampling location was measured during each test run.
The inlet and outlet stack pressures were measured with a
47
-------
Table 3-l.IMPACTOR LOCATION
RUN NO.
IMPACTOR LOCATION
la, 2a, 3a, 4a
Ib, 2b, 3b, 4b
5, 6, 7
8
9, 10
Inlet of 11, 12, 13, 14
Outlet of 11, 12, 13, 14
in-stack
ex-stack
in-stack
ex-stack (heated probe)
ex-stack (heated probe)
ex-stack (heated probe)
ex-stack (heated probe)
48
-------
U-tube manometer. Barometric pressures were determined
before each run from an aneroid barometer. The stack gas
humidities were determined by dry and wet bulb thermometer.
Gas flow rate was measured by means of a calibrated S-type
pitot tube traverse.
IMPACTOR CALIBRATION
Since different types of cascade impactors were used
in a simultaneous inlet and outlet sampling, the agreement
between these impactors was determined before making any
interpretation of the data. Four sets of samples from the
scrubber inlet gas stream were taken simultaneously with
an in-stack U. W. impactor and an ex-stack Andersen Sampler.
These sets were taken to compare and calibrate the stage cut
diameters, d 5Q, between the U. W. and Andersen impactors.
Test results were listed in Tables 3-A-l through 3-A-4 and
were plotted in Figures 3-B-l and 3-B-2.
A discrepancy was found (see Figures 3-B-l and 3-B-2) in
the measurement of particle size distribution between the
Andersen and the U. W. impactors. For consistency, the
Andersen Sampler was calibrated against the U. W. impactor,
which was assumed to be correctly calibrated. The calibrated
cut diameter , d , for each stage of the Andersen Sampler
was taken to be the diameter, based on the U. W., correspond-
ing to the mass fraction undersize measured by the Andersen.
For example, Figure 3-B-l shows that the cut diameter for
stage 7 of the Andersen was 0.36 ymA according to the manu-
facturer's calibration, but 24% undersize would correspond
to a cut diameter of about 0.58 umA based on the U. W. data
and calibration. Table 3-2 shows the calibrations for the
last 4 stages of the Andersen Sampler as given by the
manufacturer and as calibrated against the U. W.
49
-------
Table 3-2. ANDERSEN SAMPLER CALIBRATIONS
AT 0.028 m3/fliin (1 CFM)
Stage
No.
4
5
6
7
Stage Cut Diameter, ymA
Manufacturer
1.75
0.9
0.54
0.36
Field Calibration
2.0
1.5
1.0
0.6
SAMPLING RESULTS
Wet and Dry Particle Size
The scrubber outlet samples were taken either in the
duct after the exhaust fan or between the scrubber exit and
the exhaust fan. It was found that particulate matter coming
out of the scrubber was in the form of urea-water drops, be-
cause crystals and liquid drops were observed on the last two
collection plates of the Andersen Sampler. Therefore, the
particle size measured will be that of a "grown" particle and
will not correspond directly with the inlet size data for dry
particles. The actual dry particle size would be smaller and
the curve of cumulative mass vs. "d " should shift to the
Pa
left as indicated by the dotted curve on Figure 3-2.
In order to determine the nature of the grown particles
measured in the exit, some samples were taken in such a way
as to avoid evaporation of the urea-water drops. Runs 5 and
6 (see Tables 3-A-5 and 3-A-6) were done with an in-stack U. W,
impactor to sample the outlet gas stream between the scrubber
exit and the exhaust fan. Liquid drops on the collection
plates of the U. W. were observed when this sampling tech-
nique was used. Each of the impactor collection plates was
50
-------
d
pa'
Figure 3-2 - Cumulative mass concentration distribution
for Run #6.
51
-------
weighed before and after putting it in a silica gel filled
desiccator. Particle sizes "d " and drop sizes "di" are
derived from either of the following alternative assumptions.
Alternative 1 - The water collected on the cascade
impactor collection plates is in the form of water drops
when passing through the cascade impactor jets. The urea
caught on the plates is in the form of dry urea particles
when passing through the cascade impactor jets.
Alternative 2 - The material caught on the impactor
plates is in the form of drops of urea-water solution.
It was proven in test Runs 7 and 8 (see Tables 3-A-7 and
3-A-8) that Alternative 2 is more valid than Alternative i.
Test No. 7 is an outlet test run with an in-stack U. W.
Run 8 was taken with an ex-stack U. W.. attached to a
heated sampling probe. The heated sampling probe was used
to evaporate water in the sample gas before it reached the
U. Wt) so that dry particles were measured by the impactor.
Figure 3-B-3 shows that the results from both of these tests
correlate very closely when the Run 7 data are converted
to the dry basis, assuming alternative 2 is correct. There-
fore, equivalent dry particle size may be calculated from
the urea-water drop size for the outlet sampling test with
an in-stack U. W.
Once the dry particle and wet particle size relationship
was found, the prediction of particle growth based on dry
particle size was possible. Figure 3-3 is a curve showing
the dry and wet particle size relation for this particle
composition.
Scrubber Operating Condition and Particle Data
Six simultaneous scrubber inlet and outlet samples \vere
taken to determine scrubber performance related to the particle
52
-------
H
o
w
^j
u
-
H
pq
10
5
4
3
2
1.0
0.5
0.4
0.3
0.2
0.1
Run
t
£1
g -
a _
1 I
o.i
0.5 1.0 5
DRY UREA PARTICLE DIAMETER, ymA
10
Figure 3-3 Urea-Water solution drop diameter versus
Original Dry urea particle diameter.
53
-------
size. These test runs were grouped into two data sets
corresponding to different urea prill tower operating
conditions. Runs 9 and 10 were in set "A", Runs 11 through
14 were in set "B". The prill tower operating conditions
were shown in the following tabulation. Buckets are the
devices used for atomizing the urea into liquid drops which
are then solidified into prills.
Urea Prill Tower Operating Conditions
Data Set
A
B
Type of Urea
Prill Bucket
"Old"
Tuttle
Urea Prill Temp.
°C
88
71
The scrubber operating conditions during these tests were as
follows:
1. Gas parameters were shown in the tabulation below
Gas Parameters
Temperature
Pressure, cm W.C .
A m3/min
ACFM
DN m3/min 6 °C
DSCFM § 70°C
Vol. % H,0 vapor
£»
Inlet
27°C
-15.0
86
3,100
78.6
2,990
1.5
Outlet
17°C
-35.0
85
3,000
78.6
2,990
2.0
54
-------
2. Pressure drop data (in cm W.C.)
Run No.
9
10
11
12
13
14
Bottom Tray
14.2
12.2
11.2
10.7
10.7
5.3
Top Tray
5.6
7.4
5.8
5.8
5.8
6.1
Demister
0.76
0.76
0.76
0.76
0.76
3. Scrubber liquor flow rate (in m3/min)
Run No .
9
10
11
12
13
14
Inlet
0.04.7 (12.5 GPM)
0.047 (12.5 GPM)
0.06 (16 GPM)
0.06 (16 GPM)
0.06 (16 GPM)
0.049 (13 GPM)
Outlet
0.047 (12.5 GPM)
0.047 (12.5 GPM)
0.06 (16 GPM)
0.06 (16 GPM)
0.06 (16 GPM)
0.049 (13 GPM)
4. Entrainraent was not measured in this performance
test.
The particle concentration and size data which were
obtained in this performance test are presented in Tables 3-A-9
through 3-A-14. Figures 3-B-4, 3-B-5 and 3-B-6 are log-probability
plots of inlet and outlet particle size distributions for
data sets A and B. There are some variations between set
A and B in particle sizes. This is mainly due to different
55
-------
urea buckets. The mass median diameter and geometric
standard deviation for these sampling runs are listed in
the following tabulation:
Run No .
9
10
11
12
13
14
Inlet
dp50, V**
0.82
0.82
1.1
1.1
1.1
1.1
"g
1.7
1.7
1.5
1.5
1.5
1.5
Outlet
dp50, V"A
1.2
1.2
0.9
0.9
0.9
0.9
a
g
2.2
2.2
1.9
1.9
1.9
1.9
Cumulative mass concentration was plotted against aerodynamic
particle size to yield Figures 3-4 through 3-9.
PARTICLE PENETRATION
The ability of a scrubber to control particulate
emissions is interpreted in terms of "grade efficiency"
curves, which are plots of particle collection efficiency,
or particle penetration versus particle diameter. The
penetration can be described as the ratio of the outlet
cumulative mass distribution slope to the inlet cumulative
mass distribution slope, as given in equation (2-2). The
slopes, dM/d(dpi), are determined by graphical procedures
on Figures 3-4 aiid 3-9 for the scrubber inlet/outlet
samples. The particle collection efficiencies obtained by
this method were plotted in Figures 3.-10 and 3-11. It
should be noted that the penetration for particles smaller
than 1.0 umA in Figure 3-10 is highly dependent on the shape
of the cumulative mass distribution in Figures 3-4 and 3-5.
56
-------
Q
g
A
2
2
W
U
'.-:
C
U
x
w
4
H
U
; :/\t::iniot[_
-,r i ^._.
d ymA
pa'
Figure 3-4 - Cumulative mass versus aerodynamic particle
size for Run #9.
57
-------
g
z
Q
u
o
u
CO
CO
U
pa,
Figure 3-5 - Cumulative mass versus aerodynamic
particle diameter for Run #L(K
58
-------
&
2
n
M
6
z:
O
U
:-:
O
U
te
E3
U
dpa, pmA
Figure 3-6 - Cumulative mass versus aerodynamic particle
diameter for Run #11.
59
-------
24
22
I 20
18
£ 16
1 15
g 14
CJ
§
" 12
00
CO
S 10
8
6
5
4
2
0
d , ymA
pa
Figure 3-7 - Cumulative mass versus aerodynamic
particle diameter for Run #12.
60
-------
X
Q
o
II
H
2
U
J5
O
U
C/3
5
dpa, ymA
Figure 3-8 -
Cumulative mass
particle diametp
61
aerodynamic
-------
E
2:
Q
o
II
E-
CJ
2
O
oo
CO
II
H
0
d , ymA
pa
Figure 3-9 - Cumulative mass versus aerodynamic particle
diameter for Run #14.
62
-------
c
CJ
ft
^-^
X
o
II
"
H
w
H
(X
.-- ., .^'. zrT^r^.T^ri--:rT-r:r~~jL^L;^rT-,TTT:A^^.::-:-T--ifilf~±i::iiIu_I£:_.- \V=^"~^~^^" -".-^-^rrfe
0.05 -rw^minP^I
0.01
0.1
d , ymA
pa'
Figure 3-10 - Fractional penetration curves for
data set "A".
63
-------
o
o
II
H
EH
W
2
w
ex,
u
I I
H
0.05 -
0.01
0.1
0.5
1.0
d umA
pa*
2.0
10
Figure 3-11 - Fractional penetration curves for
data set "B".
64
-------
Lines for high and low computations based on different
distribution curve shapes are shown in Figure 3-10 for
Runs 9 and 10.
MATHEMATICAL MODEL
No specific performance model was available for a valve
tray so it was necessary to determine whether something suit-
able could be adapted from the available design methods. It
was hypothesized that the gas jet emerging from the slot be-
tween the valve cap and the tray orifice might give collection
comparable to the round gas jets which emerge from a sieve
plate and that the dependence on foam density might also be
comparable. The fractional collection efficiency, E , for particle
collection by inertial impaction in a sieve plate column is
given in the "Scrubber Handbook" (S. Calvert et al . , 1972)
as
E = 1 - exp [-40 F2 K ] (3-1)
"Scrubber Handbook" (S.H.B.) Eq . 4.6.4-1 and Eq. 4.6.4-3
where foam density is in the range of 0.38 < F < 0.65 and
"K " is the inertial parameter given by:
(S.H.B. Eq. 4.6.2-4)
For extensive discussion on these equations, refer to the
Section 4.6.3 of the "Scrubber Handbook". These equations
are based on sieve plate performance data. The following
data were used to calculate the predicted performance.
Top Tray
70 scrubbing elements on the tray
F = 0.33
u, = 2,000 cm/sec = gas velocity in the slot between
the valve cap and the tray
65
-------
Top Tray (continued)
p = 1.34 g/cm3 = urea particle density
d, = 0.7 cm = the width of the slot
h
Bottom Tray
27 scrubbing elements on the tray
F = 0.4
uh = 5,200 cm/sec
p = 1.34 g/cm3
d, = 0.7 cm
n
It was assumed that no particle growth occurred during
the particle collection in the bottom tray, and that all
particle growth happened between the top and bottom trays.
Figure 3-3 was used to obtain the particle diameter for the
calculation of the top tray penetration.
The results of the calculations were plotted in
Figures 3-12 and 3-13 for data sets A and B, respectively.
Actual penetration obtained by going through the graphical
procedures were also plotted on those figures as a comparison
It appears that the experimental values are slightly
higher than those calculated from the S.H.B. equations
(4.6.4-1) and (4.6.4-3) and the agreement is quite good.
CONCLUSIONS
Uncertainties in some of the particle size measurements
for submicron diameters are sufficiently large so that the
computed penetration for particles of about 0.5 ymA diameter
can vary over a range of 251 or more. A large factor is the
unreliability of the cascade impactor manufacturer's calibra-
tion. The cross-calibration of the two impactor types
used in this test provides for consistency between the two,
but did not verify the calibration of the U. W. impactor.
In later work this was done.
66
-------
E
§
w
0.01
0,1
0.5 1.0
dpa, ymA
2.0
10
Figure 3-12 - Predicted and experimental penetrations
for data set "A".
67
-------
1.0
0.5
2:
o
HH
H
U
§
I-H
a
W
2:
w
OH
w
i-3
U
i i
s
<
DH
0.051
0.01
Figure 3-13 - Predicted and experimental penetrations
for data set "B".
68
-------
As demonstrated by this work, the S.H.B. equations
(4.6.4-1) and (4.6.4-3) may be used to predict the perform-
ance of a 85 Am3/min (3,000 ACFM) two stage valve tray
scrubber. It is also noted that particle growth within the
scrubber significantly increases particle collection effic-
iency. Since the foam density, F, was approximated, an even
better fit of the experimental data can be obtained by vary-
ing the value of "F" to be used in the S.H.B. equation
(4.6.4-1).
69
-------
70
-------
APPENDIX 3-A
PARTICLE SIZE DATA
71
-------
TABLE 3-A-l PARTICLE DATA FOR RUN #1
Impactor
Stage
No.
0
1
2
3
4
5
6
7
Filter
Sample
Volume
(DNm3)
U. w. Mark III
Wcum
Og)
22.6
22.0
22.0
22.0
21.0
13.5
4.0
1.0
PS o
(pmA)
22
9.7
4.5
1.8
1.0
0.52
0.28
0.939
Andersen
Wcum
Og)
21.7
21.3
20.7
20.4
20.1
19.9
18.1
12.9
5.4
d
PS o
OimA)
fMfe.calV
9.6
6.0
4.0
2.78
1.75
0.9
0.54
0.36
0.875
TABLE 3-A-2 PARTICLE DATA FOR RUN #2
Impactor
Stage
'No
IX w
0
1
2
3
4
5
6
7
Filter
Sample
Volume
(DNm3)
U. W. Mark III
cum
Og)
14.2
13.7
13.7
13.7
13.6
9.3
3.0
1.2
d
(pmA)
22
9.7
4.5
1.8
1.0
0.52
0.28
0.640
Andersen
"cum
Og)
14.7
14.6
14.5
14.3
14.3
- 12.4
8.4
3.2
dn
pso
(umA)
(Mfg.cal.)
6.0
4.0
2.78
1.75
0.9
0.54
0.36
0.626
72
-------
TABLE 3-A-3 PARTICLE DATA FOR RUN #3
Impactor
Stage
No.
0
1
2
3
4
5
6
7
Filter
Sample
Volume
(DNm3)
U. W. Mark III
cum
(mg)
16.3
15.8
15.7
15.6
13.0
4.3
0.5
0.3
d
p 50
-CymA)
. 22
9.7
4.5
1.8
1.0
0.52
0.28
0.526
Andersen
Wcum
(mg)
18.5
18.1
17.8
17.4
16.9
16.5
12.8
5.2
1.2
d
Pso
(pmA)
fMfe.cal.}
4.0
2.78
1.75
0.9
0.54
0.36
0.519
TABLE 3-A-4 PARTICLE DATA FOR RUN #4
Impactor
Stage
'No.
0
1
2
3
4
5
6
7
Filter
Sample
Volume
(DNm3)
U. w. Mark III
W
cum
Og)
13.4
13.1
13.0
12.9
10.5
2.4
0.5
0.2
pso
(UmA)
22
9.7
4.5
1.8
1.0
0.52
0.28
0.436
Andersen
W
cum
(nig)
12.3
12.1
12.0
11.9
11.9
11.9
8.0
2.8
0.7
d
Pso
(vimA)
[MfR.cal.)
6.0
4.0
2.78
1.75
0.9
0.54
0.36
0.426
73
-------
TABLE 3-A-5 PARTICLE DATA FOR RUN #5
Impactor
Stage
No.
1
2
3
4
5
6
7
Filter
Stage
Initial
Weight
g
0.1102
0.1509
0.1503
0.1512
0.1563
0.1565
0.1507
0.1386
Stage Final Weight
wet
g
0.1107
0.1514
0.1509
0.1677
0.1973
0.1715
0.1516
0.1392
dry
g
0.1107
0.1512
0.1507
0.1534
0.1613
0.1597
0.1513
0.1392
Stage Cut Size
wet
VimA
22
9.7
4.3
1.8
1.0
0.52
0.28
dry
ymA
8.57
2.25
0.98
0.67
0.49
0.28
Sample
Volume 0.735
(DNm3)
TABLE 3-A-6 PARTICLE DATA FOR RUN #6
Impactor
Stage
No.
1
2
3
4
5
6
7
Filter
Stage
Initial
Weight
g
0.1117
0.1561
0.1575
0.1538
0.1545
0.1569
0.1540
0.1336
Stage Final Weight
wet
g
0.1119
0.1563
0.1577
0.1644
0.1852
0.1769
0.1552
0.1345
dry
g
0.1115
0.1562
0.1576
0.1552
0.1584 "
0.1603
0.1541
0.1345
Stage Cut Size
wet
|imA
22
9.7
4.3
1.8
1.0
0.52
0.28
dry
tiraA
7.76
2.35
1.0
0.62
0.285
0.28
Sample
Volume 0.612
(DNm3)
74
-------
TABLE 3-A-7 PARTICLE DATA FOR RUN #7
Impactor
Stage
No.
1
2
3
4
5
6
7
Filter
Stage
Initial
Weight
g
0.0810
0.1583
0.1620
0.1612
0.1611
0.1652
0.1591
0.1363
Stage Final Weight
wet
g
0.0811
0.1583
0.1522
0.1848
0.2117
0.1753
0.1698
0.1471
dry
g
0.0812
0.1584
0.1621
0.1673
0.1643
0.1674
0.1594
0.1370
Stage Cut Size
wet
ymA
22
9.7
4.5
1.8
1.0
0.52
0.28
dry
ymA
3.1
0.82
0.67
0.25
0.28
Sample
Volume O-55
(DNm3)
TABLE 3-A-8 PARTICLE DATA FOR RUN #8
Impactor
Stage
No.
1
2
3
4
5
6
7
Filter
Sample
Volume
(DNm3)
Stage
Initial
Weight
g
0.0789
0.1679
0.1633
0.1650
0.1654
0.1740
0.1634
0.1479
Stage Final Weight
wet
g
0.0799
0.1680
0.1634
0.1653
0.1691
0.1812
0.1656
0.1434
dry
g
0.0799
0.1679
0.1633
0.1651
0.1687
0.1800
0.1652
0.1632
Stage Cut Size
wet
ymA
22
9.7
4.5
1.8
1.0
0.52
0.28
dry
ymA
22
9.7
4.5
1.8
1.0
0.52
0.28
0.50
75
-------
TABLE 3-A-9 INLET AND OUTLET SAMPLE PARTICLE DATA
FOR SIMULTANEOUS RUN #9
Impactor
Stage
Number
0
1
2
3
4
5
6
7
Filter
Sample
Vo lume
(DNm3)
Type of
Impactor
Inlet
W
cum
(mg)
22.6
22.0
22.0
22.0
21.0
13.5
4.0
1.0
d
pc
(ymA)
22
9.7
4.5
1.8
1.0
0.52
0.28
0.94
u.w.
Outlet
W
cum
Og)
12.8
12.3
12.0
11.6
9.4
4.4
1.2
0.6
V
(ymA)
22
9.7
4.5
1.8
1.0
0.52
0.28
0.73
U.W.
TABLE 3-A-10 INLET AND OUTLET SAMPLE PARTICLE DATA
FOR SIMULTANEOUS RUN #10
Impactor
Stage
Number
0
1
2
3
4
5
6
7
Filter
Sample
Volume
(DNm3)
Type of
Impactor
Inlet
W
cum
(mg)
14.2
13.7
13.7
13.7
13.6
9.3
3.0
1.2
V
(ymA)
22
9.7
4.5
1.8
1.0
0.52
0.28
0.63
U.W.
Outlet
W
cum
(ing)
10.1
9.9
9.8
9.7
8.3
4.4
1.0
0.9
d
pc
(ymA)
22
9.7
4.5
1.8
1.0
0.52
0.28
0.61
U.W.
7.6
-------
TABLE 3-A-ll
INLET AND OUTLET SAMPLE PARTICLE DATA
FOR SIMULTANEOUS RUN #11
Impactor
Stage
Number
0
1
2
3
4
5
6
7
Filter
Sample
Volume
(DNm3)
Type of
Impactor
Inlet
W
cum
Og)
16.5
15.8
15.6
15.1
14.8
14.2
10.9
4.9
1.5
d
pc
(ymA)
2.0
1.5
1.0
0.6
0.65
Andersen
Outlet
W
cum
C^g)
9.1
8.4
7.7
7.2
6.8
5.1
0.8
0.3
V
(pmA)
22
9.7
4.5
1.8
1.0
0.52
0.28
0.60
U.W.
TABLE 3-A-12
INLET AND OUTLET SAMPLE PARTICLE DATA
FOR SIMULTANEOUS RUN #12
Impactor
Stage
Number
0
1
2
3
4
5
6
7
Filter
Sample
Volume
(DNm3)
Type of
Impactor
Inlet
W
cum
Og)
16.3
15.8
15.8
15.8
15.6
15.3
12.2
6.1
1.8
V
(ymA)
2.0
1.5
1.0
0.6
0.66
Andersen
Outlet
W
cum
Og)
6.8
6.6
6.5
6.5
6.5
4.1
1.1
0.3
d
pc
(ymA)
22
9.7
4.5
1.8
1.0
0.52
0.28
0.63
U.W.
77
-------
TABLE 3-A-13
INLET AND OUTLET SAMPLE PARTICLE DATA
FOR SIMULTANEOUS RUN #13
Impactor
Stage
Number
0
1
2
3
4
s
6
7
Filter
Sample
Vo lume
(DNm3)
Type of
Impactor
Inlet
W
cum
(nig)
19.6
19.1
19.1
19.1
19.1
19.1
14.9
5.6
1.4
d
pc
(ymA)
2.0
1.5
1.0
0.6
0.66
Andersen
Outlet
W
cum
Og)
10.9
10.9
10.9
10.8
6.6
1.4
0.4
V
(umA)
22
9.7
4.5
1.8
1.0
0.52
0.28
0.64
U.W.
TABLE 3-A-14
INLET AND OUTLET SAMPLE PARTICLE DATA
FOR SIMULTANEOUS RUN #14
Impactor
Stage
Number
0
1
2
3
4
5
6
7
Filter
Samole
Vo 3 ume
(DNm3)
Type of
Impactor
Inlet
W
cum
(ing)
14.8
14.8
14.8
14.7
14.4
14.3
13.3
7.8
2.6
V
(umA)
2.0
1.5
1.0
0.6
0.53
Andersen
Outlet
W
cum
(mg)
12.5
11.5
11.5
11.5
11.4
8.1
2.1
0.3
d
pc
CymA)
22
9.7
4.5
1.8
1.0
0.52
0.28
0.50
U.W.
78
-------
APPENDIX 3-B
PARTICLE SIZE DISTRIBUTION PLOTS
79
-------
05
w
W
hJ
u
m-^
O Run #la
A Run *lb
ORun
jRun
^Andersen 5
Sampler
2 5 10 20 30 40 50 60 70 80 90
PERCENT BY WEIGHT UNDER SIZE
Figure 3-B-l - Particle size distribution measured by the U.W.
and Andersen Cascade Impactors.
NOTE: The numbering system used here is that "a" denotes
size measured by u. W. and "b" size measured by
Andersen in a simultaneous run (designated by number)
80
-------
0
5
4
-
u
1.0
0.5
0.4
0.3
0.2
A Run #3b
1 2
10 20 30 40 50 60 70 80 90
PERCENT BY WEIGHT UNDER SIZE
98
Figure 3-B-2 - Particle size distribution, measured by the
U. W. and Andersen cascade impactors.
-------
0.2
10
20 30 40 50 60 70 80
90
98
PERCENT BY WEIGHT UNDER SIZE, DRY BASIS
Figure 3-B-3 -
Dry particle size distribution obtained with
in-stack and ex-stack U. W. cascade impactor
82
-------
w
H
i
p
H
CO,
<
PH
0.3
0.2
10 20 30 40 50 60 70 80
PERCENT BY WEIGHT UNDER SIZE
Figure 3-B-4 Particle size distribution for data set A.
83
-------
Q
W
^
U
k-H
E-
<
tl± ! ; i i . I I i 1 I ! : i -T-r-r 1
10 20 30 40 50 60 70 80 90
PERCENT BY WEIGHT UNDER SIZE
98
Figure 3-B-5 - Inlet particle size distribution
(data set B).
84
-------
3.
oi
:
f
Q
m
-
ff ^ff --'-:: -t=r;
0.2
10 20 30 40 50 60 70 80 90
PERCENT BY WEIGHT UNDER SIZE
98
Figure 3-B-6 Outlet particle size distribution (data set B)
85
-------
86
-------
VANED CENTRIFUGAL ON POTASH DRYER
(Ducon Multivane Scrubber)
SOURCE AND SCRUBBER
A Ducon Multivane scrubber was selected for the third
scrubber performance test. This scrubber is designed to
clean the exhaust gas from a rotary drier which removes the
moisture from 22,680 Kg/hour (25 TPH) of potassium chloride
crystals. The rotary drier is gas-fired with oil as a
standby fuel for periods of natural gas shortage.
The scrubber is The Ducon Company's Multivane Scrubber
Size-84 Type-L Model II (Figure 4-1). The scrubber outlet
duct is 106.68 cm in diameter and inlet is a 60.96 cm by
91.44 cm rectangular duct. The scrubber pressure drop varies
from 6.5 to 8.0 cm W.C. (or 2.7" to 3.3" H20). The scrubber
liquor flow rate is 0.12 m3/min (32 GPM) as measured by an
orifice meter on the inlet line of the scrubber liquor cir-
cuit. Liquid is introduced through spray nozzles located be-
tween the wash and eliminator turning vanes inside the scrubber
TEST METHOD
In this performance test, three types of impactors
(Andersen Sampler, University of Washington Mark III and
Brink) were used for particle measurements. Greased aluminum
foil substrates \irere used on each of the collection plates
of the Andersen and U. W. Mark III impactors and filter
papers were used on the Brink collection plates. Substrates
for the impactor plates were cut out of thick aluminum foil.
A 201 solution of silicone vacuum grease in benzene was pre-
pared. Five drops of this solution were placed on the
substrates with an eye dropper. It was then evenly spread
87
-------
GO
Figure 4-1 - Ducon Multivane scrubber.
88
-------
out on the substrates with a policeman, taking care that it
did not spread to the bottom of the substrates. These were
then placed in aluminum foil storage cups and heated in an
oven for two hours at 200°C. Then they were cooled and stored
in a desiccator for about 10 hours. Prior to each run, the
substrates and filter were removed from the desiccator, weigh-
ed with the storage cups and loaded in the impactor.
Both sampling probes in the inlet and outlet ducts were
kept at one position during the entire sampling period. The
location of the impactor was chosen such that the gas velocity
at that location is close to the average gas velocity in the
duct. The sample flow rate was also fixed during runs.
In most of the test runs, the impactor was kept in-stack,
however, in some test runs, the impactor was ex-stack. When-
ever this occurred, the impactor was put in a heated box. In
some runs, the entrainment was heavy, therefore an in-line
miniature glass cyclone was used to prevent entrained liquid
drops from entering the impactor. Details on each run's
impactor operating condition are listed in Table 4-1.
Sample flow rates were measured with the usual EPA Method 5
instruments so as to obtain isokinetic (or near isokinetic)
sampling. Scrubber inlet and outlet gas temperatures were
measured by mercury filled glass bulb thermometers. Gas temp-
erature at the sampling location was measured during each
test run. Stack pressures were measured with a U-tube man-
ometer. Barometric pressures were determined before each run
from an aneroid barometer. Stack gas humidities were deter-
mined by EPA method 4 and by dry and wet bulb thermometer.
Gas volumetric flow rate was calculated from velocity traverse
data obtained by means of a calibrated S-type pitot tube.
A total of 23 impactor test runs and 4 filter runs were
performed. Among those 7 impactor runs were purged due to
89
-------
Table 4-1. IMPACTOR OPERATING CONDITIONS
Run
No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23a*
23b*
Sampling
Location
Outlet
Outlet
Outlet
Outlet
Outlet
Inlet
Inlet
Inlet
Inlet
Inlet
Inlet
Inlet
Inlet
Inlet
Inlet
Inlet
Inlet
Outlet
Outlet
Inlet
Inlet
Outlet
Outlet
Inlet
IMPACTOR
Type
Andersen
Andersen
Andersen
Andersen
Andersen
Andersen
Andersen
Andersen
Andersen
Andersen
Andersen
Andersen
U. of W.
U. of W.
U. of W.
U. of W.
U. of W.
U. of W.
U. of W.
U. of W.
U. of W.
U. of W.
Brink
U. of W.
Location
ex-stack
ex-stack
ex-stack
ex-stack
ex-stack
ex-stack
ex-stack
ex-stack
ex-stack
ex-stack
ex-stack
ex-stack
in-stack
in-stack
in-stack
in-stack
in-stack
in-stack
in-stack
in-stack
in-stack
in-stack
in-stack
in-stack
Stage
Lining
Greased
Aluminum
ri
it
it
M
II
II
M
M
It
II
II
It
II
Tl
II
11
II
tl
M
II
II
Filter
Greased
Aluminum
Heated
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
No
No
No
No
No
No
No
No
No
No
No
Precutter
Used
No
No
No
No
No
No
No
No
No
No
No
No
Yes
Yes
Yes
Yes
Yes
No
No
Yes
Yes
No
No
Yes
Remark
Test Void (broken
probe)
Test purged (nozzle
broken)
Test void
Test purged
Test difficulty
Dryer shut-down
Test aborted
10
o
*NOTE: This is a simultaneous inlet and outlet test run.
-------
various operating difficulties and plant shut-down problems
All filter runs were conducted in-stack.
RESULTS
Scrubber Operating Conditions
There was an interruption of the test program by a
plant shutdown. The scrubber operating conditions during
the test period were as follows:
1. Gas flow rates were different before and after
the interruption. Gas parameters are listed
in the tabulation below:
Gas Parameters
Temperature
Pressure
A m3/min
ACFM
DNmVmin @ 0°C
DSCFM @ 70°F
Vol. % H_0 Vapor
Inlet
196°C (385°F)
2.3 cm W.C.
623
22,000
323
11,200
19
Outlet
78°C (172°F)
504
17,800
322
12,100
19
Gas Parameters (After Interruption)
3as Parameters
Temperature
Pressure
A m3/min
ACFM
DNmVmin @ 0°C
DSCFM @ 70°F
Vol. 1 H20 Vapor
Inlet
204°C (400°F)
1.8 cm W . C .
464.5
16,400
255
9,000
18.0
Outlet
77°C (170°F)
339.8
12,000
220
8,280
18.0
91
-------
2. Pressure drop across the scrubber is 8.0 cm W.G.
3. Liquid flow rate and parameters are as follows:
Liquid Parameter
LIQUID PARAMETERS
Temperature
m3/min
GPM
L/G
Specific gravity
Suspensed solid
Dissolved solid
Treatment
INLET
52°C (125°F)
0.12
32
1.5
1.07
-
-
OUTLET
0.12
32
1.07
-
-
"~
Particle Data
The particle concentration and size data itfhich were ob-
tained in this performance test are tabulated in Appendix 4-A.
Runs #17, 18, 21, 22, 23a and 23b were taken when the drier
and scrubber run at normal operating conditions. Among these
runs, only runs #23a and 23b were simultaneous runs. Figures
4-B-l and 4-B-2 are log-probability plots of inlet and outlet
particle size distribution for these test runs respectively.
The "actual" mass median diameter, geometric standard devia-
tion, and aerodynamic mass median diameter for the outlet
samples are listed in the following table.
RUN NO.
d rr> , urn
p50 '
a
g
d , umA
P§
18
1.4
4.3
2.1
22
2.9
3.9
4.1
23a
0.7
4.1
1.1
92
-------
Cumulative mass concentration was plotted against particle
diameter to yield Figures 4-2 through 4-4.
The rest of the test data were not plotted here because
the primary objectives of those runs were to test the equip-
ment set-up and to determine adequate sampling time.
Particle Penetration
Particle penetration was computed by the general method
described in a previous section of this report. As the first
step in this computation it was necessary to plot cumulative
particle mass versus particle diameter. Figures 4-2, 4-3,
and 4-4 are such plots for three sets of inlet and outlet
runs. Because the cyclone pre-cutter was used on the inlet
runs, the impactor stage weight gains were less than they
would have been without the pre-cutter. Consequently, the
true particle size distribution must be determined by com-
pensating the impactor data for the effect of the pre-cutter.
This compensation was performed on the basis of the approxi-
mations that the cyclone cut diameter was about 2.0 pmA and
that cyclone penetration varies exponentially with (-d* )
The dashed curves on Figures 4-2 through 4-4 are fit
by eye with the compensated data points for the inlet samples
Particle penetrations were computed from the ratio of slopes
of the outlet and inlet cumulative distributions, based on
the curves fit by eye. The penetration results are plotted
in Figure 4-5 for the three pairs of runs, in terms of
actual diameter (density = 2.0) rather than aerodynamic
diameter. The data for run no. 21 show too much scatter to
be useful for more than a general confirmation of the other
runs .
MATHEMATICAL MODEL
Preliminary computations showed that the particle
collection efficiency given by this scrubber could not be
93
-------
300
E
2
Q
E
2
E-
2
o
w
200
100
50
&
, Inlet, Un-Comp-
ensated
pre-cutter
1.0 2.0 3.0
PARTICLE DIAMETER, yra
4.0
Figure 4-2 Cumulative mass versus particle diameter
94
-------
300 r
E
R
o
H
53
W
C
52
-
2
r
u
200
100
50
A #22-, Outlet
, Inlet, Uncompensated /
;'
Q £21, Compensated for /
pre-cutter /
/
1.0 2.0 3.0
PARTICLE DIAMETER, ym
4.0
Figure 4-3 Cumulative mass versus particle diameter
95
-------
E
2
-
CO
C
5
H
2
U
O
u
w
500
400
300
200
100
0
i
I
/
J
-
I
- *-
_£
I ^
/ x^
. - r.p-
v»
;
r/
i
.
.
j A *23b, Outlet
/
-Q -f 23a,- Inlet, Un'-compen";-
sated
.
, ^
O i23a,-'Compensated for..
.pre-cutter
_^__ _- !"~" i~~"!
1.0 2.0 3.0
PARTICLE DIAMETER, ym
4.0
Figure 4-4 - Cumulative mass versus particle diameter
96
-------
1.0
0.5
2
H
W
x
EH
0.1
17 § 18
0.1
0.5 1.0
PARTICLE DIAMETER,
3 4
10
Figure 4-5 - Experimental and predicted
penetrations.
97
-------
accounted for simply by centrifugal deposition caused by the
vanes in the scrubber. Prediction of collection efficiency
based on the assumption that the scrubber mechanism was a
counter-flow spray tower were also too low to fit the data.
We then decided to test the hypothesis that water spray-
ed on top of the vanes tends to collect on the vanes and be
atomized and reentrained by the upward spiraling gas flow.
This mechanism would involve the recirculation of water at
some unknown rate and would provide particle collection
through impaction on the water drops as in a co-current, gas
atomized spray scrubber. Our approach to the computation of
collection efficiency based on this mechanism is given below.
Particle penetration for a gas-atomized scrubber can be
estimated by means of the cut diameter - pressure drop
method (Calvert, 1974). Figure 4-6 is a plot of cut diameter
versus scrubber pressure drop for several scrubber types and
conditions. Figure 4-7 is a plot of the ratio of particle
diameter to cut diameter versus collection efficiency. As
can be seen on Figure 4-6, the cut diameter for a gas-atomized
(venturi) scrubber at 7.6 cm W.C. pressure drop ranges from
2.5 ymA at f = 0.25 to 1.3 ymA at f = 0.5, These diameters
correspond to 1.7 urn and 0.85 ym actual diameter for a part-
icle density of 2.0 g/cm3.
Figure 4-5 indicates that the cut diameter was about
1.2 ym (or about i.g ymA). This corresponds to a value of
f = 0.4, which is typical for wettable particles. Pene-
trations for other particle diameters were computed for an
aerodynamic performance cut diameter, dpc, =1.8 ymA, utilizing
Figure 4-7.
98
-------
--
m
EH
P
i-
-'
u
Sieve Plate
Impingement Plate
Column
cm saddles}
0.1
10 50 100
PRESSURE DROP, cm W.C.
200
Figure 4-6 - Cut diameter - pressure drop correlations
CCalvert, 1974)
99
-------
o
o
3.
u
nl
ft
Lcted tot*
Vtj n't ii r i j '"s cTiibb c r '
0.1 0.5 1 2
10 50
PENETRATION FOR d
90 95 98 99
pa'
Figure 4-7 - Predicted particle diameter - penetration relationship
for inertia-l impaction (Calvert, 19743 .
-------
CONCLUSIONS
The general operation of this scrubber was not hampered
by any substantial problems although entrainment separation
was not very effective. Particle collection efficiency was
what would be expected for a low pressure drop scrubber and
would have to be increased to meet recent air pollution
regulations.
The unit mechanism responsible for particle collection
in this scrubber appears to be collection on drops, rather
than in curved conduits. Penetration can be accounted for
by means of a gas-atomized spray model.
The cyclone pre-cutter which was used on the inlet
samples had too low a cut diameter and it substantially
reduced the impactor stage catches. Consequently, it is
not possible to compute penetrations for particles larger
than about 1.5 ym diameter with much accuracy. The
experience of this test showed the advantage of using a
pre-cutter with a cut diameter greater than 5.0 ymA.
101
-------
102
-------
APPENDIX 4 - A
PARTICLE DATA
103
-------
RUN #1 (Outlet)
Impactor
Stage No .
Precutter
0
1
2
3
4
5
6
7
Filter
Sample
Volume
(DNm3)
Type of
Impactor
IV
cum
(mg)
33 .6
33.6
26.5
17.2
10.6
7.7
d *
pc
(urn)
4.7
3.3
2.1
1.1
0.65
0.43
0 .460
Andersen
RUN #3 (Outlet)
Impactor
Stage No.
Precutter
0
1
2
3
4
5
6
7
Filter
Sample
Volume
(DNm3)
Type of
Impactor
W
cum
(rag)
' 88.4
88.0
87.4
85.3
83.1
' 68.6
49.4
30.2
21.7
V
(vim)
11
7.0
4.7
3.3
. 2.1
1.1
0.65
0.43
0.868
Andersen
*NOTE: Particle diameters were computed from the aerodynamic
cut sizes based on a particle density of 2.0 g/cm3 and
the appropriate C'.
RUN #4 (Outlet) RUN #5 (Outlet)
[mpactor
Stage No.
Precutter
0
1
2
3
4
5
6
7
Filter
Sample
Volume
(DNm3)
Type of
Impactor
W
cum
(mg)
98.2
97.3
96.8
96.1
92.5
76
49.7
35.2
27.2
V
Um)
11
7.0
4.7
3.3
2.1
1.1
0.65
0.43
1.020
Andersen
Impactor
Stage No .
Precutter
0
1
2
3
4 -
5
6
7
Filter
Sample
volume
(DNm3)
[Type of
jlmpactor
W
cum
(mg)
69.1
68.9
68.8
67.5
63.9
52
36.5
25.7
20
d
pc
(ym)
11.0
7.0
4.7
3.3
2.1
1.1
0.65
0.43
0.481
Andersen
104
-------
RUN #8 (Inlet)
RUN #9 (Inlet)
Impactor
Stage No .
Precutter
0
1
2
3
4
5
6
7
Filter
Sample
Volume
(DNm3)
Type of
Impactor
IV
cum
Og)
22.2
15.5
11.2
5.5
2.8
1.8
1.4
V
(urn)
10.2
6.1
3.7 '
2.2
1.1
0.65
0.0144
Andersen
Impactor
Stage No .
Precutter
0
1
2
3
4
5
6
7
Filter
Sample
Volume
(DNm3)
Type of
Impactor
W
cum
Og)
22.8
17.8
11.3
9.8
6.5
3.4
1.3
d
pc
(ym)
12.4
7.9
5.3
3.7
2.4
1.3
0.0141
Andersen
RUN #11 (Inlet)
RUN #12 (Inlet)
[mpactor
Stage No.
Precutter
0
1
2
3
4
5
6
7
Filter
Sample
Volume
(DNm3)
Type of
Impactor
W
cum
Og)
77.8
75.7
74.5
72.7
10.8
6.8
4
3.4
.3.0
d
pc
(ym)
11
7.0
4.7
3.3
2.1
1.1
0.65
0.43
0.0568
Andersen
Impactor
Stage No .
Precutter
0
1
2
3
4
5
6
7
Filter
Sample
Volume
(DNm3)
Type of
Impactor
W
cum
Og)
20.4
20.0
18.4
17.0
14.7
10.3
6.1
4.6
3.9
d
pc
(ym)
11
7.0
4.7
3.3
2.1
1.1
0.65
0.43
0.033
Andersen
105
-------
RUX #15 (Inlet)
RUN #16 (Inlet)
Inpactor
Stage No .
Precutter
0
1
2
3
4
5
6
7
Filter
Sample
Vo lume
(DNm3)
Type of
Impactor
W
cum
(mg)
2949.3
200.6
188.2
187.5
125.4
51.9
35.8
27.5
15.5
V
(urn)
16.5
7.23
3.35 '
1.32
0.73
0.37
0.2
0.595
U. W. Mark III
Impactor
Stage No.
Precutter
0
1
2
3
4
5
6
7
Filter
Sample
Volume
(DNm3)
Type of
Impactor
W
cum
(nig)
1503.8
78.9
76.2
74.6
47.3
18.9
13.3
12.4
3.4
d
pc
(urn)
16.5
7.23
3.35
1.32
0.73
0.37
0.2
0.42
U.W. Mark III
RUX #17 (Inlet
RUN #18 (Outlet)
Impactor
Stage No.
Precutter
0
1
2
3
4
5
6
7
Filter
Sample
Volume
(DXm3)
Type of
Impactor
W
cum
(nig)
2669.4
84.6
79.9
78.7
61.3
36.5
23.2
15.7
9.2
d
pc
(urn)
16.5
7.23
3.35
1.32
0.73
0.37
0.2
0.179
U. W. Mark III
Impactor
Stage No .
Precutter
0
1
2
3
4 '
5
6
7
Filter
Sample
Volume
(DNm3)
Type of
Impactor
W
cum
(n\g)
117.7
112.7
105.7
82.1
51.1
39.4
24.2
12.7
d
pc
(ym)
18.1
7.85
3.67
1.45
0.80
0.41
0.22
*-*
0.336
*
U. W. Mark III
'
106
-------
RUN #19 (Outlet)
RUN #21 (Inlet)
Impactor
Stage No .
Precutter
0
1
2
3
4
5
6
7
Filter
Sample
Volume
(DNm3)
Type of
Impactor
W
cum
(ing)
133.3
119.5
104.3
75.6
44.1
29.4
21.1
11.8
V
(ym)
17.67
7.65
3.53 -
1.42
0.78
0.397
0.21
0.343
U. W. Mark III
Impactor
Stage No.
Precutter
0
1
2
3
4
5
6
7
Filter
Sample
Volume
(DNm3)
Type of
Impactor
W
cum
(nig)
3225.1
136.8
104.3
99.5
59.1
21.2
17.8
13.9
8.7
V
(ym)
16.8
7.36
3.44
1.35
0.73
0.37
0.2
0.368
U. W. Mark III
RUN #22 (Outlet)
RUN #23a (Outlet)
Impactor
Stage No.
Precutter
0
1
2
3
4
5
6
7
Filter
L "
Sample
Volume
(DNm3)
Type of
Impactor
W
cum
(mg)
65.6
60.9
53.4
35.6
17
9.6
6.5
6.0
d
pc
(ym)
17.1
7.5
3.47
1.38
0.756
0.387
0.208
0.280
U. W. Mark III
Impactor
Stage No .
Precutter
0
1
2
3
4
5
6
7
Filter
Sample
Volume
(DNm3)
Type of
Impactor
W
cum
(mg)
17.5
13.3
10.8
8.4
7.2
4.7
d
pc
(ym)
2.21
1.29
0.87
0.45
0.188
0.04
Brink
107
-------
RUN #23b (Inlet)
Impactor
Stage No .
Precutter
0
1
2
3
4
5
6
7
Filter
Sample
Vo lume
(D.Mra3)
Type of
Impactor
W
cum
(mg)
2388.8
148.8
145.3
141.3
99.1
60.4
42.2
31.7
18.8
V
(yra)
19.90
8.72
4.03 '
1.61
0.90
0.46
0.253
0.211
U. W. Mark III
108
-------
APPENDIX 4-B
PARTICLE SIZE DISTRIBUTION PLOTS
109
-------
II
ii
o
-:
U
r-
0.2
0.1
0.1 0.512 5 10 20
PERCENT BY WEIGHT UNDER SIZE
Figure 4-B-l - Inlet particle size distribution
110
-------
10
5
4
=
-
w
E-H
W
w
t-H
u
2;
1.0
0.6
0.5
0.4
0.3
0.2
0.1
- - -~ - - - .--
10 20 30 40 50 60 70 80
PERCENT BY WEIGHT UNDER SIZE
90 95
Figure 4-B-2 . Outlet particle size distribution.
Ill
-------
112
-------
MOBILE BED ON COAL-FIRED BOILER
(T.C.A. Scrubber)
SOURCE AND SCRUBBER
A model 6,700 TCA (Turbulent Contact Absorber) scrubber
designed by UOP, Air Correction Division, was the subject of
the fourth performance test. This type of scrubber utilizes
mobile (fluidized) beds of 3.8 cm (1.5") diameter polypro-
pylene spheres as the three contacting stages. (See Figure
5-1). The scrubber is equipped with a chevron mist elimina-
tor made of fiberglass reinforced plastic.
This system was installed to clean the exhaust gas from
an electrostatic precipitator used to control the particulate
emission from a 165 M.W. utility steam boiler. The boiler
is a Babcock § Wilcox Radiant Boiler (Built in 1961) with a
steam capacity of 517,560 kg/hr at design pressure of
/- «
1.51x10 kg/m^ and steam temperature of 540.6°C. An analysis
of the coal burned in the boiler is given in Table 5-A-l.
The boiler flue gas passes from the boiler to the elec-
trostatic precipitator, through 2 fans in parallel, and then
to a presaturator spray inside the scrubber body. The gas
is cooled in the presaturator from about 143.0°C (290°F) to
about 57.0°C (135°F). From the presaturator stage the gas
passes through three parallel scrubber compartments, each of
which has a series of 3 stages of fluidized balls. The
compartments divide the gas stream in the proportions of
20%-60%-20%.
After the scrubber, the gas passes through a chevron
type entrainment separator and then to a reheater. Pressure
drop through the scrubber is about 30.0 cm W.C. (12"),
5.0 cm (2") through the entrainment separator, and 9.4 cm
113
-------
GAS our,-ET
RECISCULATIOH
OUTLET HOZZtE
Figure 5-l~ Mobile Bed Scrubber
114
-------
(3.7") through the reheater. The gas is reheated to about
85.0°C (185°F) in order to provide buoyancy and to prevent
condensation in the stack.
The scrubber has four parallel inlet ducts (two into
each fan) and three parallel outlet ducts. The inlet ducts
are all 3.37 m x 1.03 m (132.5"x40.5") rectangles and the
center outlet duct is 4.58 m x 2.29 m (15'x7.5'). Figure
5-2 shows these ducts and their sampling point locations.
The inlet duct west of center and the center outlet duct
were used for sampling.
Test Method
The most essential part of the performance test was
the determination of particle size distribution and
concentration (loading) in the inlet and outlet of the
scrubber. A modified E.P.A. Method 5 train with an
in-stack University of Washington (or Pilat) cascade
impactor was used for particle measurements. Gas flow
rate was determined by means of type "S" pitot tube
traverses along with the necessary temperature and pressure
measurements. Sample flows were measured with the usual
E.P.A. train instruments so as to obtain isokinetic sampling.
Two series of tests were made; one during July, 1973
and the second from September 10 through September 14, 1973.
Plant problems caused an eventual shut-down and abortion of
the first series of tests. The second series of tests
consisted of three inlet and three outlet samples, which
were taken at different times (i.e., not simultaneous
inlet-outlet pairs). While not ideal, the taking of separate
inlet and outlet samples appeared acceptable in view of
steady plant operation and the fairly consistent data which
were obtained, based on preliminary computations.
115
-------
FOUR INLET DUCTS
THREE OUTLET DUCTS
(NOTE: Grids show sampling areas)
Figure 5-2 - Duct arrangements.
116
-------
The inlet sampling point was located upstream of the
fan and the pre-saturator. The impactor was kept at one
position during the entire sampling period. Because the
particle concentration was so low (due to the electro-
static precipitator upstream) it was not necessary to use
a pre-cutter ahead of the U. W. impactor. Outlet samples
had to be taken after the reheater because of sample port
location, so it was not possible to make any measurements
of liquid entrainment and it was not necessary to heat
the impactor.
Possible sample bias due to particle inertia effects
(segregation) \\ras not very significant in this test for
two reasons. First, any errors in the sampling of large
particles will not affect penetration for the fine particles.
Second, the particles entering the scrubber have been pre-
cleaned by the electrostatic precipitator and thus are
fairly small, except for the large particles reentrained
during rapping.
Operating Conditions
The scrubber operating conditions during the test
period were as follows:
1. Gas flow rate computed as 4 times the rate
measured in one duct was 18,000 A m3/min (630,000
ACFM) at about 143.0°C (290°F), 61.5 cm Hg (24.2"Hg)
pressure and 5% H20 vapor. The flow rate computed
as 1.67 times the 10,000 A m3/min (360,000 ACFM)
measured at the center scrubber compartment outlet
was 17,000 A m3/min (600,000 ACFM) at about 85°C
(185°F), 64 cm Hg (25.2"Hg), and 181 H20 vapor.
Data provided by the power plant personnel from
previous tests were 545,000 ACFM at 272°F (at
150 MW load) and 640,000 ACFM (at 165 MW). These
117
-------
are in good agreement with the data from this
test. Gas flow rate through the scrubber is
9,000 DN m3/min @ 0°C, 76.0 cm Hg (or 340,000 DSCFM
8 70°F, 14.7 psia).
2. Slurry flow rate to the scrubber was reported by
the power plant as approximately 113.0 m3/min
(30,000 GPM). Makeup water is introduced into
the pre-saturator at a rate of 1.44 m3/min (380 GPM] .
The total pre-saturator spray rate was around
6.8 m3/min (1,800 GPM).
3. Gas velocity in the scrubber is usually maintained
at 2.8 m/sec (9.2 ft/sec) ±20% in order to keep
the balls fluidized.
4. Entrainment could not be measured but is known to
be excessive because it causes plugging of the
gas pre-heater.
PARTICLE DATA
The data obtained on particle concentration and size
are presented in Tables 5-A-2 and 5-A-3. Size distributions for
these runs are shown in Figures 5-B-l and 5-B-2, log-probability
plots of the inlet and outlet data. As shown on the plots,
the inlet particles have a mass median diameter, d of
pg'
about 3.0 ymA and a geometric standard deviation, a , of
about 2.5. The outlet particles have d =0.5 umA and
pg . H
°g - 6-°-
Cumulative mass concentration was plotted against aero-
dynamic size to yield Figures 5-3 and 5-4. The solid curves
are for the third degree polynomials fit by the least squares
method. The dashed lines are the curves fitted by eye. The
polynomial fit for the outlet samples is obviously unrealistic
above 2 pmA particle diameter, especially if one plots all of
118
-------
160
E
p
e
W
C
_
i
EH
<
_
u
urve Fit
3rd Order
Polynomial Fit
O Run #5
A Run #6
i Run #7
Figure 5-3 - Inlet cumulative mass concentration
size distribution.
119
-------
20.0
Q
be
Pi
H
CO
to
15.0
.
by Eye
7 -0~
Polynomial Fit
A
Run #2
"
A
:
10.0 .
1.0
2.0
pa
3.0
4.0
Figure 5-4 - Outlet cumulative mass concentration si~e
distribution.
120
-------
the points for larger sizes. If the points for larger sizes
were included in the least squares regression, however, the
fit would be poor at small particle diameters. Since we are
most concerned with penetration for fine particles, the
option for a better fit at the small end was taken.
Particle Penetration
Particle penetration was computed by the following two
methods:
1. The third order polynomials describing the inlet
and outlet cumulative concentration distributions
were differentiated and the ratio of first deriva-
tives with respect to particle diameter was
computed at several values of diameter. The ratio
of outlet to inlet derivatives is, as discussed in
the section on the computation method, the pene-
tration at that "particle diameter.
2. The slopes of the eyeball fit curves in Figures 5-3
and 5-4 were measured by a graphical technique at
several values of particle diameter. The ratios of
outlet to inlet slopes were computed to yield pene-
trations at the several diameter values.
Penetrations are computed by the two methods are plotted
against particle diameter (aerodynamic) in Figure 5-5. It
can be seen that there is a slight discrepancy between the
two methods at diameters above 1.0 ymA. As discussed pre-
viously, this is an obvious consequence of the curve-fitting
computation and can be readily compensated for.
ECONOMICS AND OPERATING PROBLEMS
Points of information on economics and operating
problems for the scrubber system are listed below.
1. The approximate installed cost of the scrubber
system, for particle removal only, is $3,900,000;
121
-------
1.0
E-
U
E-
W
0,
W
y 0.05
1
0.01
0.1
Figure 5-5- Particle penetration versus aerodynamic
particle diameter for T.C.A. scrubber.
122
-------
or $23.60/KW. It was estimated that the addition
of S02 control features would cost about $10.00/KW
more. The scrubber alone accounts for about 101 of
the equipment cost. Power consumption for the
scrubber system is about 4% of the gross output of
the plant.
2. Scaling has not been much of a problem, due to
ash properties.
3. The entrainment separators (mist eliminators) have
not been sufficiently effective. Two horizontal
layers of zigzag baffles, containing 4 -90° turns
in each, were used.
4. Frequent plugging of the reheater has been caused
by carry-over from the entrainment separator.
5. Ball wear in the scrubber has been rapid and
various materials have been tried.
6. Suitable materials of construction for scrubber
internals are 316 L stainless steel or rubber
lining. Types 304 and 308 stainless are not satis-
factory. Pumps are rubber lined. The stack is
acid proof lined and has a double wall for insula-
tion.
7. Dry fans are preferred because of problems with
wet fans following inefficient entrainment
separators.
MATHEMATICAL MODELS
A major objective of the scrubber performance test pro-
gram is the validation and/or further development of mathe-
matical models which can be used for the prediction of
performance. The data on particle penetration as a function
of particle size and scrubber parameters have not been
123
-------
available prior to this time. Where a design model has been
presented before, our first approach is to determine whether
the available model(s) fit the data. If this does not prove
to be the case, then it will be necessary to develop a model
\vhich works properly.
The only model we know of for particle collection in a
T.C.A. is the semi-empirical relationship presented by
Bechtel Corp. in a June, 1971 report on the Shawnee project
for E.P.A. and cited by Calvert et al. (1972)
(0.3 6
rii = 1 - exp
- 2.18xlO-18
(5-1)
where, n- = collection efficiency for particle diameter "d."
L = Liquid rate, kg/hr-m2
PT = Liquid density, kg/ft3
lj
G = Gas rate, kg/hr-m2
PG = Gas density, kg/ft3
K. = Inertial impaction parameter for "d.", average
gas velocity through bed void space, and ball
diameter as the collector diameter.
Z = Static bed depth, cm
D, = Ball diameter, cm
This correlation is of very dubious value because it is
based on the premise that collection efficiency is due to
inertial impaction on the balls. We may note that the im-
paction parameter has a value of about SxlO'1* for a gas
velocity of 10 ft/sec, ball diameter of 1.5 inches, and aero-
dynamic particle diameter of 1.0 ymA. The collection effic-
iency for a sphere is 0% for values of the impaction parameter
smaller than about 0.1; consequently it is impossible to attri-
bute high collection efficiency to this mechanism. Collection
efficiency due to flow through the curved passages between the
balls would be comparably low.
124
-------
An attempt to explain the observed penetrations by
treating the stages of fluidized balls as sieve (or froth)
plates was also unsuccessful. In order to have a
penetration of 0.1 for three sieve plates in series, the
inertial parameter based on perforation diameter would have
to be about 0.77 (for foam density =0.7). If the aerody-
namic particle diameter is 1.0 ymA, corresponding to Pt = 0.1,
the ratio of velocity through the perforation to perforation
diameter would be (for standard air properties):
K
u,
- h = 0.77
9 (1.8x10-")
* 125,000 (sec -1) (5-2)
Even if the effective perforation diameter, d^, were
0.5 cm (0.2 in.), which is smaller than seems probable, the
gas velocity would have to be about 62,000 cm/sec (2,000
ft/sec), which is not possible.
The attempted rationalization of the observed T.C.A.
performance in terms of either a counter-current or co-
current gas-atomized spray scrubber was also not fruitful.
In both of these cases the cut diameter predicted was much
larger than observed.
Observation of a 30 cm (1 ft) diameter mobile bed column
in operation revealed that the balls near the column wall
.move downward. Therefore, there must be channeling in the
bed such that the balls in the middle move upward. Based on
this clue, computations of collection efficiency for a co-
current spray were made for a gas velocity higher than the
average superficial velocity in order to allow for gas flow
channeling. Assumptions which were explored are as follows:
1. Gas velocity is 2 times the average superficial
xrelocity.
125
-------
2. Gas velocity is 4 times the average in order to
account for bed porosity of about 50% and gas flow
channeling.
3. Liquid (drop) flow rate within the bed is that
which would cause from 1/3 to 2/3 of the bed
pressure drop of about 10 cm W.C. per stage. The
remainder of the pressure drop would be due to
static head.
4. Drop size is determined by gas atomization of the
liquid.
5. Drop shatter within the mobile bed causes drop size
to be smaller than that from gas atomization.
None of the above assumptions, alone or in combination,
would account for the observed scrubber performance. The
predicted penetration of 1.0 ymA diameter particles ranged
mainly from 0.85 to 0.9 at 3.0 cm W.C. assumed for liquid
acceleration and from about 0.7 to 0.85 at 6 cm W.C. Three
stages would result in predicted penetrations ranging from
0.61 to 0.73 for 3.0 cm W.C. and 0.34 to 0.61 for 6.0 cm W.C.
The experimentally observed penetration at 1.0 ymA was about
0.1 for three stages, which would require a single stage
penetration of 0.46, or less.
If one compares the observed mobile bed performance
(i.e., a cut diameter of about 0.4 ymA at 25 cm W.C. pressure
drop) with other types, as shown in Figure 4-6, there is
an obvious descrepancy. One would expect that a scrubber
utilizing inertial impaction only would require 200-400 cm
W.C. pressure drop to provide a 0.4 ymA cut diameter.
At least a partial explanation of the high efficiency
lies in the fact that SO,, and some H-SO, were present in the
flue gas. Any H2S04 (or S03) which adsorbs on the fly ash
particles will cause the condensation of water on the
particles. This will occur even when the relative humidity
126
-------
is considerably lower than 100%. The consequent growth of
the particles in the saturated scrubber atmosphere will
cause them to be collected at higher efficiency than the
dry particles. In the absence of any data on particle
collection efficiency for non-hygroscopic particles in a
mobile bed, we are unable to evaluate the relative import-
ance of condensation and other mechanisms.
Another factor to consider is that the mobile bed
follows an electrostatic precipitator and there may be a
particle charge effect. Ho\^ever, there is evidence that
this is not a significant factor. For one thing, the E.P.A.
tests of the mobile bed scrubber at the Shawnee Plant, which
does not follo\vT an electrostatic precipitator, show pene-
trations comparable to those found in the present test. For
another thing, Public Service Co. of Colorado has found that
there is no difference between mobile bed performances when
the units follow cyclones rather than electrostatic precipi-
tators.
CONCLUSIONS
The data obtained for particle penetration as a function
of particle size will provide a useful and important basis
for the development of a realistic mathematical model and
design method. There is scatter in the data and it is obvious
that additional data for simultaneous inlet and outlet samples
would be very important in providing a more precise basis for
design method development. The scrubber reliability has not
been good, although the plant personnel felt that continual
progress is being made toward its improvement. We strongly
recommend that additional performance tests be made on mobile
bed scrubbers and that these include further investigations
of operating problems and methods of coping with them.
127
-------
Further ivork must be done to establish a rational
mathematical model for particle collection in a mobile
bed. It would be best to start from some reliable data
on a system free of condensation effects.
128
-------
APPENDIX 5-A
PARTICLE AND COAL DATA
129
-------
Table 5-A-l. COAL ANALYSES
PROXIMATE ANALYSIS
Moisture
Ash
Volatile
Fixed Carbon
Btu
Sulfur
As Received
9.51
8.99
36.08
45.42
100.00
11028
0.43
Dry Basis
XXXX
9.93
39.87
50.20
100 .W
12187
0.48
ULTIMATE ANALYSIS
-Moisture
Carbon
Hydrogen
Nitrogen
Chlorine
Sulfur
Ash
Oxygen (diff)
As Received
9.51
63.08
4.53
1.38
0.00
0.43
8.99
12.08
100.00
Dry Basis
XXXX
69.71
5.01
1.52
0.00
0.48
9,93
13.35
100.00
130
-------
Table 5-A-2. INLET SAMPLE PARTICLE DATA
u.w.
Stage
1
2
3
4
5
6
7
Filter
Sample
Volume
(DN m3)
RUN #5 j -RUN £6
W *
cum
(ing)
38.0
33.3
31.1
23.2
9.0
3.6
1.6
1.5
d **
pc
(yraA)
18.0
8.3
3.9
1.5
0.83
0.42
0.2
0.168
W
cum
(nig)
29.7
27.6
25.7
20.1
6.9
1.6
0.3
0.3
V
(ymA)
18.0
8.0
3.7
1.4
0.7
0.4
0.2.
0.135
RUN #7
W
cum
fag)
35.7
34.7
31.7
21.9
5.9
0.9
0.1
0.1
d
pc
(pmA)
19.0
8.2
3.8
1.5
0.8
0.4
0.2
0.13
NOTES:
* W
cum
** d
pc
umA
Cumulative mass collected on that stage
and those below.
Cut diameter (aerodynamic) for that
stage .
microns, aerodynamic = d (C'p ) l
131
-------
Table 5-A-3. OUTLET SAMPLE PARTICLE DATA
u.w.
Stage
i
i
1
2
3
4
5
6
i
7
Filter
Sample
Volume
(DN m3)
RUN #2
W
cum
(mg)
30.5
24.1
24.1
24.1
22.1
16.0
12.6
10.9
d
pc
(ymA)
14.0
6.2
3.0
1.2
0.62
0.33
0.2
2.23
' RUN #3
W
cum
Og)
28.8
28.2
28.1
27.6
24.2
16.4
12.0
10.1
d
PC
CvmA)
14.0
6.2
3.0
1.2
0.62
0.33
0.2
2.12
RUN #4
W
cum
Og)
24.2
23.3
22.2
21.2
19.5
14.7
10.5
8.3
d
pc
(yiuA)
20.0
8.6
4.0
1.6
0.9
0.45
0.25
1.1
132
-------
APPENDIX 5-B
PARTICLE SIZE DISTRIBUTION PLOTS
133
-------
10.0
5.0
4.0
3.0
2.0
1.0
:~
_ ::j±ti
L - L- _ -
i- - 3.0 -urn A
P%
2.5
0.5
0.4
0.3
0.2
0.1
0
Run J6
0.5 1 2 5 10 20 30 40 50 60 70 80
MASS PERCENT UNDERSIZE
Figure 5-B-l - Inlet particle size distribution
134
-------
10.0
5.0
4.0
3.0
2.0
a
1.0
0.5
0.4.
0.3
0.2
0.1
A
d K 0 .5 ym A
PE
a « 6.0
10 20 30 40 50 60 70 80 90 95 99
MASS PERCENT UNDERSIZE
Figure 5-B-2 - Outlet particle size distribution
135
-------
136
-------
VENTURI SCRUBBER ON COAL-FIRED BOILER
(Chemico Venturi)
SOURCE AND SCRUBBER
A Chemico "Venturi" scrubber operating on the flue gas
from a coal-fired utility boiler was chosen for test no. 5.
The scrubber might more accurately be described as a variable
annular orifice type because its throat is formed by a movable
"plumb bob" concentrically mounted in a conical "dental bowl",
as shown in Figure 6-1.
A chevron type mist eliminator, washed with sprays, is
mounted within the scrubber body, as shown in Figure 6-1.
Three scrubbers in parallel handle the-flue gas from the one
boiler. The boiler is a 330 M.W. net, 360 M.W. gross Com-
bustion Engineering unit fired with low sulfur western coal.
The coal analysis varied from day-to-day, as sho\vn in Table
6-A-l. Sulfur content varied from 0.34% to 0.75% (as received)
during the months of September and October, when our test
team was at the plant. While no deliberate effort was made
to control S02 emissions, the calcium oxide content of the
fly ash normally varies from 10% to 26% and there is as much
as 40% removal of S02 by the scrubber with the pH running
around 5.0 to 6.0.
The induced draft fan following the scrubber has a water
spray at its inlet to wash off the solids carried over from
the mist eliminator. Disengagement of the mist from the fan
occurs in the stack, which has an epoxy lining. Reheating
is not used after the fan.
Additional information on the scrubber system is as
follows:
1. Pressure drop (gas phase) was 25 cm K.C. (10" W.C.)
during the test period.
137
-------
ONE-Sf AGE VENTURI
(CKEMtCO)
fUKKtl
FlOV.5 BOB SHAFT
MNCIWIW HUNS
G/.S ouiin
SPRM
KOI CAS IMU
SP8AV W022U 8«Retl
|FC« PLUMB GOEI
-OEKIAL BOWL"
"HUMS BOB"
KI5TMWRAIMS
Figure 6-1 - Chemico Venturi
138
-------
2. Liquid flow rate to the scrubber is 24.6 m3/min
(6,500 G.P.M.), of which about 381 is return from
a pond (and treatment with lime) and the remainder
is recycled from the scrubber bottom.
3. Makeup water is introduced via the fan sprays.
4. The venturi throat is stainless steel and the vessel
is epoxy lined.
There were two interruptions of the testing program. A
major disruption occured in September when the boiler was
shut down for repairs, due to plugging of the air preheater.
This caused a one month postponement and the loss of the
data collected during the first period. A minor data loss
occured when the boiler had to be operated at reduced load
during one test run.
TEST METHOD
Determination of particle size distribution and concentra-
tion (loading) in the inlet and outlet of the scrubber provides
the basis for computation of performance characteristics. A
modified E.P.A. type sampling train equipped with an in-stack
University of Washington (Pilat) cascade impactor was used
for the inlet and a similar train with an in-stack Brink
cascade impactor for the outlet. The U.W. impactor was used
on the inlet (at about 1/4 the usual sampling rate) instead
of the Brink in order to provide more dust collection capacity
for the heavy load of grit which was encountered. Ordinarily
the Brink ivould have been used following a cyclone pre-cutter,
but in this case the cyclone had broken during an early test.
Both cascade impactors were allowed to heat up to stack
temperature before the sample was taken and the outlet sampler
was also heated with an electric resistance wrapping. The
filters following the impactors were in-stack and loaded with
139
-------
Gelman type "E" glass fiber paper. The impactor stages were
covered with greased aluminum foils which were treated and
weighed in accordance with our usual procedure (as i^ere the
filters).
Isokinetic (or near-isokinetic) sampling was used, with
the sampler being held at one position in the duct. This is
generally an adequate technique for obtaining good samples
of particles smaller than a few microns in diameter because
they are well distributed across the duct. It does not
provide a good sample of the large particles when the nozzle
inlet is close to a flow disturbance; as in the case of the
inlet sample, which was taken downstream from a butterfly
valve. Thus, the total scrubber inlet loading is uncertain
because of the one position sample but the inlet fine
particle concentration is representative of the entire gas
stream.
Gas velocities in the ducts were measured by means of
type "S" pitot tube traverses, along with the necessary
temperature and pressure measurements. Sample flows were
measured with a dry gas meter and an orifice meter.
Several independent inlet and outlet samples were taken
by means of both the cascade impactors and a total filter
until there was consistency between the two methods of
measurement. A series of four simultaneous inlet-outlet
tests were then made and one of these iiras discarded, as
discussed previously. The inlet sample was taken in a
3.7 m (12 ft) diameter duct and the outlet sample in a
3.7 m x 4.4 m (12 ft x 14.5 ft) rectangular duct between
the scrubber and the I.D. fan.
140
-------
OPERATING CONDITIONS
The scrubber operating conditions during the test
period were as follows:
1. Gas flow rates were as shown in the tabulation below:
Duct Inlet Outlet
Temperature 163.0°C(325°F) 54.0°C(130°F)
Pressure during pitot run 60.0 cm Hg 60.0 cm Hg
A mVmin 13,400 12,700
ACFM 4.75xl05 4.5xl05
DN m3/min 6,300 7,150
DSCFM 2.4X105 2.7xl05
% HO vapor (vol.) 6-0% 15.01
Li
The flow rate measured by the outlet velocity traverse
is judged to be more reliable because the velocity
distribution was much more regular than at the inlet.
Based on 7,120DNm3/min (2.7xl05DSCFM), the inlet flow rate
would be 15,300Am3/min (5.4xl05ACFM), which is 81 higher
than the design flow rate of 14,200Am3/min (5xl05ACFM).
2. Slurry flow rate to the scrubber was reported by the
plant as approximately 24.6 m3/min (6,500 GPM) .
3. Entrainment is known to occur between the scrubber
and the fan but was not measured in this test series.
PARTICLE DATA
The particle concentration and size data which were
obtained in this performance test are presented in Tables
6-A-2, 6-A-3, and 6-A-4 for the three simultaneous inlet and outlet
samples. Figure 6-B-l shows log-probability plots of inlet
and outlet particle size distributions and Figure 6-B-2 is a
similar plot for the large diameter end of the inlet distribu-
tion. The inlet particles have a mass median diameter, dpg,
141
-------
of about 38.0 ymA and a geometric standard deviation, a , of
about 5.0, while the outlet particles have d =0.15 ymA and
r »
a * 4.6.
g
Cumulative mass concentration was plotted against aero-
dynamic particle size to yield Figure 6-2 and 6-3, for inlet
and outlet samples, respectively. The solid lines are for the
third degree polynomials which were fit by the least squares
method. It is obvious that the inflection of the curves
between about 1.5 and 2.5 umA is not physically realistic and
that lines with continuously positive slopes are to be expect-
ed. The fits at smaller particle diameters are better, how-
ever, and this is the more crucial region in view of our
primary interest in fine particles.
Particle Penetration
Particle penetration was computed by the following two
methods:
1. The third order polynomials describing the inlet
and outlet cumulative concentration distributions
were differentiated and the ratio of first deriva-
tives with respect to particle diameter was
computed at several values of diameter. The ratio
of outlet to inlet derivatives is, as discussed in
the section on the computation method, the penetra-
tion at that particle diameter.
2. The slopes of eyeball fit curves for the points in
Figures 6-2 and 6-3 were measured by a graphical
technique at several values of particle diameter.
The ratios of outlet to inlet slopes were computed
to yield penetrations at the several diameter
values.
142
-------
300F
-
Z
C
_
U
pq
>
hI
H
200 ^
:
--j -^';-'7^ ;;~
~V ;:^rzr^
1 _ , _ _ .. _~~
. i- « i 'ii--~---f-~r=-'i i -
100 ^3:
0.5
1.0
2.0
3.0
pa
Figure 6-2 ' Inlet cumulative mass concentration
distribution.
143
-------
O
hH
H
w
u
LO
150
100
d (ymA)
pa
Figure 6-3 Outlet cumulative mass concentration
distribution.
144
-------
Penetrations computed by the two methods are plotted
against particle diameter (aerodynamic) in Figure 6-4. It
can be seen that the two methods are in close agreement below
about 1.0 ymA particle diameter. We are more inclined to
trust the penetrations based on the eyeball fits; especially
those for runs #1 and #2. The unbelievably high plateau for
the run #4 dashed curve is due to the high concentration at
2.4 ymA, as Figure 6-3 shows.
ECONOMICS AND OPERATING PROBLEMS
Several points relating to the economics and operating
problems for the scrubber system are listed below:
1. The initial capital cost of the scrubber system
including mechanical equipment, stack and erection
costs but excluding ash ponds, development costs
since initial operation, environmental monitoring
and Owners' cost \vas $8,247,600. Subsequent
development work and modifications have increased
this amount significantly.
2. Maintenance labor was estimated to require about
four men on three shifts. Operating labor requires
less than one man per shift.
3. The boiler had been kept on line in recent months
at about 65% load factor and plant personnel were
hopeful that this would continue to improve.
4. There have been serious problems due to solids
accumulation on the scrubber above the plumb bob,
in the scrubber bottom (from material falling from
higher points), and on the fans. Deficiencies in
the fan wash system during initial stages of opera-
tion resulted in fan buildup problems. However the
wash system was modified in June 1973 and problems
with buildup on the fan since that time have been
minimal.
145
-------
1.0
o
II
E-
w
Sulid jlliret
are byi
Figure 6-4 - Particle penetration versus
aerodynamic diameter.
146
-------
During the period the tests were conducted, lime
was added to the pond return at a rate of about
100 Ib/hr. However, this proved unsuccessful as
a means of reducing scale formation. As a result,
lime addition to the pond return was discontinued
and lime is now added to each vessel at rates of
up to 1000 Ib per hour per vessel in an attempt to
reduce scaling. Thus far, there has been only
limited success in scale reduction through lime
injection. The relative benefits of lime addition
are still under investigation.
Operating experience has indicated that entrainment
separation has been satisfactory although dust does
collect in the demisters. However, this is attribu-
table to problems with the demister wash system and
the scaling tendency of the system; not to poor
entrainment separation. The ID fans have to be run
with \tfash water to prevent deposits and unbalancing.
The need for wet fans was anticipated and included
in the original design. However, entrainment is not
necessarily responsible for the dust buildup on the
fans. The plant management believe it is likely due
to penetration of the scrubber by fine particles and
subsequent deposition on the fan due to compression
and condensation of the water vapor in the gas stream
as it passes through the fan. Problems encountered
with buildup on the fan blades during the initial
stages of operation (prior to June of 1973) were
corrected by modifying the original fan wash system.
Experience with the fan wash system has shown that
spray nozzles are necessary and are being used to
introduce the fan wash water into the fan inlet
gases.
147
-------
7. The slurry pumps last about 1.5 years, pumping 2%
solids at about 2 or 3 atm. (30-40 psi) pressure.
8. The scrubber system uses about 1.71 of the gross
power (i.e., about 6 M.W.).
VENTURI MODEL
Venturi scrubber performance for the conditions of this
test was predicted by means of the method described in the
"Scrubber Handbook". Because gas velocity in the throat is
not known, it was necessary to compute a throat velocity
from the liquid to gas ratio (1.75 i/m3) and the pressure
drop. The use of SHE equation (5.3.6-10) gives a velocity
of 38 m/sec (125 ft/sec) but this will be lower than actual
because the equation predicts pressure drop about 15% high.
Consequently, the throat velocity would probably be around
42 m/sec (138 ft/sec), corresponding to a predicted pressure
drop of 30 cm W.C.
The design equation (SHB 5.3.6-5) Includes an empirical
constant, f, which has a value of about 0.25 for hydrophobic
materials and about 0.5 for hydrophyllic. Penetrations were
predicted for 1.75 £/m3, 42 m/sec; with f = 0.25, f = 0.4,
£ = 0.45 and f = 0.5 and some of the results are tabulated
below and plotted on Figure 6-5:
PREDICTED CUT DIAMETER FOR
COMBINATIONS OF "f" AND "AP"
AP = 20 cm W.C.
AP = 200 cm W.C.
V at * =
0.25
1 .5 ymA
0.52 ymA
0.4
1.1 ymA
0.37 ymA
0.5
0.95 ymA
0.29 ymA
148
-------
1.0
O
H
u
<
&
H,
H
pq
PL,
W
0.01
Figure 6-5 - Predicted and experimental penetrations
for venturi.
149
-------
PREDICTED* PENETRATION AT SEVERAL RATIOS
OF PARTICLE TO CUT DIAMETER
Pt
dpa/dpc
0.99
0.3
0.95
0.44
0.90
0.54
0.7
0.78
0.5
1.0
0.2
1.4
0.1
1.6
0.05
1.9
0.01
2.4
*for 1.75 £/m3
Figure 6-5 is an overlay of two prediction lines on Figure
6-4, which presented the experimental data. It can be seen
that the prediction for f = 0.4 has a cut diameter of 0.9 umA
and the curve is generally higher than the experimental
results. Apparently a value of f = 0.5 is about right for
the experimental data because it yields a cut diameter of
about 0.7 ymA.
CONCLUSIONS
Particle penetration data based on the measurements
made in this test appear to be reliable and the agreement
among the three runs is fairly good. The.venturi scrubber
performance is good while it is running but operating
problems have caused the scrubber system to be inoperative
about 1/3 of the time. Solids accumulation has been the
major cause of unreliability.
Sampling apparatus failures and deficiencies which
were experienced in this test have led to our subsequent
development of improvements.
Considering all of the uncertainties in the model and
the experimental data, the agreement between experiment and
prediction is good. It is reasonable that the fly ash acts
like a hydrophyllic material because of the presence of
sulfur oxides and the consequent high wettability of the
particle surface. Because the contact time in the venturi
is short and the particles do not have much opportunity for
growth before entering the collection zone, one would not ex
pect flux force/condensation effects to be very pronounced.
150
-------
APPENDIX 6-A
PARTICLE AND COAL DATA
151
-------
Table 6-A-l. COAL ANALYSIS (AS RECEIVED)
Day
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Moisture
27.8
27.7
28.4
28.4
27.6
27.6
27.8
29.0
28.2
26.0
34.7
29.6
29.3
28.5
NO SAMPLE
29.0
Volatile
Matter
34.2
31.1
31.6
31.1
. 30.5
29.9
29.7
31.2
31.3
33.2
30.1
31.3
31.3
31.8
31.1
Fixed
Carbon
29.2
29.3
26.5
33.5
32.8
31.3
32.1
33.4
32.2
27.6
25.2
30.0
29.3
29.6
30.6
Ash
8.8
11.9
13.5
6.9
9.1
11.2
10.4
6.4
8.3
13.1
10.0
9.1
10.1
10.1
9.3
kcal/kg
4212
4081
4048
4252
4150
4116
4150
4270
4664
4148
3993
4200
4168
4343
4304
Sulfur
.37
.37
.41
152
-------
Table 6-A-2 .
INLET AND OUTLET SAMPLE PARTICLE DATA
FOR RUN #1
INLET
W *
cum
Og)
562.9
348.0
148.5
68.9
27.1
10.0
5.1
4.8
d **
pc
(ymA)
27.0
12 .0
5.7
2.2
1.2
0.63
0.34
Filter
Sample
Volume n 1Q
(DN m3) U-1
OUTLET
W
cum
(mg)
5.8
5.3
5.1
4.8
4.5
3.9
V
(ymA)
2.4
1.45
0.9
0.5
0.36
Filter
0.051
NOTES:
* W
cum
**d
pc
Cumulative mass collected on that stage
and those below.
Cut diameter (aerodynamic) for that
stage.
153
-------
Table 6-A-3.
INLET AND OUTLET SAMPLE PARTICLE DATA
FOR RUN #2
INLET
Wcum
Og)
421.0
240.0
103.0
56.9
27.7
13.5
4.9
3.9
d
pc
(ymA)
33.0
15.0
7.0
2.8
1.5
0.8
0.44
Filter
Sample
Volume 0.11
(DN m3)
OUTLET
W
cum
Os)
3.6
3.5
3.5
3.2
3.0
2.0
d
pc
(ymA)
2.5
1.5
0.95
0.52
0.38
Filter
0.067
Table 6-A-4.
INLET AND OUTLET SAMPLE PARTICLE DATA
FOR RUN #4
INLET
W
cum
Og)
548.0
226.0
108.0
54.8
26.6
12.5
5.0
3.5
d
pc
CumA)
34.0
15.0
7.0
2.8
1.6
0.8
0.45
Filter
Sample
Volume 0.11
(DN m3)
OUTLET
W
cum
Og)
10.8
10.6
10.1
9.7
9.1
8.2
d
PC
(ymA)
2.4
1.5
0.9
0.5
0.36
Filter
0.072
154
-------
APPENDIX 6-B
PARTICLE SIZE DISTRIBUTION PLOTS
155
-------
10.0
5
4,
3
cd
ft
j
0
: : ;_: .
OUTLETS
> 0.!l5vlymA-;
j_u.
0.5 1 2
10 20 30 40 50 60 70 80 90 95 98 99
MASS PERCENT UNDERSIZE
Figure 6-B-l -
Inlet and outlet particle size distributions
(log-probability) .
156
-------
3-
*'
ci : « 3SfpmA:
-
_. -^ i
i
_
A RUN' #2 '
GRUN s;4 :
ml
12 5 10 20 30 40 50 60 70 80 90 95
MASS PERCENT UNDERSIZE
Figure 6-B-2 Inlet particle size distribution
(log-probability)
157
-------
158
-------
WETTED FIBROUS FILTER ON SALT DRYER
SOURCE AND SCRUBBER
A 1,700 m3/min wetted fiber scrubber designed by Encont
Corporation was chosen for the sixth scrubber performance
test.
The scrubber is installed to clean exhaust gas from
two salt dryers, one K.D. Mill and one Vacuum Mill device.
The larger size dryer dries 45 tons per hour common salts.
This dryer is equipped with one Sturtevant Planovane size
60, 382 m3/min (13,500 CFM) @ 4.8 cm W.C. and 735 RPM
exhauster. The dried salts are supplied to the K.D. Mill
system. The smaller salt drier dries 15 TPH salts and
is equipped with Sturtevant Planovane size 50, 140 m3/mn
(5,000 CFM) @ 5.1 cm W.C., 600 RPM exhauster. The dried
salts are supplied to the Vacuum Mill equipment. The
exhaust gas from this equipment is collected by means of
tunnel collectors and supplied to the scrubber. There is one
induce draft fan at the outlet of the scrubber. The fan
was made by Buffalo Forge Co. It is model 890H-36 rated
at 1,440 m3/min (51,000 CFM) § 41 cm W.C., 1,420 RPM with
dish control damper.
This scrubber has two filter pads, each containing
three layers of plastic filter medium in its inlet duct.
Following these is one entrainment separator, which is
made from the same type of materials as the inlet filter
pad, see Figure 7-1. Originally, this scrubber was equipped
with two layers of entrainment separators, but later the
scrubber user learned that the efficiency was higher with
one pad.
159
-------
ELEVATION
GAS IN
r
1
! ,
I
i
j
i
i
!
1
|
i
i
1
!
i
!
! i i
r i .
1 ! !
! S !
1 !'
! ' I
i i i
j i
-L
OUT
PLAN
Figure 7-1 - Schematic Diagram of Wet Fiber Scrubber
160
-------
Water is sprayed onto the filter media by means of
18 nozzles. The flow rate of the water is approximately
0.38 m3/rain (100 GPM) under a pressure of 1.76 Kg/cm2 (25 psi).
The sprayed water is collected at the bottom of the scrubber
and resupplied to the spray nozzles. Due to entrainment,
0.019 - 0.038 m3/min (5-10 GPM) of water is added to the
scrubber to make up the losses.
TEST METHOD
Determination of particle size distribution and concentra-
tion (loading) in the inlet and outlet of the scrubber provides
the basis for computation of performance characteristics.
A modified EPA sampling train with in-stack University of
Washington Mark III cascade impactor was used for
particle measurements. Greased aluminum foil substrates
were used on each of the collection plates of the impactors.
47 mm Gelman type "A" binderless glass fiber filters were
used in the impactor as backup filters.
Substrates for the impactor plates were cut out of
thick aluminum foil. A 201 solution of silicone vacuum grease
in benzene was prepared. Drops of this solution were
placed on the substrates with an eye dropper. It was then
evenly spread out on the substrates with a policeman,
taking care that it did not spread to the bottom of the
substrates. These were then placed in aluminum foil
storage cups and heated in an oven for two hours at 200°C.
Then they were cooled and stored in a desiccator for
about 10 hours. Prior to each run, the substrates and
filter were removed from the desiccator, weighed with the
storage cups and loaded in the impactors.
161
-------
Both impactors in the inlet and outlet ducts were
kept at one position during the entire sampling period.
The location of the impactor was chosen such that the
gas velocity at that location is close to the average
gas velocity in the duct. The sample flow rate was also
fixed during runs.
An in-line precutter was used on the outlet runs
to prevent entrained liquid drops from entering the
impactor. The impactors were heated above the stack
gas temperature with heating tapes. A thermocouple was
placed in the inlet sampling probe downstream of the
impactor. The impactor heating was controlled with a
variac such that the thermocouple reading was about
10°C above the stack gas temperature. Sample flow rates
were measured with the usual EPA train instrument so as
to obtain isokinetic (or near isokinetic) sampling.
Scrubber inlet and outlet gas duct temperatures
were measured by mercury filled glass bulb thermometers.
Gas temperature at the sampling location was measured
during each test run. The inlet and outlet stack
pressures were measured with a U-tube manometer. Barometric
pressures were determined before each run from an anaroid
barometer. The stack gas humidities were measured by dry
and wet bulb thermometers. Gas flow rate was determined
by means of a calibrated S-type pitot tube traverse.
The inlet gas flow rate was determined from 49 point
pitot tube traverses in the 0.8 m x 1.8 m duct. The outlet
gas flow rate was determined similarly from 48 point pitot
traverses in the 1.1 m x 1.3 m duct.
The scrubber liquid temperature was measured at the
recirculating pump with a mercury filled glass bulb
thermometer. The inlet liquid line pressure was measured
162
-------
with a pressure gauge. The flow rate was then determined
from the spray nozzle characteristics as reported by the
manufacturer.
OPERATING CONDITIONS
The scrubber operating conditions during the test
period were as follows:
1. Gas flow rates and conditions were as shown in
the tabulation below:
Gas Parameters
Temperature
Pressure during pitot run
k m3/min
A.CFM
DN m3/min
DSCFM
% Vol. H20 vapor
Inlet
38°C (100°F)
-11 cm WG
1,590
56,300
1,360
50,190
5.8
Outlet
32°C (90°F)
-30 cm WG
1,630
57,400
1,410
52,550
4.71
2. Liquid parameters were listed in the following
table.
Liquid Parameters
Temperature
Pressure
m3/min
GPM
Suspend Solids
Dissolved Solids
Treatment
Inlet
32°C
1.76 Kg/cm3
0.38
100
None
Outlet
32°C
0.35
90
None
Makeup
15°C
_ _.
0.019 - 0.038
10
-
163
-------
3. Entrainment is known to be excessive but was
not measured in this test series. Water balance
data provided by the plant indicate an entrainment
flow rate out of the scrubber of about 0.019
- 0.038 m3/min (5-10- GPM) .
PARTICLE DATA
A total of 10 simultaneous sampling runs were
conducted. Runs #1 and #2 were purged due to severe
entrainment problems. The remaining 8 runs were grouped
into two data sets (Runs #3-5 as set "A" and Runs #6-10
as set "B") corresponding to different impactor locations
in duct cross-section.
The particle concentration and size data which were
obtained in this performance test are presented in
Tables 7-A-l through 7-A-8. Figures 7-B-l and 7-B-2 show log
probability plots of inlet and outlet particle size
distributions for data sets A and B respectively. There
are some variations in particle sizes. This is mainly
due to the unsteady nature of the milling and drying
processes. The mass median diameter and geometric
standard deviation for these sampling runs are listed in
the following table.
Run
No.
3
4
5
6
7
8
9
10
INLET
V umA
5.2
2.05
4.25
3.7
10
10
10
10
ag
3.2
2.1
2.2
2.4
4.8
4.8
4.8
4.8
OUTLET
d , ymA
Pg'
0.23
0.31
0.34
0.46
0.46
0.46
0.96
1.4
ag
2.3
2.4
3.8
2.8
2.8
2.8
2.3
1.8
164
-------
Cumulative mass concentration was plotted against
aerodynamic particle size to yield Figures 7-C-l to 7-C-8.
PARTICLE PENETRATION
The ability of a scrubber to control particulate
emissions is interpreted in terms of "grade efficiency"
curves, which are plots of particle collection efficiency,
or particle penetration versus particle diameter.
The eyeball judgement was used here rather than the least
squares method because we could not find a simple function
that would fit the experimental data. The slopes of the
eyeball fit curves in Figures 7-C-l to 7-C-8 were measured by
a graphical technique at several values of particle diameter.
The ratios of outlet to inlet slopes were computed to
yield penetrations at the several diameter values. The
results were plotted in Figures 7-2 and 7-3.
ECONOMICS AND OPERATING PROBLEMS
The approximate installation cost of the scrubber,
including the costs of blower and duct work is $60,000. The
annual power cost is approximately $13,000 and the annual
maintenance cost is estimated at $1,000.
There are no unusual operating problems.
MATHEMATICAL MODEL
A method of performance prediction for a dry filter bed
was presented in the "Scrubber Handbook: (Calvert et al 1972)
In the following treatment, as recommended in the Handbook,
we will assume that collection efficiency for the dry fibers
is not affected by the presence of washing water.
The "Scrubber Handbook" (S.H.B.) gave the following
equation for the prediction of penetration of a bed of clean
fibrous packing on particles of a specified size.
165
-------
- I-.-)..,. i V. J. «
Ites
. :ft:-;:ri rbiri:::h4=^:^t---
;iu^-_|.-:-- .:..TX71IL'AT^r 'V a- "' [ r '^-'r~ 'r ' i i" .','.' -T-rr-
:
na
pa
Figure 7-2 - Penetration versus particle diameter (data set "A")
166
-------
l.Q
c
u
2
_
~
2
m
Z;
~
w
u
dpa, ymA
Figure 7-3 - Penetration versus particle diameter (data set "B")
167
-------
Pt = 1 - E = exp(-nsS) (7-1)
(S.H.B. Eq. 3.4-1)
where, S is the solidarity factor of the filter bed and
n is the effective collection efficiency of a single
fiber by all collection mechanisms.
Based on the fiber pad sample obtained from the scrubber
user, the filter pad solidarity factor, S, was estimated to
be 1. There were six layers of this pad in the inlet filter
bed. Therefore, the total solidarity factor for the filter
bed was 6.
In this scrubber, impaction was the most important
collection mechanism, so that we assumed impaction was the
only unit mechanism occuring in the filter bed.
The fibers in the filter pad were ellipsoid in shape
with longer axis normal to direction of gas flow. Its
collection efficiency should lie somewhere between the
collection efficiencies of a ribbon and a cylinder.
Penetration was predicted for ribbon and cylinder with
U =1.8 m/sec (undisturbed upstream air velocity) and the
results were plotted on Figures 7-4 and 7-5. Experimental
results were also plotted on these figures. It can be seen
that the penetration for a ribbon fiber has a cut diameter of
1 umA and is close to the experimental average of around
0.75 ymA. For cylinder fiber, the predicted cut diameter was
about 1.5 umA which was two times larger than the experimental
value.
Particle size data presented in Tables 7-A-l through
7-A-8 were for dry particles because impactors are heated. The
theoretical prediction was accordingly based on dry particles.
However, in actual scrubber operation, particles were wet
and common salt particles were highly hygroscopic. According
to Junge (1963), the radius of the particle will increase
to about five times that of the dry salt particle at
168
-------
1.0
~
u
-
o
'
H
<<
rt
EH
n-
-_
fX
pq
nJ
U
Cylinder fibe.r
0.01 0.2 0.3 0.4 0.5 1.0
d ymA
pa,
345
10
Figure 7-4 - Predicted and experimental penetrations
for fiber filter bed (data set "A")
169
-------
o
I-H
&
w
h-5
u
hH
FiTi'er; . 3
-
I I I
0.2 0.3 0.4 0.5
d ymA
pa'
Figure 7-5 - Predicted and experimental penetration
for fiber filter bed (data set "B")
170
-------
high humidity and about double at 751 relative humidity. The
predicted cut diameters are about two times higher than the
measured ones (for cylinders). This seems consistent
\tfith Junge's prediction for NaCl particles.
It is reasonable the predicted cut diameter will have
a lower value when we take into account other collection
mechanisms. Even if we ignore these mechanisms and based
on impaction alone, the agreement between experiment and
prediction is good.
CONCLUSIONS
The wetted fiber scrubber performs satisfactorily
for this application and presents no substantial operating
problems. Particle collection is enhanced by growth due
to the condensation of water. Performance prediction
by means of the mathematical model for fibrous filters is
satisfactory if particle growth is taken into account.
Better prediction of penetration would be possible if one
could predict particle growth with more accuracy for
hygroscopic materials.
171
-------
172
-------
APPENDIX 7-A
PARTICLE DATA
173
-------
TABLE 7-A-l - INLET AND OUTLET SAMPLE PARTICLE DATA FOR RUN #3
IMPACTOR
STAGE
NUMBER
1
2
3
4
5
6
7
Filter
INLET
Wcum
Og)
19.0
14.7
14.1
10.6
1.6
1.4
0.9
0.7
V
(yinA)
27
11.8
5.55
2.2
1.22
0.64
0.36
Sample 7.889
Volume
(DNm3)
OUTLET
w
cum
Og)
4.2
4.1
3.6
3.2
V
(ymA)
1.18
0.63
0.35
20.247
TABLE 7-A-2 - INLET AND OUTLET SAMPLE PARTICLE DATA FOR RUN #4
IMPACTOR
STAGE
NUMBER
1
2
3
4
5
6
7
Filter
INLET
cum
Og)
10.8
10.8
10.7
9.9
4.4
2.2
1.1
0.8
V
(ymA)
26.5
11.5
5.4
2.17
1.2
0.63
0.36
Sample 8.689
Volume
(DNm3)
OUTLET
W
cum
. Og)
-
5.0
4.9
4.8
4.0
3.3
V.
(ymA)
2.35
1.3
0.7
0.39
20.7
174
-------
TABLE 7-A-3 - INLET AND OUTLET SAMPLE PARTICLE DATA FOR RUN #5
IMPACTOR
STAGE
NUMBER
1
2
3
4
5
6
7
Filter
INLET
Wcum
(ing)
15.2
14.6
13.6
9.8
2.9
0.8
0.4
0.3
V
(ymA)
26.5
11.5
5.6
2.2
1.2
0.64
0.35
Sample 8.76
Volume
(DNm3)
OUTLET
W
cum
Og)
2.1
2.0
2.0
1.7
1.4
1.4
V
(ymA)
5.8
2.35
1.3
0.7
0.38
20.17
TABLE 7-A-4 - INLET AND OUTLET SAMPLE PARTICLE DATA FOR RUN #6
IMPACTOR
STAGE
NUMBER
1
2
3
4
5
6
7
Filter
INLET
Wcum
(mg)
17.2
16.3
15.8
12.5
4.2
1.3
1.1
1.0
dpc
(ymA)
26.5
11.8
5.5
2.2
1.23
0.64
0.36
Sample 11.21
Volume
(DNm3)
OUTLET
W
cum
(mg)
2.9
2.7
2.5
1.8
1.5
V
(ymA)
2.25
1.25
0.66
0.37
21.67
175
-------
TABLE 7-A-5 - INLET AND OUTLET SAMPLE PARTICLE DATA FOR RUN #7
IMPACTOR
STAGE
NUMBER
1
2
3
4
5
6
7
Filter
INLET
Wcum
Og)
36.9
18.5
18.1
15.6
7.0
2.9
1.3
1.0
V
. OmA)
26.8
11.7
5.5
2.2
1.22
0.64
0.36
Sample 10.329
Volume
(DNm3)
OUTLET
W
cum
Og)
S.I
5;0"
5.0
4.9
4.4
3.0
2.4
dpc
CumA)
12
5.7
2.25
1.28
0.67
0.38
21.32
TABLE 7-A-6 - INLET AND OUTLET SAMPLE PARTICLE DATA FOR RUN #8
IMPACTOR
STAGE
NUMBER
1
2
3
4
5
6
7
Filter
INLET
cum
Og)
40.9
24.6
22.4
17.3
6.4
2.4
0.8
0.6
V
(pmA)
26.5
11.6
5.5 '
2.18
1.22
0.64
0.36
Sample 7.781
Volume
(DNm3)
OUTLET
W
cum
. Og)
-
7.3
6.9
5.7
4.6
3.6
V.
CumA)
2.25
1.28
0.67
0.38
21.146
176
-------
TABLE 7-A-7 - INLET AND OUTLET SAMPLE PARTICLE DATA FOR RUN #9
IMPACTOR
STAGE
NUMBER
1
2
3
4
5
6
7
Filter
INLET
Wcum
Og)
23
14.7
13.8
8.4
3.7
1.2
0.3
0
dpc
(ywA)
26.5
11.7
5.45
2.2
1.22
0.64
0.355
Sample 7-824
Volume
(DNm3)
OUTLET
W
cum
Og)
5.7
4.9
3.1
1.9
0.9
PC
(ymA)
2.25
1.25
0.67
0.37
21.76
TABLE 7-A-8 - INLET AND OUTLET SAMPLE PARTICLE DATA FOR RUN #10
IMPACTOR
STAGE
NUMBER
1
2
3
4
5
6
7
Filter
INLET
cum
Og)
- 42
27.3
24.3
16
8.9
4.2
1.2
0.3
V
CymA)
26.8
11.7
5.5
2.2
1.22
0.64
0.36
Sample 7-88
Volume
rnMrn3! _
OUTLET
W
cum
. Og)
8.6
8.5
6.8
3.4
0.9
0
V
CymA)
5.7
2.25
1.28
0.67
0.38
21.40
177
-------
178
-------
APPENDIX 7-B
PARTICLE SIZE DISTRIBUTION PLOTS
179
-------
30
20
10
0.5
0.2
LU. ._Lt, ii -';-.,I
UtM. " v*r=
*
i . [ ^ i . J 1 1^ 1,, i j*-; J ^^ - --- : -T- -
10
30 40 50 60 70 80
MASS PERCENT UNDERSIZE
90
Figure 7-B-l
180
-------
mm^
10
20 30 40 50 60 70 80 90
MASS PERCENT UNDERSIZE
95
Figure 7-B-2 _ Inlet and outlet particle size distribution
(data set "B")
181
-------
182
-------
APPENDIX 7-C
CUMULATIVE MASS DISTRIBUTIONS
183
-------
m
O
f-l
X
«n
g
o
bO
2
O
100 F
co
CO
I-H
u
E-
0 2 4 6 8 10 12 14 16
PARTICLE DIAMETER, pmA
Figure 7-C-l - Cumulative mass distribution for Run #3
c
H
6
P
feO
o
,
EH
CJ
a
o
CO
CO
PO
E-
W
184
-------
x
e
2
P
o
II
e-
o
u
CO
pq
u
pq
2 4 6 8 10 12 14
PARTICLE DIAMETER, ymA
16
Figure 7-C-2 Cumulative mass distribution for Run #4
185
-------
I
Q
2
O
w
u
2
O
u
w
u
H
W
6 8 10 12
PARTICLE DIAMETER, ymA
14 16
Figure 7-C-3 - Cumulative mass distribution for Run #5
186
-------
2 4 6 8 10 12 14 16
PARTICLE DIAMETER, ymA
Figure 7-C-4 - Cumulative mass distribution for Run #6
187
-------
2 4 6 8 10 12 14 16
PARTICLE DIAMETER,
c
r
x
PI
e
bo
*i
O
Ii
E-i
H
pq
O
u
s
,-q
:=>
s
£
u
H
W
:
-
c
Figure 7-C-5 - Cumulative mass distribution for Run #7
188
-------
140-
120
c
§ 100
00
H
p
CO
80
60
-
-
~> 40
-
20
-r_x....|.,..u , . | 4 ; 1.1 . ( --t^H^' :
X
t
2
GO
t
O
M
E-i
W
U
>
O
u
Cfl
OT
f-
H
J
U
H
pq
H
O
0 2 4 6 8 10 12 14 16
PARTICLE DIAMETER, umA
Figure 7-C-6 - Cumulative mass distribution for Run #8
189
-------
1001. -
X
= ^F~. 4
2
I
Q
bfl
o
h-
H
EH
S3
W
CJ
2;
c
u
CO
5
u
W
0 O
0 2 4 6 8 10 12 14 16
PARTICLE DIAMETER, yraA
Figure 7-C-7 - Cumulative mass distribution for Run #9
190
-------
120
100
6
x
n
to
o
:
H
H
2:
W
o
u
I I
EH
u
H
pq
I
c
M
r^
H
W
O
u
CO
pq
>
M
H
U
H
m
o
0 2 4 6 8 10 12 14 16
PARTICLE DIAMETER, ymA
Figure 7-C-8 - Cumulative mass distribution for Run #10
191
-------
192
-------
IMPINGEMENT PLATE TEST
(Impinjet)
SOURCE AND SCRUBBER
The impingement plate scrubber selected for the seventh
performance test was an Impinjet wet scrubber. This scrubber
was installed to control the emission from a gas fired rotary
salt dryer. The gas emitted from the dryer contains common
salt particulates and combustion by-products (carbon monox-
ide, methane, etc.).
The scrubber (see Figure 8-1) was designed and manu-
factured by W. M. Sly Manufacturing Company for a maximum
gas capacity of 230 m3/min (8,100 CFM) at 121°C (250°F).
In actual operation, it treats 141 DNm3/min (5,380 DSCFM) of
gas which has particulate loadings about 0.0036 Kg/DNm3.
Additional information on the scrubber system (see
Figure 8-2) is as follows:
1. Gas emitted from the dryer is supplied to the
scrubber by means of a 40 HP fan.
2. Water flow rate to the scrubber is 0.035 m3/min
(9 GPM) at 20 psig sprayed onto the bottom of the
first impingement stage and 0.038 m3/min (10 GPM)
at free flow to the second stage.
3. Pressure drop (gas phase) was 30 cm W.C. (12"W.C.)
during the test period.
TEST METHOD
The performance of the scrubber is determined by
analyzing the particle size distribution, mass loading of
the inlet and outlet gas sample. Therefore, the most
essential part of the performance test was the determination
of particle size and mass loading distribution.
193
-------
4
FLANGES
I ;
: '0 .64cm-
FLANGES
MIST
M.E.
PLATE WATERS!
10 GPM AT
FREE FLOW
137.:?. cm I'.Q'.
r-*
7.6
oc
INLET ^
230m3/min ^
@121°C
O
' ''^,
SPRAYS
i
ACCESS
- H *-
7i o
. / \i . g
cm ».
SUPPORT
r .
w
-o
cm
Figure 8-1 - Two stage No. 245 Sly Impingjet Wet Scrubber
Shell - 0.035 cm
194
-------
35.6 cm dia.
-137 cm-*-
6
CM
Water Inlet
Water 35.6 cm dia.
Rotary
Gas
Water Out
Figure 8-2 - Schematic Diagram o£ Scrubber System
195
-------
A modified E.P.A. Method 5 train with an in-stack
University of Washington (or Pilat) cascade impactor was
used for particle measurements. Gas flow rate was deter-
mined by means of type "S" pitot tube traverses along with
the necessary temperature and pressure measurements. Sample
flows were measured with the usual E.P.A. train instruments
so as to obtain isokinetic or near isokinetic sampling.
Both inlet and outlet sampling impactors were kept at
one position during the entire sampling period. The inlet
impactor was allowed to heat up to stack temperature before
the sample was taken and the outlet impactor was electrically
heated. The total filter following the impactor were in-
stack in the inlet sampling and ex-stack in the outlet
sampling. The ex-stack filter was heated with an electric
resistance wrapping to prevent condensation. The impactor
stages were covered with greased aluminum foils which were
treated and weighed in accordance with our usual procedures.
The back-up total filter was Gelman type "E" glass fiber
paper. Due to liquid entrainment, a pre-cutter was neces-
sary to be used ahead of the impactor in the outlet.
Two independent sampling data sets (several independent
simultaneous inlet and outlet sample runs in each set) were
obtained. These two sets were taken at different locations
across the duct cross-section. Both inlet and outlet ducts
were 35.6 cm (14") in diameter.
SCRUBBER OPERATING CONDITIONS
The scrubber operating conditions during the test period
were as follows:
1. Gas parameters were as shown in the following
tabulation:
196
-------
Gas Parameters
Temperature
Pressure during
pitot run
A m3/min
ACFM
DNm3/rain
DSCFM
Vol.% H20 Vapor
Inlet
85°C
31.3 cm W.G.
238
8,400
169
6,416
14
Outlet
38°C
2.2 cm W.G.
263
9,300
220
8,377
7.2
2. Liquid parameters were as shown in the tabulation
below:
Liquid
Parameters
Temperature
Pressure (Kg/m2)
m3/min
GPM
Suspended Solids
Dissolved Solids
Treatment
Inlet
Bottom Plate
14,000
0.035
9
-
-
Top Plate
Free Flow
0.038
10
-
-
*
Outlet
-
0.073
19
-
-
3. Liquid entrainment was not measured in this test
series although it is known to be excessive.
PARTICLE DATA
Sampling data which were obtained in this performance
test are presented in Tables 8-A-l through 8-A-6 for the six
simultaneous inlet and outlet samples. Runs 1, 2, 3 and 4
197
-------
were taken at the same location of the duct cross-section
and was designated data set "A". The remaining two runs
were sampled at different locations and were grouped into
data set "B". Run #6 was purged due to leakage in sampling
lines.
Figures 8-B-l through 8-B-4 shows log-probability plots of
inlet and outlet particle size distribution. The mass median
diameter and geometric standard deviation are not revealed
by these figures because most of the particles are big
particles (larger than 20 ymA).
Cumulative mass concentration was plotted against
aerodynamic particle size to yield mass loading distribution
curves (Figures 8-C-l through 8-C-6). In some of these figures,
the outlet curve crosses the inlet curve. This may be caused
by breakdown of large particles. Another possibility is
particle growth due to condensation. The gas is cooled from
82°C (180°F) to 38°C (100°F) by water.
PARTICLE PENETRATION
Particle penetration was computed by taking the ratio
of the outlet to the inlet cumulative concentration distri-
bution slopes at various particle diameters. The slope can
either be obtained by a graphical technique or by fitting
the data with a mathematical function and then calculating
the slope analytically. The first approach is used here.
Penetrations were computed for each simultaneous run
and the results are plotted in Figures 8-3 and 8-4.
ECONOMICS AND OPERATING PROBLEMS
The scrubber's original purchase cost was $6,700 (June,
1969) and the operating costs are estimated at $100 per year.
The operating costs consist of po\ver required to pump
approximately 0.31 m3/min (80 GPM) of water and the exhaust
19S
-------
1.0
§
~
u
2
o
H
pq
pq
-:
U
:-
OS
<
PH
0.1
0.01
0.1
1.0
PARTICLE DIAMETER, ymA
10
Figure 8-3 - Penetration versus particle diameter
(data set "A").
199
-------
1.0
H
W
0.1
DH
0.01
0.1
1.0
PARTICLE DIAMETER, ymA
10
Figure 8-4 - Penetration versus particle diameter.
(data set "B")
200
-------
blower. Maintenance of the scrubber consists of periodic
inspections, clean and occasionally replacement of spray
nozzles and piping and its costs are estimated at $300 per
year.
There are no unusual operating problems.
MATHEMATICAL MODEL
Section 5.3.2.2 in the "Scrubber Handbook" gave a
design equation (Eq. 5.3.2-6) for Impingement type scrubber.
This equation uses information on jet hole diameters, number
of holes and gas flow rate to predicted d 50- However, the
scrubber user did not have any information on the construc-
tion of the jet plate, which makes it impossible to use this
equation.
An alternative method to predict scrubber performance
is the cut diameter approach as described in Chapter 2.
Based on this method, for a pressure drop of 29 cm W.C., the
cut diameter for impingement type scrubber is 1.4 ymA. Pene-
tration for other particle diameters is based on the ex-
ponential variation of penetration with d* . The predictions
so obtained are shown as "prediction 'A'" in Figures 8-5
and 8-6 along with experimental results. It can be seen
that the test data indicate a cut diameter of about 1.0 ymA.
The density of sodium chloride is about 2.1 g/cm3 so the
diameter of a dry salt particle equivalent to 1.0 ymA is
about 0.6 ym. As discussed in the preceeding section on the
fibrous filter, the salt particle diameter should increase
at least 2 times, and as much as 5 times due to water conden-
sation. A salt solution particle 1.2 ym diameter would have
a density of about 1.1 g/cm3 and an aerodynamic diameter of
about 1.3 ymA.
An alternative method was used to predict penetration,
assuming that particle growth occurs in the first impingement
201
-------
1.0
01
0.1
0.5 1.0
PARTICLE DIAMETER,
10
Figure 8-5 - Predicted and experimental penetration.
(Data set "A")
202
-------
1.0
0.5
x
c
--
u
~
I I
E-i
:-
_
X
U
0.1
0.05 -T-T
0.01
0.1
1.0
PARTICLE DIAMETER, ymA
10
Figure 8-6 - Predicted and experimental penetration
(Data set "B").
203
-------
plate and that penetration based on dry particle size is
therefore lower on the second plate. The line labelled
"prediction 'B'" on Figure 8-5 is the result of the
alternative prediction with the assumed effect of condensa-
tion being to double aerodynamic diameter as the particles
flow from plate #1 to plate #2. The penetration for each
plate was predicted for a pressure drop of 14 cm W.C. . This
corresponds to an increase of roughly 3 times in actual
diameter. It can be seen in Figure 8-5 that the penetra-
tion so predicted is somewhat lower than the experimental
results. A particle size increase of about 2 times (actual)
would yield a prediction more in accord with the experi-
mental data.
CONCLUSIONS
This scrubber system is generally satisfactory and has
presented no significant operating or maintenance problems.
Performance appears to be in line with predictions based on
the cut diameter-scrubber pressure drop correlation if
allowance is made for salt particle growth. The major
uncertainty in the prediction method is related to particle
groivth by condensation in near-saturated gas.
204
-------
APPENDIX 8-A
PARTICLE DATA
205
-------
Table 8-A-l. INLET AND OUTLET SAMPLE PARTICLE DATA
FOR RUX #1.
IMPACTOR
STAGE
NUMBER
1
2
3
4
5
6
7
Filter
Sample
Volume
(DNra3)
INLET
"cun,
(nig)
22.4
3.2
3.2
3.2
2.7
2.2
2.0
1.7
dpc
(ymA)
23.0
10.2
4.75
1.9
1.05
0.55
0.30
0.22
OUTLET
W
cum
(mg) .
8.6
8.5
8.4
8.4
8.3
8.1
8.1
V
(ymA)
11.2
5.25
2.10
1.18
0.615
0.34
0.64
Table 8-A-2. INLET AND OUTLET SAMPLE PARTICLE DATA
FOR RUN 92.
TMPACTOR
STAGE
NUMBER
1
2
3
4
5
Q
Filter
INLET
cum
(rag)
17.8
5.4
4.6.
4.3
4.0
3.9
3.8
3.7
V
Cum A)
20
8.7
4.1
1.62 .
0.9
0.465
0.25
Sample
Volume 0 31
(DNm3) U' L
OUTLET
Wcum
Og)
4.5
4.2
4.0
V
(pmA)
0.64
0.355
0.42
206
-------
Table 8-A-3.
INLET AND OUTLET SAMPLE PARTICLE DATA
FOR RUN #3.
IMPACTOR
STAGE
NUMBER
1
2
3
4
5
6
7
Filter
INLET
Wcum
Og)
103.2
12.4
4.4
2.8
2.6
2.2
2.0
2.0
dpc
. (ymA)
20
8.7
4.1
1.62
0.9
0.465
0.25
Sample
Volume 0 .,,
(DNm3) °'31
OUTLET
W
cum
Og)
35.7
35.2
34.5
34.2
dpc
(ymA)
1.28
0.62
0.34
1.00
Table 8-A-4.
INLET AND OUTLET SAMPLE PARTICLE DATA
FOR RUN #4.
IMPACTOR
STAGE
NUMBER
1
2
3
4
5
6
7
Filter
INLET
cum
Og)
66.8
3.8
2..0
1.2
0.9
0.8
0.7
0.7
V
(ymA)
20.5
9.5
4.2
1.66
0.83
0.475
0.25
Sample
Volume n 21
(DNm3) U'Z1
OUTLET
W
cum
. Og)
3.6
3.5
3.5
3.4
3.2
2.3
1.8
dpc
(ymA)
11.0
5.25
2.1
1.15
0.61
0.335
1.03
207
-------
Table 8-A-5. INLET AND OUTLET SAMPLE PARTICLE DATA
FOR RUN #5.
IMPACTOR
STAGE
NUMBER
I
2
3
4
5
6
7
Filter
Sample
INLET
Wcum
(rag)
22
2.4
0.6
0.6
0.5
0.5
0.5
0.5
V
(ymA)
21
9.2
4.3
1.7
0.97
0.5
0.265
Volume 0.18
(DNm3)
OUTLET
yj
cum
(rag)
1.1
1.0
0.9
0.3
0
V
(ymA)
2.05
1.13
0.6
0.33
1.07
Table 8-A-6. INLET AND OUTLET SAMPLE PARTICLE DATA
IMPACTOR
STAGE
NUMBER
1
2
3
4
5
6
7
Filter
INLET
cum
(mg)
179.8
6.3
5.8.
4.0
3.3
2.9
2.7
2.7
V
CumA)
21
9.2
4.3
1.7
0.97
0.5
0.265
Sample
Volume 0.39
(DNm3) .
OUTLET
Wcum
(mg)
61.3
61.2
60.4
48.4
V
(ymA) _
1.1
0.58
0.33
1.14
m-L
208
-------
APPENDIX 8-B
PARTICLE SIZE DISTRIBUTION PLOTS
209
-------
w
H
w
Q
W
CJ
t-H
H
O.S 1 2 5 10 20 30 40 50 60 70 80
MASS PERCENT UNDERSIZE
Figure 8-B-l - Inlet particle size distribution for
data set "A".
210
-------
w
-
<
'
~
u
20 30 40 50 60 70 80 90 95 98 99 99.8
MASS PERCENT UNDERSIZE
Figure 8-B-2- Outlet particle size distribution for
data set "A".
211
-------
<
H
s
1I
Q
w
o
l-H
H
.512 5 10 20 30 40 50
MASS PERCENT UNDERSIZE
Figure 8-B-3 Inlet particle size distribution
for data set "B".
212
-------
0.2
20 30 40 50 60 70 80 90 95 98 99 99.8
MASS PERCENT UNDERSIZE
Figure 8-B-4 - Outlet particle size distribution for
data set "B".
21
-------
214
-------
APPENDIX 8-C
CUMULATIVE MASS DISTRIBUTIONS
215
-------
o
iH
X
I>
-------
O
H
X
I
bO
55
o
r-
35
W
U
fe
c
U
>
M
^
B
u
01234567 10
d ymA
pa*
Figure 8-C-2 - Mass concentration distribution
for Run #2.
217
-------
«
o
,-H
x
e
z
Q
ec
2:
O
z:
fri
u
2
O
CO
CO
II
E-
<
J
3
2
3
U
16
14
12
10
-±r+:ri^T±rr
TA^/V"
A '
_.;. i. ..:. _
"-3T-t : i i
^
-a
=1=
-
"^
-
S
^
A -
Z_i :.
te-t^
Lle.L.
'i;-;.:..
rrrrpi
/
4
d
pa'
Figure 8-C-3- Mass concentration distribution
for Run #3.
218
-------
c
rH
X
E
X
bo
cq
u
o
u
CO
pq
d umA
pa'
Figure 8-C-4 - Mass concentration distribution
for Run #4.
219
-------
o
r-i
x
O
u
o
CO
CO
u
_ |' __ ^2'
q:ip:±rr-drfa
dpa'
Figure 8-C-5 - Mass concentration distribution
for run #5.
220
-------
o
X
55
M
c
''
H
u
^
o
CJ
CO
<
:
o
H
X
/\
CO
E
n
o
M
H
H
53
W
u
HH
u
H
pq
':
0
7 8
d , ymA
pa'
Figure 8-C-6- Mass concentration distribution for
Run #7.
221
-------
222
-------
VENTURI ROD SCRUBBER ON CUPOLA
SOURCE AND SCRUBBER
An Environeering Venturi-rod scrubber (Model
A33 Hydro-Filter) was the subject of performance
test no, 8. This device consists primarily of several
parallel rods which are positioned in a duct with some
space between the rods so that gas can flow between them
(see Figures 9-1 and 9-2). Water is introduced upstream
from the rod bed and is atomized by the gas stream as it
flows between the rods. The basic operating mode (or unit
mechanism) of this scrubber is essentially the same as for
a venturi scrubber.
The scrubber is installed to control the emissions from
an iron melting cupola. The gas from the operating cupola
is drawn through the offtake and into the selector box.
This gas is pulled into the inlet of an air-to-air heat
exchanger, where it is cooled by an air stream on the outside
of the exchanger tube. After passing through the heat ex-
changer the gas enters a quench-dropout box where pre-clean-
ing and quenching take place. All the large particles drop
out in the quench section and they are vrater washed down
the quench-dropout box drain and deposited in one of the
sludge tanks.
This scrubbing system consists of:
1. A quench section to reduce temperature of offtake
gases prior to their entry into the Hydro-Filter.
2. A venturi-rod section to provide high energy
contact of particulate with scrub liquor.
3. A demisting section for removal of scrubbing
liquor drops from the cleaned gas stream.
223
-------
From cupola
and heat
exchanger
Water spray
Baffle
Demister
Rod bed
Venturi-rod scrubber
1
Figure 9-1 - Schematic diagram of scrubber system
224
-------
V
a
Di\
< -Flow
Venturi-Rod
40.64 cm
erting
le&
Area-
^. A
itU
6.2 cm
SECTION "AA"
Figure 9-2 - Schematic diagram of venturi-rod bed
225
-------
The quenched gas is drawn through the venturi-rod where
the high energy scrubbing takes place. The particulate
laden water is then washed down the Hydro-Filter drain and
deposited into the recycle section of one of the sludge
tanks.
The scrubbed gas is drawn through the stainless steel
demister vanes, where the free water is removed from the
gas. The gas is then drawn into the primary and secondary
fans and discharged up the stack. In the stack, the
scrubbed gas is mixed with the heat exchanger cooling air
(at 200°C). This reduces the steam plume and it improves
stack appearance.
TEST METHOD
The performance characteristic of the scrubber is
determined by analyzing the particle size distribution
and mass loading of the scrubber inlet and outlet gas
sample.
A modified E.P.A. Method 5 train with an in-stack
University of Washington Mark III (or Pilat) cascade
impactor was used for particle measurements. The impactors
in the inlet and outlet were allowed to heat up to stack
temperature before the samples were taken. Gas flow rate
was determined by means of type "S" pitot tube traverses
along with the necessary temperature and pressure measure-
ments. Sample flow rates were measured with the usual
E.P.A. train instruments so as to obtain isokinetic
sampling.
The inlet sampling point was located between the
quencher and the venturi-rod scrubber. Outlet samples
had to be taken after the fan because the negative pressure
after the demister was too high (-280 cm W.C.) for the
226
-------
sampling system to handle. The temperature and pressure
of the gas in the duct after the demister were measured.
A total of 13 simultaneous sampling runs were conducted
and four of these were discarded due to cupola shutdown
during sampling. The remaining 9 runs were grouped into
three data sets, namely A, B, C (Run No. 1, 2, 3 as set
"A", Run No. 7, 9, 10 as set "B", and Run No. 11, 12, 13
as set "C") , corresponding to different operating conditions,
as discussed later. All runs were sampled isokinetically
with the sampler being held at one position in the duct.
This is generally an adequate technique for obtaining
good samples of particles smaller than a few microns
diameter because they are well distributed across the duct.
It does not provide a representative sample of the large
particles when the nozzle inlet is close to a flow distur-
bance; as in the case of the outlet sample, which was taken
3 feet downstream of a bend. Thus, the total particulate
loading is uncertain because of the one position sample but
the fine particle concentration is representative of the
entire gas stream.
OPERATING CONDITIONS
The scrubber operating conditions during the test
period were as follows:
1. Gas flow rates were as shown in the tabulation
below:
DUCT
Temperature
Pressure during
pitot run
A m3/min
ACFM
DN m3/min
DSCFM
Vol. % H20 vapor
INLET
(190°F)
-6.5 cm H20
1,274
45,000
780
29,270
19
OUTLET
After demister After fan
(150°F)
-280 cm H20
(160°F)
+13 cm H20
16
227
-------
2. Water flow rate to the Hydro-Filter system was
reported by the plant as approximately 1.0 m3/min
(265 GPM) sprayed in the quencher and 3.0 m3/min
(800 GPM] sprayed in the venturi-rod bed. Make-up
water was estimated at 0.26 m3/min (70 GPM) which
consisted of 0.19 ma/min (50 GPM) evaporated and
0.076 m9/min (20 GPM) blow down. The temperature
of the spraying water was 24°C (75°F). The temp-
erature of the sludge washed out of the quencher
box was 71°C (160°F) and temperature of the
venturi-rod bed sludge holding tank was 65°C
(150°F).
3. Entrainment is known to be excessive, because it
causes fan unbalance, but was not measured in this
test series. Scrubber user indicated the slurry
blow down was 0.076 m3/min (20 GPM).
PARTICLE DATA
Three sets of data (3 simultaneous pairs in each set)
were obtained. These data sets were obtained at different
sampling locations and different plant operating conditions.
Particle concentration and size for these runs are presented
in Tables 9-A-l to 9-A-9. Size distributions for these runs are
shown in Figures 9-B-l to 9-B-3. The run numbering system used
here is that "a" denotes the inlet sample and "b" is assign-
ed to outlet sample in a simultaneous sampling run (designated
by the number).
Data sets "A" and "C" were taken under the same plant
operating conditions (melting ductile iron) but at different
sampling locations. Data set "B" was obtained when the
scrubber user was melting gray iron.
228
-------
As seen in Figures 9-B-l to 9-B-3, particles have the
following mass median and geometric standard deviation.
DATA SET
"A" (ductile)
"B" (gray)
"C" (ductile)
INLET
d (ymA)
Pg
0.92
1.15
0.94
' ag
2.0
1.7
2.1
OUTLET
d (ymA)
Pg
0.69
0.62
0.62
a
g
2.1
2.0
1.8
A diffusion battery was used to obtain information about
the size distribution of particles smaller than 0.3 ymA.
This was done by connecting the diffusion battery to the
outlet of the U.W. impactor (without backup total filter).
The diffusion battery was kept outside the stack and was
heated to stack temperature with heating tapes. The arrange-
ment is shown in Figure 9-3. Particle number concentration
in the inlet and outlet stream of the diffusion battery was
counted by a Gardner condensation nuclei counter. In this
performance test, particle number concentration was so high
that the condensation nuclei counter was overloaded even
with 5 to 1 clean air dilution (5 parts of clean air to 1
part sample gas), and the Gardner CNC was not stable. For
these reasons, only one run, namely 4b, was successful. The
inlet and outlet number concentration of the diffusional
battery for this run were 7xl03 particles/cm3 and 3xl03
particles/cm3 respectively. This gives a penetration of
0.43. Based on our design of the battery, this corresponds
to a 0.05 ym particle diameter cut point, or an aerodynamic
diameter cut point of about 0.2 ymA if particle density is
about 3 g/cm3. Unfortunately, the impactor portion of this
run was purged due to cupola shutdown during sampling.
229
-------
Stack
Impactor
(without back-up
filter)
to
Gardner CNC |
1 Valve
±__r£^
Total Filter
H
*
. i
\Charge
j Neutralizer
^iri^ir^r/Sj Diffusion
battery
Total i
Thermometer '
r
i
j Flowmeter
, Filter;
1 i
1
1
i !!
j ]! Pressure
jj gauge
Heated box
Figure 9-3 - Diffusion battery assembly
Impii
etc,
230
-------
Several simultaneous total filter runs were conducted.
However, due to the clogging of the filter by water droplets
even with a precutter ahead of the filter (the cut diameter
of the precutter is 10 ymA), no total particulate loading
data were available.
Plume opacity was 15% for runs #1, 2, 3, 11, 12, 13 and
10% for runs #7, 9, 10. All opacity readings were taken by
visual observation method by an observer trained in a
California Air Resources Board "smoke school".
PARTICLE PENETRATION
Particle penetration \vas computed and is shown in Fig-
ures 9-4 through 9-6. It was calculated by the following
method:
Cumulative mass concentration vs. aerodynamic particle
diameter data were fitted with a curve by eyeball method.
The slopes of these curves (Figs.9-C-l to 9-C-9) were measured by
a graphical technique at several values of particle diameter.
The ratios of outlet to inlet slopes were computed to yield
penetrations at the several diameter values.
ECONOMICS
The cost of installing and operating air pollution
equipment is a function of many direct and indirect cost
factors. These factors can be grouped into two cost cate-
gories; initial costs and annual costs.
1. Initial costs
The initial installed costs of the scrubber system
(1970) are listed below:
Scrubber Purchase Cost
a. F.O.B. Wet Scrubber $ 14,154
b. Freight 1,104
Total $ 15,258
231
-------
i.o r
H
U
O
I-H
E-H
W
O,
W
U
DJ
0.1
01
0.001
0.1
1.0
PARTICLE DIAMETER
10
- Particle penetration versus dia.eter for venturi ro
scrubber (data set "A")
232
-------
-I r- - - Htr : M-:f
0.001
0.1 1.0
PARTICLE DIAMETER, ymA
10
Figure 9-5 - Penetration versus particle diameter
for venturi-rod scrubber (data set "C")
233
-------
0.001
0.1
1.0
PARTICLE DIAMETER, ymA
10
Figure 9-6 - Penetration versus particle diameter
for venturi-rod scrubber (data set "B")
234
-------
Scrubber Auxiliaries
a. Fans, motors and motor starters $119,046
b. Ducting 78,086
c. Liquid and solid handling 38,090
and treatment
d. Instrumentation 6,750
e. Electrical material 10,694
Total $252,666
Scrubber Installation Cost
a. Site preparation, including $ 16,465
foundations and supports
b. Modifications in existing 88,764
processes
c. Installation 75,147
d. Start-up and modification 72,951
e. Engineering 56,000
Total $289,327
Total Initial Cost $557,251
(1970 prices)
Annual Costs
Annual Costs include the following factors
Operating Costs
a. Utilities $ 80,000
b. Labor 20,000
c. Supplies and Materials 9,000
d. Treatment and Disposal 4,000
Total $113,000
Maintenance Costs
a. Labor $ 44>000
b. Materials 6,000
Total $ 50,000
235
-------
Plant overhead, space, heat, $ 7,000
light, insurance, etc.
Total Annual Costs $170,000
OPERATING PROBLEMS
Several points relating to operating problems for
the scrubber system are listed below:
1. Entrainment separation has not been satisfactory
with the result that it causes fan unbalancing
even with chemical spray to minimize deposits
on fan blades.
2. Scrubbing water is recycled to reduce its
consumption. However, the sludge holding tank
is too small to allow solids to settle and clarify
the scrubbing water. Therefore, some solids
are recycled to the scrubber and eventually
cause the nozzle plug-up problem.
3. The pressure drop across the venturi-rod bed is about
260 cm W.C. (100"W.C.) and the throat velocity is
very high. This induces erosion problems,
especially on tube supports, scrubber bottom
and back plates.
4. Some quencher liquids are washed down to the venturi-
rod scrubber. The liquor contains large amounts
of solids which settle in the scrubber.
MATHEMATIC MODEL
A major objective of the scrubber performance test
is to compare the measured particle collection with that
predicted by mathematical models and/or further improve-
ment of the models which can be used for prediction of per-
formance .
236
-------
The venturi-rod scrubber is essentially several
Venturis or orifices connected in a parallel arrangement.
A method for prediction of performance for gas atomized
scrubbers was presented in "Scrubber Handbook" and a
further development is described in Chapter 2.
The operating conditions of the venturi-rod scrubber
were
1. Total gas flow rate is 21 m3/sec which gives
a throat velocity of 196 m/sec (643 ft/sec)
2. The liquid flow rate is 50 £/sec, corresponding
to QL/QG = 2.4 5,/m3
3. Pressure drop across the rod-bed is 275 cm W.G.
The "Scrubber Handbook" presented an equation (Eq. 5.3.6-10)
for estimating the pressure drop through a venturi scrubber
by assuming that all energy is used to accelerate the
liquid to the throat velocity of the gas. Based on this
equation, we calculated the pressure drop across the rod-
bed would be 916 cm W.G. However, measurements showed
the drop was only 275 cm W.G. This is an indication
that the liquid droplets are only accelerated to about
half the gas throat velocity. This is probably due to
the short exposure time of the droplets in the "throat"
between the pipes. At a velocity of 197 m/sec (647 ft/sec)
the gas moves a distance of 1 pipe diameter (6.2 cm) in
about 0.0003 sec.
Because the mathematical model given in the "Scrubber
Handbook" was based on the assumption that the liquid
is accelerated to the throat velocity, it seemed less
appropriate than the cut diameter - pressure drop correlation
The latter relationship should account for the opposing
237
-------
effects of higher relative velocity and shorter exposure
time in the venturi rod than in the venturi scrubber.
Consequently, cut diameters were predicted by means of
Figure 9-7 for a pressure drop of 275 cm W.C. and values
of f = 0.25, 0.4, and 0.5; corresponding to the particles
being more wettable as "f" increases.
Penetration for other particle diameters is based on
the exponential variation of penetration with d2 . The
predictions so obtained are shown in Figures 9-8 through
9-10 along with experimental curves. It can be seen that
the data are fairly well represented by predictions for
"f" between 0.4 and 0.5 in the size range of 1.0 ymA.
Smaller particles have lower penetration than predicted,
based on inertial impaction, and this can be explained in
part by diffusional collection.
As an estimate, we can assume that Brownian diffusion
in the venturi rod scrubber, entrainment separator, and
two blowers would account for the same efficiency as in a.
3 sieve plate column. The predictions for 3 sieve plates
(S. Calvert, 1974) are converted to aerodynamic size
(assuming p = 3.0), as tabulated below, and plotted on
Figure 9-11. The dashed line on Figure 9-11 is the esti-
mated combined effect of inertial impaction and Brownian
diffusion on particle penetration.
V ym
d , ymA
Pt
0.015
0.1
0.22
0.032
0.15
0.5
0.053
0.2
0.65
0.11
0.3
0.8
238
-------
1.0
<
o
PH
^ 0.5
w 0.4
Q
E-H
U
0.3
0.2
= 0.25
20 30 40 50 100
PRESSURE DROP, cm W.G.
20Q 300
Figure 9-7 - Predicted particle cut diameter versus
pressure drop for venturi scrubber.
239
-------
0.001
0.1
1.0
d umA
pa'
10
Figure 9-8
- Predicted and experimental penetration
for venturi-rod scrubber (data set "A"
Ductile).
240
-------
1.0
Figure 9-9 - Predicted and experimental penetration for
venturi-rod scrubber (data set "B1' - gray
iron) .
241
-------
0.001
0.1
1.0
10
pa>
Figure 9-10 - Predicted and experimental penetration for
venturi-rod scrubber (data set "C" - Ductile)
242
-------
Inertia,
£ = 0.5
>
pa,
Figure 9-11 - Predicted penetration by Brownian Diffusion
and Inertial Impaction
243
-------
As Figure 9-11 shows, Brownian diffusion can have a
very significant effect on particle penetration for sizes
smaller than a few tenths micron (aerodynamic). However,
the predicted penetration at about 0.4 ymA is still
higher than computed from experimental data. Additional
points to consider are as follows:
1. Flux force effects are unlikely to be significant
because the scrubber water is recycled and little
heat transfer can be obtained in the system.
2. The high velocity and extreme turbulence in the
two blowers in series may be quite effective
in causing particle collection; especially in
the presence of the entrained liquid which reaches
the blowers.
3. Particle size determination and the efficiency
computation at the 0.4 ymA diameter region are
dependent on the last impactor stage measurements
and are subject to error. The one diffusion
battery data point indicates a penetration of
43% by number for particles smaller than about
0.05 ym, corresponding to about 0.2 ymA. Because
the penetration drops off so rapidly for smaller
diameters, one would expect the penetration based
on particle mass to be of about the same magnitude,
although smaller than the number penetration.
CONCLUSIONS
Particle penetration data based on the measurements
made in this test agrees with prediction in the size
range of 1.0 ymA. Smaller particles have lover penetration
than predicted based on inertia impaction. Improvement
244
-------
of the model is needed for this type of scrubber and
probably has to account for the wet fans in series with
the scrubber.
The venturi-rod scrubber performance is good while
it is running but numerous operating problems forced the
system to be shut-down about half of the time.
High particle concentration and the stability problem
of the CNC have interfered with most of the diffusion
battery measurements. An efficient diluter and a more
stable CNC are necessary for future work.
245
-------
246
-------
APPENDIX 9-A
PARTICLE DATA
247
-------
Table 9-A-l. INLET AND OUTLET SAMPLE PARTICLE DATA FOR RUN #1
IMPACTOR
STAGE
NUMBER
1
2
3
4
5
6
7
Filter
Sample
INLET
M *
cum
(g/DNm3)
2.336
2.280
2.275
2.219
2.122
1.491
0.804
0.219
d xx
pc
(urnA)
28
12.3
5.7
2.25
1.31
0.68
0.375
0.060
Volume . DNm3
OUTLET
M
cum
(g/DNm3)
0.0487
0.0487
0.0487
0.0484
0.0447
0.0210
0.0210
0.0036
V
(umA)
23
10
4.7
1.85
1.03
0.54
0.295
0.27
M
cum
Cumulative mass collected on that stage and those
below
** d = Cut diameter (aerodynamic) for that stage
pc
ymA
Microns, aerodynamic = dp(c'Pp)1/2
248
-------
Table 9-A-2. INLET AND OUTLET SAMPLE PARTICLE DATA FOR RUN #2
IMPACTOR
STAGE
NUMBER
1
2
3
4
5
6
7
Filter
INLET
M *
cum
(g/DNm3)
2.280
2.231
2.209
2.157
2.068
1.767
0.722
0.308
d **
PC
. (umA)
28
12.3
5.7
2.25
1.31
0.68
0.375
Sample 0.041
Volume, DNm3
OUTLET
M
cum
(g/DNm3)
0.0437
0.0437
0.0437
0.0426
0.0404
0.0374
0.0174
0.0067
V
(umA)
23
10
4.7
1.85
1.03
0.54
0.295
0.27
Table 9-A-3. INLET AND OUTLET SAMPLE PARTICLE DATA FOR RUN #3
IMPACTOR
STAGE
NUMBER
1
2
3
4
5
6
7
Filter
INLET
M *
cum
(g/DNm3)
1.644
1.622
1.605
1.563
1.489
1.252
0.461
0.221
d KK
pc
(vimA)
28
12.3
5.7
2.25
1.31
0.68
0.375
Sample 0.041
Volume, DNm3
OUTLET
M
cum
(g/DNm3)
0.0515
0.0515
0.0508
0.0493
0.0457
0.0393
0.0202
0.0065
V
(ymA)
23
10
4.7
1.85
1.03
0.54
0.3
0.28
249
-------
Table 9-A-4. INLET AND OUTLET SAMPLE PARTICLE DATA FOR RUN #7
IMPACTOR
STAGE
NUMBER
1
2
3
4
5
6
7
Filter
INLET
M *
cum
Cg/DNm3)
3.522
3.508
3.505
3.481
3.377
2.936
0.532
0.124
d **
pc
(ymA)
23.5
11
5.15
2.04
1.16
0.6
0.33
Sample 0.041
Volume, DNm3
OUTLET
cum
(g/DNm3)
0.0213
0.0213
0.0208
0.0208
0.0202
0.0197
0.0115
0.0038
dpc
(ymA)
27.5
10.8
5.05
2.0
1.15
0.58
0.32
0.18
Table 9-A-5. INLET AND OUTLET SAMPLE PARTICLE DATA FOR RUN #9
IMPACTOR
STAGE
NUMBER
1
2
3
4
5
6
7
Filter
INLET
M *
cum
(g/DNm3)
3.019
3.005
2.995
2.963
2.623
2.004
0.218
0.014
d **
pc
(umA)
26.5
11.7
5.4
2.3
1.24
0.64
0.36
Sample
Volume, DNm3
OUTLET
Mcum
(g/DNm3)
0.0119
0.0119
0.0119
0.0119
0.0119
0.0103
0.0060
0.0027
V
(ymA)
24.2
10.8
5.0
1.97
1.13
0.58
0.32
-
250
-------
Table 9-A-6. INLET AND OUTLET SAMPLE PARTICLE DATA FOR RUN #10
IMPACTOR
STAGE
NUMBER
1
2
3
4
5
6
7
Filter
INLET
M *
cum
(g/DNm3)
2.532
2.467
2.460
2.418
2.122
1.526
0.225
0.029
d **
PC
(ymA)
26
11.6
5.3
2.1
1.22
0.63
0.35
Sample 0.14
Volume, DNm3
OUTLET
cum
(g/DNm3)
0.0096
0.0096
0.0096
0.0096
0.0091
0.0059
0.0021
0.0011
V
(ymA)
5.15
2.04
1.16
0.6
0.33
0.19
Table 9-A-7. INLET AND OUTLET SAMPLE PARTICLE DATA FOR RUN
IMPACTOR
STAGE
NUMBER
1
2
3
4
5
6
7
Filter
INLET
M *
cum
(g/DNm3)
2.837
2.837
2.798
2.793
2.726
2.507
0.791
0.577
d X57
pc
(ymA)
26.5
11.7
5.4
2.3
1.24
0.64
0.36
Sample 0.02
Volume, DNm3
OUTLET
Mcum
(g/DNm3)
0.0387
0.0387
0.0387
0.0387
0.0381
0.0327
0.0174
0.0054
V
(ymA)
24.5
]O.S
5.1
2.6
1.14
0.58
0.32
0.18
253
-------
Table 9-A-8. INLET AND OUTLET SAMPLE PARTICLE DATA FOR RUN #12
IMPACTOR
STAGE
NUMBER
1
2
3
4
5
6
7
Filter
INLET
M *
cum
(g/DNm3)
3.535
3.483
3.471
3.414
3.214
2.354
0.473
0.155
d *x
pc
. (ymA)
27
11.8
5.5
2.4
1.27
0.65
0.36
Sample 0.019
Volume, DNm3
OUTLET
Mcum
(g/DNm3)
0.0302
0.0296
0.0279
0.0156
0.0034
V
(ymA)
5.1
2.03
1.15
0.595
0.325
0.18
Table 9-A-9. INLET AND OUTLET SAMPLE PARTICLE DATA FOR RUN #13
IMPACTOR
STAGE
NUMBER
1
2
3
4
5
6
7
Filter
INLET
M *
cum
(g/DNm3)
2.522
2.443
2.438
2.374
2.201
1.859
0.661
0.493
d **
pc
(ymA)
26.5
11.7
5.4
2.3
1.24
0.64
0.36
Sample 0.042
Volume, DNm3
OUTLET
Mcum
(g/DNm3)
-
0.0205
0.0188
0.0172
0.0100
0.0039
V
(ymA)
2.05
1.16
0.6
0.33
0.18
252
-------
APPENDIX 9-B
PARTICLE SIZE DISTRIBUTION PLOTS
253
-------
10
A Run
-
/ /
5 10 20 30 40 50 60 70 80 90
MASS PERCENT UNDERSIZE
98
Figure 9-B-l-Inlet and outlet size distribution (set "A")
254
-------
10
-
1.0
0.5
0.2
GT>, * 7 n
Run - / a"
;
A
O .Run £
!
I
T^
i .
A
-iO*-
CD Run *7b
10 20 30 40 50 60 70 80 90
MASS PERCENT UNDERSIZE
98
Figure 9-B-2 - Inlet and outlet size distribution (set "B")
255
-------
10
p«
31.0
0.5
0.2
Run jf11a
Run'#12a
.
O Run *13a !
, -
'
m
Run Ill-t?-
^-Outlet
10 20 30 40 50 60 70 80
MASS PERCENT UNDERSIZE
90 98 99
Figure 9-B-3- Inlet and outlet size distribution (set "C")
256
-------
APPENDIX 9-C
CUMULATIVE MASS DISTRIBUTIONS
257
-------
I
bO
C/5
CJ
-
-
-
Q6
05
.04
02
p
^>
u
w
-
o
.01
012345 10 15
AERODYNAMIC DIAMETER, ymA
Figure 9-C-l ' Cumulative mass concentration for Run #1
258
-------
2.4
0.06
2.0
0.05
=
Z
to
0.04
1.5
1.0 -
0
0
0
0.1 - -
0 1 2
345 10
AERODYNAMIC DIAMETER, ymA
15
p
W>
en
K_
r:
S
u
'
r-
::
C
Figure 9-C-2- Cumulative mass concentration for Run #2
259
-------
I
bC
CO
CO
s
II
E-
<
g
2
P
u
-
pq
_
2
0
D
D
0
'
05
-0.04
2
O
M
CO
CO
0.03
,02
u
w
-!
E-
b
o
01
012545 10
AERODYNAMIC DIAMETER, ymA
15
Figure 9-C-5- Cumulative mass concentration for Run #3
260
-------
?T3 0.04
e
G
W>
in
CO
Q 1 2 3 4 5 10
AERODYNAMIC DIAMETER
15
~
IS
U
H
S
::,
Figure 9-C-4 - Cumulative mass concentration for Run #7
261
-------
3.5
3.0
I
bO
*
CO
CO
f-
ffl
014
012
010
,008
g
o
ba
n
CO
006
E-
004 o
.002
012345 10
AERODYNAMIC DIAMETER, ymA
Figure 9-C-5 - Cumulative mass concentration for Run #9
26;
-------
012
; i .
012345 10 15
AERODYNAMIC DIAMETER, ymA
Figure 9-C-6 - Cumulative mass concentration for Run #10
263
-------
3.0
2 .5
bO
A
C/5
2.0
U
0.5
-.::-
- - '- -
ATI
^A
-
tern
r.i _i, r
*irlb
~-r
T-retrtrle-t1:
:.
f
- - \__
^----'--'---T-----\---
0.06
0.05
I
0.04
0.03 >
0.02
U
yj
0.01
012345 10
AERODYNAMIC DIAMETER , .
15
ymA
Figure 9-C-7 - Cumulative mass concentration for Run #11
264
-------
e
p
M
CO
-
u
H
-
-
Z
1.0
0.5
0.035
0.03
0.025
0.02
-
Z
w>
n
CO
- 0.015
0 .01
: :
I
_:
H
ID
C
0.005
012345 10
AERODYNAMIC DIAMETER, ymA
Figure 9-C-8 - Cumulative mass concentration for Run #12
265
-------
ao
CO
CO
gl
U
H 1
W
F
*^*
tt .. .
0.03
0.025
0.02
6
^
n
tX>
ua
CO
0.015 >
0.01
_
I
0.005
012345 10 15
AERODYNAMIC DIAMETER, ymA
Figure 9-C-9 - Cumulative mass concentration for Run #
266
-------
REFERENCES
S. Calvert, J. Goldshmid, D. Leith, and D. Mehta. "Scrubber
Handbook", A.P.T., Inc. Riverside, California. EPA Contract
No. CPA-70-95. August 1972. 'PB-213-016.
S. Calvert, J. Goldshmid, and D. Leith, "Scrubber Performance
for Particle Collection", A.I.Ch.E. Symposium Series 7_0
(137):357(1974).
C. E. Junge . "Air Chemistry and Radioactivity". Academic
Press. 1963.
C. W. Lapple, and H. J. Kamack. "Performance of Wet Dust
Scrubbers". Chem. Eng. Prog. 151(3) : 110-121, March 1955.
K. T. Semerau. J. Air Pollution Control Assoc. 10, 200
(1960) .
M. Taheri, and S. Calvert. "Removal of Small Particles From
Air by Foam in a Sieve Plate Column". J. Air Pollution
Control Association. 18:240-245, 1968.
267
-------
NOMENCLATURE
A = a constant in eq. (2-4) and (2-5)
A = Constant
a
A = total outside surface area of drops in scrubber, cm2
d
B = constant defined by equation (2-5)
C1 = Cunningham correction factor, dimensionless
d = diameter, m or cm
d, = bubble diameter, cm
b
d = packing diameter (nominal), cm
d, = drop diameter, cm
d, = sieve plate hole diameter, cm
n
d = geometric mean particle diameter, ym
p /
d = aerodynamic particle diameter, ymA ='d (C' P )/
pa f f
d . = particle diameter, ym
Pi
d = geometric mean particle diameter, ym
PS
d = diameter of particle collected with 50% efficiency, ym
p50
d = performance cut diameter (aerodynamic), ymA
PC v
d = required cut diameter, ymA
RC
d = differential mass, g
w
D = diffusivity, cm2/sec
D, = ball diameter, en
b
D = particle diffusivity, cm2/sec
P
DNm3 = dry standard cubic meter, at 0°C and 760 mm Hg
E = efficiency, fraction or %
E = particle collection efficiency
268
-------
£ = ratio of drop velocity relative to gas velocity,
empirical constant
F = froth density, g/cm3
G = gas rate Ib/hr-ft^ or kg/hr-m
h = height of scrubber, cm
h. = inertial impaction parameter
i = Van't Hoff factor
K. = inertial impaction parameter for "d."
K = particle inertial impaction parameter
d2 v, x 10-8
pa h
9 "G dh
L = liquid rate, Kg/hr-m2 or lb/hr-ft2
m = mass, kg or g
mg = milligram
M = cumulative mass collected on that stage and those
cum below, g
Pt = penetration (one minus efficiency), fraction or percent
Ft = overall penetration
AP = pressure drop, cm W.C. or atm.
Q = heat transferred per unit cross-section area of
column, cal/cm2
Q = gas volumetric flow rate, m3/sec
b
Q = liquid volumetric flow rate, m3/sec or £/sec
Li
S = solidarity factor
uon = particle deposition velocity for Browiiian diffusion
BU
u~ = gas velocity relative to duct, cm/sec
269
-------
u, = gas velocity through sieve plate hole, cm/sec
u = gas velocity, cm/sec
upn = deposition velocity, cm/sec
U = undisturbed upstream air velocity, m/sec
U, = hole velocity, cm/sec
W = cumulative mass, g
W = total collected mass, g
Z = static bed depth, m or cm
Greek
E = summation
n = efficiency due to unit mechanism, fraction, or percent
H. = collection fficiency for particle diameter "di"
n = effective collection efficiency of a single fiber
s by all collection mechanisms
8 = penetration time, sec
a = geometric standard deviation of particle size
£ distribution
y = viscosity, g/cm-sec
u,, = gas viscosity, centipoise
b
ym = micron (micrometer)
ymA = aerodynamic diameter = d (C1 p )/2,ym (g/cm3)/2
p = density, kg/m3 or g/cm3
PG = gas density, g/cm3
p = liquid density, lb/hr-ft3 g/cm3
Li
p = particle density, g/cm3
270
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-650/2-74-093
2.
3. RECIPIENT'S ACCESSION-NO.
4. TITLE AND SUBTITLE
Fine Particle Scrubber Performance Tests
5. REPORT DATE
October 1974
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
Seymour Calvert, Nikhil C. Jhaveri, and
Shuichow Yung
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
A. P.T. , Inc.
P.O. Box 71
Riverside, California 92502
10. PROGRAM ELEMENT NO.
1AB012; ROAP 21ADJ-037
11. CONTRACT/GRANT NO.
68-02-0285
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
NERC-RTP, Control Systems Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Final
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
is. ABSTRACT-PI^ reporf gives results of fine particle scrubber performance tests on
industrial installations and the comparisons of experimental data with mathematical
models. Particle size and concentration in the inlet and outlet scrubber gas streams
were measured by means of cascade impactors and other apparatus. Tests were
completed for a valve-type tray on a urea prilling tower, vaned centrifugal on a
potash dryer, mobile bed on a coal-fired boiler, venturi on a coal-fired boiler,
wetted fibrous filter on a salt dryer, impingement plate on a salt dryer, and venturi
rod on a cupola. Performance is reported as particle penetration as a function of
particle diameter. Mathematical models are satisfactory for all the scrubbers
tested except the mobile bed. Information on costs, operating problems , mainten-
ance, and other system characteristics are reported.
7.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Air Pollution
Scrubbers
Performance Tests
Mat hemati cal Models
Measurement
Particle Size Distribution
Operating Costs
Ai r Pol 1 ut i on Cont rol
Stationary Sources
Fine Particulate
13B,
07A
14B
12A
14A
18. DISTRIBUTION STATEMENT
Unlimi t ed
19. SECURITY CLASS (This Report)
Unclassifi ed
21. NO. OF PAGES
271
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
271
------- |