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	1—g£
     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
I—I
H
2
—
U
J5
O
U
C/3
5
                             dpa,  ymA
Figure 3-8 -
Cumulative mass
particle diametp

      61
                                         aerodynamic

-------
E
2:
Q
o
I—I
E-
CJ
2
O
oo
CO
I—I
H
      0
                              d  , ymA
                               pa


       Figure  3-9  -   Cumulative mass versus  aerodynamic particle
                      diameter for Run #14.
                             62

-------
c



CJ



ft
^-^


X

o
I—I

"—




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
I—I
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
   i—i
   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

>
h—I

H
    200 ^
                     :

                                     --j -^';-'—7^	;—;—~
                                            —
                                                                  ~V ;:^rzr^
                                1  —— _ , _	_ .. _~~	
                                    . • i- «   •—i • '—i—i--~---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
I—I
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—-r—r-
                                          :
                                   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
       I—I

       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
                                                              I—i

                                                              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
    1—I
    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
I—I
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-d—rfa
                              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
      I—I
      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

-------