EPA-600/2-76-035
February 1976
Environmental Protection Technology Series
^ ^ + s* %'t*^\f+ $">~vtu^*n %/'* % ,^ *
„ * * < X i "-.%»•? ,? V *"** "*<**s> ^ " * ^
, , rt i 4\ \^ * t * *H;V\* -* - <°f % \
EVALUATION »ffiKt
ION
m »rpT»i. - ^T^j y-y >; , , <
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into five series. These five broad
categories were established to facilitate further development and application of
environmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The five series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
This report has been assigned to the ENVIRONMENTAL PROTECTION
TECHNOLOGY series. This series describes research performed to develop and
demonstrate instrumentation, equipment, and methodology to repair or prevent
environmental degradation from point and non-point sources of pollution. This
work provides the new or improved technology required for the control and
treatment of pollution sources to meet environmental quality standards.
EPA REVIEW NOTICE
This report has been re viewed by the U.S. Environmental
Protection Agency, and approved for publication. Approval
does not signify that the contents necessarily reflect the
views and policy of the Agency, nor does mention of trade
names or commercial products constitute endorsement or
recommendation for use.
This document is available to the public through the National Technical Informa
tion Service, Springfield, Virginia 22161. *<-nnicai miorma-
-------
EPA-600/2-76-035
February 1976
EVALUATION OF EIGHT NOVEL
FINE PARTICLE COLLECTION DEVICES
Douglas W. Cooper, Richard Wang,
and Daniel P. Anderson
GCA Corporation
Burlington Road
Bedford, Massachusetts 01730
Contract No. 68-02-1316, Task 8
ROAP No. 21ADL-004
Program Element No. 1AB012
EPA Project Officer: Dale L. Harmon
Industrial Environmental Research Laboratory
Office of Energy, Minerals, and Industry
Research Triangle Park, NC 27711
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Washington, DC 20460
-------
ACKNOWLEDGMENTS
This work has been facilitated by the cooperation of the other
organizations which made the original evaluations: Southern Research
Institute, Midwest Research Institute, and Air Pollution Technology,
Inc. The Project Officer, Dale Harmon, was instrumental in obtaining
the materials which were central to this report. This help is grate-
fully acknowledged.
11
-------
CONTENTS
Page
Acknowledgments ii
List of Figures iv
List of Tables Vll
Acknowledgments vlii
Section
I Conclusions 1
II Recommendations 2
III Introduction 3
IV Aronetics Two-Phase Jet Scrubber 5
V Braxton Sonic Agglomerator 24
VI Centrifield Scrubber ' 38
VII Cleanable High Efficiency Air Filter (CHEAF) 48
VIII Dynactor Scrubber 55
IX Lone Star Steel Steam-Hydro Scrubber 75
X Mystaire Scrubber 91
XI Pentapure Scrubber 102
XII University of Washington Electrostatic Droplet Scrubber 115
XIII Efficiency Comparison 124
XIV Cost Comparison 136
XV Power Comparison 145
XVI Applications 149
XVII Collection Efficiency Theory 156
XVIII Appendix 176
XIX References 179
iii
-------
FIGURES
No.
1 Generalized Two-Phase Jet Scrubber Nozzle 6
2 Generalized Two-Phase Jet Scrubber System 7
3 Variation of Water Droplet Velocity From Two-Phase Jet
Nozzle With Water Temperature 8
4 System Pressure Rise as a Function of Water Flow Rate 10
5 Optical and Diffusional Sizing System 15
6 General Layout of Furnace and Scrubber System Showing
Locations 16
7 Average Inlet Size Distribution on a Cumulative Weight
Percentage Basis 19
8 Fractional Efficiency of the Aronetics Scrubber Based on
Optical, Diffusional, and Inertial Sizing 20
9 Fractional Efficiency of the Aronetics Wet Scrubber Calculated
From the Weighted Averages of 30 Inlet (Brink) and 6 Outlet
(Andersen Mark II) Impactor Tests 21
10 Schematic Diagram of Test of Braxton Sonic Aggloraerator 27
11 Collection Efficiencies Versus Aerodynamic Diameter for
Cyclone, Agglomeratorv and Cyclone Plus Agglomerator 31
12 Centripetal Vortex Balances Gas Velocity Against Centrifugal
Force 39
13 Schematic Diagram of the Asphalt Plant and Scrubber Layout
and the Locations of the Sampling Points Used in the Tests 44
14 Fractional Efficiencies as Determined by the Four Methods
Used in the Test Program 45
15 Fractional Efficiencies of the Entoleter Centrifield
Scrubber as Determined From Cascade Impactor Data 46
16 Diagram of CHEAF Installation 51
17 Collection Efficiency Versus Aerodynamic Diameter, CHEAF 52
18 Inlet and Outlet Cumulative Aerodynamic Diameter Distribu-
tions, CHEAF Tests 53
iv
-------
FIGURES (Continued)
No. Page
19 Single-Stage Dynactor Diffusion System Cross Sectional View 56
20 Two-Stage Dynactor 58
21 Test System for Dynactor Two-Stage Scrubber Evaluation,
Including Filter Samplers (F), Thermometers (T), and Pressure
Gauge (P) 63
22 Details of Aerosol Concentration and Size Distribution Mea-
surement Sections 64
23 Summary of Results of 16 Collection Efficiency Tests
(Factorial Design) 68
24 The Lone Star Steel Steam-Hydro Air Cleaning System 77
25 Optical and Diffusional Sizing System 83
26 Inlet and Outlet Particle Size Distributions Measured Using
Optical and Diffusional Techniques 85
27 Fractional Efficiency of the Lone Star Steel Steam-Hydro
Scrubber 86
28 Collection Efficiency Versus Aerodynamic Diameter, Lone Star
Steel Steam-Hydro Scrubber 87
29 Power Consumption Versus Flow Resistance for Lone Star Steel
Steam-Hydro Scrubber 89
30 Schematic Diagram of the Mystaire Laboratory Scrubber 92
31 Estimated Mystaire Collection Efficiency Versus Particle Size 100
32 Schematic Diagram of Purity Corp. Impingement Scrubber, the
Pentapure Scrubber 103
33 Schematic of Sampling Positions 106
34 Inlet Particle Size Distribution 108
35 Collection Efficiency Versus Aerodynamic Diameter, Data
Means for Pentapure Scrubber 110
36 Data and Approximate Theory Predictions (f = 0.25, 0.50;
QL/Qg =0.57 x 10-3) for Pentapure Scrubber 113
v
-------
FIGURES (Continued)
No. Page
37 Schematic Diagram of Electrostatic Droplet Scrubber 116
38 Calculated Particle Collection Efficiencies for a Single
200-ym Diameter Droplet With a 100-cm/s Undisturbed
Fluid Velocity 121
39 Particle Collection Efficiency of Electrostatic Spray
Droplet Scrubber as Function of Particle Size 122
40 Hypothetical Cumulative Mass Distribution (Log Normal, Mass
Median = 1, Geometric Standard Deviation = 2) 125
41 Hypothetical Control Device Collection Efficiency Versus
Particle Diameter 127
42 Treatment of Waste Waters to Effect Various Degrees of
Contaminant Removal 141
43 Relationship Between Total Water Cost and Treatment 142
44 Collection Vehicle Hauling Costs 144
45 Predicted Aerodynamic Cut Diameter Versus Pressure Drop and
Power Consumption (Adapted from Scrubber Handbook) 147
46 Criteria for Selection of Gas Cleaning Equipment 150
47 Geometry for Collector Analysis 157
48 Flow Streamlines and the Limiting Trajectories 163
48A Combined Collection Efficiency Based on Experimental Results
nICD Versus (6Sc~2/3 Re-l/2 + SRe1/2 R2) 168
49 Theoretically Calculated Migration Velocities for Four
Electrostatic Mechanisms Versus Particle Diameter 171
vi
-------
TABLES
No. Page
1 Weighted Average Inlet and Outlet Loadings, Aronetics Tests 18
2 Fractional Collection Efficiencies for the Braxton Sonic
Agglomerator System, Cyclone, and Agglomerator/Cyclone
Combination 30
3 Sonic Agglomerator System Operating Parameters 32
4 Comparison of Inlet and Outlet Dust Loading, Cyclone Catch,
Cyclone Undersize Dust and Overall System Collection of All
Particle Sizes 33
5 Summary of Results of 16 Collection Efficiency Tests for
Dynactor (Factorial Test Design) 67
6 Efficiency Averages Versus Particle Aerodynamic Diameter 69
7 Significance of Effects of Flow, Dust, Temperature, and
Concentration on Scrubber Collection Efficiency 70
8 Individual and Overall Particle Collection Efficiency Due
to Impaction, Interception, and Diffusion 98
9 Mean Collection Efficiency, Impactor Data 109
10 Example of Calculation of Predicted Mass Collection
Efficiency 129
11 Aerodynamic Cut Diameters for Several Novel Control Devices 132
12 Average Fine Particle Efficiencies for the Novel Control
Devices 134
13 Power Consumption Figures for Novel Control Devices
Evaluated for IERL-RTP 146
14 Advantages and Disadvantages of Collection Devices 152
15 Advantages and Disadvantages of Wet and Dry Collectors 153
16 Industrial Process and Control Summary 155
17 Calculated m Compared With Data 166
18 Approximate Particle Charge (Cochet Equation) (Conducting
Particles; E = 10 kV/cm = 33.3 esu) 170
vii
-------
SECTION I
CONCLUSIONS
The novel control devices evaluated for the Industrial Environmental
Reseach Laboratory - Research Triangle Park, of the Environmental Pro-
tection Agency displayed a wide range of efficiencies and power con-
sumption rates. Not all were capable of 90 percent or better collection
efficiency on particles smaller than 3 ym, an approximate criterion
for being a high efficiency collector of fine particulates. Most of
the novel control devices were related to the spray scrubber, includ-
ing two which utilized waste heat for improved scrubbing efficiency
and one which added electrostatic forces to the collection mechanisms.
Two of the novel devices were essentially wetted filters, with the
cleaning being done by the flow of water rather than by mechanical dis-
lodgement. In comparing the novel devices with each other and with
conventional control devices, one must consider the collection effi-
ciency needed and the efficiency attainable; a rational economic deci-
sion involves comparison of the devices on the basis of total annualized
cost, where possible.
-------
SECTION II
RECOMMENDATIONS
The primary goal of this work was to present, rather than evaluate,
the results of the novel control device evaluations sponsored by the
Industrial Environmental Research Laboratory - Research Triangle Park.
The prospective user of particulate control equipment should determine
the nature of the particulate he must control with regard to its
chemical nature and particle size distribution, and the nature of the
gas in which it is borne, especially its actual and standard flow rates,
its temperature, and its chemical composition. From those novel or con-
ventional control devices which would be suited to his requirements, the
user can choose among them on the basis of their collection efficiencies
and total annualized costs, methods for doing both of which are indi-
cated herein, in Sections XIII to XVI on efficiency, power, cost, and
applicability.
-------
SECTION III
INTRODUCTION
As one of its approaches to the reduction of the emission of fine par-
ticulates (those with diameters smaller than about 3 ym), the Industrial
Environmental Research Laboratory - Research Triangle Park of EPA has
sponsored the evaluation of a number of control devices presented by
their designers as having important novel features in comparison with
conventional technology (filters, scrubbers, electrostatic precipita-
tors). These evaluations were of two general types: primarily theoret~
ical or primarily experimental. Only the Mystaire and the University
of Washington scrubber were of the former, nonexperimental type of
evaluation. Presented here are synopses of these evaluations, along
with methods for comparing efficiencies, costs, power consumption, and
applicability.
The devices are presented in alphabetical order, by the name usually
associated with them. The principal features of the devices are as
listed below:
• Aronetics Scrubber - uses waste heat to produce a high
energy spray which cleans and moves the gas.
• Braxton Sonic Agglomerator - employs high energy sound
to agglomerate fine particles on droplets and on coarser
particles, thus facilitating their removal by low-energy
collectors such as cyclones.
• Centrifield Scrubber - creates vortex flow to improve the
collection efficiency of its spray.
• Cleanable High Efficiency Air Filter - produces collection
efficiencies on the order of 95 percent with a porous
medium that is continually cleaned by water.
-------
• Dynactor Scrubber - incorporates a proprietary nozzle
design to enhance collection efficiency in a spray
scrubber that cleans and moves the gas.
• Lone Star Steel Scrubber - uses waste heat to produce
steam to drive a scrubber similar to a venturi scrub-
ber, producing efficiencies higher than 99 percent
for particles larger than about 0.3 vim, and providing
motive power to the gas.
• Mystaire Scrubber - combines spray scrubbing and fil-
tration to produce particle collection on wetted wire
mesh, which is cleaned by a combination of spray flow
and cross-current water flow.
• Pentapure Scrubber - produces particle agglomeration
and collection on spray droplets by passing the gas/
droplet/particle mixture through a converging section
which directs the mixture jet at a collecting target.
• University of Washington Electrostatic Spray Droplet
Scrubber - adds electrostatic force collection to the
usual collection mechanisms in spray scrubbing to im-
prove collection efficiency at a given level of power
consumption.
The individual device summaries are followed by sections showing how to
estimate efficiencies for a particular dust given the control device
efficiency curves; how to compare cost and power consumption; the typical
applications for devices similar to either filters, scrubbers, or elec-
trostatic precipitators; the underlying theory of particle collection on
obstacles, which is relevent for these scrubber-type or filter-type col-
lectors; and an appendix on the definition of the particle aerodynamic
diameter, a term frequently used in particulate material control
technology.
-------
SECTION IV
ARONETICS TWO-PHASE JET SCRUBBER
MANUFACTURER'S DESCRIPTION1
"A pressurized, heated liquid when passed through a properly designed
nozzle will produce a two-phase mixture of vapor and liquid droplets
that is an excellent cleaning medium. The droplets can be accelerated
to extremely high velocity as a result of the expansion force created
by a portion of the liquid being converted to vapor. The general con-
figuration of this type of scrubber is shown in Figures 1 and 2. The
proper arrangement of components allows a draft to be induced which
eliminates or drastically reduces fan power requirements. The two-
phase jet scrubber produces water droplet velocities which vary with
the temperature of the scrubbing fluid as shown in Figure 3. It is
Aronetic's experience that jet velocities in the range of 1000 feet per
second are quite satisfactory for particulate removal in the size
range down to 0.10 microns. The velocity in the region immediately
downstream of the nozzle is probably substantially supersonic since
there is considerable evidence that sonic velocity in a two-phase
mixture may be as low as 350 feet per second. However, Aronetics
believes that inertial impaction is the controlling mechanism in clean-
ing and that the existence, or absence, of the shock phenomena associated
with supersonic flow is not an advantage in the cleaning effectiveness.
Thus, the velocity in feet per second is the controlling parameter
rather than the Mach number, or relationship of velocity to the local
speed of sound. The device operates at a very low noise level. The
sound frequency and level is similar to that generated by a garden hose
nozzle.
-------
GAS AND
LIQUID OUT
Figure 1. Generalized two-phase jet scrubber nozzle
-------
e
0)
4J
CO
>>
CO
"§
h
o
CO
-------
o0'
1
8
o
to
ro
O
O
o
V
CM
CO
E
o
00
o
eg
8
o
8
00
8
o
o
•o
M
o>
•u CL,
0) 6
i-4 d)
d, JJ
O
n n
T3
-------
"With respect to the draft producing capacity of this type of
scrubber system, it should be evident that this is a pure momentum
transfer mechanism; therefore, the amount of draft is a function
of the amount of fluid passed through the nozzle and the degree to
which this fluid is accelerated. The velocity is a function of
initial water temperature, as shown in the previous figure. The
effect of water flow rate on system pressure rise is shown in
Figure 4 for water at a temperature of 400°F. It should be empha-
sized that this is a rise in pressure across the scrubber and should
not be confused with the pressure drop which is associated with the
venturi type scrubber. In most applications, the pressure rise pro-
duced by the two-phase jet is sufficient to overcome the pressure
drop in other components of a complete system.
"The most direct application of the two-phase scrubbing system is
in the control of emissions from processes which generate high tem-
perature gas, laden with submicron particulate. Typical examples
are the various metallurgical furnaces and processes. Figure 2
showed schematically the general arrangement of the components as
tested in this report. An economizer type of heat exchanger is used
to transfer thermal energy from the high temperature process exhaust
gas to pressurized hot water which is delivered to the heat exchanger
by a pump. Water exiting the heat exchanger is delivered directly to
the nozzle in its liquid state. For most applications, the water
temperature is approximately 400°F and the water pressure is approxi-
mately 350 psi, or high enough to ensure that the fluid remains in
the liquid state until it has passed the nozzle throat. A properly
dimensioned mixing section must be provided for intimate contact
between the accelerated water droplets and the particle laden gas.
The final component in the scrubbing system train is a separator which
will remove the dirty water droplets and allow the clean gas to be
discharged. Water drained from the separator is passed to water
treatment equipment which may be used to remove substances scrubbed
-------
o
o
o
o
o
o
ro
ew
E
O
O
O
CM
o
o
o
o
(M
(0
CM
00
•5
*
0)
u
UJ
(O
CO
to
UJ
CO
CM
T3
0)
T3
CO
4J
a
•H
0)
4J
tfl
M-l
(U
4J
§
•rl
en
cfl
a>
CO
•H
(-1
3
CO
CO
a)
S-i
0)
CO
C/l
-------
from the gas and to prepare the scrubbing liquid for recycling. In
the present instance, the water is used for approximately three passes
through the system before final discharge.
"If the process off-gas is at a temperature level above 1200 F, an
additional option becomes available in the selection of components.
The economizer type of heat exchanger may still be used to deliver
heated water directly to the nozzle, or a steam boiler may be used
as an intermediate step in the heating of water. If sufficient energy
is contained in the gas, it is possible that a quantity of steam may
be produced which is greater than the demands of the scrubber. This
steam is then available for other possible plant applications. Other
elements of the system remain essentially the same with the exception
of the addition of a method to transfer thermal energy from steam to
water. The choice between the steam boiler and the economizer type
of system for the very high temperature gases is dictated by local
conditions at the site in question."
POTENTIAL APPLICATIONS
This scrubber would be applicable to situations in which other high-
energy scrubbers are logical candidates such as steel mills, foundries,
refineries, paper mills, and industries with sticky or liquid emissions.
It is substantially more economical if waste heat is available than it
would be without waste heat. It is now manufactured by Chemical Air
Pollution Control Co. of New York City.
Particle Characteristics
Particle Size - Collection efficiency was found to be greater than
50 percent for particles larger than 0.15 |j.m aerodynamic diameter,
greater than 90 percent for particles larger than 0.3 p.m aerodynamic
diameter, and greater than 99 percent for particles larger than
about 1 (am aerodynamic diameter. (See Appendix for definition of
aerodynamic diameter.)
11
-------
Particle Phase - The scrubber should be capable of removing solid or
liquid aerosol particles.
Particle Resistivity - Scrubbing action should be independent of
resistivity.
Particle Abrasiveness - Because the design produces high relative
velocities between particles and drops without producing high gas
velocities, abrasion should be less a problem than for a venturi
scrubber.
Gas Characteristics
Volume Rate of Flow - Because the relative velocity between drops and
particulate pollutants is produced by the nozzle rather than by the
gas, changes in volume flow rate should not have as much effect upon
collection efficiency as for constant-throat venturi scrubbers.
Gas Temperature - Higher gas temperatures mean that the gas is more vis-
cous, thus making particle collection by impaction somewhat more difficult
This is overbalanced by the advantage of having available waste heat
to drive the scrubber. Sufficient heat must be available to raise
the water temperature in the scrubber to near the 204 C (400 F) design
value. The manufacturer suggests using an economizer for temperatures
less than 650°C (120QOF), noting that in the range 650 to 2200°C (1200°F
to 4000°F) gas temperature it may be advantageous to use a steam boiler
to generate steam, only part of which would be needed for the scrubber.
THEORY OF OPERATION
The mechanisms by which particles are captured by the high-speed spray
are those discussed in Section XVII on collector theory. The water
will generally be at a lower temperature than the stack gas, which
12
-------
means that thermophoresis will be & force for collection of particles;
evaporation from the droplets will tend to reduce their collection
efficiency, however. The data below are consistent with the hypothesis
that impaction predominates for particulate matter as small as a few tenths
microns, and diffusive collection predominates for the particles
still smaller.
The generation of the spray is described in the manufacturer's des-
cription at the beginning of this section. The droplets are acce-
lerated through the nozzle by the pressure in the tank, and then
accelerated still more by the evaporation/disintegration process that
occurs due to the high vapor pressure of the water at 200 C (400 F) or
more. The resulting kinetic energy is the product of the nozzle effi-
ciency factor and the change in enthalpy of the steam/water mixture,
the two-phase flow. Calculations by the developers indicate drop velo-
cities of about 300 m/s (1000 ft/s) are attained. If the droplets
were roughly 100 pm in diameter, a typical value for other scrubbers
that use different atomization techniques, then impaction could well
be important even for particle diameters of 0.2 ym.
The droplets and the particulate matter captured by them are collected
in the cyclone that is part of the system. Particles on which water
has condensed may also become sufficiently massive to be collected in
the cyclone.
Gas flow is induced by the two-phase flow from the nozzle, so no fans
are required, but the pressure in the heating se<
(350 psi) and this must be maintained by a pump.
6 2
are required, but the pressure in the heating section is 2.3 x 10 N/m ,
COLLECTION EFFICIENCY
The particle concentrations in various size intervals for measuring
the efficiency as a function of particle size were determined by three
13
-------
different methods by Southern Research Institute. For particles
with aerodynamic diameters between 0.3 and 5 p.m, cascade impactors
were used. For particles with sizes less than 1 p.m, an optical par-
ticle counter (aerosol photometer) and the combination of diffusion
battery and condensation nuclei counter (CNC) were used.
The optical and diffusional system shown in Figure 5 included a
diffusion battery for size distribution, a condensation nuclei counter
for concentration and an optical counter for particle size and concen-
tration, with a dilution device which reduces the concentrations by a
2
factor of about 10 and diffusional driers to eliminate water conden-
sation effects.
Due to its complexity, only one set of equipment was available. The
data for the inlet and outlet were obtained in two separate attempts
with the assumption that the furnace process was sufficiently
repetitive.
The total mass concentrations for the inlet and outlet were sampled
by conventional (EPA Method 5) techniques. The locations for sampling
systems and the general layout of furnace and scrubber system are
presented in Figure 6.
The tests were conducted at a 7.5 MW submerged arc ferro-alloy furnace
at Chromasco, Inc.'s, Memphis, Tennessee facility, which operates
continuously, producing around 31 metric tons/day of ferrochrome.
The operations for a batch cycle are: charging, stoking, normal
operation, and tapping, within a nominal 120 minutes period. The
emission rates vary throughout the cycle, and this causes some dif-
ficulty in measurement and interpretation. However, the trends in the
fractional efficiencies and the fractions of the influent material
that penetrate the scrubber are estimated by the experimenters to be
reliable to within a factor of two.
14
-------
Flowmeters
Cyclone Pump
Process
Exhaust
Line
Neutralizer
Flowmeter
Particulate
Sample Line
Diffusional Dryer
(Optional)
Charge
Neutralizer Pressure
Balancing
Line
Recirculated
Clean Dilution
Air
Filter
Pump
Bleed
Figure 5. Optical and diffusional sizing system
15
-------
to
a
o
o
o
00
a
CO
CO
>>
CO
CD
,0
•§
VI
o
co
a
n)
cu
o
cfl
a
14-1
o
o
>>
rt
rH
td
J-l
cu
§
o
vO
cu
16
-------
The general configuration for the scrubber system has been shown in
Figure 2.
Waste gas at 590 C (1095°F) passing through the heat exchanger brings
up the water temperature from 43°C (109°F) to 207°C (404°F). The gas
leaves the heat exchanger at 104°C (219°F). Water at 207°C (404°F)
5 2
and 24 x 10 N/m (350 psi) jets out of the nozzle and meets the gas
stream in the mixing tube. The resulting spray jet acts as a fan to
produce the draft for the entire furnace and scrubber system. The
unevaporated droplets collect particulates and enter the separator,
from which clean air leaves at 62 C (143 F) and water with captured
particulate goes into treatment section for recycling or discharging.
Because of the relatively small duct dimensions and the relatively
fine particulate being sampled, single point sampling was used, at
isokinetic conditions.
Inlet and outlet mass loadings were different by from about 50:1 to
380:1. Therefore, it took 8 hours sampling time for outlet and about
5 minutes for inlet. The weighted average for all runs corresponding
to each size interval for both inlet and outlet are shown in Table 1.
Figure 7 (from the SoRI report) shows the average inlet particulate
size distributions on a cumulative basis. (That is, the graph is
total concentration smaller than a given diameter.) In Figure 8 (again
from SoRI) the inlet and outlet concentration data have been used to
obtain collection efficiency as a function of particle diameter. The
diameters used were equivalent diameters; for optical sizing, the
diameter of a polystyrene sphere giving the same light scattering; for
3
the impactor sizing, the diameter of a sphere of density 4.5 g/cm .
The impactor data were also presented by SoRI in terms of efficiency
as a function of aerodynamic diameter, in Figure 9, a useful presenta-
tion for those who would extrapolate these results to particulate
material of another density. The aerodynamic diameter at which the effi-
ciency was 50 percent was roughly 0.3 ym, obtained by correcting the
17
-------
Table 1. WEIGHTED AVERAGE INLET AND OUTLET LOADINGS, ARONETICS TESTS
Inlet
Size interval
(ym)
Loading
(mg/DSCM)a
Outlet
Size interval
(ym)
Loading
(mg/DSCM)
<0.5
91.5
<0.67
5.56
0.5
-1.2
75.5
0.67
-0.92
0.92
1.2
-1.6
57.2
0.92
-1.5
0.62
1.6
-3.0
34.3
1.5
-3.2
0.43
3.0
-4.3
27.5
3.2
-4.9
0.33
4.3
-7.2
29.7
4.9
-6.9
0.28
7.2
-13
61.8
6.9
-11.1
0.29
>13
199
11.1
-15.8
0.28
>15.8
0.39
Mg per dry standard cubic meter.
18
-------
CO
cc
UJ
t z,
i °
< 0
fO
E
o
en
<
cr
o
in
o
a:
UJ
UJ
ul^
I
o
oc Sec
UJ 5 <
< Q Q.
o +
E
cc
UJ
UJ
<
Q
UJ
O
p fe
00
en
to
o
oo
o
CM
CO
•H
CO
a)
§
O
1-1
(U
00
•H
-------
o
d
E
a.
UJ
UJ
o
UJ
o
•H U)
4-1 M
04 0)
O 4-1
0)
d e
o rt
•H
T3 T3
(U
CO 4->
Cfl C
Xi 01
^H
t-i cd
o >
3 o4
H CO
O> CJi O>
o> o>
en
o
01
o
CO
00
CVJ
0) «
rH
rH tS
2 C
d o
O -rl
•H CO
4J 3
CJ M-l
oo
a)
M
00
•H
20
-------
1 1 i
1 II 1 ' 1 ' 1
— •-
I
c
^ ^ ^
1
-
o
O -
>
o
_
.
^o
V
\^
""" 'P "™"
\
\
1
o
1
1 1 1
1 II 1,1,1
01
co cd
0)
60 CO
cd 01
)-J N
0) -H
> co
td
01
13 rH
01 CJ
1 I ,_J
60 )-l
•H cd
CD PU
o a, .
c- J3 CO
C 4-1 4J
•4 CO
•*• So)
t> O 4J
|_l
K M-l }-l
UJ 0
UJ 01 O
^ 4-J Cd
5 3 -HrH^
OrH ^^ ^*\
__ cd Mro
5 o H e
z QJ >-< e
O 'S rH
QC ^ C
UJ U 0) II
rf co co
M >i
CO 01 4-1
UJ U T3 -H
N 33 a
w g-«
0 0 4J ^
'-' UJ r4 0)
— _) <4 rH CO
0 w ^
b: 0) 3 0)
1- ,C O 4J
(K 4-1 0)
of 14-1 ^ cd
O T3 -H
f^ cd
0 CJ
C ^""* "iH
01 4^ S
•rl C Cd
CJ -H d
•rl H ?•>
14-1 pq -a
0) V4
4-1 0)
rH 01 Cd
a c w
O iH cd
•H
4J 0 C!
cj n 01
cd >
M 14-1 -H
Pn O Ml
_ ai
» OCOtOOCOiOQO O O ° 5
". 0j0>Oi0>0)0>0>OD (0 ^ 5P
> o> o <*> E
o/ * i fNnnioi i n LISNI i *%"n^-t/%/N
21
-------
mobility diameter for 50-percent efficiency, 0.15 ym, by the square root
of the particle density to yield the approximate aerodynamic diameter
of the 0.15 ym diameter particles.
Total mass collection efficiencies were obtained by Guardian Systems,
Inc., under subcontract to SoRI. Average efficiencies were 95.1 per-
cent for this dust, which has a mass mean diameter near 3 ym and a
mass mean aerodynamic diameter of about 6 ym. (See Figure 7.) The
geometric standard deviation appears to have been approximately 10, so
that there was a very substantial fraction (~30 percent) of the
particulate mass contained in particles with geometric diameters
smaller than 1/2 micron, which are often very difficult for particu-
late pollution control systems to remove from the gas stream.
POWER CONSUMPTION
The scrubber uses power from two sources: electrical energy to drive
the pump which supplies water to the heater and thermal energy from
the gas to provide the temperature increase in the water heater. If
the latter is obtained from heat normally wasted, then in a sense it
does not count as power consumption.
The power, P, consumed by the pump is
P = QLAPL/n
where P = power, watts
Q = liquid volume flow rate, m /s
2
ApT = liquid pressure, N/m
j_i
n = pump mechanical efficiency, 0.6.
22
-------
3 -33
For air flows of 0.47 m /s (1000 cfm), 26.5 x 10 m /min (7.0 gal/min)
f\ 9
water spray is needed; the pressure drop was 2.41 x 10 N/m , so
P = ((26.5/60) x 10~3 m3/s) (2.41 x 106 N/m2 /0.6J
P = 1.78 kW = 2.38 hp.
The intrinsic power (not including pump efficiency) is 1.43 hp/1000
cfm or 2.'.
gas flow.
cfm or 2.27 kW/(m /s) or 23 cm WC equivalent pressure drop for the
The waste heat power used P , can be obtained from the product of the
Wn
gas heat capacity, the gas mass flow rate and the change in temperature
of the gas in going through the boiler. Southern Research Institute
calculated this to be 17,000 Btu/1000 cfm which is 400 hp/1000 cfm or
630 UW/(m3/s).
SUMMARY
The Aronetics spray scrubber produces particle collection and gas motion
by a spray which derived its power from pumping and waste heat recovery.
The test results obtained by SoRI can be used with the efficiency and
cost comparison methods of SectionsXIII and XIV as part of an engineer-
ing analysis of applicability (Section XIV).
23
-------
SECTION V
BRAXTON SONIC AGGLOMERATOR
DESCRIPTION
3
The 7 m /s (15,000 cfm) prototype unit developed by Braxton Corpo-
ration and tested by GCA/Technology Division2 incorporated an electro-
magnetically driven piston to produce sound intensities of approxi-
2
mately 165 decibels (3.0 W/cm ) in a chamber 0.75 m (2.5 ft) in
diameter and 4.3 m (14 ft) in length, through which the contaminated
gas flows. Water is added to this chamber at its top (it stands with
its axis vertical) through Sonicore nozzles said to deliver spray
with droplets ~ 20 ^m. The sound frequency (-v370 Hz) and the chamber
length are tuned to produce a standing wave 4.5 wavelengths long.
The sound moves the particulate material with respect to the spray
and with respect to itself, causing sonic agglomeration of particulates.
The agglomerated material can more easily be removed than the unagglom-
erated material, thus the Braxton device is a pretreatment for fine par-
ticulate emissions to allow the use of a control device, such as a
cyclone, which is not an efficient collector of particles smaller than
a few microns.
POTENTIAL APPLICATIONS
The Braxton sonic agglomerator was conceived as a means of allowing
the use of cyclones or low energy scrubbers in the control of fine
particulates, obviating the use of high energy scrubbers, fabric fil-
ters or electrostatic precipitators.
24
-------
Particle Characteristics
Particle Size - The agglomerator is most useful for particles too small
to be collected by cyclones, <5 pm, yet not so small that they do not
impact upon other particles and drops available for agglomeration,
meaning particles >0.5 pm.
Particle Phase - This should not be important, although the likelihood
of agglomeration may be affected by certain surface properties.
Particle Resistivity - This should be immaterial.
Particle Abrasiveness and Corrosiveness - These will influence the
choice of construction material.
Gas Characteristics
Gas Flow - Variations in flow velocities should not be important, as
changes in residence time by a factor of two did not significantly
change performance of the agglomerator in the tests.
Gas Temperature - It would not be expected to be a major factor,
although the increased viscosity of high temperature gas would tend
to inhibit agglomeration and cyclone collection somewhat.
THEORY OF OPERATION
The GCA report gives a more detailed explanation of the theory than
2
presented here.
The standing acoustical wave set up by the sound generator produces
gas oscillation throughout the chamber. The smallest particles
oscillate with the gas, the largest particles and the drops remain
25
-------
almost unaffected by the gas oscillation. This produces a relative
velocity between the smallest particles and the larger particles and
drops, and these small particles are captured by the larger consti-
tuents through the various mechanisms discussed in Section XVII on the
general theory of obstacle collectors. Impaction, interception, and
diffusion would all be expected to play roles. The larger particles
and the drops grow at the expense of the smaller particle fraction,
then the larger particles and drops can be collected within the
agglomerator or by a control device (such as a cyclone) downstream
from the agglomerator. Because these fine particles are hardest to
collect, in general, the use of the agglomerator might require a less
costly method of particle control downstream from it.
COLLECTION EFFICIENCY
In this subsection we present the results of measurements and analysis
done by GCA/Technology Division as part of their evaluation of the
2
Braxton sonic agglomerator.
The sonic agglomerator is intended primarily as an exhaust precondi-
tioner to be used in conjunction with another collecting device such
as, in the present case, a cyclone. Because of the physical obstruc-
tion it presents to the gas stream, however, the agglomerator itself
will also have a certain collection capacity irrespective of its pre-
conditioning capabilities. A proper assessment of the effectiveness
of the agglomerator thus requires an investigation of the collection
efficiency of the whole system, as well as an investigation of the
collection efficiencies of its individual components; namely, the
agglomerator and the cyclone.
The method used to study the collection efficiency of the Braxton Sonic
Agglomerator System is shown schematically in Figure 10. Identical
sampling systems, consisting of appropriate sampling probes, Aerotec
26
-------
DRAIN
INLET
IM FACTOR
IM FACTOR
SAMPLER #2
Figure 10. Schematic diagram of test of Braxton Sonic
agglomerator
27
-------
model 1-1/2 cyclones, and Andersen six-stage cascade impactors, moni-
tored the gas streams at the inlet and outlet of the sonic agglomera-
tor. The larger-size dust fraction was separated by the cyclone,
allowing the more interesting undersized portion to be analyzed with
the cascade impactor. More details are available in the original
report.
The different particle concentrations obtainable from the test con-
figuration were:
• Total mass concentration
• Mass concentration of particles larger than cyclone cut-
off size
• Mass concentration of particles smaller than cyclone
cut-off size
• Size distribution of particles smaller than cyclone cut-
off size
These were available upstream and downstream from the sonic agglom-
erator.
The inlet aerosol concentration for a given particle size is N . The
concentration which reaches impactor No. 1 is N :
NT = (1 - E ) N
1 co
where E is the cyclone collection efficiency at that particle size.
The cyclone collection efficiency as a function of particle size
(Aerodynamic diameter) was taken from its manufacturer's specifications,
The concentration at the outlet of the agglomerator is N where
N = (1 - E ) N
a a o
and E is the collection efficiency of the agglomerator, again at the
a
particular particle size.
28
-------
The concentration which reaches impactor No. 2 is N.:
Concentrations as functions of particle size (aerodynamic diameter)
were obtained from the cascade impactor measurements. Thus, N?(d )
would be the mass concentration as determined by each stage of the
outlet cascade impactor, with d being a particle size which is
characteristic of the material caught on each stage. The collection
efficiency as a function of d for agglomerator plus cyclone is calcu-
lated from the ratio N./tT where N-, is the stage-by-stage mass con-
centration as calculated from the results of sampling with the cas-
cade impactor on the inlet. Because the impactors at the inlet and
outlet of the agglomerator were the same, the collection efficiency
of the agglomerator (for particle size d ) is simply N-/N.., with N?
and N, being the stage-by-stage mass concentrations from the impactors
No. 2 and No. 1.
In the field study, the data were generated using a cyclone efficiency
curve based upon a particle specific gravity of 2.5 and a cascade
impactor cut-off size based on a particle specific gravity of 2.0.
To facilitate comparisons, the data have here been converted to aero-
dynamic diameter size intervals. These data are summarized in
Table 2.
Table 2 lists several collection efficiencies versus particle aero-
dynamic size intervals traceable to the cascade impactor sizing inter-
vals. The measured values are the direct basis for collection effi-
ciency of cyclone plus agglomerator. The manufacturer's data was in-
terpolated to give the cyclone efficiencies shown, and the efficiency
29
-------
of the agglomerator only was calculated from the other two values
(the penetration through both agglomerator and cyclone being the pro-
duct of the penetrations through each). Figure 11 presents these in
graphical form. From this information we can deduce that the
agglomerator-cyclone combination had 50 percent efficiency at particle
aerodynamic diameters around 2 ym; cyclone 50 percent efficiency
occurred near 3 pm.
Table 2. FRACTIONAL COLLECTION EFFICIENCIES FOR THE BRAXTON
SONIC AGGLOMERATOR SYSTEM, CYCLONE, AND
AGGLOMERATOR/CYCLONE COMBINATION
Size interval
aerodynamic diameter
(in urn)
<0.5
0.5-1.0
1.0-2.0
2.0-3.2
3.2-5.5
5.5-9.2
9.2-14.1
Agglomerator
and
cyclone
59.407o
42.90
47.09
55.28
88.21
87.80
96.91
Agglomerator
only
56.35%
29.95
20.44
13.99
65.00
26.08
43.89
Cyclone
only
7.0%
18.5
33.5
48
66.3
83.5
94.5
Some other results of these tests are worth noting. Tables 3 and 4
show the conditions of the tests and the collection efficiencies with
respect to total mass and to the fine fraction escaping capture in
the cyclone. Reduction in fine particles (d < 3 urn) ranged from 14
P ~~
to 83 percent. The following correlation coefficients were found for
the reduction in fine particles using sound, water, and steam:
sound - 0.68
water spray - 0.65
steam - 0.06
30
-------
99.9
99.8
o
oc
UJ
0.
o
z
UJ
o
u.
u.
UJ
z
o
CJ
UJ
_l
_J
o
o
99.5
99
98
95
90
80
50
10
SONIC AGGLOMERATOR
x PLUS CYCLONE
• AGGLOMERATOR ONLY
O CYCLONE ONLY
I I I I I
O.I
0.2
0.5
1.0
2.0
UJ
Id
a
5.0 <
H-
UJ
UJ
a.
10
20
50
100
23456789
AERODYNAMIC DIAMETER , pm
10 II 12 13 14
Figure 11. Collection efficiencies versus aerodynamic diameter
for cyclone, agglomerator, and cyclone plus
agglomerator
31
-------
S
w
PM
O
H
W
>H
w
ttJ
O
H
$
g
O
inOm-sfo m oo
OOOO-j^-*aooo
cMCMCMCM^t-j'-Jtnvnin
a>
0)
•o
V
CMCMCMCMrMCMCM^-^Nj w
CMCMCMCMCMCMCMCMC^CMQ
OOOOOOOOOOO
0.0.0.0.0.0.0.0.0.0.0.
33333333333
cncomm mmmmcn
-^-^^^r^-^----— ^.-^«^'^.^^.-^.CM
^ O O
0 OJ 2 " °
CM CM CM 00 CM
oo r- r* ,
efflclem
cabined
e
«
i
b
o
2
§
41
id
jQ
O.
U.
1
o
1
U
I
•
o
.0
*
•rt
U
IM
IM
«
S
2
S
O
U
r«»
"8
•
1
|
2 i.
32
-------
a
w
o
Ed
£J
S
*\
g
S8
3 N
M en
£
, CYCLO
\RTTCLE
O PM
h j
§[ "]
^^
(J fa
o
H
co 2;
•— , r*j
Q M
H^
W W
iJ n4
H *1
D O
O O
<3 H
CO
H >-"
W CO
LI T
^
fa 2
E?
§°
CO Q
M g
B< <
SS H
SCO
J33
O O
•
-4"
,— 1
J3
CO
H
0
§
4J
u
3
•D
W
(A
1;
c '
1 Reduce Ion
r-4
0)
3
a*
y-o
•H V
-M b
s«2
4J
n«M
«D
M 4J
Sa
y
>-« u
j a
o
u
M
V
-4
U
r4
n a)
a u
u
4
M
M
0
fc*
<
c
ai£
percent
•^ x 100
V
8
M
8
M
Cj
s «
E
V O
0 4J
u a
u ^
1
U
(0
U M)
y
C
O
c
M O
i-l U
a
0 w
y «
>> M
c
o
4J
8*i
u
1
3SSSSSRS2SSR8SS8
I 1
^223K5S§K55^S22S
g£332S£233£KS£SS
ooooooooooo-^oooo
1
22o2S°!OS22°oSo8S
ISiR22SS£SSSs§S5
33
-------
The first two were statistically significant. GCA also put this in
the form of a multivariate regression formula:
R = -15.1 + 6.5 W + 8.3 S
where
R = percentage reduction in fine particle mass
— S *^
W s spray water flow, 10 nr/s
S = sonic generator power, kW
The -15.1 in the equation reflects the interesting observation that
although the sonic agglomerator collected, by itself, from 42 to 59
percent of the particulate matter entering it (fly ash, cupola dust)
even without steam or sound or water, the agglomerator increases the
fraction of fine particles under these conditions, presumably due to
shear de-agglomeration.
POWER CONSUMPTION
Most of the field tests on the prototype model were conducted with
2
gas flows on the order of 2.4 m /s (5000 cfm). Practical applications
will require higher gas flows; the power requirements for this system
o
have been calculated per 1000 cfm (0.47 m /s).
Power consumption is oased on the relationship:
P = Q AP
in which P is the power; Q the gas flow rate; and Ap the pressure
o
drop across the system. Since the flow rate is assumed (0.47 m /s) ,
the only variable needed to calculate the power consumption is the
pressure drop, ^p, which can be estimated from cyclone design theory.
The following calculations of the pressure drop for the scaled version
of the cyclone system to be used with the Braxton Sonic Agglomerator are
34
-------
3
based on a recent review of cyclone design theory by Leith and Mehta.
In cyclone theory, the pressure drop is often expressed in terms of
inlet velocity heads AH:
in which K is a constant (K = 16 for a standard tangential inlet);
a is the gas inlet height; b is the gas inlet width; and D is the
cyclone exit duct diameter. If one of the dimensions of the cyclone
is known (e.g., b, the width of the inlet duct), the others can be ob-
tained from standard dimension ratios. For a Stairmand type cyclone,
a/D =0.5, b/D =0.2 and D /D = 0.5, where a, b, and D are as defined
above and D is the diameter of the inlet duct. !
The gas inlet width, b, can be calculated from:
1/2
d
pc 2 ftp v N
^ O
where d is the particle aerodynamic cut-off diameter, the aerodynamic
diameter (Appendix) for which the cyclone penetration is 50 percent;
Vi is the gas viscosity (1.8 x 10~ poise); p» is the particle density:
O
v is the gas velocity and N is a constant (usually 5). The gas
velocity, v, can be calculated from the relationship:
v = Q/A
O
in which Q is the gas flow rate (m /s) and A is the area of the inlet
2
duct (note - from the standard design dimension ratios, A = ab = 2.5 b )
Using the Aerotec cyclone manufacturer's value for the cut-off diameter,
d =2.76 ym, the inlet width, b, can be calculated from the equations.
Knowing b as well as standard design dimension ratios permits AH to
be calculated. The pressure drop in terms of inlet velocity heads, AH,
can then be converted to static pressure AP, required for the power
calculation, by the following expression:
35
-------
v2 p AH
-3 3
in which v is the gas velocity; p , the gas density (1.2 x 10 g/cm );
8 2
g, the acceleration of gravity (980 cm/s ); p , the reference liquid
3
density (1 g/cm ).
Using the last four equations, the pressure drop, Ap, required to pass,
3
say, 4.7 m Is (10,000 cfm) of gas through a single Stairmand cyclone
5 2
with a cut-off diameter of 2.76 ym is 7.5 x 10 N/m (7.5 atm), which
is impractically high. However, multiple cyclones, for example a
3
group of 10, each capable of handling 0.47 m (1000 cfm), can be
arranged to provide the equivalent throughput with a much lower pres-
sure drop. For this arrangement, these equations predict a pressure
drop of 7.5 x 103 N/m2 (75 cm WC), which would be 4.75 hp/1000 acfm
or 7.5 kW/(m3/s), still quite high.
A low-power venturi scrubber with a particle cut diameter of 3 ym
would give roughly equivalent collection efficiencies if used rather
2
than the cyclone; its pressure drop would be 300 N/m (3 cm WC)
4
approximately, much less than the cyclone.
In addition to the cyclone power consumption, the power requirements
of the sonic generator as well as the power needed to move gas through
2
the agglomerator should be considered. In the prototype tests, for
3
gas flows of 2.2 m /s the sonic generator operated typically at 3 kW.
3
Using a simple multiplier scaling, the power necessary for 0.47 m /s
should be on the order of 0.6 kW or 0.8 hp/1000 cfm.
Operating the agglomerator itself requires a pressure drop of approxi-
2 2
mately 750 N/m (7.5 cm WC), which at the flow rate of interest
(0.47 m3/s), would require, 0.35 kW or 0.47 hp/1000 cfm.
36
-------
The overall power requirements then for whole system: the high effi-
ciency cyclone, the sonic generator, and the agglomerator should be
3
about 4.5 kW at 0.47 m /s or 6.0 hp/1000 cfm, equivalent to 96 cm WC
pressure drop; substitution of the low-energy scrubber for the cyclone
3
lowers this consumption to about 2.4 kW/(m /s) (1.5 hp/1000 cfm),
equivalent to a pressure drop of 24 cm WC.
WASTE DISPOSAL
The material caught by the agglomerator itself will flow out of its
liquid exhaust stream as a sludge, which would be handled as scrubber
sludges are. The material collected by the control device used in con-
junction with the agglomerator would be handled as it normally is.
37
-------
SECTION VI
CENTRIFIELD SCRUBBER
DESCRIPTION
This material is taken from the manufacturer's literature as cited by
Southern Research Institute in its evaluation.^ "Modern, high-efficiency
wet scrubbers for the removal of entrained particulates from industrial
exhaust gases depend, almost without exception, on spray contact for
particle interception. The atomization and acceleration of the scrubbing
liquid into contacting zone can be performed by a wide variety of mechan-
ical arrangements. A new high-performance contactor, the centripetal
vortex contactor, is now available.
"The centripetal vortex principle is illustrated in Figure 12. For
any point on radius r and tangential gas velocity v , a particle may
exist whose terminal velocity toward the vortex periphery exactly equals
radial gas velocity v . If the particles consist of liquid droplets,
and the peripheral inlet vanes are of the proper configuration, the vortex
selectively develops a specific droplet diameter distribution within the
rotating field, as determined by r , v., and 0. The average droplet di-
ameter created by a field of this type is substantially smaller than that
by a conventional spray contactor at the same energy level. Droplet di-
ameter is one of the basic parameters of spray contactor efficiency.
"Gas scrubber performance is most conveniently described by the relation-
Y r i
ship: Nt = aP , where Nt = loge I/(1-collection efficiency)! and
P = energy input per unit of gas treated. For a given dispersoid, an
38
-------
CENTRAL RISER
STATIONARY
VANE CAGE
DROPLET
FORMATION
A-A
INLET VELOCITY
VECTORS
INLET VANES
LIQUID INLET
Figure 12. Centripetal vortex balances gas velocity
against centrifugal force
39
-------
improvement in gas scrubber design results in an increase in a. The
centripetal vortex principle should theoretically increase a by approx-
imately 20 percent over a conventional contactor. Laboratory and field
tests show .an increase of 15 to 25 percent. This is a significant
improvement in scrubber performance in relation to energy input.
"The droplet formation within a spray contactor is inherently in co-
current flow in respect to the gas stream, which limits its gas absorption
efficiency to a maximum of one theoretical plate per stage. Nevertheless,
spray contactors, in single or multiple stage arrangements, are often the
only feasible mass transfer device, especially in applications where in-
soluble particulate is present, or where the product of the absorption
process tends to foul conventional packing and sieve plates. Mass transfer
efficiency in a spray contactor is, as would be expected, a function of
L/G ratio and specific droplet surface area. Since specific surface area
varies inversely with droplet diameter, it is reasonable to expect that
the centripetal vortex, with its unusually small equilibrium droplet diam-
eter, would exhibit superior mass transfer capabilities. Laboratory and
field data have confirmed this expectation. Under circumstances where an
active reagent is involved, mass transfer efficiencies approaching equi-
librium have been recorded at nominal pressure drops."
POTENTIAL APPLICATIONS
The Centrifield scrubber would be expected to be applied to situations in
which other scrubbers would also be candidates such as steel mills, found-
ries, refineries, paper mills and other industries with fine, sticky or
liquid emissions. The device tested for EPA was in use on an asphalt
plant.
Particle Characteristics
Particle size - The scrubber was found to give 99 percent or better col-
lection efficiency for particles with aerodynamic diameters larger than
40
-------
3 to 4 /im, but less than 80 percent for particles with aerodynamic diam-
eters from 0.1 to 1 fxm.
Particle Phase - Liquids or solids should be acceptable.
Resistivity - Collection should be independent of resistivity.
Abrasiveness - Special materials may be necessary in certain cases.
Corrosiveness - Similar considerations apply as for abrasiveness.
Gas Characteristics
Gas Flow Rate - For the same geometry, increasing the volume flow rate
will linearly increase the inlet gas velocity and increase by the square
the centrifugal force on the droplets; the equilibrium droplet size would
be expected to change as the inverse square root of the volume flow rate,
not a strong dependence. The relative radial velocity of the gas and
droplets would increase with the gas volume flow, enhancing impaction.
The rate of droplet evaporation will increase due to the ventilation
factor. The foregoing sketches some of the complexities involved in
predicting the effect of flow rate. Presumably, the scrubber is designed
to be maximally efficient at a given flow rate, with somewhat degraded
performance when this optimum is not met.
Gas Temperature - Almost all particle collection mechanisms, except dif-
fusion (significant only for particles ^0.1 /xm diameter), are inhibited
by higher gas temperatures, and the problem is aggravated in scrubbers
by the change in volume flow rate and the increased droplet evaporation,
both generally to the detriment of collection efficiency.
41
-------
THEORY OF OPERATION
The droplets formed in the Centrifield Scrubber collect particles by
mechanisms such as impaction and diffusion, as outlined more fully in
Section XVII on general theory of collectors. The flow geometry of the
scrubber is said to be such that the droplets are formed so as to be
smaller than those of conventional scrubbers, for the same power input.
Inertial impaction and diffusion are both enhanced by decreases in drop-
let size, other things being equal. The device also is said to maintain
a high concentration of droplets in the scrubbing section by balancing
the velocity which would occur due to "centrifugal force" against the
radial component of the air flow velocity. This would also be expected
to increase scrubbing efficiency. Offsetting some of these advantages,
however, is the fact that the droplets which remained in the scrubbing
section would reach the same tangential velocity as the gas/particle mix-
ture and the collection due to relative motion (impaction and interception)
would be then due only to relative radial velocities. These potentially
important details aside, the device acts basically in a manner similar
to that of other spray scrubbers in collecting particulates and, to a
degree, gases.
COLLECTION EFFICIENCY
Southern Research Institute (SoRI) tested the Centrifield Scrubber in
operation on a four-ton asphalt plant equipped with a dryer having a
capacity of 250 ton/hr and burning No. 2 fuel oil. The plant was oper-
ated intermittently. Total mass efficiency measurements were made by
taking simultaneous representative samples upstream and downstream by
standard methods (EPA "Method 5") giving values of 99.50 and 99.73 per-
cent collection efficiency on the particulate material, which had a
mass mean diameter around 100 um. Efficiencies as functions of particle
size were obtained by using the following instruments upstream and
downstream:
42
-------
• impactors
• optical counter
• diffusion battery/condensation nuclei counter
• electrical mobility analyzer
The impactors were used at inlet and outlet test points (Figure 13) during
the same run; the other equipment was used either on the inlet or the
outlet during the run. Rather rapid variations in concentration due to
variations in the process made accurate assessment of true mean particle
concentrations difficult, leading SoRI to estimate the final precision
to be about a factor of two.
Figure 14 is taken from the SoRI report and shows the collection ef-
ficiency as a function of particle size for the Centrifield scrubber they
tested. The particle sizes determined by optical measurements were con-
verted to geometric sizes assuming the light scattering to be that due
to polystyrene latex of the size indicated; the geometric size was
estimated from impactor measurements by assuming spherical particles
3
of density 2.5 g/cm . Minimum collection eff:
a few tenths micron in diameter, as expected.
3
of density 2.5 g/cm . Minimum collection efficiency occurs for particles
Figure 15 is another figure taken from the SoRI report on the Centrifield
scrubber. Here efficiency is graphed against aerodynamic diameter as
determined from measurements using cascade impactors at the inlet and
outlet of the scrubber. The curve of efficiency versus aerodynamic diam-
eter is directly comparable to other such curves obtained for novel con-
trol devices in other evaluations in this series. The aerodynamic diam-
eter at which the collection efficiency is 50 percent appears to be
between 0.6 and 0.7 (j,m.
-------
EXHAUST DUCT ADDED
FOR TEST. PROG RAM
FLOW '
STRAIGHTENERS
KILN
AGGREGATE,
SIZING.WEIGHING,
AND MIXING
INLET
TEST POINTS
Figure 13. Schematic diagram of the asphalt plant and scrubber layout
and the locations of the sampling points used in the tests
44
-------
— T i 1 i -" " rnr -_
a 1 1
>» — **
«•
™^» O *"••
^_
o
. •*•
o
o —
o
— o •—
I ° * ° -
a
^^
40
O
o
o
^^
i— ~z
~ • _
— ~ j "^
- si^
_ 0£ tjr ^ J3 ^
1— —» ^j ^— . ^^
^ li_ L QC . ^"^
"™" Lu A ^ """
iu 3 o 5
1 1 1 1 °
- o O < o — '
i r i i i ii
S
C 0)
••-1 M
•O
a) con
CO 4J 0
3 cfl o
•o -•»
CO CO
-a i-i B
o o S
4-1 O 60
01 CO
M i-l CM
O 0) <4-l
»W J3 0
01 >>
O J3 H 4->
-J 4J O i-l
M-l CO
1 £g S
* 5 -^
• "O O
pC 0,0, C> M
% 'AON3IDUJ3
45
-------
99.99
99.9
99.8
jo 99
>T 98
o
z
<3
u.
u.
UJ
| 80
O
UJ
90
o 60
40
10
I I f I I
0.5
1.0 10
AERODYNAMIC PARTICLE DIAMETER,*!™
Figure 15. Fractional efficiencies of the Entoleter Centrifield
Scrubber as determined from cascade impactor data
46
-------
POWER CONSUMPTION
The Centrifield scrubber tested by SoRI has a pressure drop of 17 cm WC
or 6.7 incl
1000 acfm.
3
or 6.7 inches WC. This is the equivalent of 1.66 kW/m Is or 1.06 hp/
-33 3
Water consumption was measured to be about 0.5 x 10 m water per m
3
air or 4 gal/1000 ft .
WASTE DISPOSAL
As with any scrubber, the waste will be in the form of waterborne solids,
which will either be sent to a receiving body or treated. More detail
is given on the section on waste treatment, within Section XIV.
47
-------
SECTION VII
CLEANABLE HIGH EFFICIENCY AIR FILTER (CHEAP)
DESCRIPTION
The CHEAP operates by drawing gas and particulate material at high
velocities through a porous polyurethane foam structure which is wetted
by a water spray. The system has recently been purchased from the Johns-
Manville Corporation by Andersen 2000, Inc. of Atlanta, Georgia.
POTENTIAL APPLICATIONS
Based on our limited information, the following considerations would
apply to applications of the CHEAF system.
Particle Characteristics
Particle Size - Collection efficiency in the tests performed for the
Industrial Environmental Research Laboratories - Research Triangle Park,
EPA, showed essentially no effect of particle size on efficiency. It
is possible that preferential collection efficiency with respect to
particles larger than a micron were offset to a degree by penetration
of agglomerated materials of nearly the same size.
Particle Phase - Presumably the CHEAF would find insoluble viscous drop-
lets to be more difficult to clean than other aerosol types.
Resistivity - This should have no influence on the CHEAF*s performance.
48
-------
Abrasiveness - By locating the air-mover downstream from the control
device, its abrasion can be minimized.
Corrosiveness - This is potentially a problem in any system, and can
usually be handled through suitable choice of materials.
Gas Characteristics
Gas Volume Rate of Flow - The collection efficiency would be expected
to be somewhat dependent upon flow rate.
Gas Temperature - The CHEAP can be operated with a quenching system to
reduce the temperature to within its operational range.
THEORY OF OPERATION
The collection mechanisms would be those described in Section XVII on
general theory of obstacle collectors. The porosity of the foam is
90 percent and the characteristic collector dimension of the 65 pore per
2
inch foam (26 pore per cm) is on the order of 10 ym. The ratio of the
-4
water spray rate to the air volume flow rate is about 10 (about 1 gal/
1000 cfm). As part of an expanded evaluation of the CHEAP, GCA/Tech-
nology Division is testing it on laboratory-generated iron oxide aerosols
and formulating a model to describe its collection mechanisms.
COLLECTION EFFICIENCY
The evaluation was done by Air Pollution Technology, Inc. (APT).
Cascade impactors were used upstream and downstream from the CHEAP
3
while it was in operation on a 53.2 m /s (25,000 acfm) diatomaceous
earth calcining and drying process. The CHEAF was downstream from a
cyclonic precleaner with water sprays. Saturated gas (at 63 C) went
49
-------
from the cyclonic precleaner through flow straighteners, past the inlet
sample port to the CHEAP, then to a blower and to the main stack, where
the outlet ports were. This is shown schematically in Figure 16. APT
used heated University of Washington Mark III impactors for the size
distribution work and attempted to use diffusion batteries with con-
densation nuclei counting as well, without success. The total mass
concentrations were measured with the standard equipment and procedures
usually referred to as EPA Method 5.
Even with a device used to remove the spray droplets at the inlet of the
impactors, there was quite a bit of difficulty due to moisture. Several
successful tests were achieved, however, and Figure 17 is derived from the
APT results, showing collection efficiency as a function of particle
aerodynamic diameter. As APT noted, the efficiency is almost independent
of particle size in the range measured.
Because the diatomaceous material is hygroscopic, there may have been
considerable growth in particle size due to moisture addition. It is
typical of such growth that the final particle sizes of the particles
are much more nearly the same than were the dry particles. The particles
were dried before sizing, masking any growth. This might be the explana-
tion for the nearly uniform collection efficiency versus particle size.
This is to be studied further in the tests underway by GCA.
Figure 18, from the APT report, shows inlet and outlet size distributions,
in cumulative form - the percent smaller than a given aerodynamic versus
the aerodynamic diameter. The distributions are nearly identical, with
mass median aerodynamic diameters of 0.8 jim and geometric standard devia-
tions of 3.8. The total mass efficiencies on this aerosol, according to
the Method 5 sampling results, were 95.6 percent and 94.7 percent.
50
-------
v>
c
o
•rl
4-1
«
rH
[0
C
•H
g
M-l
O
I
M
00
rt
•H
O
vO
3)
TT / // /
j)
O
o
S
2
^
51
-------
99.9
99.8 -
99.5
o
o
99
Ul
o
ac
UJ
a.
i 9e
UJ
o
O
Ul
90
80
50
10
0.
0 I 2
5 6 7 8 9 10 II
AERODYNAMIC DIAMETER , jim
12
O.I
0.2
0.5
1.0
2.0
o
a:
UJ
a.
5.0 2
ui
10
20
50
15 14
100
Figure 17. Collection efficiency versus aerodynamic
diameter, CHEAP
52
-------
E
a.
(£.
LU
Ul
S
O
o
o
-------
POWER CONSUMPTION
Pressure drop across the CHEAP system was reported to be about 53 cm WC
(or 21 inches H-O). This corresponds to 3.3 hp/1000 acfm or 5.2 kW/(m /s).
This is the hydraulic power, the actual electrical power consumption would
be obtained by dividing the hydraulic power by the efficiency factor for
fan and motor combinations, usually about 0.6.
WASTE DISPOSAL
We have very little information regarding the wastes produced by the
CHEAP system, but it is likely that they would be similar to the wastes
created by scrubbers and would be handled similarly.
54
-------
SECTION VIII
DYNACTOR SCRUBBER
MANUFACTURER'S DESCRIPTION7
The Dynactor uses a proprietary nozzle design to produce a water spray
which serves as an air mover and as a scrubber, thus cleaning and pro-
pelling the gas simultaneously. A description of the Dynactor, written
by its manufacturer (RP Industries, Hudson, Massachusetts), follows:
"The continuous gas/liquid contactor, the Dynactor, is a proprietary
development of RP Industries, Inc. Figure [19] is a cross section of a
single stage Dynactor diffusion contactor. Liquid entering the system
under a pressure of 140 to 200 pounds per square inch (typical) is
atomized into thin films and droplets of average thickness or diameter
less than 1/64 inch. This liquid discharge diffuses or expands into the
reaction chamber causing air or gas to be aspirated by being trapped
within the moving shower of films and particles. The resulting mixed
fluid then continues to travel down the reaction column with intimate
contact maintained between gas and liquid. This causes physical and
chemical equilibria to occur by the time the mixed fluid exits from the
reaction column into the separation reservoir. The Dynactor can be
viewed as a macroscopic diffusion pump which makes use of diffusion
principles in order to aspirate large volumes of air per volume of motive
liquid. By utilizing diffusion rather than Bernoulli principles, the
Dynactor aspirates up to 4,800 standard volumes of gas per volume of
motive liquid. In comparison, venturi eductors will aspirate not wore
than 100 volumes of gas per volume of motive liquid.
55
-------
LIQUID INPUT, 140 TO 200psi
AIR INPUT, LOW VELOCITY, AMBIENT PRESSURE
HIGH VELOCITY, SUB-AMBIENT PRESSURE
SHOWER OF THIN FILMS AND PARTICLES
REACTION COLUMN
TURBULENT
GAS OUTPUT
LIQUID LEVEL DETER-
MINING TRAP
LIQUID OUTPUT
Figure 19. Single-stage Dynactor diffusion system
cross sectional view
56
-------
"Because there are no venturi or other constrictions in the Dynactor,
energy requirements are considerably lower than for conventional jet
or venturi eductor systems. If the gas carries small solid particles
along with it, such as activated carbon or powdered neutralizing and pre-
cipitating agents, such particles are wetted and captured by the liquid
throughout the entire length of the reaction chamber. By contrast,
venturi wet scrubbers make effective contact between gas and liquid only
in the constricted throat region. Contact time, therefore, in the
Dynactor is about 20 times longer than in venturi devices.
"Just as in oil and mercury diffusion vacuum pumps, it is also possible
to construct Dynactors having multistage gas inputs." Figure 20 is a
drawing of the two-stage Dynactor diffuser system employed in these
studies. "The internal configuration was constructed to maximize gas/
liquid turbulence and contact throughout the length of the 6-foot long,
12-inch diameter reaction column."
The unit tested by GCA/Technology Division was one of the smaller units
3
of its type, a two-stage device with a nominal rating of 0.47 m /s
(1000 cfm), Model DY 12 F2. Water was recirculated from the reservoir
through pumps (one for each stage) to nozzles and was changed only after
each test.
POTENTIAL APPLICATIONS
The Dynactor scrubber would be a candidate for application in those areas
where scrubber technology has been shown useful, such as metallurgical
operations, refineries, paper mills and industries which have fine,
sticky or liquid emissions. Dynactors are used in textile processing to
collect chemicals and plasticizers, in food processing to remove water
soluble odors, in the plastic industry to collect chemicals and plasti-
cizers, to condense hydrocarbons and remove polymer odors, to purify kiln
outputs after cyclone preseparation, and to hood cupolas in foundries.
57
-------
Nwmoo
o
4-1
NOIID3NNOO
33VISH3INI
NWniOD
H
S
Q
0)
t>0
CO
4-1
(0
O
O
CM
(1)
M
3
60
•H
•*^«
58
-------
Applicability is discussed in more general terms next, with respect to
particle types, gas characteristics, etc.
Particle Characteristics
Particle Size - The Dynactor was found to give efficiencies greater than
99 percent for particles larger than 4 |_im aerodynamic diameter, but sub-
stantially less than 90 percent for those smaller than 1 urn.
Particle Phase - Solids were used in the tests for evaluation (fly ash
and iron oxide), but it is expected that liquids (droplets) would be
collected equally well.
Resistivity - Resistivity does not affect collection efficiency.
Abrasiveness - Abrasive aerosols may have a detrimental effect on water
pumps and nozzles if they appear as suspended solids in recirculated
water. The test series using abrasive aerosols (fly ash and iron oxide)
did not noticeably effect pumps or nozzle after a total operating time
of about 40 hours.
Corrosiveness - RP Industries constructs Dynactors from stainless steel,
polypropylene, PVC, and fiberglass reinforced epoxy and polyester.
Other construction materials are possible for difficult materials
handling problems.
Gaa Characteristics
Volume Rate of Flow - The largest Dynactor now available will move
3
24 m /s (50,000 cftn). Dynactors may be connected in parallel. RP
Industries plans to develop a 470 m /s (10 cfm) unit.
59
-------
Temperature - The Dynactor was tested at 93°C (200°F) with an insig-
nificant decrease in collection efficiency. Hotter gases, fed
directly into the Dynactor, may be more difficult to scrub. In one
application, 830°C (1500°F) effluent gas is prechilled with water
spray before flowing to the Dynactor.
Pressure - For proper Dynactor operation, the inlet gas should not be
2
in vacuum in excess of 125 N/m (1/2 " HO). This would not be
troublesome for stack gas which is pressurized compared to atmosphere,
but may present difficulties for some applications.
For some Dynactor installations, existing ducting may have to be
shortened and/or widened to reduce flow resistance.
THEORY OF OPERATION
Particle Collection
As discussed in Section XVII on general theory of collection devices,
this scrubber works primarily due to the motion of the collecting drop-
lets with respect to the particle-laden gas stream. Details o-f this
process in the Dynactor are proprietary or unknown. The two most sig-
nificant collection mechanisms are expected to be impaction, which
improves with increasing particle size, and diffusion, which improves
with decreasing particle size: The interaction of these two mechanisms
is expected to give a collection minimum ~ 10 fxm, as observed. Other
mechanisms expected to contribute to collection: sedimentation, inter-
ception, agglomeration, condensation.
60
-------
Air Induction
The spray from the Dynactor nozzles entrains air and produces air motion,
helping to overcome resistance to air flow in the Dynactor and the duct-
work to which it is connected. The nozzle hydraulic power is the product
of its flow rate and pressure:
Pn
where
P = nozzle hydraulic power
Q = water flow rate
*
Ap = nozzle pressure
a
At most, all this power would be transmitted to the air so that
Q
The subscript a is for air.
Test measurements indicated:
Q = 2 x (9.2 gal/min) = 2(0.58 x 10"3 m3/s)
Ap = 200 psig = 12.8 x 105 N/m2
n 3
Q Ap = 1.48 x 10 W
& n
Qa = 0.25 m3/s = 530 cfm
Ap = 2.7 x 102 N/m2 - 1.1 " H,0
3. £•
Q Ap = 67 W
3 3
61
-------
The difference is due to various losses and increases with increasing
flow rates. Still, this represents a boost for the flow rather than a
flow impediment.
COLLECTION EFFICIENCY
Test Procedure
The evaluation of the Dynactor scrubber was with respect to its mass
collection efficiency as a function of particle size, the effect of
several parameters on this efficiency, the air-moving capability and
power consumption of the device, and its cost. Measurements included:
• Air flow and pressure gain versus spray nozzle pressure
• Electrical power consumption versus spray nozzle pressure
• Mass collection efficiency as a function of particle aero-
dynamic diameter at two levels of flow, temperature, and
concentration, for two different dusts, in a balanced
test matrix
• Mass collection efficiency as a function of particle size
for several additional sets of conditions
• Total mass collection efficiency at the conditions noted
above
Test Equipment
Most of the experimental work was done with the equipment shown in
Figures 21 and 22. This equipment allowed measurement of collection
efficiency as a. function of particle size and the effects of differ-
ences in flow, concentration, temperature, and pressure drop which
prevailed during the efficiency tests.
62
-------
M
0)
00
C
•H
-d
t-l
o
a <-»
O PM
cd cu
3 oo
H 3
cd cd
> oo
cu
CU
M
cu
JO
•s
n
o a.
co
co
ca
cu
a) C
00 ct)
(0
4J «.
CO X-N
I H
4-> CO
i-i
t-l CU
O 4J
4J QJ
o e
cd o
c s
P cu
X
l-t -U
CO CO
>» v<
ca cu
r-\
CX
CO
a)
H
en
CM
cu
M
oo
•H
63
-------
CO
a
o
u
-------
Figure 21 gives the overall picture of the test setup. Dust from a
screw feed was picked up by an air ejector/aspirator and blown into
ducting leading to the Dynactor.
Upstream, the turbulent mixture was sampled five or more duct diam-
eters from the dust feed by a heated Andersen Model III cascade
impactor, usually run with isokinetic flow, and by a filter assembly
which always was operated isokinetically. An identical sampling com-
bination was used downstream, as described more fully in Figure 22. The
temperature and pressure of the mixture entering the Dynactor was
measured as was, sometimes, the temperature of the mixture leaving the
Dynactor. The pressure drop or gain across the device was measured.
Upstream and downstream from the Dynactor, the Andersen Model III in-
stack impactor was used to size-fractionate the aerosol. The impactors
were used with the glass fiber media impaction substrates designed for
them. The temperature of the impactor at the Dynactor outlet and its
drying section (Figure 22) was kept about 20°C above the temperature of
the Dynactor exhaust stream to produce drying of the droplets present
in the exhaust. To obtain enough material on the downstream impactor
for weighing and yet not too much material on the upstream impactor
(to prevent rebound and reentrainment), it was necessary to run these
impactors for different total durations, the downstream impactor sampling
for about 10 times as long as the upstream impactor. The total filter
samples had the same durations as their impactor counterparts. To pre-
vent material from being captured by the probes upstream when a sample
was not being taken, these probes were blocked with removable baffles.
The concentrations were determined from weight changes in the filters
and impaction substrates. Substrate material was dessicated before
making the tare weighing as well as before making the weighing with the
captured particulate material.
65
-------
Test Results
Results are presented for collection efficiency, by mass, as a function
of particle aerodynamic diameter. Sizing was done by using Andersen
Model III In-Stack Impactors at 14 jfpm flow rate (0.5 cfm), and the
material on a given stage of the impactor was classified as being of an
aerodynamic diameter greater than or equal to the aerodynamic cutoff
diameter of the impaction stage and less than the aerodynamic cutoff
diameter of the impaction stage immediately upstream. The cutoff diam-
eter is the diameter for which the collection efficiency of the given
impaction stage would be 0.50. The cutoff diameters were taken from
information supplied by the impactor manufacturer.
Fractional Efficiency of Collection
A factorial test design was used to determine the Dynactor efficiency
and the effects, if any, of flow rate, dust concentration, temperature,
and dust type. The design allowed testing whether or not each of the
parameters were significant, using standard analysis of variance, and
allowed estimates of the experimental error of measurement.
The results of the factorial test are summarized in Table 5 and plotted
on Figure 23. The mean efficiencies for the tests and the corresponding
uncertainties are listed with the aerosol fraction to which they corres-
pond. The aerosol fraction is either the size interval on the impactor
stages or the total size range which reached the total filter.
Tables 6 and 7 present a much more detailed picture of the results of
the efficiency tests. These are listed in Table 6:
• The aerosol fraction
• The mean of all the efficiency tests in the factorial design
• The means of the eight tests each at two different levels
of flow, temperature, and concentration, and two different
dusts
66
-------
Table 5. SUMMARY OF RESULTS OF 16 COLLECTION EFFICIENCY TESTS
FOR DYNACTOR (FACTORIAL TEST DESIGN)
Aerosol fraction
Total filter
Iron oxide
Fly ash
2.5 - 4.0 urn
1.3 - 2.5 urn
0.8 - 1.3 urn
0.54 - 0.8 jam
< 0.54 \m
Mean
efficiency
96.04 %
93.11
98.97
98.37
93.00
75.4
27.4
47.4
Number
of
tests, n
16
8
8
16
16
16
16
16
Estimated
uncertainty
(+ cr /s/n)
v_ e' /
+ 0.32 %
0.46
0.46
0.12
0.21
2.1
6.6
3.7
67
-------
99.8
99.5
_ 99
UJ
o
DC
UJ
0.
>- 98
o
Ul
o
u.
u.
UJ
1 95
I—
u
UJ
_J
o
o
80
50
10
0
<
— —
X
— _
—
- X I
- -
— —
X
^_ -
-x
— 1 1 1 1 1 1 1 1 1 1 1 1 1 ~
312349878910 II 12 11^
O.I
0.2
0.5
1.0
20
5.0
20
50
100
DIAMETER.pw
Figure 23. Summary of results of 16 collection efficiency tests (fac-
torial design)
68
-------
»3
M
H
p
o
§
o
>-<
§
IX
co
w
o
vO
(U
rH
•8
H
c
o
-H
4-1
td
t.
rH
4-1
c
o
o
c
o
0
O
rH CO
P4 **^-..
^S
Is
5*
CO
S
> 00
O rH
rJ •
O
•CO
"ob-^
•H 00
S3 H
i
4J o
o in
S3 ON
4J
C 0
(DO
•H O
43 CM
1
<- 9>
°3
M *
" O
_J rn
l""1 UJ
Pn «
r- o
st 0
• o
o , |
3°
0 £,
(3
0
•H
4J
O
cd
V-i
•4-1
rH
o
co
o
M
0)
<:
0
vO
ON
p>.
00
•
•*J
ON
rH
CM
£;
ON
"^
in
ON
o
vO
ON
rH
rH
CO
ON
ON
CO
ON
o
st
in
ON
ON
vO
vO
^
M
01
4-J
rH
•H
U-t
rH
0)
4-J
O
H
rH
CO
ON
CM
vO
•
rH
ON
ON
m
ON
__,
o
CO
o
st
rH
oo
ON
00
o
vO
1^.
ON
CO
CO
ON
m
o
-------
Table 7. SIGNIFICANCE OF EFFECTS OF FLOW, DUST, TEMPERATURE,
AND CONCENTRATION ON SCRUBBER COLLECTION EFFICIENCY
Aerosol fraction
Total filter
Total impactor
2.5-4.0 pirn
1.3-2.5 p.m
0.8-1.3 [am
0.54-0.8 [am
< 0.54 urn
Significance level
Flow
~ 0.90
--
> 0.95
> 0.99
~ 0.90
--
--
Dust
> 0.99
> 0.99
> 0.99
> 0.95
--
--
~ 0.90
Temp.
> 0.99
--
> 0.95
--
--
--
Cone.
> 0.95
> 0.99
> 0.95
~ 0.90
--
__
> 0.99
Table 7 shows whether these differences are statistically significant
as a result of the standard F-test analysis of variance. The analysis
of variance allows one to determine what likelihood there is that the
differences noted between measurements come from differences in the
parameters under study rather than from extraneous variations. The
"significance level" is the probability that one would be correct in as-
cribing a difference in the results to a difference in the level of the
parameter under test, here flow, dust, temperature, and concentration,
making the usual statistical assumptions about normal populations. The
above information in Tables 6 and 7 allow the following conclusions:
1. The lower flow rate yielded higher efficiencies for
all size fractions and the differences were usually
statistically significant.
2. Fly ash was collected with greater efficiency than
iron oxide for all but one size fraction, and these
differences were usually statistically significant.
70
-------
3. The lower temperature produced greater efficiencies than
did the higher temperature, in seven of eight aerosol
fractions, but this was statistically significant in
only two fractions.
4. Higher concentrations were collected with greater
efficiency than lower concentrations and this was sta-
tistically significant in most fractions.
Discussion of Experimental Results
As noted in the results, there were statistically significant effects
on collection efficiency due to flow, dust, temperature, and concentra-
tion of particulates. These effects are discussed and linked with
results of others.
Flow
Increasing the flow rate will increase the velocity gradients, which
would be expected to increase deposition due to impaction and intercep-
tion and to increase the turbulent eddy diffusivity, which is a linear
Q
function of the Reynolds number (Calvert et al. ) and thus increase
the rate of mass transfer to the droplets. Increasing the flow rate
will also decrease the residence time, which would give less time for
the collection mechanisms to act, significant for the smallest particles,
where diffusion would predominate the collection mechanisms, and for
the largest particles, for which settling would become important. Some-
what higher collection efficiencies at the lower flow rate were found,
0
as Lancaster and Strauss found in their experiments with spray
scrubbers, using ZnO particles with a number median diameter of 1.0 pra.
Dust
Particles of fly ash were generally collected with greater efficiency
than particles of iron oxide having the same aerodynamic diameter in
our tests. Different aerodynamic behavior by particles having the
71
-------
same aerodynamic diameter is unexpected, but perhaps the particles
differed in the likelihood with which a particle/droplet collision pro-
duced capture or in the degree to which they served as nuclei during
condensation or the degree to which high humidities facilitated their
agglomeration. The iron oxide powder seemed less hydrophilic than
did the fly ash, because water droplets beaded up on iron oxide layer
and absorbed into a fly ash layer, thus the iron oxide may have been
less wettable and more difficult for the water droplets to entrap.
Lohs found that making hydrophobic polystyrene particles into hydro-
philic ones, by coating their surfaces with a wetting agent, increased
the capture of these particles by a spray scrubber. From their experi-
ments with venturi scrubbers, Calvert, Lundgren, and Mehta concluded
that particle wettability enhanced collection efficiency.
Temperature
Temperature can influence collection efficiency in a variety of ways.
Higher temperatures means higher viscosity for gases; for example, as
air goes from 20°C to 100°C, its viscosity increases a factor of 1.20
12
(Bird et al.), which increases its resistance to particle motion,
hindering the various collection mechanisms. For the submicron particles,
this can be offset by the increases in the particle diffusivity due to
Brownian motion. Our experiments showed a statistically significant
decrease in collection efficiency for 1.3-2.5 fim aerodynamic diameter
particles as temperature increased from 20 C to 95 C, as well as a gen-
eral trend toward decreased efficiencies for all the aerosol size frac-
g
tions. Lancaster and Strauss measured a decrease in efficiency in
going from 20°C to 30°C with a spray scrubber operating on water-saturated
air containing particulate material.
72
-------
Concentration
As particle concentration increases, particle agglomeration increases
due to coagulation. Increased agglomeration means an aerosol having
larger mean size, which generally enhances collection efficiency in
spray scrubbers. Increased concentrations yielded higher collection
efficiencies in all the aerosol size fractions in our tests. Lancaster
Q
and Strauss, among others, reported increased efficiency with
increased mass loading. In the tests with the Dynactor, the improve-
ment in collection was most dramatic for the smallest particles, indi-
cative of coagulation.
Water Vapor Addition
A discussion of the net effect that diffusiophoresis and thermophoresis
have in this type of system is included in the GCA study, but is omitted
here because the water vapor addition tests with the Dynactor were
inconclusive.
POWER CONSUMPTION
The Dynactor tested used two high-pressure pumps to supply the spray
nozzles. Measurements showed the total power consumed by both pumps
was 4.6 VW = 6.2 hp regardless of nozzle pressure. A basic estimate
of power consumption, expended by the nozzles, is the product of pres-
sure drop across the nozzles and the volume rate of water flow, which
equaled 1.48 kW (2.00 hp) for the highest pressure drop (200 psig).
Thus, the efficiency of the pumps was 35 percent, but larger models
with larger pumps are expected to yield 60 - 70 percent efficiency.
At 60 percent power efficiency, the power consumption would be 3.3 hp/
1000 cfm = 5.2 kW/(m3/s).
73
-------
WASTE DISPOSAL
Waste water from the Dynactor may be treated the same way as waste
water from any other scrubber. RP Industries manufactures settling
basins and high rate settling columns for use with their scrubbers
when solids loading is heavy. Waste disposal is treated in Section XIV
on costs.
74
-------
SECTION IX
LONE STAR STEEL STEAM-HYDRO SCRUBBER
MANUFACTURER'S DESCRIPTION13
"The system utilizes a high-speed steam drive with injected water to
perform an extremely efficient scrubbing action. The heart of the system,
which contains no moving parts, consists of a steam nozzle, water injector,
mixing tube and twin cyclones. System operation is simple and easily
controlled.
"Normally, the system operates on energy produced by waste heat captured
from the process being controlled. The heat is used to generate steam
in a waste heat boiler. In installations where heat energy is low, sup-
plemental heat may be provided. In many cases, a package steam boiler
may supply all the energy.
"In addition to driving the system, the steam nozzle creates draft which
draws contaminated gases into the system.
Atomizer Chamber
"First stage of cleaning is done in an optional atomizing chamber with
water sprays that may be employed to cool the gas stream and remove heavy
particulate. Most processes do not require this chamber but it can be
installed as a first-phase cleaner for certain difficult effluents. A
75
-------
negative pressure is maintained in this chamber. A process occurs
where steam joins small particulate for second-phase removal in the
mixing tube.
Mixing Tube
"Collision between injected water droplets and the particulate
(including acidic gases if present), encapsulation, nucleation, and
droplet growth take place in the mixing tube. Collisions occur
between particulate and billions of high-speed water droplets. Parti-
culate is encapsulated and a growth process begins to bring submicron
particulate to manageable size for disposal through low-pressure-drop
cyclones. To insure positive capture of all particulate, a shock
wave pattern is created in the mixing tube. Massive turbulence
created by the shock wave pattern subjects encapsulated particulate to
a sudden and violent scrubbing action'-'.
Cyclones
"Separation of particulate from the gas is achieved by entering low-
pressure-drop cyclones with appropriate velocities and particulate
which has grown to a size matched to the system. Centrifugal energy
in the cyclones is maintained by force imparted from the mixing tube."
Figure 24 shows the Lone Star Steel Steam-Hydro Air Cleaning System,
13
from the Southern Research Institute (SoRI) evaluation.
POTENTIAL APPLICATIONS
The Steam-Hydro system, manufactured and marketed by Lone Star Steel,
provides >99 percent collection efficiency over a wide range of
particle size, those >0.3 urn. However, intrinsic power consumption
is quite high, an order of magnitude greater than an ejector venturi
scrubber. If fuel must be purchased to make the steam, the Steam-
Hydro system is not an economic alternative to such systems as fabric
76
-------
Outlet sampling
locations
Mixing tube
Injection water-
Steam nozzel
inlet
Particle
accelerator
Cyclones
Atomizer water
Cyclone
slurry
Inlet duct
Inlet sampling
locations
Rue gas from waste
heat boiler. Fed by
open hearth furnance
Atomizer slurry
Figure 24. The Lone Star Steel steam-hydro air cleaning system^
77
-------
filtration or electrostatic precipitation. If sufficient waste heat is
available, the Steam-Hydro system becomes economically feasible. It
should be noted that this system creates its own draft, thus obviating
the need for fans, blowers, etc. The Steam-Hydro system has been in-
stalled in open hearth furnaces, cupolas, iron ore kilns, paper mills,
sinter plants, ammonium nitrate fertilizer plants, and copper smelters.
Other possible application sites are basic oxygen furnaces, electric arc
furnaces, lime and cement kilns, and municipal incinerators.
Particle Characteristics
Particle Size - The SoRI study reports very high collection efficiency
for sizes down to about 0.5 jum, with a minimum efficiency of 70 percent
at around 0.05 p.m (based on diffusional results). The Steam-Hydro system
can be expected to clean particulate laden gas streams to high efficiency
(about 99.9 percent for the SoRI test with 1.8 /xm mass median aerodynamic
diameter aerosol) unless the particulate is extremely fine.
Particle Phase - Solid or liquid particles are expected to be collected
equally well. The manufacturer indicates that vapors can be effectively
controlled with additional energy input.
Resistivity - Particle resistivity is not expected to influence collection
efficiency.
Abrasiveness - The Steam-Hydro system appears well suited to controlling
abrasive aerosols because of the absence of moving parts, fabric filters,
etc.
78
-------
Corrosiveneee
Steam-Hydro systems have been built from stainless steel, carbon steel,
coated materials, and fiberglas. Difficult materials handling problems
can presumably be solved by choosing proper construction material.
Gas Characteristics
q
Volume Rate of Flow - Modules can be designed to handle from about 2 m /s
(4 x 103 scftn) to 150 m3/s (3 x 105 scfm).
Temperature - An optional atomizing chamber with water sprays may be
employed to cool the gas stream and remove heavy particulate. With the
proper wasteheat boiler, even the highest temperature industrial gas
streams may be cleaned.
Flow Velocities, Magnitudes, and Variations - There are three controls
for the Steam-Hydro system, the proper adjustment of which should pro-
vide maximum collection efficiency. These are steam pressure at the
ejection nozzle inlet, gas flow rate, and cyclone accelerator position.
The first two adjustments are self-explanatory and the last determines
flow pattern in the cyclone to affect maximum droplet removal.
Stack Characteristics
A most important consideration towards deciding if a Steam-Hydro system
should be installed is the availability of steam. If waste heat boilers
are already in use for a certain process, significant capital expense
will be saved. In a given situation, the Steam-Hydro system will probably
not require all the steam that is generated. For example, the SoRI
study of the Lone Star Steel open hearth operation found that the Steam-
Hydro system used 0.92 kg/s (7300 Ib/hr) of 18 x 105 N/m2 (250 psig)
saturated steam, whereas a single open hearth may provide as much as
79
-------
42 kg/s (150,000 Ib/hr). Given that steam is available, ducting and
space requirements should not be much different for the Steam-Hydro sys-
tem compared to other alternatives such as venturi scrubbers, and should
be less than fabric filtration units (bag houses). Another variable
worth mentioning is the operating condition of the waste-heat boilers.
If the waste heat boilers are not designed so that they can be kept clean
the steam requirement is greater to overcome the resistance caused by
particulate buildup in the heat transfer sections. A dirty waste heat
boiler will not be as efficient for heat exchange as well.
THEORY OF OPERATION
The scrubber uses high pressure steam to move the gas through the system
through eduction. The rapidly moving steam entrains gas, particles, and
droplets. The particles collect on the droplets by mechanisms described
in our section on the general theory of obstacle collectors; the principal
mechanism is believed to be impaction for particles larger than a few
tenths micron. Diffusion and the condensation of the steam onto the drop-
lets (entraining particles) are also thought to be significant. Once the
particles have collected on the droplets, the drop/particle mixture is
caught in the centrifugal collectors (cyclones) immediately downstream
from the scrubber mixing section. The energy required to move the gas,
particulates, and droplets through the system is supplied not by fans or
pumps but by steam produced in boilers or economizers operating on the
waste heat of the process. The wastewater sludge effluent from the
cyclones must be handled as other scrubber waste water is.
Collection Efficiency
Southern Research Institute conducted tests on the Steam-Hydro system;
13
excerpts from their report are in quotation marks. Their report
is our basic source for the Lone Star Steel system. "A total of four
80
-------
measurement techniques were used during the tests. These were: (1) dif-
fusional techniques using condensation nuclei counters and diffusion
batteries for determining concentration and size distribution on a number
basis for particles having diameters less than approximately 0.2 /urn,
(2) optical techniques to determine concentrations and size distribution
for particles having diameters between approximately 0.3 j^m and 1.5 jjm,
(3) inertial techniques using cascade impactors for determining concentra-
tions and size distributions on a mass basis for particles having diameters
between approximately 0.25 Jim and 5 ptm, and (4) standard mass train mea-
surements for determining total inlet and outlet mass loadings."
"The useful concentration ranges of both the optical counter and the con-
densation nuclei counters are such that extensive dilution of the gas
streams being sampled was required. Dilution factors of about 65:1 were
used for the outlet measurements and about 500:1 for the inlet measurements.
In order to insure that condensation effects were minimal and that the
particles were dry as measured, the diluent air was dried and filtered,
and diffusional driers were utilized in the lines carrying the diluted
samples to the various instruments."
"Because of the size and complexity of the optical and diffusional mea-
suring systems, and the fact that only one set of equipment exists for
measurements of this type, it was not possible to obtain simultaneous inlet
and outlet data with these methods. The system was first installed at the
outlet sampling location, the scrubber was tuned, and all the outlet data
were obtained. Subsequently, the equipment was moved to the inlet and the
necessary inlet data were obtained. For the purposes of calculating the
efficiency of the scrubber, the assumption was made that the open hearth
process was sufficiently repetitive that the inlet data, as obtained above,
were a valid representation of that which would have been obtained during
the time the outlet measurements were made. Accuracy in the diffusional
measurements was limited by process variations and the efficiencies derived
from these data are rather uncertain. However, the trends in the fractional
81
-------
efficiencies derived from the data are probably real and the fractions
of the influent material that penetrate the scrubber are probably
correct to within a factor of two to three."
"The optical data are presented on the basis of equivalent polystyrene
latex sizes and the indicated sizes can differ from the true sizes by
factors as large as two to three. Data obtained using this method were
primarily intended as a means of real time monitoring of process changes
and the results of changes in the scrubber operation, but also serve as
rough checks on the data obtained with the cascade impactors. The
sampling system used for obtaining the optical and diffusional data is
illustrated diagrammatically in Figure [25]. "
The investigators reported that the use of the rapid-response instruments,
the condensation nuclei counter and the optical counter, facilitated
optimization of the operating parameters.
The tests were difficult for a number of reasons, requiring several engi-
neering compromises. The great disparity between the inlet loadings (about
1 gr/ft =2.3 g/m3) and the outlet loadings (~ 0.02 g/m ) meant that
two different impactor types were used, a Brink impactor upstream and an
Andersen in-stack impactor downstream. Furthermore, the sampling times
were very different, about 6 hours downstream and 6 minutes upstream.
Only one set of the equipment, shown in Figure [25] was available, so that
upstream and downstream sampling were done at different times, average
size distributions being obtained for comparison from measurements at
various stages in the process. Large and rapid concentration fluctuations
added uncertainty to the average values obtained from the short-term
samples for total mass measurement, impactor sizing, optical sizing,
and diffusional sizing.
82
-------
Flowmeters
Cyclone Pump
Process
Exhaust
Line
\
Cyclone
(Optional)
Manometer
Diffusional Dryer
(Optional)
Particulate
Sample Line
Orifice
**^ * ft nil lit-inn \
Dilution
Recirculated
Clean Dilution
Air
Filter
Pressure
Balancing
Line
Pump
Bleed
Figure 25. Optical and dlffusional sizing system1^
83
-------
"Inertial sizing was accomplished using Brink cascade impactors for inlet
measurements and Andersen impactors for outlet measurements. Sampling
was done at near isokinetic rates. Errors due to deviations from isoki-
netic sampling should be of little consequence for particles having aero-
dynamic diameters smaller than 5 pirn or physical diameters smaller than
3
2 pm for an assumed density of 5.2 gm/cm . Further, because the sampling
was at near isokinetic rates, the calculated collection efficiencies for
larger particles are probably reasonably close to the true values.
Because of the relatively small duct dimensions as compared to the sizes
of the impactors, single point sampling was used in the ducts with the
inlet impactors at flue gas temperature ( 515°F). The outlet impactora
were heated to about 40 above flue gas temperature to insure that no
condensation took place within the impactor. Such condensation might
cause operational difficulties or lead to incorrect sizing."
In Figure 26 are shown the inlet and outlet size distributions obtained
by measurements during the oxygen lance part of the process cycle. The
instruments used were the optical counter and the diffusion battery plus
condensation nuclei counter combination, giving optical equivalent diam-
eters and mobility equivalent diameters, respectively. These measure-
ments were combined with the impactor measurements to give efficiency as
a function of particle geometric diameter, shown in Figure 27, on the
3
assumption that the dust had a density of 5.2 gm/cm .
Figure 28 shows fractional efficiency using impactor data only, adjusted
to aerodynamic diameter. The points are for optimum Steam-Hydro operating
parameters (cyclone accelerator position, steam pressure, and gas flow).
The relationship used to go from Figure 27 to Figure 28 is
84
-------
I I I I I I I I I I I I I I I I I I
0.01
O.I 1.0
PARTICLE DIAMETER, M">
Figure 26. Inlet and outlet particle size distributions measured
using optical and diffusional techniques 13
85
-------
% *AON3IOUd3 NOI103T100
n —
8ot n o o o ""• 3
0> O> IO X to — O O o
J . , , V
4
* ;
+
+
fc
+ u. ,_ 2
5 o •*
~~
4 O D +
+
"I" ~
•f
O
+
•f
a
•
o
~
_
o
"
o
o
1,1111 ,1 1 , 1 II II 1
0
0)
3
o
0
•«
IB
e
«»
O (U
- - w
1 "
rH
o>o>° ^
< o ~ o w^epi^Sb
3 O> » -r<
UAI iw\j i -aura J e/ «" Cu
86
-------
99.999
99.99
t-
UJ
u
tt
m
•>
u
5 "-9
0
it
u.
UJ
z
o
p
5
u
99.0
on
•» •
- X
- X . "-
X
X
.
.
I x x ;
~x x x -
— —
'X
-
—
1 1 1 1 1 1 1 1 1 1 1 1 1
0.001
- 0.01
- O.I
I
- 1.0
I 2 3 4 5 6 7 8 9 10 II 12 13 I4IO'°
AERODYNAMIC DIAMETER , pm
Figure 28. Collection efficiency versus aerodynamic diameter,
Lone Star Steel Steam-Hydro Scrubber
87
-------
where d = aerodynamic diameter
ae
C = Cunningham factor for particles in test
3
P = density of test particles, 5.2 g/cm
3
P = density of water, 1.0 g/cm
d = reported diameter (from Figure 27)
C = Cunningham correction factor for aerodynamic diameter
particle.
An iterative solution was used. Aerodynamic diameter is more fully
discussed in the Appendix.
Mass train measurement obtained by Guardian Systems, Inc. of Anniston,
Alabama under subcontract to SoRI showed an average collection effi-
ciency of 99.86 percent for 9 tests, with a range of 99.78 percent to
99.95 percent. Inlet mass median aerodynamic diameter was calculated
to be 1.8 ym from SoRI Brink impactor data.
POWER CONSUMPTION
Energy requirements to achieve various outlet loadings are plotted on
Figure 29 as calculated by the manufacturer. Energy used in the SoRI
3
study to achieve outlet loadings consistently less than 0.002 g/m
(0.001 gr/scf) were 395 x 103 J/kg calculated as (170 Btu/lb) of air from
the manufacturer's estimate of 475 x 103 J/m3 (12,750 Btu/1000 SCF) for
3 3
a system back pressure of 15.2 cm H_0 (6 in. H20), and 348 x 10 J/m
(150 Btu/lb) of air from the test conditions of 0.92 kg/s (7300 Ib/hr)
of 1.72 x 103 N/m2 (250 psig) saturated steam and 6.1 m3 (13,000 scfm).
Therefore, the point on Figure 29 denoted by an asterisk has been exper-
imentally verified.
One scale has been added to the right hand side of Figure 29 to show equiv-
alent electric energy in hp/1000 scfm, another for kW-hr/1000 scf. Note
that 1 hp/1000 scfm equals 1.58 kW/(m3/s).
88
-------
360
320-
280-
240-
200-
* 170-
v 160-
l 150
K.
UJ
*
O
D.
120-1
80-
40H
I
-2
i i i i
-4 -6 -8 -10
FLOW RESISTANCE , H20
-7
-6
— 5
-4 8
C.
I
—3
— 2
h-300
— 250
— 200
•150
—100
— 50
-12
Figure 29. Power consumption versus flow resistance for Lone Star
Steel Steam-Hydro Scrubber1-*
89
-------
WASTE DISPOSAL
The waste generated from the Steam-Hydro system is a wet slurry and ref-
erence should be made to the section on wastewater disposal for better
understanding of the problems involved. For the case of the open hearth
operation studied by the SoRI, a clarifier was used to drop the solids
concentration to less than 50 ppm and the waste was lime treated to reduce
acidity caused by S02 collection. Most waste water is recirculated.
A notable consideration of this system used at the Lone Star Steel open
hearths is the fact that zinc recovery from the wet slurry is enough to
provide zinc requirements for the entire galvanizing operation at that
plant and is valued at $1-1/2 million per year. Lead is also recovered.
Lone Star Steel reports that the high collection efficiency, especially
at small particle size, of the Steam-Hydro system makes ZnO recovery
greater than recovery from other lower-energy cleaning systems.
90
-------
SECTION X
MYSTAIRE SCRUBBER
14
This section is based upon the theoretical evaluation done by Midwest
TM
Research Institute (MRI) of the Mystaire scrubber, manufactured by
Heat-Systems Ultrasonics, Inc. of New York.
DESCRIPTION
Details of the Mystaire scrubber construction are proprietary. Sketchy
information is available from the manufacturer and some further infor-
mation has been developed by MRI and by GCA/Technology Division.
Figure 30 shows a bench-scale model of the device. Air containing par-
ticulate and/or gaseous contaminants enters the scrubber, passing through
a spray and traversing a multi-layer wetted wire mesh screen, the Water-
TM
web . Gases are captured to an extent in the wetted screen by diffusion,
and particulate material is caught by impaction, interception, and dif-
fusion, as outlined in the section in this document on general theory of
particle capture. The captured material is drawn off by the drains and
removed from the water, some of which may be recirculated.
POTENTIAL APPLICATIONS
This device is similar to scrubbers in that much of the particle and gas
capture will be occurring on water surfaces having dimensions of the order
2
of 10 urn. To the degree that the captured material is indeed removed by
the water flow, the waste disposal situation will be very similar to that
of scrubbers. It differs from filter systems primarily in the method
91
-------
-a-
•-I
n
•§
^
u
o
JJ
(0
M
o
A
a
-------
of cleaning (by water rather than by the dislodgement of material caught
on dry fibers). Pressure drop is much more like that for fabric filters
than it is for scrubbers with comparable efficiencies, thus power savings
could result.
The promotional literature notes: "The Mystaire scrubber has proven
successful through field installation or independent laboratory testing
on the following 'difficult1 pollutants: IkS, SO^, NO^, hydrofluoric
acid, bacteria, ammonium chloride, titanium tetrachloride,formaldehyde,
paint bake oven fumes, food processing odor, and mineral oil fumes."
Particle Characteristics
Particle Size - The MRI evaluation predicted particle collection effi-
ciencies of 94 percent for particles of diameters 0.5 and 1.0 urn, with
efficiencies of 99.9 percent for particles larger than 3 um or smaller
than 0.1 ^.m. As discussed in the following sub-section, we have arrived
at somewhat lower efficiency predictions, with a minimum in the same range.
Particle Phase - Liquid particles presumably would be more readily car-
ried off than solid particles in the water flow which cleans the filter
mat. With solids, the development of dry spots and the buildup of mate-
rial thereon, if it occurred, could be quite problematic. Water soluble
solids should also be easily removed from the mesh. Solid particles have
been successfully treated with the Mystaire scrubber, its makers report,
and we have no information to the contrary.
Resistivity - This should have no effect upon performance.
Abrasiveness - The device does not employ relatively high velocities,
thus abrasion is lessened.
Corrosiveness - Successful application for TiCl, and for hydrofluoric
acid indicates corrosive materials can be handled.
93
-------
Gas Characteristics
Volume Rate of Flow - Collection efficiency will vary with flow rate.
It is expected that more rapid flow will enhance capture of particles
larger than about 0.5 urn, at the expense of increased pressure drop
across the system and at the expense of the decreased collection effi-
ciency for particles much smaller than 0.5 p,m.
Temperature - Higher temperatures will generally decrease the collection
efficiency for particles larger than a few tenths microns and increase
the collection efficiency for those substantially smaller. Pressure drop
will increase at higher temperatures for the same flow rate in standard
3
cubic feet per minute (scfm) or standard m /s.
THEORY OF OPERATION
The collector operates with the various collection mechanisms described
in Section XVII. The water acts as a medium for cleaning the mesh.
The mesh is made up of about 100 layers. Each layer has a penetration
which can be described by the equations
P^ = 1 - (TJ Ai/AQ)
or
- (wA*/Q)
where TJ = single target efficiency for the mesh layer;
A. = cross-sectional area of the single mesh layer;
£
A. = surface area of the single mesh layer;
w = migration velocity toward collection surface;
94
-------
A = cross-sectional area of the duct or flow channel
in which the layer is placed;
Q = V A , volume rate of flow through the layer;
V = face velocity of the flow through the layer.
If each layer is the same, and if collection on each layer is independent
of the others, then the penetration through n such mesh layers is
Pn - (1.- nA/nAo)n
where A = total mesh cross-sectional area.
This can also be written
Pn = (1 -wA*/nQ)n
*
where A is total surface area.
For n» 1 and T|A/nA « 1, the formula for penetration approaches the
exponential equation used in Section XVII:
Pn = exp (- wA*/Q) = exp (- t]A/A ).
For example, for t) A/A = 3 and n = 100, the exponential equation gives
a penetration of 0.050 and the first expression gives 0.048. Note that
the single target efficiency and the effective migration velocity are
related by
*
VolA = wA ;
the product of velocity, target efficiency, and cross-sectional area
equals the product of the total surface area and the effective migration
95
-------
velocity. To predict the collection efficiency, one must calculate
either n or w and know the geometry and flow rate. In the next section
we will present the calculated efficiencies for the Mystaire scrubber,
based partly on the MRI evaluation.
COLLECTION EFFICIENCY
The MRI evaluation centered on the collection ability of the Waterweb
filter immediately downstream from the spray in the Mystaire scrubber,
as simple calculations indicated very little collection of particles
smaller than 3 nm occurred due to the spray itself.
MRI used formulas suitable for knitted mesh separators, assuming that
the fibers in the waterweb were increased by about 5 percent by their
water coating. The penetration was calculated from
Pn = (1 - nA/nAQ)n
The cross-sectional area of a layer, A/n, was calculated as the product
of the total length of strands in the layer, their diameter, and the frac-
tion of strands in effective positions, taken as 0.67 from the work on
wire-mesh separators by Carpenter and Othmer. The product of the strand
22 22
length and diameter was taken as 290 m /m of projected area (290 ft /ft ),
and was increased by an assumed 5 percent due to wetting. This seems
somewhat in error, because MRI indicated, "The manufacturer did supply us
with a web sample and the information that every square foot of projected
web area has an internal area of about 290 sq ft." The internal area
would be higher than the cross-sectional area by the ratio of surface
area to cross-sectional area for a cylinder, T. Thus the MRI areas are
high, we believe, by a factor of ^, which will be seen to make a sub-
stantial difference.
96
-------
The combined single target collection efficiency was obtained from the
target efficiencies due to several mechanisms by the formula (after
16N
Strauss ) :
which assumes that these mechanisms act independently; by definition:
T] = impaction single target efficiency;
T\ = interception single target efficiency;
c
T) = diffusion single target efficiency.
The impaction efficiency was obtained from an evaluation of the
impaction parameter,
* = CPPdp V18
where C = Cunningham slip correction factor (set equal
to 1 by MRI);
C = 1 + 0.17 |Wd at 20°C and 1 atm
pressure;
d = particle diameter, cm;
3 3
p = particle density, g/cm (set equal to 2.8 g/cm
by MRI as typical of foundry)
V = gas velocity, cm/s, approximately the face velocity;
-4
y = gas viscosity, 1.8 x 10 poise;
-4
d = diameter of collecting strand, 203 x 10 cm dry,
C 213 x 10"4 cm wet.
The impaction efficiency on a cylinder as a function of i/iwas obtained
from the book by Strauss.
97
-------
Impaction is due to the inertia of the particles, causing them to deviate
from the gas streamlines flowing around the strands. Even those particles
which follow the streamlines will be captured to some extent because their
paths come within a particle radius of the obstacle. Such collection is
said to be due to interception, and the single target efficiency of inter-
ception given by Strauss and used by MRI is
R> '
where R = ratio of particle diameter to collector diameter
(strand diameter).
Brownian motion of the particles produces collection by diffusion to the
strands; the rate of diffusive deposition becomes greater for smaller
particles, higher temperatures and, at equal total areas of collecting
surface, for collectors with smaller dimensions. MRI modified equations
given by Strauss. Table 8 gives the MRI results, with the final column
being our recalculation of total efficiency, using the factor of IT
correction to take internal area to cross-sectional area.
Table 8 . INDIVIDUAL AND OVERALL PARTICLE COLLECTION EFFICIENCY DUE
TO IMPACTION, INTERCEPTION, AND DIFFUSION3
Particle
size (jam)
0.01
0.10
0.5
1
3
5
7
10
1l
0
0
0
0
0.01
0.10
0.32
0.56
ic
0.00004
0.0010
0.0046
0.0094
0.0280
0.0465
0.0646
0.0917
ID
5.8 x 10'1
3.1 x 10"2
8.5 x 10~3
3.9 x 10"3
2.3 x 10"3
1.6 x 10"3
1.2 x 10"3
7.7 x 10"4
'llCD
0.58
0.032
0.0131
0.0133
0.0399
0.1432
0.3647
0.6007
E (?.)
100.
99.9
93.5
93.5
99.9
100.
100.
100.
E <7.)b
100.
87.8
57. A
58.0
92.8
100.
100.
100.
8A11 but last column comes from MRI report.14
Corrected efficiency, see text.
98
-------
Figure 31 shows the results of both MRI and GCA calculations of collec-
tion efficiency, graphed versus aerodynamic diameter. (The change to
aerodynamic diameter produces appreciably different results only for
3 |om and 5 |im.) The effect of the factor of IT is striking, suggesting
that minor errors in evaluating (1 - T] A/A ) can produce major errors
in the penetration and efficiency due to the n = 100 power to which
this is raised. We quote the MRI report on the limitations of their
analysis:
Limitations of Analysis
The main limitations in the analysis are:
"1. The simple model we have used may not be applicable. However, little
work has been done on wetted fibers or meshes and a suitable theory
or model is not available.
"2. Assuming that the water layer does in fact increase the effective
surface area of the mesh by increasing the strand diameter, the
value of 5 percent, assumed by us, is strictly optional and was
made only to quantify this effect.
"3. The diffusional efficiencies obtained from Ref. [l6] may not be
strictly applicable to our strand diameter and gas velocity.
However, in the absence of other data we were left with no other
choice.
"4. The presence of the water layer could favorably alter the adhesion
characteristics of the mesh surface and affect the collection
efficiency even more. Unfortunately, no theory has been developed
for quantitatively establishing this effect even though such an
effect is reported as being very favorable towards particle
collection."
Davies, C.N., Air Filtration, Academic Press, New York (1973)
99
-------
79.7
99.8
99.5
2 99
Ul
u
cc
(L
*
>- 98
z
UJ
u
OL
U.
UJ
§ 95
T~
r—
O
U
-1
_l
O
0 90
80
50
10
+ = MRI
o = GCA
— ' —
-
— "
~" ^~
.
_
— —
""+ + ~"
O
__ __
D
— —
"o o
*^^
~ l l l 1 l l l i l l l i l
U.I
0.2
0.5
1.0
2.0
5.0
10
20
50
inn
°0 1 2 3 4 5 6 7 8 9 10 II 12 13 14""
i-
Ul
u
a
kl
a.
z
o
H
2
»-
UJ
u
a.
AERODYNAMIC DIAMETER , pm
Figure 31. Estimated Mystaire collection efficiency versus particle size
100
-------
POWER CONSUMPTION
The sprays will use electrical power equal to the hydraulic power
(pressure times volume flow rate) divided by pump/motor efficiency
(~ 0.6). The spray power consumption was not indicated. The lab-
o
oratory-scale device used 3 gal/min per 50 ft min or 60 gpm/1000 acfm,
much higher than most scrubbers use; in metric terms this would be
-33 3
8 x 10 m spray per m of air.
The electrical power used by the flow through the filter web will again
be the hydraulic power divided by the fan/motor efficiency. The
pressure drop is about 12.5 cm WC (5"WC) which is 0.8 hp/1000 acfm or
3
1.27 kW per m /s. Electrical power would be 1.3 hp/1000 acfm or
o
2.1 kW per m /s. This is much lower than that for conventional scrubbers
having similar collection efficiency.
WASTE DISPOSAL
The waste disposal problems will be quite similar to those involved
for high efficiency scrubbers, and will depend upon the fraction of the
rather high water rate which is involved in makeup and blowdown water
(see Section XIV).
101
-------
SECTION XI
PENTAPURE SCRUBBER
DESCRIPTION
A diagram of the Purity Corporation "Pentapure" scrubber is shown in
Figure 32. Dust-laden air mixes with water spray. The mixture
accelerates in the converging section causing particle-droplet col-
lisions and perhaps particle-particle collisions as well. The mixture
is directed in a jet at the impingement surface and undergoes rapid
deceleration, causing more collisions and causing the droplets, along
with particles they have caught, to strike the surface, from which
they run off and form a slurry. The air exhaust contains any un-
captured particulate material and droplets.
POTENTIAL APPLICATIONS
This scrubber is very similar to a venturi scrubber, thus its applicabi-
lity should also be quite similar, such as foundries, refineries, paper
mills, industries emitting sticky particulates, and some applications
needing gas and particle removal.
Particle Characteristics
Particle Size - The 50 percent efficiency point was reached for particles
with aerodynamic diameters between about 2 and 4 (am. Efficiency would be
expected to increase for larger particles, decrease for smaller particles
(to a minimum for particles ~ 10 um).
102
-------
J-
co
tree
X
u
COUJ
n)
4J
§
PL.
(U
2 O
UJ<
O^1
So
Q. CO
CD
0
o
*>•
UJ <
I- t£
< a.
£co
\ * 1 / '
v
\-
CM
a
•H
103
-------
Particle Phase - The scrubber should be suitable for either liquid or
solid particulates.
Resistivity - This device should not be affected by particle resistivity.
Abrasiveness - High velocity air/particle mixtures could cause abrasion
in some instances, although special materials should be able to circum-
vent this problem.
Corrosiveness - Again, special materials may be required in special
cases.
Gas Characteristics
Gas Volume Rate of Flow - Increased velocities would improve the col-
lection efficiency of individual spray drops, but there would be fewer
drops per unit gas volume, so that the effects of volume rate of flow
changes somewhat offset each other. Because the collection efficiency
due to impaction is strongly dependent upon velocity differences, de-
creased volume flow rates will probably lead to decreased efficiencies
(as well as decreased pressure drops across the scrubber, varying
approximately with the square of the velocity at the throat).
Gas Temperature - The higher the temperature, the higher the gas vis-
cosity (proportional to approximately the 0.6 power of absolute tem-
perature), lowering collection efficiency. High temperature, un-
saturated air can also produce evaporation, lowering efficiency, unless
the scrubber is equipped with conditioning sprays to reduce temperature
and raise humidity, as the manufacturers did at the foundry installation
at which the Pentapure was tested.
Flow Velocities - Increased flow velocities should increase collection
efficiency at the expense of increased pressure loss, as noted. It
104
-------
can be shown that the more uniform the velocity distribution in the
collection zone, the higher the efficiency.
THEORY OF OPERATION
The gas flow is accelerated in the converging nozzle. Spray droplets
are introduced and the spray in the gas accelerates less rapidly than
the gas and particles. The particles are captured by the droplets,
as described in the section on general theory, due primarily to the
impaction caused by the relative motion of droplets and the air/par-
ticle mixture. The mixture is directed at the impingement surface,
causing more relative motion between particles and droplets, thus more
collection; some of the droplets are captured by the impingement
surface as well.
COLLECTION EFFICIENCY
GCA tested the Pentapure at a gray iron foundry. The tests were
designed to give the fine particulate (<3 ^m diameter) collection
efficiency of the scrubber section of the Pentapure installation.
Figure 33 shows schematically the installation at the foundry and
the location of the sampling equipment. Two kinds of efficiency de-
terminations were made: total mass efficiency on the aerosols that
entered the scrubber and mass efficiency as a function of particle
aerodynamic diameter on the same aerosols, using Andersen in-stack
impactors.
The EPA Method 5 minimum of 5 minutes per traverse point for the 40-point,
\
two-diameter traverse on the inlet required 200 minutes, longer than the
plant operation period, so one diameter at the inlet was done one day
while one diameter was being done at the outlet. The next day, the other
pair of diameters were traversed. A complete cross-section is the sum of
the two one-diameter traverses.
105
-------
SAMPLING
POSITIONS
SPRAYS
CUPOLA
SEPARATION
CHAMBER
PENTAPURE
SCRUBBER
TM
Figure 33. Schematic of sampling positions
106
-------
The mean mass percentage efficiency was 10.0 percent. The standard
deviation of this number was 2.5 percent. Thus, the efficiency of
the scrubber on the rather fine aerosol which reached it after tra-
versing the sprays and the separation chamber was:
= (10.0 ± 2.5) %
The particles were primarily submicron, on a mass basis, with a
mass median aerodynamic diameter at the inlet of about 0.5 p.m. Such
fine particles are relatively easy to sample, but they are difficult
for conventional scrubbers to collect.
To measure the collection efficiency as a function of aerodynamic
diameter, impactors (Andersen In-Stack Impactors) were used simul-
taneously upstream and downstream from the scrubber. A pre-impactor
was used to remove spray droplets larger than 15 u.m aerodynamic
diameter and heating tapes were used to dry those remaining. The
-4 3
impactors were operated at 14 liters per min, 2.4 x 10 m /s
(0.5 cfm). Dry air flow was measured with an orifice meter and the
amount of moisture collected in the condenser noted.
Figure 34 shows the particle size distribution for the inlet aerosol
as the cumulative mass percentage having aerodynamic diameters as
large as or larger than the values on the abscissa. The mass median
diameters were 0.5 ym at the inlet and 0.4 ym at the outlet, definitely
a submicronic aerosol. The fineness of the aerosol made it difficult
to evaluate collection efficiencies outside the fine particle range
(0 to 3 ym) because the collection stages in the finest sizing intervals
would approach the maximum recommended loadings (20 mg) before much
material accumulated in the larger sizing intervals. The mean collection
efficiency is shown in Table 9. Graphically these results are depicted
in Figure 35. The device was not an efficient collector of fine partic-
ulates, which is not an unexpected result considering the low pressure
drop across the scrubbing section, an average of 6.06 in. H~0 = 1.51 x 10
N/m2.
107
-------
a;
u
TEST NO.
o
A
+
X
a
99
98
95
90
•f
a.
a
| 80
* 70
I 60
« 50
iS 40
z
30
20
10
5
2
6 INLET, HORIZONTAL TRAVERSE, 21 NOV. '74
7 INLET, VERTICAL TRAVERSE, 22 NOV. '74
8A INLET, VERTICAL TRAVERSE, 25 NOV. '74
8B INLET, HORIZONTAL TRAVERSE, 25 NOV. '74
9A INLET, HORIZONTAL TRAVERSE, 26 NOV.
9B INLET, VERTICAL TRAVERSE, 26 NOV.
'74
'74
1
1 1 1 1 1
1
0
+
a
°
X
D
i
I
li
I
0
a
1 I
I
1
I I l i I
-P
x
1 1
.4 .5 .6 .7 .8 .9 1.0
2JO
10 40 5.0
AERODYNAMIC DIAMETER, D
Figure 34. Inlet particle size distribution
9
1 1 1
10.0
108
-------
Table 9. MEAN COLLECTION EFFICIENCY, IMPACTOR DATA
Aerodynamic
diameter,
ym
3.0
1.75
0.98
0.62
0.25
Mean
ef f iciencyf
"A
43.8
44.1
15.8
-38.5
-30.0
Possible explanations of the negative efficiencies (greater concen-
tration of mass at the outlet in a size interval than at the inlet)
include: agglomeration, deagglomeration, or other change in par-
ticle size distribution without increase or decrease in total mass;
production of particles, such as by spray introduced between inlet
and outlet; random data errors in the vicinity of zero efficiency;
undetected systematic error.
The Scrubber Handbook gives the following formula for the penetration
of a venturi scrubber:
Pn = exp (2
u p. D, F(K , f)/55 Q u )
g KL d p' xg >V
where
F(K
Q_ - volume flow of scrubbing liquid, cm /s
u a gas velocity at scrubber throat, cm/s
3
p = scrubbing liquid density, g/cm
Dd = scrubbing droplet mean diameter, cm
f) s inertial collision function
3
Q = gas volume flow, cm /s
O
H = gas viscosity, poise
109
-------
yy
98
O
Z
UJ
S£ 95
u.
u_
UJ
z
O
P 90
O
UJ
_J
o
0 80
UJ
o
f-
UJ
o
a:
u 50
Q.
0
-40
-100
-
-
(
-
X X
X
-
> x '.
1 J 1 1 1 1 1 1 1
I.U
2.0
5.0 §
L_
^
QC
H
Ul
10 S
a.
Ul
o
20 5
Ul
U
a:
Ul
a.
50
100
140
200
o
2345678
AERODYNAMIC DIAMETER,
10
Figure 35. Collection efficiency versus aerodynamic diameter, data
means for Pentapure scrubber
110
-------
The function F(K , f) depends upon a Stokes parameter, K , given by
KP = C PP "g dP ' 9 'g °d
where C = Cunningham correction
C = 1 + 0.17 x 10"4 cm/d
d « particle diameter
P
p = particle density.
F depends upon the parameter f, which is determined empirically and in-
cludes effects of mean gas droplet relative velocity and droplet-
particle wetting. In the range where collection is appreciable (d > 1),
Q IT
an approximate form given in the Scrubber Handbook for F is
F(K , f) = -0.156 K f2
P P
The resulting expression for the penetration becomes:
Pn = exp [ - (6.3 x 10~4) f2 QL u2 PL C p d2 / Q u 2 J
The factor f is usually between 0.25 and 0.5. Here 0 /Q was <
3 -3 L 8 ~ ,
85 gal/20,000 ft . 0.57 x'10 . The gas flow at the throat was 9.4 m /s
3 9
(20,000 ft /min) through a 0.6 m (2 ft) orifice; thus, ng was 32.4 x 10
o
cm/s (6,400 fpm). The scrubbing liquid density was 1 g/cm and the air
viscosity 2.1 x 10 poise. For f s 0.5 and /C p d six 10~4 cm
~~
Pn = exp
(. (6.3 x 10"4)(0.5)2(0.57 x 10"3)(10.5 x 1Q6)(1 x Ip"4)2 J
(4.41 x 10"8) /
Pn = 0.808
111
-------
Figure 36 gives the non-negative impactor data and the penetrations
calculated using this formula for both f = 0.25 and f r 0.50. These
calculations are approximate, partly because the scrubber geometry is
somewhat different from that of the usual venturi, though not so dif-
ferent that their major collection mechanism differs: the impaction
of particulate material on droplets, followed by the capture of the
droplets.
The scrubber is not identical to a venturi scrubber, but basically
both types use the acceleration and deceleration of the air to pro-
duce relative velocity gradients between the airborne particulates
and the collecting droplets. The Pentapure scrubber collection effi-
ciency is expected to be related to the pressure drop across the
scrubbing section in much the same way the venturi scrubber effi-
ciency is. The data are in agreement with theory.
The final result of a theoretical analysis performed on this type of
18
scrubber by Midwest Research Institute was that the particle size for
3
which the efficiency would be 50 percent was 3 um for a 2.8 g/cm
density particle, which corresponds to a 5 ^m aerodynamic diameter
(K.P. Ananth, MRI, personal communication, December 1974).
In summary, the GCA theoretical analysis and the experimental results
indicated that the Pentapure scrubber collection efficiency reaches
50 percent between 2 and 4 um aerodynamic diameter and increases for
the larger particle sizes, but decreases for the smaller sizes, at
least down to a few tenths microns.
POWER CONSUMPTION AND COSTS
The power used by the scrubber is predominantly the pressure drop of
the air going through the scrubber (Ap) times the flow rate of the
air (Q ) in actual volume terms the hydraulic power.
O
112
-------
NON-NEGATIVE
IMFACTOR DATA
23456789
AERODYNAMIC DIAMETER,^
Figure 36. Data and approximate theory predictions
(f = 0.25, 0.50; QL/Q
Pentapure scrubber
0.57 x 10~3) for
113
-------
Flow and pressure drop were:
Q = 20 x 103 ft3/min = 9.4 m3/s
O
AP = 6.06" H20 = 1.51 x 103 N/m*
Thus, the air hydraulic power consumption in watts was an average
1402 kW. A standard estimate of the electrical power necessary to
achieve this air flow power is to assume combined efficiency for
pumps and fans of 0.6, which would make the electrical power con-
sumption for this operation 23.8 kW. The fan and motor combination
at the site was rated at 75 hp - 5.6 x 10 W = 56 kW. The Pentapure
actual power consumption, including 60 percent power efficiency, becomes
1.6 hp/1000 cfm or 2.5 kW per m3/s.
114
-------
SECTION XII
UNIVERSITY OF WASHINGTON ELECTROSTATIC DROPLET SCRUBBER
INTRODUCTION
Midwest Research Institute (MRI) used what theoretical and experimental
information was available to make a preliminary evaluation of the poten-
tial applicability of the electrostatic droplet scrubber for fine par-
ticle control.
DESCRIPTION
The scrubber consists of two spray chambers, the first being counter-
current flow and the second chamber cocurrent. The scrubber is elec-
trostatically augmented by charging of the droplets and the particulates
to opposite polarities using inductive charging and corona charging,
respectively. The scrubber configuration, with aerosol charging chamber,
19
is shown schematically in Figure 37. The device was developed at the
University of Washington, under Dr. Michael Pilat.
POTENTIAL APPLICATIONS
The electrostatic droplet scrubber, if successfully scaled-up to indus-
trial sized units, would have many potential applications. Any industrial
source of fine particles might find the performance of the electrostatic
droplet scrubber attractive. The characteristics of the particles and gas
which affect the operation of the scrubber have not yet been determined,
but it would seem clear that the present applications for spray scrubbers
115
-------
CHARGED
SPRAY
MIST
ELIMINATOR
GAS
OUTLET
CHARGED
SPRAY
WATER
OUTLET
WATER
OUTLET
AEROSOL
GENERATOR
BLOWER
CORONA
CHARGER
Figure 37. Schematic diagram of electrostatic droplet scrubber
116
-------
would be likely candidates for the enhanced fine particle removal capabil-
ities of the electrostatically augmented scrubber. Some typical users of
spray scrubbers are: coal-fired boilers, lead sinter crushers, phosphate
rock dryers, coal dryers, electric arc furnaces, carbon black manufactur-
ing, refineries, paper mills.
Particle Characteristics
Particle Size - The efficiency of the scrubber was reported by its devel-
opers to be greater than 95 percent for particles between 0.5 and 4 um,
with the efficiencies of greater than 85 percent for particles greater
than 0.3 ym but less than 0.5 ym.
Particle Phase - Liquid particles of dioctyl phthalate (DOP) were utilized
for the testing of the system. Solid particles should be equally
collectible.
Resistivity - The resistivity of the particles does not affect the col-
lection efficiency of the scrubbers without electrostatic augmentation.
The resistivity of the particles may be reflected in the corona charging
of the particles however, and thus may be related in a minor way to the
overall collection efficiency.
Abrasiveness - The liquid DOP particles used were not abrasive; however,
in an industrial application of the electrostatic droplet scrubber the
abrasiveness would have to be considered due to its effect upon spray
nozzles. It is not anticipated that particle abrasiveness would interfere
with the overall collection efficiency of the scrubber.
Corrosiveness - Again, the liquid DOP particles used are fairly inert and
would not have caused any corrosive action on the scrubber. While corro-
sive particles need to be considered in the choice of materials when de-
signing a specific system, it is not expected that corrosive particles
will have any effect on the overall collection efficiency of the scrubber.
117
-------
Gas Characteristics
Volume Rate of Flow - The two electrostatic droplet scrubbers thus far
constructed and tested by the University of Washington have had nominal
—2 3 "^
capacities of 6.61 x 10 m /s (140 cfm) and 0.472 m /s (1000 cfm). The
3
larger unit has been operated only at 0.33 m Is (700 cfm), while the
smaller unit has been operated at the stated capacity. The effect of
flow on collection efficiency will be discussed under "Theory of Operation."
Temperature - Collection efficiency is predicted to decrease as temperature
increases, due to increased gas viscosity, for d » 0.1 vim.
Stack Characteristics
It is conceivable that a spray scrubber could be modified to incorporate
the basic elements necessary to make an electrostatic spray scrubber. The
spray nozzles of an existing scrubber could be either electrified or
replaced with more easily electrified spray nozzles, which would leave
only the addition of a corona charging section at the scrubber inlet
to approximate the electrostatic droplet scrubber.
THEORY OF OPERATION
The basic idea is to augment the collection processes usually associated
with spray scrubbing (primarily impaction) with electrical collection
forces as well.
As outlined in Section XVII on general theory of collection, the penetration
of aerosol particles of a given size will be approximately
Pn = exp (- wA/Q)
118
-------
in which w is an effective relative velocity, A is the collection area
and Q is the volume flow rate. In addition to the relative velocity pro-
duced by particle inertia and spray motion, there is an added velocity
toward the droplet due to electrostatic attraction. With the particles
charged to one electrical polarity and the droplets charged to the other,
the primary force will be the Coulomb force:
FC
where F = electrical force, dynes
q = particle charge, stat-coul (3.3 x 10~ coul)
E = electrical field, stat-volt/cm (300 volt / cm) .
The electrical field at the surface of the drops will just be
E* = qd/Rd2
where q, = droplet charge, stat-coul
R, = droplet radius, cm
and the field will decrease with the square of the distance between drop-
let and particle.
If the particles are exposed to the corona charging action sufficiently
long to approach equilibrium charge levels, they will have a charge given
approximately by,
q = 3 E d 2/4
P c p
where E = corona charging electric field, stat-volt/cm.
C
119
-------
The droplet charge is more difficult to calculate, but a plausible
assumption is that they have a potential, V (stat-volt), at their sur-
22
face equal to that of the nozzle, which seems to have been about 5 kV,
V = q./R,.
-7 3
The measured charge on the droplets was about 6 x 10 coul/g or 1.8 x 10
stat-coul/g. Their number median diameter was about 50 |_im.
The electrical contribution to the particle effective collection velocity
can be estimated by calculating the electrical force and multiplying that
force by the particle mobility, B. More information is available in the review
by Cooper.^ (See Section XVII.) The University of Washington researchers,
under Dr. Michael Pilat, made somewhat more elaborate calculations.
22
In the work reported thus far, Pilat et al., used a simple exponential
(Kleinschmidt or Deutsch-type) model to estimate the difference in
collection efficiency due to the addition of charge effects. This is the
exp (-wA/Q) model we present above and in the section on the general
theory, but w is obtained by a more sophisticated combining of impaction
22
and electrostatic migration. Figure 38 is from the paper by Pilat et al.,
based in turn on the work done by Sparks,^ in which collection effi-
ciency was gotten from particle trajectories calculated by numerical
integration of the particle equations of motion, considering diffusion,
electrostatics, and particle inertia. The droplets were assumed to be
200 p.m in diameter, moving at 100 cm/s, and the particles were either
uncharged or carried charge equivalent to that induced by corona charging
in an electric field of 1 kV/cm. This figure indicates that the minimum
collection efficiency for the charged aerosol should be the maximum effi-
ciency for the uncharged aerosol under the conditions considered. This
was tested experimentally, but the measured efficiencies were much
less than predicted by the simple model for both cases, especially for
120
-------
300
200
^Chorged particles(-)
and droplets (+)
01
01 I
Po'lide radius (microns)
Figure 38. Calculated particle collection efficiencies for a
single 200 um diameter droplet with a 100-cm/s
undisturbed fluid velocity^
the smaller particle sizes. The model for the charged aerosol predicts
that collection efficiency should increase as particle size decreases*
and exactly the opposite was measured.
COLLECTION EFFICIENCY
The mass collection efficiency of the electrostatic droplet scrubber has
been determined as a function of aerodynamic particle size, utilizing
22
simultaneous cascade impactor measurements. Equipment utilized to test
the two electrostatic droplet scrubbers consisted of: University of
Washington Mark III Source Test Cascade Impactors, a Zeiss particle size
analyzer, and a charge measuring circuit.
The fractional mass efficiency versus aerodynamic diameter was determined
using simultaneously operated impactors, the results of which are given
in Figure 39. Efficiency data for the scrubber operated with and without
particle/droplet charging was determined to assess the effect of particle/
121
-------
100
80
60
40
20
Droplets and
particles
char ged
oppositely
Liq /gas = \5 ('gal /lOOOocf
Mean drop diameter by
number - 50 microns
Drop geometric stand dev = 19
.
2 46 810 2 4 6 6 10
Particle diameter (microns)
Figure 39. Particle collection efficiency of electrostatic
spray droplet scrubber as function of particle
size
droplet charging on collection efficiency. As previously discussed, the
particle collection does not agree well with the theoretically predicted
results. The theoretically derived curve for single particle collection
efficiency for oppositely charged particles and droplets, Figure 38, pre-
dicts that the efficiency should decrease with increasing particle size
from 0.01 - 1.0 um and then level off (up to lOpm). The experimentally
observed relationship between particle size and collection efficiency
displays an increase in collection efficiency with increasing particle
size from 0.3 - 1.0 |j.m which then approximately levels off.
i
The overall particle collection efficiency increased from 68.8 percent at
uncharged conditions, to 93.6 percent at charged conditions, indicating a
substantial improvement in collection efficiency due to the electrostatic
augmentation. The collection efficiency increased from 50 percent over-
all to 69 percent overall by increasing the water consumption rate from
7.14 to 15.7 gallons/1000 acf, or 1 to 2 x 10~3 m3/m3, without charging.
With charging, the efficiency increased from 95 percent to 98 percent for
the same increase in water consumption rate.
122
-------
The development of the electrostatic droplet scrubber is in its early
stages, and very little data are as yet available for a more thorough
analysis of the collection efficiency and the effect of other parameters.
POWER CONSUMPTION
The electrostatic droplet scrubber consumes power at four distinct por-
tions of the equipment: fan energy to provide the pressure drop across
the scrubber, the pumping energy to supply the water pressure for the
spray nozzles, corona charging of the particulate, and inductive charging
of the water droplets. The theoretically calculated power consumption
for these are: 49 watts, 409.5 watts, 118 watts, and 22 watts respec-
3
tively, per 1000 acfm (0.47 m /s). This adds up to a total of approxi-
o
mately 600 watts (0.8 hp) per 0.47 m /s (1000 acfm), which is much
less than the predicted power consumption of conventional venturi
scrubbers approximately matching the collection efficiency for fine
particles.
WASTE DISPOSAL
The waste water from an electrostatic droplet scrubber would require the
same treatment that would be required from any scrubber used to clean the
same off-gas. Costs associated with the treatment will depend primarily
on the volume of water to be treated, and the type of treatment which
will be required. Present estimates are that the water consumption is
3 3
approximately 2.1 liters/m (15 gallons/1000 ft ). This water consump-
tion is higher than most scrubbers and would add significantly to the
cost of water treatment if it cannot be reduced.
123
-------
SECTION XIII
EFFICIENCY COMPARISON
INTRODUCTION
Each of the novel control devices has had presented with it a graph
of collection efficiency as a function of particle size (aerodynamic
diameter), E(d ). In this section is given a simple procedure for con-
verting this collection efficiency curve into total mass collection
efficiency for a particular control application. This allows a potential
user to estimate the emissions reduction achievable by the application
of the control device to his dust. The total efficiency (one minus the
ratio of particulate mass outflow to particulate mass inflow) will be
strongly dependent upon particle size for most control devices, thus it
has to be calculated specifically for each application.
SIZE DISTRIBUTIONS
The concentration of particulate material is often presented as a function
of particle size by means of the cumulative mass concentration, M(d ),
which is the mass concentration of particles having diameters
-------
G(dp)
/ ^
1.0-
ci 0.8-
tj
z
x
I-
cc
Ul
n ~ _,
0.6 -
z
o
o 0.4
<
a:
CO
to
0.2-
0.0
i
3
I
4
PARTICLE DIAMETER , dp , pm (I0"4cm)
Figure 40. Hypothetical cumulative mass distribution (log normal,
median = 1, geometric standard deviation = 2)
125
-------
In order to estimate collection efficiency for a particular process and
control device, we must know the size distribution at least approxi-
mately. Often it is obtained from cascade impactor measurements.
CONTROL DEVICE EFFICIENCY VERSUS PARTICLE SIZE
Figure 41 is a hypothetical collection efficiency curve, such as those
presented here for novel control devices. It shows collection efficiency
E(d ) as a function of particle diameter, d .
P P
The essence of the method presented here is: obtain average efficiencies
for each of a number of size intervals; multiply these average efficien-
cies, size-interval-by size-interval, by the mass fractions contained
within the intervals; sum these products.
Let G(d ) be the fraction of the mass concentration contained in particles
P
-------
1.00
o
2
UJ
O
u.
u.
Ill
o
UJ
o
o
Z 3 . 4
PARTICLE DIAMETER , dp
Figure 41. Hypothetical control device collection efficiency versus
particle diameter
127
-------
The size intervals do not have to be all the same width; make them
narrower where the efficiency is more curved, wider where it is nearly
straight. The intervals must be the same for the efficiency curve and
the size distribution curve, however. The dp values which are the
interval demarkations are dpi. The fraction of the mass in the interval
from dp^ to dp^ + i is just the difference between the cumulative mass
distribution values at these two diameters, G^ + ^ - G^. The efficiency
value to use on this interval is just the efficiency at the mid-point,
= E (dmi), where
d . = (d . . . + d . } /2.
pmi \ pi + 1 pi /
Then the mass efficiency can be obtained from summing the products of the
mid-point efficiency and the fraction of mass on the intervals:
N
i = 1
Emi[Gi
The error in such a technique can be estimated by cutting the intervals
in half (using twice as many intervals) and repeating the process; the
difference between the two estimates will normally be larger than the
difference between the second trial solution and the true solution.
As an example, we will compute the efficiency for the hypothetical dis-
tribution and the hypothetical control device efficiency given in Fig-
ures 40 and 41. This is done in Table 10. The total mass efficiency is
K, = 0.220 + 0.211 + 0.092 + 0.035 + 0.15 = 0.573
M
. = 57.3%
M
128
-------
W
W
H
U
W
O
U
C/3
O
Pd
H
U
H
P
P-i
fn
O
O
H
53
o
o
W
o
I—I
0)
C8
H
c
o
05 vH
mi t
4-1
rt o •
2
HH
0)
>,
•H .
cd ^
rH •
3 1
S •)
3
U •
1
4
•t
1
•
M
f«
ID
a
•W
rH fi
cfl -H
> O
V^ Oi
01
4J
C
M
•iH
/-. 1
•iH •— 1
g
•<£ +
TJ
•— ' -H
W O
X
•H
rH O
cd 1
- 1 **
M H
JJ
M >r^
-H C)
i— i
-< *T"
_i ,_j
j O
3
3
H
b4
_)
J}
H -rl
3 O
-1 rS
3 U
-1 C ^
3 0) -H
i -H 0
1 t ) Q .
1 W V-M
a -H -d
0)
4J /^v
C 8
•H 3.
O ^-^
o.
1 -H
T3 0
•H 0,
+ <~*
•H 1
T3
x— s
•H S
13 v-X
•H
rH
0) cd
N >
•H M
co oi
C
•H
o I-H CM in m
CM i-H CTv CO •-!
CM CM O O O
o o o o o
m m
o> o\
o o o o o
m
st sj- r^
o m oo CTV ON
• • • • •
o o o o o
>J- CM CM O O
-------
This is the efficiency we would expect for the control device operating
on the specific aerosol. (Intuitively, we would expect approximately
this result; most of the aerosol is between 0 and 2 ym and the efficiency
curve seems to average about 60 percent in this range.)
To summarize the procedure:
a. Segment the curves for efficiency and cumulative particle mass size
distribution, using particle size intervals which are close
enough that the curves are nearly straight lines between the
particle sizes chosen.
b. Obtain mid-point values for efficiency.
c. Calculate the product of the mid-point efficiency E (dpm^) and
mass fraction on each interval (G-^ + 1 - G±).
d. Sum these products.
e. The mass efficiency for the control device on this aerosol
is the result.
f. To estimate the error, redo the calculation with half as many
intervals. The difference is approximately the error.
In mathematical terms, to obtain the total mass efficiency, we have
numerically integrated the product of efficiency and fractional mass
concentration. We estimate the error by the rate of convergence as
the intervals are made smaller.
If only fine particle (d <_ 3 ym) efficiency is desired, form G(dp) from
M(dp) using M* = M(3 pm) and repeat the process above with an upper
size limit at d = 3 ym.
130
-------
CUT DIAMETER COMPARISONS
4
Calvert compared several different scrubbers on the basis of their cut
diameters, the particle diameter at which their efficiencies are 50 per-
cent. This has the following advantages:
1. A single number is obtained for comparing collection
efficiency curves of different devices, quite useful if
the dominant collection mechanisms are the same.
2. If the curves are similar in shape, then this number is
quite informative about relative efficiencies for
particle collection, and for certain particle size dis-
tributions, the efficiency can be predicted. .
It has the following disadvantages:
3. If the curve has a minimum or a maximum within the range,
there may be two particle sizes at which the efficiency
is 50 percent.
4. Some devices may be so efficient that there is no particle
size for which the collection efficiency is 50 percent.
5. If the collection efficiency versus particle size curves
of the two devices being compared have quite different
shapes, then the use of the cut diameter may be very
misleading with regard to their overall efficiency on
particular aerosols.
6. Choice of diameter (aerodynamic, geometric, mobility, etc.)
is a problem. For those novel collection devices evaluated
here for which such a treatment seemed sensible, we have
listed in Table 11 the aerodynamic cut diameters.
131
-------
Table 11. AERODYNAMIC CUT DIAMETERS FOR SEVERAL NOVEL CONTROL DEVICES
Control device
Aerodynamic cut diameter3
Aronetics
Braxton with cyclone
or scrubber
Centrifield
CHEAP
Dynactor
Lone Star
Mystaire
Pentapure
University of
Washington
0.3 ym
2.0
0.65
0.55
3.0
aSee Appendix for definition of aerodynamic diameter. This is the
largest aerodynamic diameter for which the control device has an
efficiency of 50 percent.
AVERAGE FINE PARTICLE EFFICIENCY
Another method will now be presented for obtaining a single efficiency
number with which to compare control devices. Recently, interest has
focused on the fine particle range, particles with diameters less than
or on the order of 3 |j.m, because these particles persist in the atmo-
sphere longer than larger particles, generally, and they tend to escape
25
the human body's respiratory defenses. (See Lippman's and also see
f\ r — —
Friedlander's articles for example, for more details.) The average
fine particle efficiency for a control device in the range 0 to 3 urn
would just be
E = (3
E(dp) d(d),
the integral of the efficiency curve from 0 to 3 ym divided by 3 um.
weights all particle sizes equally in the range 0 to 3 ym.
This
132
-------
It has the following advantages:
1. A single number is obtained for comparing the collection
efficiency of different devices.
2. If the collection efficiency curves are similar, or not, this
number is quite informative about relative efficiencies.
3. The number obtained is a kind of efficiency rather than a
particle diameter.
4. The number is unambiguous, whether or not the efficiency
curve has a maximum or minimum in the range, whether or not
efficiency of 50 percent occurs in the range.
5. Regardless of collection efficiency curve, the number will
be informative, though not definitive, about fine particle
collection efficiency.
It has the same disadvantage as the cut diameter concept:
6. Choice of diameter for the collection efficiency curve is
still problematic. (But we shall use aerodynamic diameter,
as does the cut diameter method.)
To calculate the average fine particle efficiency by detailed integration
would be tedious, and often not warranted by the accuracy of the data, so
we use the following numerical method for its calculation:
a. Subdivide the collection efficiency curve from 0 to 3 urn
into six equal sub-intervals, 1/2 ym wide.
b. Evaluate the efficiency at the mid-points of these sub-
intervals (0.25 urn, 0.75 urn, 1.25 ym, 1.75 urn, 2.25 ym
and 2.75 ym).
c. Add these six efficiency values and divide by six.
d. The result is the approximate average fine particle
efficiency.
Parenthetically, we add that the method is just the mid-point quadrature
evaluation of the integral, and the absolute error can be shown to be
less than I/24th the maximum absolute value of the curvature (second
133
-------
derivative with respect to particle size) of the efficiency curve times
2 2
(1/2 ym) or 0.25 ym . See a numerical analysis text for more information.
In Table 12, we list the average fine particle efficiencies for the novel
control devices as determined from this method and the curves given in
the sections on the devices themselves. (The true fine particle effi-
ciencies will, of course, depend upon the aerosol to be controlled, just
as is true with the cut diameter estimates.)
- *
Table 12. AVERAGE FINE PARTICLE EFFICIENCIES FOR THE NOVEL CONTROL DEVICES
Control device
Average fine particle
efficiency
Aronetics
Braxton with cyclone
Centrifield
CHEAP
Dynactor
Lone Star
Mystaire
Pentapure
University of Washington
97.4
47. 3*
72.3*
95.2l
79.2
a
99.89
76.3b
97.8C
29.0
94.5
a
a
Includes extrapolated values
GCA calculations
"MRI calculations
134
-------
SUMMARY
The total mass efficiency can be calculated for the novel control devices
for aerosols other than those on which they were tested by using the
methods of this section, thus facilitating a choice as to which might be
applicable to one's specific process. Two methods are presented for
comparing control device efficiencies without knowing the particle size
distribution. The average fine particle efficiency number appears
superior to the particle cut diameter approach.
135
-------
SECTION XIV
COST COMPARISON
INTRODUCTION
The goal of this section is to indicate how cost comparisons can be
made for various particulate control options, including the novel
control devices discussed in this document. The approach is based
27
upon that developed by Edmisten and Bunyard while employed by EPA's
predecessor, the National Air Pollution Control Administration. In
brief, the goal is to develop a single cost parameter, the total
annualized cost, by which different air pollution control devices
can be rationally compared with respect to cost. This is quite
useful because, for example, electrostatic precipitators have rela-
tively high initial costs and relatively low operating costs in com-
parison to scrubbers of comparable collection efficiency.
27
Edmisten and Bunyard divided the costs into three categories:
• Capital investment cost - This includes the control
hardware cost, the cost of auxiliary equipment and
the cost of installation, including initial studies.
• Maintenance and operating costs - These are taken on
a yearly basis, averaged over the life of the
equipment.
• Capital charges - These are what it costs to borrow
the money equivalent to the capital investment,
usually 6 to 12 percent per year, plus taxes and
insurance.
136
-------
ANNUALIZATION OF COSTS
To convert these various costs into a single number, the total an-
nualized cost, one sums the annual capital investment depreciation,
the operating and maintenance costs, and the capital charges (CC).
The usual method of depreciation in such contexts is to assume
straight-line depreciation. One estimates the life of the equipment,
27
Edminsten and Bunyard suggested 15 years, and figures the yearly de-
preciation as the capital investment divided by the life expectancy.
Thus, the total annualized cost (TAG, $/yr) is given by
TAG s CI/L + MO + CC
where
•
CI = capital investment, $
L = lifetime, yr
MO a yearly maintenance and operating costs, $/yr
Generally MO and CC will be nearly proportional to CI.
For more details on conventional control device costs, see the article
27
by Edmisten and Bunyard. The literature on filters, scrubbers, and
electrostatic precipitators often will provide additional information:
For scrubbers, see also the articles by Hanf and
MacDonald,28 by Fraser and Eaton,29 and the
Scrubber Handbook by Calvert et al.8
For fabric filters, see also the article by Fraser
, ri'f-t' T 1
and Foley30 an(j the report by McKenna .-*1
For electrostatic precipitators, see also the
articles by Schneider et al,^^and by Benson and
Corn.33
137
-------
POWER COSTS
Although much of the necessary information on costs will have to be
obtained from the manufacturers for a specific application, one can
estimate power costs from the information in this document.
Generally the power consumption figures are given in terms of kW per
3
m /s flow rate or hp per 1000 acfm flow rate. (Note: 1 hp s 0.746 kW,
o
1000 acfm = 0.47 m /s.) Where the power is given as hydraulic power
(pressure drop times volume flow rate) a pump/fan/motor efficiency
factor must be used (as a divisor) to convert to actual electrical
power; this efficiency factor is generally about 0.6, whether fans are
moving gas or pumps are moving liquid. The power cost (PC, $/yr) is
given by:
PC r Q x (P/Q) x CP x TT
where
3
Q = volume flow rate, m /s
3
P/Q = power consumption per unit flow rate, kW/(m /s)
CP = cost of power, $/kW-hr
TT s total yearly operating time, hr.
Certain forms of power may be nearly free: the recovery of waste heat
is free with regard to operating costs, although it will add to the
capital investment and thus to the costs associated with capital
investment.
As with other costs, the power costs will vary considerably from
situation to situation. The P/Q information presented for the novel
devices should aid in the estimation of these costs.
138
-------
Using the methods of this section, a single cost parameter can be
developed by which to compare control system alternatives. This
parameter is the total annualized cost.
Although specific circumstances will also greatly affect waste dis-
posal costs, we present a brief discussion of these costs to highlight
that having captured the airborne material is not the final stage in
air pollution control. Generally, that final phase is to convert the
captured material into a solid for such uses as landfill or to find
a method for recycling some or all of the captured material. The
scrubbers will produce waste water which must be handled properly;
dry collectors and wastewater treatment will produce solid wastes
which must be used or disposed.
WASTE WATER
A cost incurred by wet scrubbing which estimates should take into
account is the cost of cleaning the waste water generated by this
type of device. Depending on the type of contaminant and the degree
of treatment necessary before discharge, the cost can vary an order
of magnitude. It can be more expensive than the raw water cost.
If scrubber waste water is not discharged to a municipal sewer, it
must be brought up to sufficient quality so that it can be discharged
to a receiving body, or disposed of in some manner such as lagooning
or deep-well disposal. The quality to which the water must be brought
for discharge to receiving bodies such as streams, lakes, estuaries,
coastal and marine waters, etc., depends in general on the uses
(such as swimming, recreational boating, fish and wildlife, etc.)
designated for the receiving body. Such information may be obtained
from the state agency responsible for water quality standards or water
139
-------
pollution control. In many instances, scrubber waste water will be
processed through the plant wastewater treatment facility.
So that proper treatment may be specified, it is necessary to charac-
terize the effluent from the scrubber. Figure 42 diagrams treat-
34
ment processes for various degrees of contaminant removal. For dis-
charge to municipal sewers or to the plant treatment facility, dis-
charged waste water must not disrupt such processes as biological
oxidation, or overload such processes as carbon filtration, or ion
exchange. Acids and alkalies must be neutralized before biological
oxidation. Toxic materials such as cyanides, metallic ions, barium
chloride, chromic acid, aldehydes, germicides, dye stuffs, radio-
active isotopes, etc., have rigid discharge standards and must be
treated appropriately. Phenols, which result in tastes and odors in
water and are rigidly controlled, are often encountered in waste
water from foundry scrubbers because coke and core binders contain
phenols. Shock discharges (such as emptying a recirculation reservoir
for a scrubber) should be equalized so that trickling filters, sludge
basins, etc. downstream are not flushed with the resultant discharge
of raw effluent to the receiving body. Thermal discharges, perhaps
from scrubbers used to cool as well as clean effluent gas, reduce
dissolved oxygen and may kill sensitive species in receiving bodies.
Thermal discharges can be controlled by ponding, spray systems, or
cooling towers. In certain applications, cleaning scrubber waste water
may result in product recovery where expensive solvents are condensed
or valuable materials enter effluent gas.
Costs may be estimated from Figure 43 after the degree of treatment
35
has been determined. The cost, plotted on the ordinate of Figure 43
should be multiplied by a factor for the increase in the Marshall and
Stevens Equipment Cost Index (found in Chemical Engineering) from 1970
(M & S = 303.3) to the present (M & S = 444). When planning a scrubber
installation,the plant engineer responsible for wastewater treatment
140
-------
li
si
s;
- s
.5
|«
£ °
I1
1]
2!
Sf
C O
si
N
E |
u. *
*<»
ll
* E
1 *
0.
H
i S
«
* i £
312 i
3 - ™ £
o a 2 £
> c •= J2
m .0 5
00 « -O
o tfi
*
il
P
s- •»
5 •a
"I
in
•g
0 m
" S
* 1
m e
|S
i ,--.
\ r~-
9
2.
3
•o
0
. M * o
*5 a 5 g|
o i li
o » §*
_ t _ *
l i
1 §8 -!
1 I! 11
1 51 *|
« 4
u
t *
1
I if
1 "1
4
— .J
a)
O
B
4J
c
O
u
en
V
(0
a
O
•H
J-l
«
U
01
H-l
4-1
(U
(0
V-i
-------
50
o
0>
O
O
O
v.
in
o>
o
V)
O
O
30
20
10
Ion
exchange
Adsorption _
Filtration
20 40 60 80
TREATMENT,%
100
Figure 43. Relationship between total water cost and
treatment 5
142
-------
should be consulted and outside technical help sought if necessary. For
more information about wastewater treatment, consult a text such as
34
that by Eckenfelder, Industrial Water Pollution Control.
SOLID WASTES
Just as disposal of waste water from scrubbers results in a cost to
the user, so does disposal of solid wastes from dry collectors, such
as cyclone precipitators, baghouses, wall filters, rotary drum filters,
etc., result in some expense.
Solid waste disposal cost can be broken down into.costs of hauling and
costs of disposal. Hauling costs are dependent on the type of equip-
ment, length of hauls, type of route and traffic encountered, and the
number of employees necessary. Figure 44 shows how this type of
calculation was made, for residential collections. The data are inten-
ded for the purpose of demonstrating the method, but appear acceptable
for rough estimates. Costs of disposal usually means the cost of a
sanitary landfill, although other disposal methods such as ocean dumping
or mine refilling are possible. Components of sanitary landfill cost
are cost of site, degree of compaction, and cost of developing such
things as access roads, water supply, fences, landscaping, water runoff
diversion facilities, etc. Experience shows a cost of $2.00 to $7.00/ton
of refuse handled. In many instances, a particular industrial facility
will contract to have solid wastes removed. It is hoped that the above
comments will be helpful in determining price.
143
-------
2 6
o:
ui
CL
CO
O
o
10 20 30 40 50 60 70 80 90 100
ROUNDTRIP TRANSPORT TIME (minutes)
Figure 44. Collection vehicle hauling costs
36
The following data was used to derive the graph above:
(1) time-based transportation costs, yearly
• vehicle amortization
• driver's salary and fringe benefits
• collector's salary and fringe benefits
• vehicle insurance, licenses, taxes
total time-based transportation cost per minute
(2) mileage-based transportation costs
• fuel, oil, tires cost per mile
• maintenance and repair cost per mile
total mileage-based transportation cost per mile
direct haul
$ 5260
10625
9375
1500
26760
$0.214
0.080
0.050
0.130
144
-------
SECTION XV
POWER COMPARISON
Most of the novel control devices tested used power in pumping scrubbing
liquid or in moving the gas or both. The Aronetics and the Lone Star
scrubbers derived power from waste heat, which in some sense is "free"
and should not be treated in the same manner as purchased power. The
major parameters by which control systems can be compared are their col-
lection efficiencies as functions of particle size and their total annual-
ized costs. A significant factor in the latter is power consumption.
Moreover, the nation is becoming more conscious of the need to conserve
energy. For these reasons we have presented in this brief section
Table 13 which contains a summary of the power requirements of the control
devices evaluated. The first column lists the devices. The next four
columns have power consumption per 1000 acfm of gas treated, separated
into blower power, pump power, other power, and total power consumption.
(To obtain the metric equivalents, kW/m Is, multiply hp/1000 cfm values
by 1.58.)
Figure 45 contains theoretical/empirical lines presented by Calvert for
various scrubber types. The aerodynamic cut diameter is graphed against
various measures of power consumption (hydraulic power per volume flow
rate). (There is a bit of a difference between Calvert's aerodynamic
cut diameter definition and the standard definition. This difference
has been neglected in the drawing of the curves and the addition of the
data from the novel device evaluations. (See Appendix.) Point A is for
the Aronetics scrubber if the waste heat is free and not counted as power
consumed. Point B is for the Braxton device used with a low energy
145
-------
Table 13. POWER CONSUMPTION FIGURES FOR NOVEL CONTROL
DEVICES EVALUATED FOR IERL-RTP
Control system
Aronetics Two-Phase
Scrubber
Braxton Sonic
Agglomerator
Centrifield Scrubber
Cleanable High
Efficiency Air Filter
Dynactor
Scrubber
Lone Star Steel
Steam- Hydro Scrubber
Mystaire
Pentapure Scrubber
U. of W. Electro-Static
Spray Scrubber
Power consumption, hp/1000 acfm
Air blower(s)
0.0
0.5
5.3e
0.83g
2.8
5.5
0.0
0.0
1.3
1.6
Water pump(s)
2.4
f
f
f
f
f
3.3
f
f
f
Other
400°
0.38
0.38
0.38
0.0
0.0
0.0
200-300°
0.0
0.0
Total
26.
2.4d
0.9
l.'2?8
2.8
5.5
3.3
200-300
0.0d
1.3
1.6
~ 0.5
a1.58 kW/(m /s) = 1.58 kj/m = 1.00 hp/1000 cfm.
Blower or pump plus motor assumed to have 60 percent overall power efficiency.
c
Waste heat.
Omitting waste heat consumption.
elncluding very high-efficiency cyclone, Ap = 30"H20 = 75 cm WC.
f
Unspecified, assumed negligible.
or
Including low-energy scrubber, Ap = 3 cm WC.
146
-------
P E
o
8
o u
o
w
E
UJ
K
3
>
in
UJ
a:
o.
q
ro
O
«NJ
CO
O
to
6
10
o
co O
3
01 0)
CO k
•i-l O
T3 CO
CO O
4J
oo IH co
CU
CO
0)
co
CO
3
O
H-)
,0
CN
,
C r-l
CO CO CO
•^ >
CO -i-l
C»1 CN *J
CJ
co co co
C co ft.
•rl C CO
i-< •!-! CO
iro
B cu co C o
c o. co 3
>> CO r-l r-l O
T3 T3 Pu O
O co U •>
M ^^ TJ CN
CO C CO CO •
O -rl A! O
•O V-) CO CJ
CO 4J CO II
4J CX >-l O.
CJ S O H-l
•rl 3 «W CO
T3 w ,e
CO C CO Vl 4-1
rl O rl O -rl
fL| CJ CO 4-1 S
-------
scrubber itself having a cut diameter around 3 urn. Point C is for the
Centrifield. Point D is for the Dynactor. Point P is for the Pentapure.
The power is in hydraulic terms for purposes of comparison (same as
assuming motor/fan/pump efficiency of 100 percent) where fan or pump
power is considered. The Centrifield and Dynactor are comparable in
performance to a well-designed venturi with regard to power consumption.
The Pentapure and Braxton use more power than other scrubbers for the
same efficiency. The Aronetics uses less, if waste heat is "free."
If the waste heat energy is included, the Lone Star and Aronetics de-
vices do not differ greatly from the extrapolated cut diameter versus
power consumption shown in Figure 45, although such extrapolation is
questionable.
148
-------
SECTION XVI
APPLICATIONS
In this section we present some of the considerations that enter into
the applicability of a control device. This information is important
for potential users of such devices and for those who wish to assess
the potential significance of innovations in fine particle control
technology.
CONTROL SYSTEMS SELECTION CONSIDERATIONS
Figure 46 is from the HEW document Control Techniques for Particulate
37
Air Pollutants. It summarizes the considerations which apply to the
choice of pollution control equipment for particulate emissions reduc-
tion. The control goals are usually set by law or by policy of the
facility operators.
The controlled emissions will have a mean concentration equal to the
mean overall mass efficiency of the control device (See Section XIII)
times the mean mass concentration of the uncontrolled emissions. The
collection efficiency required and the particle size distribution of
the uncontrolled emissions will often limit the control device alter-
natives. Combustible material may be handled with an afterburner;
particulates predominantly larger than 5 or 10 um can often be con-
trolled to 90 percent or better with a cyclone (dry centrifugal col-
lector); finer particles will generally require either a wet collector
(spray, packed bed, etc.) or an electrostatic precipitator or a fabric
filter. The alternatives still feasible after evaluation with regard
149
-------
EMISSIONS AND EMISSIONS
STANDARDS
DETERMINES COLLECTION EFFICIENCY
.CONTROL EQUIPMENT ALTERNATIVES
DRY
CENTRIFUGAL
COLLECTOR
ELECTROSTATIC
WET
COLLECTOR
GAS STREAM
CHARACTERISTICS
PARTICLE
CHARACTERISTICS
VOLUME
TEMPERATURE
MOISTURE CONTENT
CORROSIVENESS
ODOR
EXPLOSIVENESS
VISCOSITY
PROCESS
IGNITION POINT
SIZE DISTRIBUTION
ABRASIVENESS
HYGROSCOPIC NATURE
ELECTRICAL PROPERTIES
GRAIN LOADING
DENSITY AND SHAPE
PHYSICAL PROPERTIES
WASTE TREATMENT
SPACE RESTRICTION
PRODUCT RECOVERY
PLANT
FACILITY
WATER AVAILABILITY
FORM OF HEAT RECOVERY
(GAS OR LIQUID)
ENGINEERING STUDIES
HARDWARE
AUXILIARY EQUIPMENT
LAND
STRUCTURES
INSTALLATION
START-UP
COST OF
CONTROL
< •
POWER
WASTE DISPOSAL
WATER
MATERIALS
GAS CONDITIONING!
LABOR
TAXES
INSURANCE
RETURN ON INVESTMENT
SELECTED
GAS CLEANING SYSTEM
DESIRED EMISSION RATE
Figure 46. Criteria for selection of gas cleaning equipment
37
150
-------
to collection efficiency will then have to be compared with respect
to process parameters, available facilities, control costs. Once the
evaluation has been made, a cleaning system type may emerge as best.
The same kind of evaluation would then be repeated for various candi-
dates of this type, probably employing methods for cost, power, and
efficiency analysis similar to those in Sections XIII, XIV and XV.
COMPARISONS
37
From the same HEW reference is Table 14 which gives brief comparisons
of the advantages and disadvantages of settling chambers (gravitational
collectors), cyclones, wet collectors, electrostatic precipitators, and
fabric filters. Table 15 comes from Strauss; it makes a somewhat
more general comparison between those devices which collect particles
dry (most filters, cyclones, most electrostatic precipitators) and
those which collect them wet (scrubbers, some electrostatic precipita-
tors, water-washed filters). If one type of control device were
clearly superior for all applications, manufacturers of the other
types would not not stay in business for long; the continued use of
all these control types is evidence that for specific applications
each has advantages that make it preferable to the other alternatives.
The potential user should investigate the various alternatives rather
fully.
Most of the novel control devices are variations on the theme of wet
9
scrubbing. As Calvert et al. note in the Scrubber Handbook, scrubbers
as a group have the following advantages and disadvantages, generally:
capital cost is relatively low; explosions are unlikely; they are small
and simple; the particulate material is removed continuously. On the
other hand, operating costs are relatively high; often the efficiencies
are substantially less than 90 percent on particles between 0.1 and 1 pm;
water treatment is required; clogging, corrosion, and erosion can be
problems; cooling the gas and increasing its volume (water vapor) may be
151
-------
oi o
rH rH
U MH
>rf
*» -g
Vi 0
is co
a
a
•o
I
u a •
to »
a
M>
3
I
O
U
s
I
o"
0)
U
01
U
CO
O
4J
3
-o
oi
«
co -i
a. u
•o ._
a a
H <
•rl O i -
3 iH f>
0* *W V
0> t4 -rl
rl 0) 14
B fl >
O T4 O
r< 4J 4J
O
•a oi o)
CO rH >
£W
O "H
o
M
a 9
S1*5'
00
•o -g u
rH e o
O M w -rJ
to
O
o
n
•rl <«
01 O
-we
§,
rH
01
i g "e S
C tH 00 O CO
•rl CO S -rt > „ _,-
S 4J N O rl -O 4J
4-1 Q 0) 3 01 C CO
§01 01 -O 4J 3 rH
H 0) CO 0)
DON > o p
B .rl i) rH .rl
10 CO > 4J -H C T5 0)
U CO CO SI -rl
•rl —I O*3 U O
rH O IH (3 3 -rl
•H B >4 CO -O OH
.43 O 0) *W
«C O (4 [4
•rl
CO
rr
0)
CD
O
IH
0)
O,
•D
0)
U
U
01
a
a
u
13
4J
s;
•o
o)
rH
U
•r4
4J
tO ^1
&S.
£i
u
0)
BS
S i
(U U
£
• -rl CO CO
u E O. co oo
U 0) rl
O O 0) 00 01 O
*•* d ^ C P.I 4J
O CO -rl vH O
0) C to a. o
rH 0) 01 O 01 O
rH 4J .e E ja m
O B •O m
O -iH CO ^ f* ^^
S -£S
u
a
2
V
I
ta
U
a
•rl 4J
-------
Table 15. ADVANTAGES AND DISADVANTAGES OF WET AND DRY COLLECTORS
16
Wet collectors
Dry collectors
Advantages:
1. Can collect gases and par-
ticles at the same time.
2. Recovers soluble material,
and the material can be pumped
to another plant for further
treatment.
3. High temperature gases cooled
and washed.
4. Corrosive gases and mists can
be recovered and neutralized.
5. No fire or explosion hazard
if suitable scrubbing liquor
used (usually water).
6. Plant generally small in size
compared to dry collectors
such as baghouses or electro-
static precipitators.
Disadvantages:
1. Soluble materials must be re-
crystallized.
2. Insoluble materials require
settling in filtration plant.
3. Waste liquids require disposal
which may be difficult.
4. Mists and vapours may be en-
trained in effluent gas
streams.
5. Washed air will be saturated
with liquid vapour, have high
humidity and low dew point.
6. Very small particles (sub-mi-
cron sizes) are difficult to
wet, and so will pass through
plant.
7. Corrosion problems.
8. Liquid may freeze in cold
weather.
Advantages:
1. Recovery of dry material may give
final product without further
treatment.
2. Freedom from corrosion in most
cases.
3. Less storage capacity required for
product.
4. Combustible filters may be used
for radioactive wastes.
5. Particles greater than 0.05 microns
may be collected with long equip-
ment life and high collection
efficiency.
Disadvantages:
1. Hygroscopic materials may form
solid cake and be difficult to
shake off.
2. Maintenance of plant and disposal
of dry dust may be dangerous to
operatives.
3. High temperatures may limit means
of collection.
4. Limitation of use for corrosive
mists for some plants (e.g., bag-
houses) .
5. Creation of secondary dust problem
during disposal of dust.
153
-------
undesirable. The emphasis of most innovations in scrubber technology
presented here as novel control devices has been to increase collection
efficiency or decrease operating cost (power) or both.
EXTANT APPLICATIONS
Assuming that current users of control devices have generally made
rational choices in the systems employed, one could gauge the applica-
bility of a control method on the uses to which it is now being put.
Table 16 displays a list of ten major industries with the control
37
devices used predominantly. Clearly, the inventors of novel control
devices would hope that their innovations would extend the applicability
of their device to new industries or make it the method of choice from
among similar methods in an industry already employing particulate pol-
lution control. The prospective user will have to evaluate control
devices for his particular application.
SUMMARY
The principal considerations which apply to the selection of a control
system have been listed and discussed briefly. The major features of
the different types of control devices have been compared to help narrow
the field of candidate devices for a given application. Some examples
of current use are presented, with the caution that the prospective user
will want to assess the wisdom of the choice made by others, as well as
the degree to which his requirements resemble theirs, and the extent to
which the prospective control device resembles those already in use.
No choice should be made without a thorough engineering analysis.
154
-------
CO
CO
CO
Pd
Pn
Pi
H
co
g
^D
r-H
Q)
Cfl
IH
rH
0
g
o
UH
0
•n
.1
4J
S
IH
Of
4J
4-1
01
4J
CO
rH
3
U
•rH
4J
IH
CO
ssions
•g
0)
Source of
CO
CO
QI
8
g.
Ll
0
Industry
„
CO
u
*4
O.
. (j
CO U
SlH
0. •
a to
O U Ll
J! -rl O
00 4J 4J
CO CO U
rO 4J 0)
01 rH
« O rH
ui Li O
§4J O
U
O 41 *J
rH rH >H 00
C (X co
O > O S
0 -rl S
C U 4J
0) CO
O to (•?
UH Li A
UH 4J .£>
01 U 3
1
rH
W
01
4J
10
3
•n
rH
CO
O
•H
1
J3
O
n
rl 01
•rl rl 01
3 4J «rl
»H g x
r-l S O
3
CO
CO
rH
" rf
60 CU
0 43 '
•rl -rl 00
rH UH 0
•O -rl
C - M
CO CO 3
.fl cu o
Raw materials
glass furnac
forming and
Ll
JB
•rl
UH
CO
CO
CO
rH
00
rH
U
g
•H
Ll
J3
co
MH
M
CO
Ll
terburne:
UH
CO
* 01
01 IH
82
O rH
O *W
^
,
A 01
CO tj
CO
M
* oc
CO f*
rH *r|
0 0
CO Ll
O 3
IH ft
0}
CO UH
rH CO
•rl A
0 0
UH O
CO UH
U
•t
n
H -
01 CO
>, IH
M 0)
T3 rH
iH
!r. O
Roasters, spra;
waste heat b
00
a
.jj
CO
CO
0)
8
rt
SI
5
s
t
01
*J
OJ
3
*O
0)
0)
UH
Q
U
•o
CO
Ll
"X3
8P
•H
SS
coolers, con
equipment.
155
-------
SECTION XVII
COLLECTION EFFICIENCY THEORY
Each of the novel devices tested in the IEKL program is of the group of
particle collection devices which work by using the collection of par-
ticles on obstacles placed in the flow: e.g., spray scrubbers, high
porosity beds, high porosity filters. Another group is more suitably
viewed as doing their collection in channels: electrostatic precipi-
tators, cyclones, sedimentation chambers, and low porosity filters and
beds. The analysis which follows is for the first group, and to sim-
plify description we have called the device a "scrubber" and labelled
the collection obstacles as "droplets," but the generality of the
approach should not be ignored.
38
This analysis, influenced by those of Knettig and Beeckmans and
9
Calvert et al., starts with the volume of particle-laden gas
per unit time, dQ , which sweeps past droplets having a number concen-
8
tration n, and individual cross-sectional areas A, contained in the
d d
small volume A dx, where dx is a distance measured parallel to the flow
(see Figure 47). This volume rate of "sweeping" or scrubbing is:
dQg = nd Ad A dx (1)
where u = gas velocity, cm/s
O
u, = droplet velocity, cm/s.
156
-------
ud
n
:•
•J ^ L.
>i
duct
Figure 47. Geometry for collector analysis
157
-------
The rate of particle removal due to this "sweeping" is:
A u dc = - T] c dQ
s
c (u - u ) n A, A dx
g d d d
where T] = single droplet collection efficiency
.3
c = particle concentration, cm
38
or (Knettig and Beeckmans):
dc
u
8
From this point, the derivation can be done with various degrees of
sophistication:
1. The calculation of the single droplet efficiency T) can
include some or all of the following mechanisms:
impact ion
interception
diffusion
electrostatic interactions
diffusiophoresis
thermophores is
2. The velocities ug and u^ can be calculated in detail
including their dependence on position, measured parallel
to and perpendicular to the mean gas flow.
3. The spatial variation of n, can be taken into account.
4. Various averages of droplet areas, Ad, can be used, or
a functional form for their distribution employed.
158
-------
If_ the various quantities in the right-hand-side of equation (2) are
independent of position, then:
" T)
u - u ,
R d
u
g
nd Ad L
t
(3)
where c = concentration at x = 0
o
c = concentration at x = L
Droplet Concentration
One model for the concentration of the droplets, n,, is just the volume
of droplets supplied per unit time, Q,, divided by the average volume
per droplet V,> divided also by the volume rate of gas flow Q :
Qd/vd
(4)
(For packed beds, Q,/Q is replaced by the solids fraction). For spher-
ical droplets of a single diameter (Dj)> n. A, becomes:
nd Ad = ~
r V Q
6 d xg
4d
(5a)
If this is combined with an empirically obtained factor, f, for
(u - u,) / u , one has:
g d g
= exp
(5b)
159
-------
It can be shown that in calculating penetration when one uses an average
value for the argument of the exponent in equation (3), the ratio c/c
thus calculated is less than when C/CQ values are calculated from each
value of the argument and then averaged (which is much closer to the
actual physical situation). The greater the dispersion of the values
appropriate for the argument of equation (3), the greater the difference
between the value calculated by using an average and the value appropriate
to the situation. For a scrubber, this means that for droplet distribu-
tions having an optimal average cross-sectional area, the distribution
having the smallest standard deviation of size would be most efficient.
The same could be said about the advantage of flow and droplet concen-
tration uniformity.
The concentration of collectors in a bed or a filter is fairly uniform,
but this is often not true in a spray scrubber. Although the droplet
concentration is usually calculated by using equation (4), which
involves the ratio of the spray volume flow to the gas volume flow,
this is not strictly correct. In steady state, the volume flow of
droplets through the complete cross-sectional area of the duct, A*,
is equal to the injected spray volume flow, Q,:
Qd = A*"o~"d~V
the product of the area, the average droplet velocity, the average
number concentration, and average single droplet volume. This means
that the average concentration at A* is:
nd = V A* ud vd (7a)
= Qd y Qg ud vd (7b)
160
-------
not
V A*ugvd °*Qd/Qgvd (7c)
which is from equation (4).
Thus, where the droplets are moving faster than the gas, their concen-
tration will be less than equation (4) predicts, and where they are
moving more slowly than the gas, their concentration will be greater.
In general, this indicates that, other things being equal, scrubbing
becomes more intense when the gas is accelerating and less intense
where the gas is decelerating (because the droplet motion lags behind
the gas motion) .
The incorporation of equation (7a) into equation (3) yields:
c/c = exp
- 1}
(8)
Relative Velocities
In some circumstances we may be able to calculate I" (u - u,)/u 1 , but
often it is just estimated.
Droplet (Obstacle) Areas
If all droplets have the same cross-sectional area, A,, then this is
simple. If not, one may want to use a distribution of droplet cross-
sections f(A.)
foo —
where / f (A,) A, dA, = A, (9)
/OuuQQ
the average cross-sectional area.
161
-------
Single Droplet Efficiency
As noted the single droplet efficiency T) involves a number of physical
mechanisms. It is defined as:
flow area cleaned per collector (10a)
~ collector cross-sectional area
TI = A /A, (10b)
c a
This is shown in Figure 48. The dotted lines are representations of
the trajectories of particles just barely captured, the limiting
trajectories.
The calculation of the single particle efficiency, f], depends in part
upon the flow past the collector (droplet, fiber, etc.). There are
two flow models commonly in use: viscous flow and potential flow.
Viscous flow is an appropriate model when the obstacle Reynolds number
is small; that is, when:
Red = p (ug - ud) Dd /u « 1
-3 3
where p = gas density = 1.2 x 10 g/cm for air at room temperature
and pressure
-4
H = gas viscosity = 1.8 x 10 poise for air at room temperature
and pressure.
Although the motion of the dust particles in the gas stream often meets
this Reynolds number criterion, collectors usually do not. A flow
of 10 cm/s (20 fpm) past a fiber or droplet 100 urn in diameter gives
Re, ~1. The model of potential flow is derived for Re, »1, but even
in this regime it is appropriate only up to near the point on the
collector surface where the flow separates and forms a wake, which
trails behind the obstacle. Although we will work with the potential
flow approximation, its limitation should be appreciated. For those
162
-------
T-
COLLECTOR WIDTH
Figure 48. Flow streamlines and the limiting trajectories
163
-------
phenomena in which the collection occurs on the upstream half of the
obstacle, it should work well; for those with a contribution from the
downstream side, such as the case for electrical attraction, the model
39
is not strictly appropriate, though in use.
The single particle efficiency can be calculated for various collection
mechanisms separately and then combined as though the mechanisms acted
38
independently (Knettig and Beeckmans) , but this is not generally
correct. The better, though more difficult, manner for handling the
contributions of several collection mechanisms is to solve the particle
trajectory equations in the appropriate flow field including the collec-
40
tion forces and mechanisms, as done by Sparks and George and
39
Poehlein, among others. The disadvantage of the correct approach
is that it would require a great many such calculations, one each for
every specific situation of interest. This disadvantage is partially
circumvented by doing such calculations for a number of combinations of
parameters of interest and then interpolating from the results. Where
such calculations are available, one can use them; otherwise one must
employ other methods for combining the effects of different mechanisms,
simplifying the procedure whenever possible by neglecting mechanisms
which are much smaller than the primary collection mechanisms for the
particular situation. A justification for using equation (3) even though
its assumptions are not met in practice is that the accuracy to which the
other components in its argument are known is not great, so that using
them in a more accurate expression than equation (3) would be gilding
the lily, and more difficult than the situation warrants.
Impaction - When a dust particle strikes the collection surface because
of its inertia and its inability to follow the gas streamlines, the
collection is said to be due to impaction. Particles in a moving stream
will impact upon a deflecting surface provided that the inertia of the
particles is sufficient to overcome the drag exerted by the air stream
164
-------
deflected by the impaction surface. The impaction process can be
characterized by the impaction parameter i|r which is defined by the
following expression:
(12)
18 ^ Dd
where p = particle density
v = gas velocity
d = particle diameter
\i = gas coefficient of viscosity
Dd = droplet diameter
C = Cunningham correction factor defined empirically as:
C = 1 + 2.492 X/d + 0.84 (X/d ) exp (-0.435 d/X)
where X is the molecular mean free path.
The following expression has been found, empirically, to approximate
Q
the single droplet efficiency for impaction:
Tjj - ^2/ (Y + 0.35)2 (14)
In Table 17 are listed several values of T) along with the calculated
41
values from equation (14) and the results reported by May and Clifford.
The agreement seems sufficient for most purposes.
165
-------
Table 17. CALCULATED nT COMPARED WITH DATA
t
0.01
0.03
0.1
0.3
1.0
3.0
10.0
30.0
11
0.0008
0.0062
0.049
0.213
0.549
0.802
0.934
0.977
41
Experimental data
-
-
0.03
0.3
0.6
0.75
0.94
- •
Interception - Particles whose centers move along flow streamlines may
still strike the obstacle (droplet, fiber) because of their width per-
pendicular to their motion, in which case collection is said to be via
interception. The parameter which indicates the magnitude of this effect
is N , the ratio of particle radius to collector radius (spherical or
R
cylindrical collectors). The incremental efficiency due to this effect
lies between 2N and 3N for potential flow around a sphere and between
N and 2N for potential flow around a cylinder (Fuchs) for inertial-
less and highly massive particles, respectively.
Diffusion - The Brownian motion of particles causes there to be a net
diffusive flux of particles from regions of higher concentration to
regions of lower concentrations and from a gas stream to adjacent sur-
faces, if some or all of the particles adhere to the surfaces. The
Peclet number, Pe, measures the relative amount of convective transport
to diffusive transport for a moving gas stream:
Pe = u D,/ D
g d
Re, Sc
d
(15)
166
-------
2
where D = diffusivity, cm /s
= kTC/3rt(j.d for d > 0.01 urn
P P
k = Boltzmann constant
T = absolute temperature
Sc = Schmidt number = |i/Dp
Diffusion is substantial when Pe « 1.
As discussed more fully by Strauss, combining these mechanisms
and forming an appropriate single particle efficiency T] for their
J.l»L)
combination is difficult, but experimental evidence supports well the
use of the following equation:
,ICD - 6 So'2'3 Re'"2 + 3KR2 Re1'2 (16)
Figure 48a is taken Jlrom the book J>y Strauss to show the^agreement
between experimental rj , and this expression. The second term on
ICD s g 1/2
the right-hand-side is equivalent to (54 N- ty) , and dominates where
impaction is important. Diffusion is rarely important for particles of
diameter > 0.5 \j.m, so this equation is worth using only for particles
smaller than 0.5 i_im. For larger particles equation (14) should suffice.
Electrostatic Attraction - The two major types of electrostatic force
significant in particle collection are the Coulomb force a charged par-
ticle is subjected to in an electric field and the induced charge (dipole)
167
-------
I01
* Wong ond Johntlor*
• Chen
•» Thomas end Yodtr
.a'
1-0
Figure 48a. Combined collection efficiency based on experimental results
^
ICD
versus (6Sc"2/3 Re"1/2 + 3Re1/2 R2).16 (R - N_.)
K-
force created by an inhomogeneous electric field. The Coulomb force is
given by the equation
(16)
where q = particle charge, coul or stat-coul
E = electric field, v/m or stat-v/cm
The force due to induced polarization in an inhomogeneous electric field
is
F = x£ V grad (E )
(17)
where XE = (3it/8) (e - l)/(e + 2) for a sphere
e = dielectric constant of particle
P
V = particle volume.
168
-------
Because the gradient of a homogeneous field is zero, this second force
is only operative in inhomogeneous fields, such as those close to a .
charged nonplanar collector.
Generally, particles are charged for electrostatic precipitation by
2
having them pass through a region of high ion concentration (N, ions/cm )
during which time, t, they approach an equilibrium charge, q (stat-coul).
The approach to equilibrium has the time dependence (1 + t /t) where
the equilibrium time at 20°C and 1 atm is
t = 1.0 x 106 N"1 seconds. (18)
6
Many particles behave as though they were conductors during the time
scales of interest. Furthermore, control device designs usually assure
sufficiently long residence times, t » t . Under such conditions
21 e
Cochet s equation for charge is:
-ill o I
(19)
-4
where A.1 = 0.1 x 10 cm.
For d > 1 (o,m,
p ~
qp = 3 E dp*/4. (20)
The above equations enable determination of the particle charge with
several different degrees of sophistication. Table 18 has been calculated
24
using the next-to-last equation above, and is taken from work by Cooper.
169
-------
Table 18. APPROXIMATE PARTICLE CHARGE (COCHET EQUATION)
(CONDUCTING PARTICLES; E = 10 kV/cm =33.3 esu)
Particle diameter,
d » (urn)
0.1
0.3
1.0
3.0
10.0
Particle charge,
q , (esu)
8.05 x 10"9
2.80 x 10"8
2.59 x 10"7
2.26 x 10"6
2.50 x 10"5
Number of
electrons
16.8
58.3
539.6
4.7 x 103
5.2 x 104
The Coulomb force is proportional to the electric field. In practice,
it is difficult to get fields that are very much stronger than 33 stat-
volt/cm (10 kV/cm) because of electrical breakdown of air at such high
fields. This value has been used to calculate reasonable upper limits
on particle migration velocities.
The induced polarization force is proportional to the gradient of the
field, which can be estimated for spheres or cylinders by the field at
the surface divided by the radius, as we have done.
23
Figure 49 (from Cooper ) shows the results of calculations of migration
velocities assuming maximum particle charges, fields, and field gradients,
as indicated above, for particles of the size indicated being collected
by a 100 |_im spherical collector. The case of charged particle and
charged collector produced the largest migration velocities, followed
by charged collector with uncharged particles; charged particles with
uncharged collector; and mutual repulsion of charged particles in the
vicinity of an uncharged collector, assuming maximum charge, mass median
diameter indicated, mass concentration of Ig/m , and particles of unit
density.
170
-------
I0
10'
6
u
g '0°
o
uj -5
•v
.3
O
o
2
10
lO'
10
'4
X'
X
'
/
fr'V
\
\
X
/
x
/ • I. COULOMB FORCE-CHARGED PARTICLE IN
A FIELD.
A 2. CHARGED PARTICLE WITH UNCHARGED
COLLECTOR.
D 3. IMAGE FORCE - CHARGED COLLECTOR WITH
UNCHARGED PARTICLE.
X 4. CHARGED PARTICLE IN SPACE CHARGE FIELD.
I I
.3 1.0
PARTICLE DIAMETER,
3.0
10.0
Figure 49. Theoretically calculated migration velocities for four
electrostatic mechanisms versus particle diameter
171
-------
The migration velocity, w, is defined so that the product of migration
velocity and collector surface area, A , is the same as the product
s
of the target efficiency and relative velocity and collector cross-
sectional area,
w A = TI (u - u ,) A,,
s g d' d
a definition which breaks down for u = u,.
8 d
For electrostatic collection, the migration velocity is independent of
u and u ; it is the terminal velocity at which the fluid drag force on
the particle equals the electrostatic force. Because A and A, are the
s a
same order of magnitude, electrostatic migration will be at least as
important as the aforementioned mechanisms where
W/T1ICD (ug - Ud> * l' (22)
For collection to be appreciable, one also needs
w (SAe) / 0 > 1. (23)
Here 2A is the total collection surface area.
s
Figure 49 allows an estimation of w.
Path Length, L
If the collectors are a fixed bed, then its length parallel to flow is used
as L in equation (8). If a spray or other moving spherical collectors are
used, then L is the net travel distance with respect to the gas under-
gone by the collectors. The net distance is determined by the drag
coefficient, CD, where
172
-------
2 <>
173
-------
An approximation scheme should suffice, using the following formulas,
depending upon initial Re,:
Re. ^ 1 t = CP d 2/18 v (31)
d p
1 < Red ^ 10 t* =0.72 t (32)
10 < Red <: 400 t* = 0.44 t (33)
For scrubbers, L increases with droplet size and with initial velocity
differences.
SUMMARY
The calculation of collection efficiency for a given system can be quite
difficult, despite the availability of the theoretical expressions.
Theory is valuable, however, in showing what factors enhance or interfere
with collection efficiency and indicating how strongly efficiency is
dependent upon the factors.
Those who wish more information on the following topics should consult
the indicated references:
42
Aerosol Behavior - Fuchs, Mechanics of Aerosols; Green and
44
Lane, Particulate Clouds; Davies, Aerosol
o , 45
Science
46
Filtration - Davies, Air Filtration; Billings and Wilder,
47
Handbook of Fabric Filtration Technology;
Journal of Air Pollution Control Associa-
tion, 24 (12), December 1974
174
-------
g
Scrubbing - Calvert et al., Scrubber Handbook; Strauss
Industrial Air Pollution Control; Journal
of the Air Pollution Control Association,
24 (10), October 1974
Electrostatic Precipitation - White, Industrial Electrostatic Precipita-
21
tion; Oglesby and Nichols, A Manual of
48
Electrostatic Precipitator Technology;
Journal of the Air Pollution Control Associ-
ation, 25 (2), February 1975
175
-------
SECTION XVIII
APPENDIX: AERODYNAMIC DIAMETER DEFINITION
In expressing particle size, we have often used "aerodynamic diameter"
rather than geometrical diameter, especially in discussing collection
efficiency as a function of particle size. For situations such as
gravitational settling and inertial impaction where the force on the
particle is proportional to its mass, aerodynamic diameter is the
particle "size" which governs its motion.
Stober discussed the origin of the term. He also gave a clear defi-
3
nition: "the diameter of a sphere of unit density (p =1.0 g/cm )
o
attaining at low Reynolds numbers in still air the same final settling
velocity as the actual particle under consideration." As he noted,
this parameter is of physical interest and can be obtained by measure-
ments in some cases where other measures of particle size and density
are not obtainable. This definition means that the particle terminal
settling velocity, v , can be used to obtain the aerodynamic diameter,
s
d :
ae
Vs ' 8 C(dae) po dae2/18 y ' (34)
for Stokes law settling (particle Reynolds number < 1).
Similarly, the aerodynamic diameter can be measured by impaction
efficiency under known conditions. If the particle is a sphere with
diameter d , then the following relationship holds:
176
-------
C(d ) p d 2 = C(d ) p d 2, (35)
ae o ae P P P
where C(d) = Cunningham slip correction factor;
= 1 + (0.16 um)/d, at normal temperature and pressure
(25°C, 1 atm),
and the appropriate diameter, d, must be used in each calculation of C.
Given the aerodynamic diameter or the particle diameter and density,
collection by gravitational settling, centrifugal sedimentation, or
inertial impaction can be readily calculated from available formulas.
It is not generally recognized, however, that going from particle
diameter and density to particle aerodynamic diameter or vice-versa
is not routine; for example, where the Cunningham correction is
appreciable,
d 4 /C(d ) p d . (36)
ae ^ p p p
p
The Scrubber Handbook, it should be noted, defines an "aerodynamic
diameter" as:
V' C< " " <37)
and seems to use this definition consistently, even though recogniz-
ing the unusual units it implies, cm(g/cm )*•'*•.
To go from particle diameter and density to aerodynamic diameter or
vice-versa must be done by solving the equation
po dae = C(V pp
177
-------
This can be done in two fairly easy ways:
a) Using the formula for the Cunningham correction given here, one
obtains a quadratic equation in powers of the diameter sought, d:
ad2 + bd = - c (39)
the solution to which is the well-known formula
d = (_b + (b2 - 4ac)1/2)/2a. (40)
b) Setting the unknown value of the Cunningham correction equal to
one, C = 1, the first approximation to the unknown diameter, d-, is
obtained from Equation (38); then this first approximation is used to
evaluate the unknown Cunningham correction factor C(d-) and a second
approximate solution, d«, is obtained, also from equation (38). A
third approximation, d , using the d_ value for the unknown diameter,
is rarely needed.
178
-------
SECTION XIX
REFERENCES
1. McCain, J. D. Evaluation of Aronetics Two-Phase Jet Scrubber.
Office of Research and Development. U.S. Environmental Protection
Agency. EPA-650/2-74-129. December 1974.
2. Dennis, R., R. Bradway, and R. Cass. Braxton Sonic Agglomerator
Evaluation. Office of Research and Development. U.S. Environmental
Protection Agency. EPA-650/2-74-036. May 1974.
3. Leith, D. and D. Mehta. Cyclone Performance and Design. Atmo-
spheric Environment, Pergamon Press. 7:527-549. 1973.
4. Calvert, S. Engineering Design of Wet Scrubbers. J APCA.
24:929-934. 1974.
5. McCain, J. D. Evaluation of Centrifield Scrubber. Office of Re-
search and Development. U.S. Environmental Protection Agency.
EPA-650/2-74-129-a. June 1975.
6. Calvert, S., J. Rowan, and C. Lake. Experimental Tests of Novel
Fine Particulate Collection Devices. Preliminary Report. Contract
No. 68-02-1496. Task Order No. 1. January 13, 1975.
7. Cooper, D. W. and D. P. Anderson. Dynactor Scrubber Evaluation.
Office of Research and Development. U.S. Environmental Protection
Agency. EPA-650/2-74-083-a. June 1975.
8. Calvert, S., J. Goldschmid, D. Leith, and D. Mehta. Scrubber
Handbook. Control Systems Division, Office of Air Programs. U.S.
Environmental Protection Agency, Research Triangle Park, N. C. 1972.
9. Lancaster, B. W. and W. Strauss. A Study of Steam Injection Into
Wet Scrubbers. Ind Eng Chem Fundamentals. 10(3):362-369. March
1971.
10. Lohs, W. Manufacture of Aerosols and Separation of Ultrafine Dusts
in Spray Washers. Staub, 29(2):43. 1969.
179
-------
11. Calvert, S., D. Lundgren, and D. S. Mehta. Venturi Scrubber
Performance. J Air Pollut Contr Assoc. 22:529-532. 1972.
12. Bird, R. B., W. E. Stewart, and E. N. Lightfoot. Transport
Phenomena. Wiley, New York. 1960.
13. McCain, J.D., and W.B. Smith. Lone Star Steel Steam-Hydro Air
Cleaning System Evaluation. Office of Research and Development.
U.S. Environmental Protection Agency. EPA-650/2-74-028.
April 1974.
14. Ananth, K.P. Evaluation of the Mystaire Scrubber. Draft of
Final Report. EPA Contract No. 68-02-1324, Task Order No. 7.,
for Control Systems Laboratory. Office of Research and Devel-
opment. U.S. Environmental Protection Agency.
15. Carpenter, C. L. and D. F. Othmer. Entrainment Removal by a
Wire-Mesh Separator. A.I.Ch.E.J. 1:549-557. 1955.
16. Strauss, W. Industrial Gas Cleaning. Pergamon, New York. 1966.
17. Cooper, D.W. Pentapure Impinger Evaluation. Office of Research
and Development. U.S. Environmental Protection Agency.
EPA-650/2-75-024a. March 1975.
18. Ananth, K.P., Evaluation of the Pentapure Impinger. Final Report.
EPA Contract No. 68-02-1324, Task Order No. 6., for Control
Systems Laboratory. Office of Research and Development. EPA.
19. Greene, F.T., L.J. Shannon. Evaluation of Electrostatic Droplet
Scrubber. Contract No. 68-02-1324, Task Order No. 16., for
Control Systems Laboratory. Office of Research and Development.
EPA.
20. Pilat, M. J. Collection of Aerosol Particles by Electrostatic
Droplet Scrubbers. J Air Pollut Contr Assoc. 25:176-179. 1975.
21. White, H. J. Industrial Electrostatic Precipitation. Pergamon,
New York, 1963.
22. Pilat, M. J., S. A. Jaasund, and L. E. Sparks. Collection of
Aerosol Particles by Electrostatic Droplet Spray Scrubbers.
Environ Sci & Technol. 4:360. 1974.
23. Cooper, D. W., Fine Particle Control by Electrostatic Augmentation
of Existing Methods. Paper No. 75-02.1 presented at the 68th
Annual Meeting of the Air Pollution Control Association. Boston,
Mass. June 1975.
180
-------
24. Sparks, L. E., and M. J. Pilat. Effect of Diffusiophoresis on
Particle Collection by Wet Scrubbers. Atmos Environ. 4:651-660.
1970.
25. Lippman, M. Respirable Dust Sampling. Am Indus Hyg Assoc J.
31:138. 1970.
26. Friedlander, S. K. Small Particles in Air Pose a Big Control
Problem. Environ Sci & Technol. 7:1115-1118. 1973.
27. Edmisten, N. G., and F. L. Bunyard. A Systematic Procedure for
Determining the Cost of Controlling Particulate Emissions from
Industrial Sources. J Air Pollut Contr Assoc. 20:446-452. 1970.
28. Hanf, E. M., and J. W. MacDonald. Economic Evaluation of Wet
Scrubbers. Chem Eng Prog. 7(3):48-52. 1975.
29. Eraser, M. D., and D. R. Eaton. Cost Models for Venturi Scrubber
System. Presented at 68th Annual Meeting of APCA, Boston, Mass.
1975.
30. Fraser, M. D., and G. J. Foley. Cost Model for Fabric Filter
Systems. Presented at 67th Annual Meeting of the APCA. Denver,
Colorado. 1974.
31. McKenna, J. D. Applying Fabric Filtration to Coal-Fired Boilers.
Environmental Protection Agency Report EPA-650/2-74-058.
July 1974.
32. Schneider, G. G., et al. Selecting and Specifying Electrostatic
Precipitators. Chem Eng. 94-108. May 26, 1975.
33. Benson, J. R., and M. Corn. Costs of Air Cleaning With Elec-
trostatic Precipitators at TVA Steam Power Plants. J Air Pollut
Contr Assoc. 24(4):340-348. 1974.
34. Eckenfelder, W. W., Jr. Industrial Water Pollution Control.
McGraw-Hill. 1966.
35. Eckenfelder, W. W., Jr. Water Quality Engineering for Practicing
Engineers. Barnes and Noble. 1970.
36. Greco, J. R. Transfer Station Feasibility is Measured Against
Direct Haul. Solid Wastes Management. 17. April 1974.
37. Control Technology for Particulate Air Pollution. U.S. Department
of Health, Education and Welfare. Washington, D.C. January 1969.
181
-------
38. Knettig, P. and J.M. Beechmans. Inertial Capture of Aerosol
Particles by Swarms of Accelerating Spheres. J Aerosol Sci.
5:225. 1974.
39. George, H. F., and G. W. Poehlein. Capture of Aerosol Particles
by Spherical Collectors: Electrostatic, Inertial, Interception, and
Viscous Effects. Environ Sci & Technol. 8:46. 1974.
40. Sparks, L. E. The Effect of Scrubber Operating and Design Para-
meters on the Collection of Particulate Air Pollutants. Ph.D.
Dissertation (Civil Engineering) University of Washington, 1971.
41. May, K. R. and R. Clifford. The Impaction of Aerosol Particles on
Cylinders, Spheres, Ribbons, and Discs. Ann Occup Hyg. 10:83-95.
1967.
42. Fuchs, N. A. The Mechanics of Aerosols. The MacMillan Company.
New York 1964.
43. Ingebo, R. Drag Coefficients for Droplets and Solid Spheres in
Clouds Accelerating in Airstreams. NACA Technical Note 3762.
1956.
44. Green, H. L. and W. R. Lane. Particulate Clouds: Dusts, Smokes
and Mists. Spon, England. 1964.
45. Davies, C. N., Editor. Aerosol Science. Academic Press, New York.
1966.
46. Davies, C. N. Air Filtration, Academic Press, London and New
York. 1973.
47. Billings, C. E., and J. E. Wilder. Handbook of Fabric Filter
Technology. Volume 1. Fabric Filter Systems Study, GCA/Tech-
nology Division. Department A, Clearinghouse, U.S. Department
of Commerce, Springfield, Va. 22151. Report Number GCA-TR-70-17-G,
APTD-0690. Contract CPA-22-69-38. PB-200-648. December 1970.
48. Oglesby, Sabert, Jr. and Grady B. Nichols. A Manual of Elec-
trostatic Precipitator Technology. Part 2: Application Areas.
Southern Research Institute, Birmingham, Alabama. August 25, 1970.
49. Weast, R. C. Standard Mathematical Tables. Chemical Rubber Co.
Cleveland, Ohio. 1968.
50. Stober, W. A Note on the Aerodynamic Diameter and Mobility of
Nonspherical Aerosol Particles. J Aerosol Sci. 2:453-456. 1971.
182
-------
TECHNICAL REPORT DATA
(Please read tiiilfucnons on the reverse before completing)
1. REPORT NO.
EPA-60Q/2-76-035
4. TITLE ANDSUBTITLE
Evaluation of Eight Novel Fine Particle
Collection Devices
7 AUTHOR(S) Douglas W. Cooper, Richard Wang, and
Daniel P. Anderson
3. RECIPIENT'S ACCESSION NO.
5. REPORT DATE
February 1976
6. PERFORMING ORGANIZATION CODE
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
GCA Corporation
Burlington Road
Bedford, MA 01730
10. PROGRAM ELEMENT NO.
1AB012; ROAP 21ADL-004
11. CONTRACT/GRANT NO.
68-02-1316, Task 8
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Task Final: 2-11/75
14. SPONSORING AGENCY CODE
EPA-ORD
15. SUPPLEMENTARY
number for this report is PB 251-621/AS.
EPA project officer for this report is D.Harmon, 919/549-8411, Ext 2925.
16. ABSTRACT
The report summarizes IERL-RTP sponsored evaluations of eight novel
control devices: Aronetics Two-Phase Jet Scrubber, Braxton Sonic Agglomerator,
Centrifield Scrubber (Entoleter Corp.), Cleanable High Efficiency Air Filter
(CHEAF, Johns-Manville), Dynactor Scrubber (RP Industries), Lone Star Steel
Steam Hydro Scrubber, Mystaire (Heat Systems-Ultrasonics, Inc.), and Pentapure
Scrubber (Purity Corp.). For each device is given a description,'potential
applications, theory of operation, collection efficiency (as a function of particle
size), power consumption, and waste disposal techniques. Methods are detailed
for comparing collection efficiencies and costs. The general theory of collection
in scrubbers and by high porosity filters is presented to indicate the important
parameters and their influence on collection efficiency.
7.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
Air Pollution
Control Equipment
Scrubbers
Filters
Acoustics
Collection
Agglomeration
Efficiency
b.IDENTIFIERS/OPEN ENDED TERMS
Air Pollution Control
Stationary Sources
Novel Control Devices
Particulate
Fine Particles
Sonic Agglomerators
c. CO3AT! Field/Group
13B ~~
14B
07A
20A
B. DISTRIBUTION' STATEMENT
Unlimited
19. SECURITY CLASS (TinsKeportl
Unclassified
21. NO. OF PAGES
190
20. SF.CURITY CLASS (Thispage)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
183
------- |