EPA-650/2-74-036
May 1974
Environmental Protection Technology Series
:*:¥:*:¥:*:*:
T>
z
UJ
o
T
>
iSsiS
;:•:•$>:¥;
•:•:•:•:¥:$
Si-i-xS
:|:i:|:iji:
i'S-iSi-ft
i;:;i;i
^^i^|^^^^|^||^i§^^|^^i|:;il
-------
EPA-650/2-74-036
BRAXTON SONIC AGGLOMERATOR
EVALUATION
by
Richard Dennis , Robert Bradway ,
and Reed Cass
GCA Corporation
GCA/Technology Division
Bedford, Massachusetts 01730
Contract No. 68-02-1316 (Task 1)
ROAP No. 21ADL-04
Program Element No. 1AB012
EPA Task Officer: Dale Harmon
Control Systems Laboratory
National Environmental Research Center
Research Triangle Park, North Carolina 27711
Prepared for
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, D.C. 20460
May 1974
-------
This report has been reviewed by the Environmental Protection Agency
and approved for publication. Approval does not signify that the
contents necessarily reflect the views and policies of the Agency,
nor does mention of trade names or commercial products constitute
endorsement or recommendation for use.
11
-------
ABSTRACT
This report presents the evaluation of a novel air pollution control
system developed by the Braxton Corporation. The alternating velocity
precipitator, or sonic agglomerator, is designed to decrease the number
and increase the size of particles in a gas stream by agglomeration in-
duced by a standing sound wave through which the aerosol moves. A proto-
type unit of the alternating velocity precipitator was tested to determine
its basic performance characteristics and to evaluate the effect of the
addition of water and/or steam to the system's performance.
Results from those tests using resuspended cupola dust indicate that
the device decreases the mass of fine particles but that the reduction is
more highly correlated to the use of water sprays than to the use of the
sonic generator. The correlation coefficients relating the use of sound
and water sprays to fine particle reduction are statistically significant
but the correlation coefficient for steam addition is not statistically
significant.
The particle size distributions of the fine particles at both the inlet
and outlet of the sonic agglomerator were determined with Andersen cascade
impactors. Although shifts between the inlet and outlet size distributions
were often observed, there was no clear trend to the changes and they could
not be correlated to system operating parameters.
111
-------
CONTENTS
Page
Abstract iii
List of Figures v
List of Tables vii
Acknowledgments viii
Sections
I Conclusions 1
II Recommendations 3
III Introduction 4
IV Sampling Methods 6
V Analysis Methods 10
VI Results 11
VII Appendices 44
IV
-------
FIGURES
No. page
1 Schematic of Sampling Train 7
2 Effect of Water Flow Rate on the Reduction
of Fine Particles 17
3 Effect of Steam Flow Rate on the Reduction
of Fine Particles 18
4 Effect of Sonic Generator Power Consumption
on Reduction of Fine Particles 20
5 Particle Size Distribution of Cyclone Undersize Dust
for Test 1 23
6 Particle Size Distribution(s) of Cyclone Undersize
Dust for Test 2 2.4
7 Particle Size Distribution of Cyclone Undersize Dust
for Test 3 25
8 Particle Size Distribution(s) of Cyclone Undersize
Dust for Test: 4 26
9 Particle Size Distribution(s) of Cyclone Undersize
Dust for Test 5 27
10 Particle Size Distribution(s) of Cyclone Undersize
Dust for Test 6 28
11 Particle Size Distribution(s) of Cyclone Undersize
Dust for Test 7 29
12 Particle Size Distribution(s) of Cyclone Undersize
Dust for Test 8 30
13 Particle Size Distribution(s) of Cyclone Undersize
Dust for Test 9 31
14 Particle Size Distribution(s) of Cyclone Undersize
Dust for Test 10 32
15 Particle Size Distribution(s) of Cyclone Undersize
Dust for Test 12 33
16 Particle Size Distribution(s) of Cyclone Undersize
Dust for Test 13 34
-------
FIGURES (continued)
No. Page
17 Particle Size Distribution(s) of Cyclone Undersize
Dust for Test 14 35
18 Particle Size Distribution(s) of Cyclone Undersize
Dust for Test 15 36
19 Particle Size Distribution(s) of Cyclone Undersize
Dust for Test 16 37
20 Particle Size Distribution(s) of Cyclone Undersize
Dust for Test 17 38
21 Particle Size Distribution(s) of Cyclone Undersize
Dust for Test 18 39
22 Effect of AVP System on Mass Median Diameters of
Cyclone Undersize Dust 41
VI
-------
TABLES
No. Page
1 Sonic Agglomerator System Operating Parameters 12
2 Comparison of Inlet and Outlet Dust Loading, Cyclone
Catch, Cyclone Undersize Dust and Overall System
Collection of all Particle Sizes 14
3 Comparative Concentration Measurements of Cyclone
Undersize Dust with Glass Fiber Filters and Andersen
Impactors 22
4 Estimated Operating Costs of AVP System 43
vii
-------
ACKNOWLEDGMENTS
The information and support provided by Mr. Ed Kent and the Braxton
staff are acknowledged with sincere thanks.
viii
-------
SECTION I
CONCLUSIONS
The prototype sonic agglomerator tested in this program collected 42 to
59 percent of the test dust entering it without the use of sound, water or
steam. The collection efficiencies using resuspended cupola dust and dif-
ferent combinations of sound, steam and water ranged from 81 to 97 percent.
These efficiencies are based on using a cyclone after the sonic agglomerator
equal in collection efficiency to the cyclone used for sampling during these
tests.
The amount of fine particles, as defined by the cyclone undersize dust
(less than about 2 ^m), was measured upstream and downstream of the chamber
and the use of sound and water sprays reduced the fine fraction at thp out-
let. The reduction in fine particles as a result of conditioning by the
sonic agglomerator system under various operating modes ranged from 14 to
83 percent.
The correlation coefficients for the small particle reduction for all
test runs with sound, water and steam individually and neglecting other
operating parameters were 0.68, 0.65 and 0.06 respectively. The coef-
ficients for sound and water are statistically significant while that for
steam is not. A multiple regression analysis of those tests using cupola
dust shows that the use of steam is not significant but that the use of
water sprays and the sonic generator are highly significant in reducing the
mass of fine particles. The regression analysis also shows that the sonic
generator has less influence on fine particle reduction than does water
sprays when using resuspended cupola dust.
Cascade impactor samples of the cyclone undersize material at the inlet
and outlet of the unit showed no consistent trend in the shift of particle
size distributions. Some tests resulted in an outlet dust with a larger mass
median diameter than the inlet dust, others resulted in an outlet dust with
a smaller mass median diameter than the inlet, and some tests showed no
change. These results do not appear to correlate with any system operating
parameters.
-------
The estimated annual operating costs for the sonic agglomerator system
using sound and water is $13.30/nr/minute ($0.38/cfm). The use of a nominal
amount of steam in addition to sound and water results in an annual operating
cost of $46.22/m3/minute ($1.32/cfm). These estimates are based upon the
costs of 1,000/nr*/minute (35,000 cfm) unit and do not include any annualized
capital costs.
-------
SECTION II
RECOMMENDATIONS
This program was limited to the field evaluation as described in this
report. It was not within the scope of the program to make specific recom-
mendations concerning the sonic agglomerator system design or ultimate use.
It is recommended, however, that if the alternating velocity precipitator ia
to become a viable part of an industrial gas cleaning system, then the sev-
eral operating parameters must be optimized. The use of steam should be
reexamined and the optimum interaction of sound and water sprays should be
determined. In an industrial sized unit the ultimate disposal or reuse of
the effluent base wash that contains whatever dust is retained in the chamber
must also be examined.
-------
SECTION III
INTRODUCTION
The purpose of this study was to determine the basic performance
characteristics of a sonic agglomerator that was designed and constructed
by the Braxton Corporation. All tests were performed on a prototype unit
at the Braxton plant in Medfield, Massachusetts.
The sonic agglomerator, or alternating velocity precipitator (AVP),
is designed to impart different velocities to particles of different sizes,
hence increasing the probability of collisions that result in agglomeration.
When sufficient coagulation occurs, the number of small particles is de-
creased through collisions with larger ones creating a dust of large enough
particle size to be capable of being efficiently collected by a low energy
gas cleaning device such as a cyclone. A more detailed discussion of the
theoretical basis of acoustic coagulation is presented in Appendix A.
Since the agglomeration rate should be enhanced by both a higher con-
centration and wider size range of the inlet particulate, the system is
designed to allow the admission of steam and/or water droplets. The steam
can be added in the acoustical chamber or, as was done in those tests in
which steam was used in this study, added in the inlet duct about 50 feet
before the chamber. This allows the dust and steam to interact, hopefully
increasing the cohesive properties of the dust and also raising the humidity
of the gas to a point where the liquid particles in the chamber will be
preserved. The water is added as a spray at the top of the chamber via
Sonicore nozzles. According to the manufacturer the Sonicore nozzles de-
liver a spray in the 10-20 p.m range.
The chamber itself is approximately 0.75 meter (2.5 feet) in diameter
and 4.27 meters (14 feet) in length. The chamber length is adjustable to
enable tuning the standing sound wave which is 4~l/2 wave lengths long, so
that nodes are located at top and bottom. The sound wave is generated by
an electromagnetically driven piston which displaces about +0.05 cm.
(+ 0.02 inches) at a frequency of 366 Hz. The sound intensity inside the
-------
chamber is 165 decibels. Any material that is precipitated from the gas
stream before exiting the chamber is washed off the base plate and out of
the system with a continuous flow of water.
The prototype unit tested was designed for a maximum flow rate of
3
7.0 m /s (15,000 cfm). Existing fan capabilities limited the flow to
3.7 m /s (8,000 cfm) but most tests were run at 2.2 m /s (5,000 cfm). At
the latter flow rate the average velocity in the chamber was 5.2 m/s (1,030
feet per minute) while at the inlet and outlet ducts (2 foot diameter) the
average velocity was 8.1 m/s (Ij600 feet per minute).
There are several factors that could lead to particulate depletion in
the system besides the agglomeration and subsequent fallout as a result of
sonic action. For example, there could be wash out from the water spray
system, the larger fraction entering the chamber may settle without further
coagulation because of the relatively low velocity in the chamber, or the
right angle bends at the entrance and exit of the chamber may remove some
particulate. The test program was designed in such a way as to first es-
tablish the effectiveness of the entire system and then determine the role
of each of the components (sound, steam, water, velocity, etc.).
-------
SECTION IV
SAMPLING METHODS
The sonic agglomerator is designed to increase the size of small parti-
cles and, as such, is a preconditioner rather than a collector. Although
some dust would be expected to remain in the sonic chamber, the effective-
ness of the device can be determined only by characterizing the properties
of the outlet dust that would make it more readily collectable than the
inlet dust. Since the Braxton system is designed to be followed by a cyclone
as a final collector, a sampling system with a cyclone followed by a filter
would allow determination of the change in mass of the cyclone undersize
material as a result of the conditioning system.
To further characterize the effectiveness of the system, cascade impactor
samples were taken of the cyclone undersize material at both the inlet and
outlet. This showed any shifts in the particle size distribution of the
fine material.
The sampling system used in this study is shown schematically in Figure 1
and calibrations for specific components are detailed in Appendix B. A
portion of the system flow was withdrawn isokinetically through a sampling
nozzle (a). The nozzles were interchangeable and a set was designed and
manufactured to match all flow combinations used during the tests. The
sampled gas stream then entered the probe (b) connecting the nozzle and the
cyclone. This probe was one inch copper pipe and the bend before the cyclone
was necessitated by space limitations at the prototype unit. The larger
sized dust fraction was then separated by an Aerotec model 1% cyclone (c)
manufactured by UOP Air Correction. The cyclone was also used to meter the
flow in the sampling system by measuring the pressure drop across it with a
manometer (d). A variable transformer on the main sampling pump was used to
maintain 746.5 n/ 2 (3 inches water) differential across the cyclone which
m 3
corresponded to 0.0085 m / (18 cfm).
The exit duct from the cyclone (e) was also one inch copper pipe with a
small port drilled in it about twenty duct diameters downstream. This
-------
m
Figure 1. Schematic of Sampling Train
-------
allowed access for the small probe (f) which led to the Andersen six stage
cascade impactor (g) . The probe was designed for isokinetic sampling and
kept as short as possible. The collection surfaces of the cascade impactcrs
were coated before use with a thin uniform layer of petroleum jelly to mini-
mize particle bounce and reintrainment. The impactor was followed by a glass
fiber filter (h) and a critical orifice (i) for flow control. Conditions
for critical flow were monitored by observing the upstream pressure with a
mercury manometer (i) and knowing the pump pressure capabilities of the model
0522-V3-G18D Cast pump (k). Laboratory measurements determined that the
Cast pumps used were capable of maintaining a flow rate of one cubic foot
per minute at an absolute suction pressure of 10.2 inches of mercury. Since
critical flow of air through an orifice is achieved when the pressure down-
stream of the orifice is less than or equal to 0.53 of the upstream pressure,
a downstream absolute pressure of 10.2 inches of mercury assures that critical
flow will occur if the upstream absolute pressure is 19.25 inches of mercury
or more. Monitoring the upstream pressure also allowed the critical flow-
rate to be corrected to standard conditions.
The cyclone undersize dust that was not diverted through the impactor
continued on to a glass fiber filter (1). The filter holder was a Precision
Scientific model high volume sampler (m) with the inlet modified to make a
smooth transition from the copper pipe to the filter area. The high volume
sampler was the main air mover of the sampling system and the flow rate
through it was controlled electrically by a variable transformer (n) so that
the pressure drop across the cyclone remained constant.
The resistance across the high volume filter increased quite rapidly and
it was found that fifteen minutes was a safe sampling time without over-
extending the capabilities of the high volume sampler. It also became appa-
rent after the first few runs that the dust concentration was apt to fluctuate
with time so that it was very important to have simultaneous upstream and
downstream sampling.
An Acrison 120 D dust feeder delivered the test dust to an air ejector
which dispersed the dust into the gas stream. Occasionally the dust in the
hopper adhered to the side walls, creating a temporary light dust loading
in the system. During one test the delivery auger on the feeder broke and
8
-------
no dust was dispersed at all. Except for that one test, however, the vari-
ations in dust concentration due to the dust feeder were minimized, if not
eliminated, by simultaneous upstream and downstream sampling.
During some of the first few tests, an attempt was made to traverse the
inlet and outlet ducts, sampling at several points in the cross section. The
severe space limitations at the sampling locations made this a very awkward
procedure. Furthermore, the sampling time restriction of fifteen minutes
per filter necessitated disassembling the train and changing the filter in
the middle of each run if meaningful traverses were to be made. It was felt
that since the sampling ports were installed at locations where velocity and
concentration gradients were expected to be minimized, centerline sampling
would give meaningful results. Velocity traverses indicated a smooth flow
profile and traverses with an Ikor continuous particle monitor showed no
large concentration gradients. The Ikor device does not measure the mass
concentration directly but measures the electron charge transfer, or tribo-
electric effect, to monitor particulate concentrations. Measurements with
this device at the inlet of the Braxton AVP indicated concentration variations
of less than 25 percent across the duct and most of those variations were
thought to reflect a temporal variation of dust loading rather than a spatial
variation.
-------
SECTION V
ANALYSIS METHODS
After each test run, both the inlet and outlet sampling trains were dis-
assembled and removed to a clean work area. Any dust deposited on the inside
of the nozzle and probe before the cyclone was brushed into a tared container.
The final net weight of this dust was added to the cyclone catch. The
cyclones were stoppered and taken back to the laboratory for cleaning. This
was necessitated by the very laborious procedures required to quantitatively
remove the dust that did not fall into the hopper but was impinged on the
inside of the cyclone.
Any dust deposited on the duct between the cyclone and the high volume
sampler was also brushed into a preweighed container and later added to the
filter catch. Occasionally some particulates deposited on the walls of the
impactor sampling probe. When this occured the dust was added to the first
stage of the impactor.
All samples were equilibrated in an atmosphere of less than 50 percent
relative humidity at about 65 degrees before weighing. All weighings were
done on a Mettler H-15 balance and recorded to the nearest tenth of a milli-
gram.
Sonic agglomerator system parameters such as steam flow rate, water flow
rate, sound intensity and power consumption by the sonic generator were
recorded by the Braxton personnel operating the unit.
10
-------
SECTION VI
RESULTS
This section summarizes the results of the test program. Many sonic
agglomerator operating parameters were varied during the evaluation procedure
and the effects of these variations were assessed by measuring several pro-
perties of the inlet and outlet aerosols. The complete matrix of operating
variables and dust properties is too vast to present as one unit so the data
have been divided into separate sections.
The first section details the system operating parameters such as volume
flow rate, sound intensity, rate of steam and water addition and pressure
drop across the unit. The next section summarizes the inlet and outlet dust
loadings in terms of total mass, cyclone fraction, reduction of total mass,
reduction in mass of fine particles and the total system efficiency if a
cyclone were to follow the sonic device. The third section shows the changes
in the particle size distribution of the fine particles. This is done to
show what changes, if any, were imparted to that portion of the dust that
penetrated the cyclone collector. The last section is an estimate of operat-
3
ing costs of a 1000 m /min. (35,000 cfm) unit if it were to display the same
operating characteristics as the prototype unit.
A. SYSTEM OPERATING PARAMETERS
Table 1 details the several sonic agglomerator system operating
parameters that were varied during the test program. Most tests were run in
3
the 2 m /s (5000 cfm) range as that was considered nominal for the prototype
device. Redispersed cupola dust and fly ash were used as test dusts. The
cupola dust was used more than the fly ash because the high degree of spher-
icity and possible poor wetting properties of the fly ash were expected to
have a negative influence on agglomeration. Preliminary sizing with Andersen
impactors of the test dusts was also performed by redispersing them in the
laboratory with an air ejector similar to the one in the Braxton dust feed
system. The results of these measurements (Appendix D) also indicated the
suitability of the cupola dust.
11
-------
Table 1.
Sonic Agglomerator System Operating Parameters
RUN
NO.
1
2
3
14
5
6
7
8
9
10
11
12
13
14
15"
16
I7b
18
19
DATE
10/17/73
10/17/73
10/18/73
10/24/73
11/2/73
11/13/73
11/13/73
11/19/73
11/19/73
11/29/73
1/2/74
1/2/74
1/4/74
1/4/74
1/22/74
1/22/74
2/19/74
2/19/74
2/19/74
TEST
DUST
cupola
cupola
cupola
cupola
cupola
cupola
cupola
cupola
cupola
cupola
cupola
cupola
cupola
cupola
fly ash
fly ash
cupola
fly ash
fly ash
VOLUME
FLOW
m3/s
2.22
2.22
2.22
2.22
2.23
2.23
2.23
2.47
2.48
2.48
PRESSURE DROP
ACROSS SONIC
CHAMBER
n/m2 x 10-3
2.96
2.96
2.96
2.96
1.09
1.09
1.09
1.09
1.09
1.09
DUST FEEDER FAILURE
0.82
3.72
3.72
2.48
0.82
2.47
2.47
2.47
0.12
2.53
2.53
1.09
0.12
1.09
1.09
1.09
FAN POWER8
KILOWATTS
12.40
12.40
12.40
12.40
4.94
4.94
4.94
5.48
5.50
5.50
0.20
19.19
19.19
5.50
0.20
5.48
5.48
5.48
STEAM
PRESSURE
n/n>2 x 10-3
3.45
3.45
48.26
3.45
3.45
20.68
3.45
3.45
FLOW
m3/s x 105
0
0
0
0
0
0
0
2.27
2.27
11.35
2.27
0
2.27
0
7.72
0
2.27
2.27
WATER
PRESSURE
n/m2 x lO'4
10.34
10.34
10.34
10.34
10.34
10.34
6.90
6.90
6.90
6.90
10.34
10.34
FLOW
m3/s x 105
0
8.20
8.20
8.20
4.42
4.42
4.42
4.42
4.42
4.42
0
0
4.42
0
0
0
4.42
4.42
SONIC GENERATOR
INTENSITY
db
0
164
164
166
165
0
165
164
0
164
c
0
161
0
162
0
164
165
POWER
KILOWATTS
3.25
2.90
2.90
3.60
3.00
3.20
3.10
3.20
3.00
3.30
3.20
3.50
• Fan power based on motor-fan combined efficiency of fifty-five percent.
b Runs 15 and 17 were to determine mass efficiency by size of sampling cyclone only.
c Microphone broken - sound Intensity unknown.
-------
The high pressure drop across the sonic chamber during tests one
through four was caused by an inlet manifold with a very small opening.
Although this tortuous configuration may have increased the dust-spray inter-
actions, the high shear forces encountered may actually break up any
agglomerates entering the system. The inlet restriction also was an added
burden on the fan capacity and was replaced by the normal inlet before test
five.
The increase in volume flow starting with test eight reflects the
installation and use of gas burners at the air intake. This was required to
keep the temperature in the system high enough to prevent the condensation of
the steam as soon as it was introduced. Runs 12 and 16 were at a low flow
rate (2000 cfm) while 13 and 14 were at the highest flow rate the system
could achieve (8000 cfm).
The fan power required at the indicated volume and static pressure
was based upon an assumed motor-fan combined efficiency of 55 percent. The
sound intensity and power demands of the sonic generator were taken from
direct readings on the control panel of the Braxton device. The steam and
water consumptions were determined by Braxton personnel with calibrated flow
meters.
B. EFFECT OF THE SONIC AGGLOKERATOR ON PARTICULATE COLLECTION
Table 2 summarizes the inlet and outlet dust loading. The relatively
large portion of the inlet dust that was captured by the cyclone indicates a
fairly coarse particulate. While this was expected to be the case with fly
ash it was unexpected for the cupola dust because laboratory dispersion tests
with a relatively high ejector air pressure (90 psi) had shown a mass median
diameter of just larger than one micrometer (Appendix D)• The Braxton dust
feed system was designed to have a similar air pressure to resuspend the
dust, but the high pressure compressor was out of service during the early
part of the tests. Therefore, the low pressure compressor at Braxton had to
be utilized and the resulting air pressure (20 psi) to the air ejector prob-
ably failed to break up the agglomerates in the bulk dust and led to a
coarser particulate than obtained in the laboratory. The high pressure
13
-------
TABLE 2. P INLET AND OUTLET DUST LOADING, CYCLONE CATCH, CYCLONE UNDERSIZE DUST AND OVERALL SYSTEM COLLECTION OF ALL PARTICLE SIZES.
RUN
NO.
1
2
3
4
5
6
7
8
9
10
12
13
14
16
18
19
INLET
A
CONCENTRATION
g/03
1 . 3309
1.5859
1.8343
2.2744
1.1202
1.1982
1.6875
1.2012
0.8792
1.5213
0.9227
2.5218
2.0105
2.5690
6.0677
5.4163
B
CYCLONE
FRACTION
.836
.809
.906
.869
.887
.794
.850
.902
.875
.886
.698
.931
.942
.983
.951
.965
C
FILTER
FRACTION
.164
.191
.094
.131
.113
.206
.150
.098
.125
.114
.302
.069
.058
.017
.049
.035
OUTLET
D
CONCENTRATION
g/m3
0.7707
0.1945
0.2629
0.3121
0.2124
0.3806
0.4639
0.1648
0.1499
0.2352
0.5824
1.0376
0.1856
0.2799
0.3936
0.4069
E
CYCLONE
FRACTION
.657
.368
.756
.568
.626
.569
.713
.391
.472
.507
.702
.756
.723
.899
.879
.922
F
FILTER
FRACTION
.343
.632
.244
.432
.374
.431
.287
.609
.528
.493
.298
.244
.277
.101
.121
.078
REDUCTION IN MASS
flF PTH7 PARTTf"7.FQ
Ur Fine rAni il if* a ,
PERCENT
(A C)-(D F)
J-T* X IUU
A C
-20.9
59.4
62.8
54.9
37.0
33.6
47.2
14.6
28.1
33.0
37.8
-45.8
55.3
34.0
84.0
83.2
REDUCTION OF
TflTAT MAQQ
1V1AL nfua 1
PERCENT
*r* l°°
A
42.1
87.7
85.7
86.3
81.0
68.2
72.5
86.3
83.0
84.5
36.9
58.9
90.8
89.1
93.5
92.5
SYSTEM EFFICIENCY
IF FOLLOWED BY
CYCLONE , PERCENT
A"(DAD E) " I0°
80.1
92.3
96.5
94.1
92.9
86.3
92.1
91.6
91.0
92.4
81.2
90.0
97.4
98.9
99.2
99.4
-------
compressor became available later in the program but it was decided that a
change in operating parameters part way through the tests would be unwise
so the low pressure air supply was utilized throughout.
In addition to the mass concentration, the cyclone fraction and the
filter fraction at the inlet and outlet, Table 2, shows the percent reduction
of total mass and the percent reduction of fine particles. The reduction of
total mass is that portion of the dust that was removed in the sonic chamber
by wash out, settling, precipitation, impaction or whatever other mechanisms
may be involved. It is a measure of the AVP as a dust collector rather than
a conditioner. An average of 86 percent of the cupola dust and 97 percent
of the fly ash is cyclone collectable before treatment by the AVP so it is
important to examine the reduction in mass of fine particles. This is merely
the percent reduction of the cyclone undersize dust at the outlet of the AVP
as compared to that of the cyclone undersize dust at the inlet.
Also computed in Table 2 is the system collection efficiency if it
were followed by a cyclone displaying the same collection efficiency as the
cyclones in the sampling trains. This takes into account both the dust lost
in the sound chamber and the increased collectability of the effluent as a
result of the reduction of small particles. It is important to keep in mind
when using the efficiencies shown that the inlet cyclone fraction shown in
column B would be the system efficiencies without any conditioning whatso-
ever.
Two tests (1 and 13) were run in which no sound, steam or water were
used to determine what effect the physical configuration of the Braxton sys-
tem has on the test aerosol. It is interesting to note that under those con-
ditions the system retains about half of the total dust entering it but the
dust that penetrates the chamber has a substantially higher portion of its
mass in the cyclone undersize range. This shows up as a negative number in
the reduction in mass of fine particles. This would indicate that the
physical configuration works against agglomeration by breaking up aggregates
that enter.
Test six was run with water spray but no sound or steam. This shows
the positive contribution of the sprays both in increasing the percent of the
15
-------
dust retained in the chamber and also in decreasing the fraction of fine
material. The positive influence of the water sprays can also be seen by
examining tests two, three and four. The average reduction in the percent
of cyclone undersize dust is considerably greater (59 percent versus 37 per-
cent) than those tests run with the same test dust with sound but with half
the water spray volume (tests 5, 7, 8, 10 and 14). Figure 2 shows the percent
reduction in the mass of fine particles as a function of water flow rate
disregarding other system parameters. Although the data show considerable
scatter, there is a trend toward more efficient reduction of fines at in-
creased water flow rates. The correlation coefficient of the 13 pairs that
showed a positive reduction in fine particles is 0.43, while the correlation
coefficient of the data for all tests, including negative reduction in fines,
is 0.65. The former correlation coefficient is not statistically significant
while the latter is statistically significant (a = .05).
Test nine was run with steam and water but no sound to compare the
results with test six without steam or sound. The steam appears to sub-
stantially increase the percent mass retained in the chamber but has a slight
negative influence on the reduction of fine particles. Figure 3 shows the
percent reduction in the mass of fine particles as a function of steam flow
rate disregarding other system parameters. Disregarding the two tests that
resulted in negative fine particle reduction, the correlation coefficient is
- 0.30. The correlation coefficient of the data for all runs, including
those with negative reduction in fines, is 0.06. Neither coefficient is
statistically significant.
Tests 16, 18 and 19 were run with fly ash and the larger particle
size distribution of this test dust is obvious. Test 19 was a repeat of
test 18 because of difficulties experienced during test 18. The data are
included for completeness but are in doubt because the outlet sampling train
nozzle was misaligned during part of the test run. The strong influence of
the water sprays can also be seen by comparing the reduction of fine particles
in run 16 (34 percent) without water but with steam and sound versus run 19
(83 percent) with water, steam and sound. This is in spite of the fact that
run 16 was at a low flow rate so the residence time in the chamber was about
two and one half times as long as for run 19.
16
-------
9O
eo
70
S 6°
o
OC.
UJ
0- SO
o
F 40
OC
UJ 30
fe 20
< 10
2 0
O
8 -10
Q
O
Q
O
O
O
O
O
a
-20
-30
o = CUPOLA DUST
Q - FLY ASH
0 -
-50
O
I
01 2345678
WATER FLOW RATE, m3/s x I05
Figure 2. Effect of Water Flow Rate on the Reduction
of Fine Particles.
17
-------
»-
UJ
UJ
CL
in
UJ
90
80
70
60
SO
I<°
*
. . 30
U. 20
O
O -10
o
UJ
tr. -20
-30 -
-40 -
-5O
O
o
o
o
o
o
I
o = CUPOLA OUST
o s FLY ASH
0 2 4 6 B 10 12
STEAM FLOW RATE, m3/s x I05
14
Figure 3.
Effect of Steam Flow Rate on the Reduction
of Fine Particles.
18
-------
The influence of the sonic generator can be seen by comparing test
6 (water, no sound, no steam) with tests 5 and 7 (water, sound, no steam).
The addition of sound increased the reduction of fine particles and increased
the amount of dust retained in the chamber. The use of sound, however, did
not appear to play as significant a role in the depletion of small particles
as the addition of water. Test 6, in which water was added but no steam and
no sound was used, showed a 33.6 percent reduction in fines. The use of
sound and water in tests 5 and 7 averaged 42.3 percent reduction in fines.
Tests 2, 3 and 4 were run with sound and twice the water rate as tests 5 and
7 and showed a 59.2 percent reduction in fines. Figure 4 shows the percent
reduction in the mass of fine particles as a function of the sonic generator
power consumption disregarding other system parameters. For the 13 tests
that showed a positive influence on the fines reduction the correlation coef-
ficient is 0.33, while the correlation coefficient of the data for all tests
is 0.68. The former coefficient is not statistically significant and the
latter is statistically significant.
Since most of the tests were run with cupola dust, a more detailed
statistical analysis was undertaken for those 13 tests. A multiple regression
analysis was used on the data of the cupola dust tests and showed the
following relationship:
% reduction in mass of fine particles = -15.1 + 6.5
(water flow, m /s x 10 j + 8.3 (sonic generator power, kilowatts)
An analysis of variance shows that the regression is highly significant (99
percent). It should be noted that the first member on the right side of the
equation is negative and thus shows the detrimental influence on fine par-
ticle reduction of the physical configuration of the unit as was demonstrated
when tests were run with no sound, water or steam. Furthermore, steam con-
sumption does not appear in the relationship because it does not contribute
significantly to the regression. The model also shows that the use of the
sonic generator has somewhat less influence on fine particle reduction than
does the use of water sprays, although the interaction of both plays a sig-
nificant role
19
-------
90
80
TO
GO
50
40
UJ
Q.
V)
UJ
O 20
h-
2
10
Ul
z
tt.
en
CO
-10
O -20
§
S -30
QC
-40
-50
O
O
O
o
O O
o = CUPOLA DUST
a = FLY ASH
0 I 2 3.4
SONIC GENERATOR POWER CONSUMPTION, KILOWATTS
Figure 4. Effect of Sonic Generator Power Consumption
on Reduction of Fine Particles.
20
-------
3 3
Increasing the volume flow rate from 2.2 m /s (5,000 cfm) to 3.7 m /s
(8,000 cfm) and hence decreasing the residence time in the chamber did not
seem to adversely affect the performance of the AVP system when run with steam,
water and sound. Run 14 was at the elevated flow rate with steam, water and
sound and the reduction in the mass of fine particles was 55.3 percent. How-
ever, the elevated flow rate in test 13 when the system was run dry (no water,
no steam, no sound) increased the mass of fine particles by 45 percent which
indicates the breakup of agglomerates that enter the chamber.
Two tests were run at 0.82 m /s (2,000 cfm) and again the effect of
residence time is obscure. Tests 12 and 16 were run at the reduced flow with
steam and sound and the average reduction of fines was 35.9 percent. This
indicates that the increased residence time with sound and steam is about as
effective in reducing the fine particles as running with water only at a
residence time of two and one half times less (test 6). In fact, running the
system at a high flow rate with sound, steam and water (test 14) produced
considerably better results than running at a low flow rate with sound and
steam (tests 12 and 16).
C. EFFECT OF THE SONIC AGGLOMERATOR ON THE SIZE OF SMALL PARTICLES
In addition to examining the change in the mass of the small particles,
it is of interest to investigate what changes, if any, are made in the size
distribution of the cyclone undersize material. Effective agglomeration would
be expected not only to decrease the mass of the small particles that penetrate
the cyclone but also to increase the particle size of that portion that
remains cyclone undersize at the outlet of the AVP.
Six stage Andersen cascade impactors were used downstream of the cyclone
in both the inlet and outlet sampling trains, except for tests one and three
when only outlet impactor samples were taken. Table 3 compares the concentra-
tion of cyclone undersize dust as measured with the glass fiber high volume fil-
ter and with Andersen impactors. The particle size distributions determined by
the impactors are shown graphically in Figures 5 through 21 and in tabular
form in Appendix C. Figures 18 and 20, corresponding to tests 15 and 17,
show the particle size distribution upstream and downstream of the sampling
21
-------
Table 3. COMPARATIVE CONCENTRATION MEASUREMENTS OF
CYCLONE UNDERSIZE DUST WITH GLASS FIBER
FILTERS AND ANDERSEN IMPACTORS
RUN NO.
lb
2a, b
3b
4a
5b
6
7
8
9
10
12
13
14
16
18
INLET
FILTER,
g/m3
0.22
0.30
0.17
0.30
0.13
0.25
0.25
0.12
0.11
0.17
0.28
0.17
0.12
0.04
0.30
CASCADE
IMPACTOR,
g/m3
_
0.23
-
0.24
0.11
0.20
0.17
0.04
0.06
0.14
0.20
0.16
0.05
0.04
0.18
OUTLET
FILTER,
g/m3
0.26
0.12
0.06
0.14
0.08
0.16
0.13
0.10
0.08
0.12
0.17
0.25
0.05
0.03
0.05
CASCADE
IMPACTOR,
g/m3
0.79
0.29
0.06
0.06
0.05
0.13
0.11
0.07
0.05
0.06
0.14
0.22
0.04
0.02
0.03
denotes different sampling periods for impactor and filter at inlet
denotes different sampling periods for impactor and filter at outlet
22
-------
10-
9-
8-
7-
6 •
3-
t-l
01
to
Q
0)
r-l
O
9
6
7
6
5
X Outlet
0.01
98 99
99.9
Figure 5.
5 >0 2O SO 40 SO 60 70 30 90 05
Percent Mass ^ Stated Particle Diameter
Particle Size Distribution of Cyclone Undersize Dust for Test 1
-------
10-
9-
8-
7-
6 •
5-
0)
S
CO
••-I
a)
u
••-I
CO
PM
D Inlet
: i
X Outlet
001
O.I
10
DO 30 4O 50 60 70
30
90
93 99
S9.9
90.D9
Percent Mass ^ Stated Particle Diameter
Figure 6. Particle Size Distribution(s) of Cyclone Undersize Dust for Test 2
-------
Ln
0)
AJ
CO
•i-l
P
0)
t-H
o
10-
9-
8-
7 -
6 •
5-
9
B
7
6
3
X Outlet
0.01 0.1 123 10 ZO JO 40 30 CO 70 00 90 93 93 99 39.9
Percent Mass ^ Stated Particle Diameter
Figure 7. Particle Size Distribution(s) of Cyclone Undersize Dust for Test 3
99.99
-------
t-l
-------
e
I
Q
4J
M
10-
9-
8-
7-
6 •
5-
9
B
7
6
3
0.01
z
0 Inlet
X Outlet
0.1
10
20 SO 40 SO 60 70 SO 9O 95 98 99
Percent Mass ^ Stated Particle Diameter
Figure 9. Particle Size Distribution(s) of Cyclone Undersize Dust for Test 5
99.9
99.99
-------
1-0
CO
l-i
HI
4-1
0)
s
n)
•1-1
Q
u
•1-1
0 Inlet
X Outlet
0.01
10 20 JO 40 SO 60 70 80 90 95
Percent Mass ^ Stated Particle Diameter
98 99
99.9
99.99
Figure 10. Particle Size Distribution(s) of Cyclone Undersize Dust for Test 6
-------
VO
0)
4-1
p
-------
u>
o
0)
e
to
o
.,-<
4-1
M
0)
P-i
10-
9-
8-
7-
6
5-
9-
8-
7
6
S
O.OI
^
d Inlet
I
i X Outlet
O.I
5 10 ZO 50 40 SO 60 70 8O 90 99
Percent Mass / Stated Particle Diameter
98 99
99.9
99.99
Figure 12. Particle Size Distribution(s) of Cyclone Undersize Dust for Test 8
-------
a
0)
4-1
S
CO
•1-1
Q
J-i
cfl
10-
9-
8-
7-
6
5-
JL
0 Inlet
X Outlet
0.01
O.I
I 10 20 JO 4O SO 60 70 80 90 99
Percent Mass ^ Stated Particle Diameter
98 99
99.99
Figure 13. Particle Size Distribution(s) of Cyclone Undersize Dust for Test 9
-------
N>
0)
a
CO
0)
r-<
u
10-
9-
8-
7-
6 •
3-
9-
8-
7
6
3
0 Inlet
X Outlet
0.01
O.I
10
20 30 40 50 60 70
90
90
93
98 99
99.9
Figure 14.
Percent Mass ^ Stated Particle Diameter
Particle Size Distribution(s) of Cyclone Undersize Dust for Test 10
-------
e
CO 4J
Co QJ
a
o
0)
i-i
o
•H
10-
9-
8-
7-
6
S-
,0
Inlet
Outlet
o.oi
O.I
10
SO 40 SO 6O 7O
80
90
95
98 99
9<».9
99.99
Figure 15,
Percent Mass ^ Stated Particle Diameter
Particle Size Distribution(s) of Cyclone Undersize Dust for Test 12
-------
0)
4-1
CO
•H
Q
OJ
i-l
o
U) -H
4> -U
Vi
CO
10-
9-
8-
7-
6
5-
;0 ; Inlet
• X Outlet
0.01
O.I
90
99
99.9
99.99
12 5 10 20 30 40 JO 60 70 00 9O S3
Percent Mass^ Stated Particle Diameter
Figure 16. Particle Size Distribution(s) of Cyclone Undersize Dust for Test 13
-------
OJ
Ul
.2
o
-------
10-
9-
8-
7-
6
5-
0 Before Cyclone
X After Cyclone
-------
U)
I
•1-1
10-
9-
8-
7 -
6 •
3-
9
8
X
X
I 0 i Inlet
i
X ; Outlet
0.01
O.I
93
58 99
99.9
12 S 10 CO 30 40 SO 6O 70 80 90
Percent Mass ^ Stated Particle Diameter
Figure 19. Particle Size Distribution(s) of Cyclone Undersize Dust for Test 16
90.99
-------
00
-------
M
01
4J
IjO £!
vo n}
b
o>
t-i
o
M
(0
10-
9-
8-
7-
6
5-
0.01
y
0 Inlet
X Outlet
0.1
lO
20
JO «O SO 60 TO
80
90
99
98 99
99.9
99.99
Percent Mass -^ Stated Particle Diameter
Figure 21. Particle Size Distribution(s) of Cyclone Undersize Dust for Test 18
-------
cyclone with flyash and cupola dust, and served as the field calibration of
the collection efficiency by particle size of the cyclone collectors as
shown in Appendix B.
The particle size distributions of the cyclone undersize dust at the
inlet and outlet show no clear trend. More often than not there is a signif-
icant shift in the size distribution but not in any consistant direction.
For example, tests 2 and 4 were run at nearly identical conditions but the
results showed opposite effects. Test 2 resulted in a substantially larger
size at the outlet but test 4 resulted in a somewhat smaller size at the
outlet. Tests 5, 6 and 7, on the other hand, resulted in practically no
change between the inlet and outlet particle diameters but test 5 and 7 were
run with sound and test 6 was run without sound.
A comparison of tests 8 and 9, however, indicates a positive influ-
ence of sound because of the larger size of the outlet cyclone undersize dust
in test 8 with sound as compared to no change in test 9 without sound.
The lack of any clear trend of a change in the particle size of the
cyclone undersize dust after going through the sonic agglomerator is also
indicated in Figure 22. This is a plot of the mass median diameter of the
fine particles at the inlet versus that at the outlet. Each point is numbered
according to the test number in order to facilitate identifying each one with
the system operating parameters in Table 1. As can be seen on Figure 22,
test 2 resulted in the greatest increase in mass median diameter of the outlet
dust. The reason for this is not understood and a repeat test at the same
conditions (test 4) failed to demonstrate any increase in the outlet mass
median diameter at all.
D. ESTIMATE OF OPERATING COSTS
An estimate of the operating costs was made for the Braxton sonic
agglomerator. The operating parameters such as steam and water consumption
and electrical requirements for each run were expanded to a 16.67 m /s
3
(1000 m /min or 35,000 cfm) system by a simple multiplier. For example, if
3
the system required 15 kilowatts at a flow rate of 2.2 m /s then 7.58 times
15 kilowatts, or 113.64 kilowatts, would be required for the scaled up unit.
The same type of scaling was done for steam and water consumption and the
40
-------
3.0
2.8
2.6
2.4
Z
a.
- 2.2
o
g 2.0
o
1.8
o
3 '••
O
O
« I 9
ijl I.C
UJ
5 1-0
en
CA
4
0.8
0.6
04
0.2
I
°2
09
O
IO
O_
I
I
O
13 O
12
16
'14
NUMBER DENOTES TEST
I
I
0 0.2 0.4 0.6 0.8 10 1.2 1.4 L6 1.8
MASS MEDIAN DIAMETER DOWNSTREAM OF INLET CYCLONE, JOM
Figure 22. Effect of AVP System on Mass Median Diameters
of Cyclone Undersize Dust.
2.0
41
-------
cost for each parameter estimated for each test run in Table 4. An estimated
annual operating cost based upon 8,000 hours operation is also included.
The estimated cost for steam consumption is based on a purchased
price of $3.10 per 1,000 pounds. Water costs are based on $0.40 per 100
cubic feet and the electricity rates used are $0.025 per kilowatt hour.
It should be pointed out that the cost estimates in Table 4 do not
include the additional power requirements for a high efficiency cyclone.
Some of the earlier performance data were based on the use of the cyclone
in the sampling train and the operating cost of a scaled up cyclone are not
reflected in the cost estimates.
Test 7 reflects normal water and power consumption while test 8 shows
the same conditions except for steam addition. The estimated annual operating
cost under the conditions in test 7 is $ 13.30/m3/minute ($0.38/cfm) and for
test 8 it is $46.22/m3/minute ($1.32/cfm). These estimates are based upon
the previously explained conditions and costs but do not include any annual-
ized capital costs.
It is clear that within the limitations of the assumptions and con-
ditions associated with Table 4 that by far the most expensive item is steam.
The effectiveness of steam addition is at best questionable and more than
likely detrimental so it is safe to assume that the costs associated with
steam use could be subtracted out.
42
-------
Table 4. ESTIMATED OPERATING COSTS OF AVP SYSTEM0
Test
1
2
3
4
5
6
7
8
9
10
12
13
14
16
18
19
Cost per 1000 m3 of Gas Treated, Dollars
Steam
0
0
0
0
0
0
0
0.0688
0.0688
0.3440
0.0688
0
0.0688
0.2339
0.0688
0.0688
c
Water
0
0.0056
0.0056
0.0056
0.0028
0.0028
0.0028
0.0028
0.0028
0.0028
0
0
0.0028
0
0.0028
0.0028
d
Electricity
0.0391
0.0494
0.0484
0.0484
0.0270
0.0155
0.0249
0.0247
0.0155
0.0243
0.0290
0.0361
0.0418
0.0299
0.0246
0.0254
Annual Cost,
Dollars6
18,771
26,400
25,920
25,920
14,304
8,784
13,296
46,224
41,808
178,128
46,944
17,328
54,432
126,624
46,176
46,560
abased on a 1000 m /min (35,000 cfm) system
based on $3.10 per 1000 pounds
°based on $0.40 per 100 ft.3
based on $0.025 per kilowatt hour
Q
based on 8000 hours of operation per year
43
-------
APPENDIX A
ACOUSTIC COAGULATION AND THE BRAXTON SONIC AGGLOMERATOR
Mednikov (1965) gave a much more advanced presentation on this subject
than did Fuchs (1964) or Green and Lane (1964), and his results are applied
here to the Braxton acoustic agglomerator . It is Mednikov's conclusion that
for industrial gas cleaning situations, the predominant mechanism of acousti-
cal agglomeration is the different vibratory velocities and phases of par-
ticles of different sizes, despite the great variety of possible mechanisms
presented in the literature.
Coagulation is generally governed by the following equation:
2
dn/dt =-K n
where
t = time, sec
n = concentration, cm
2
K = coagulation constant, cm x cm/sec.
The units for K suggest the general mechanism: particles in motion "sweep"
a given volume per unit time, the product of their effective cross sectional
area and their velocity, and the number of particles is decreased by each
effective particle-particle collision in this swept volume.
The aerodynamic diameter of a particle, or equivalently its "relaxation"
time t, governs the particle's response to velocity changes of the medium in
which it is suspended. The ratio of particle velocity amplitude Up to gas
velocity amplitude Ug in a sonic field is (Fuchs, 1964; Mednikov, 1965):
where
Up/Ug =!/[! + (2*)2 (t/t )2]^
s
2
T = Cp d /18 |a, the particle relaxation time, sec,
t = 1/f, the vibration period of the sound, sec,
S
f = sound frequency, sec or Hz
44
-------
d = particle diameter, cm,
P 3
p = particle density, g/cm ,
\JL = medium viscosity, poise
Roughly, this means that small particles, those having 2ntf « 1, follow
the gas vibrations, and large particles, those having 2nTf » 1, do not
vibrate appreciably. At the frequency of the Braxton device, ~ 400 Hz,
the critical T is
c
T = l/2*f
c
-4
T = 4.0 x 10 sec
c
which is the relaxation time of a 12-^m diameter particle of unit density.
Particles of T « T will agglomerate on particles of T » T , if they
agglomerate at all. Since it takes many "small" particles to change the
mass of a "large" particle, the primary change in the size distribution
would be a decrease in the number (and mass) concentrations for particles
having T « T and an increase for particles having T » T , reflected by:
(a) a decrease in the total number concentration,
(b) an increase in the number mean and median diameters,
(c) no change in the mass mean diameter,
(d) an increase in the mass median diameter,
(e) a decrease in the geometric standard deviation (if
a log-normal curve is fit).
Of all these effects, the most noticeable should be the change in the
"small" particle concentration, here (Braxton) the respirable fraction.
This may be the most useful test of whether or not the device is ag-
glomerating particles.
Mednikov's model subdivides the aerosol into its small particle con-
centration n^ and its large particle concentration n (assumed to be
constant with time). The coagulation equation for the aerosol becomes:
where
dn/dt = dn2/dt = - Kfll
n = n + n_
3
K = coagulation coefficient, cm /sec,
3 J.
45
-------
and
where
Kal -° (*rl V'
2
jtr, is an average cross-sectional area for the large particles
and 0(x) means "of the order of magnitude of x."
The coagulation equation here becomes
n2 " (lVo 6XP (" Kal nl C)
and the exponential time dependence has been found in many experiments
(Mednikov, 1965). Assuming n_ » n , then n = n« and n = (n^) .
Coagulation is appreciable, n/n « 1, if
Q = ni ri2 V >> L-
The gas velocity amplitude is given by
1/2
U8 = (2j/pg V
where
2 -7 2
J = sound intensity, erg/cm -sec, 10 W/cm ,
2
p = gas density, g/cm ,
O
c = gas speed of sound, cm/sec,
U = gas vibration velocity amplitude, cm/sec.
o
Coagulation will thus be improved by increasing
2 1/2 ^
nn r., J t
1 1 res
in the coagulation chamber: t is the residence time. Note the rela-
res 1/2
tively weak dependence upon power level changes, J
The Braxton device has a 5 hp acoustic power source radiating into a
3-foot diameter cylinder which is about 10 feet long. This gives a gas
velocity amplitude:
46
-------
2x5 hp x 746 watts/hp x 10 erg/watt
1.2x10 g/cm x 3.3x10 cm/sec x 6.3x10 cm
1/2
2
= 5.4x10 cm/sec
If the large particle concentration is due to the Braxton water spray
nozzles, then that concentration is
n. = m/(4n/3)p r3 V
i w
where
m = mass flow of water, g/sec
(
3
3
V = volume flow of air, cm /sec
p = water density, g/cm
w
~~3~
r = water droplet mass mean diameter, cm.
The residence time of the particles in the agglomerator chamber will
be the chamber volume divided by the flow rate or
t = V /V .
res c
the parameter governing acoustic coagulation is
2
Q = n, JT r, U t
1 1 g res
and it will have a V dependence on flow rate due to its residence time
dependence; if the large particle concentration n1 is due primarily to
•-1
the water spray, as assumed, then n.. <* V will increase the dependence
•-2 --2
of Q on V to Q a V • (This V dependence of coagulation will be some-
what offset by increased large particle deposition at larger V due to im-
paction on walls and in bends and due to coagulation produced in non-
oscillating accelerations of the air flow.) Since the spray concentra-
tion n is related to the spray mass flow and mass average spray droplet
"
.. ,x .
radius (r ) by _
3 •
m « n r V,
2 23
V r « m (r ITC ),
47
-------
the coagulation rate will be increased at constant droplet mass flow for
smaller droplets, as long as they are large enough so their T » T .
Whether or not the spray evaporates completely can be estimated.
The effect of a relative velocity (U - U ) between particle and fluid
t O
is to accelerate mass and heat transfer in comparison with the rates
appropriate to a stationary particle. For water droplets evaporating,
this multiplies the stationary mass loss rate by a ventilation factor
(Green and Lane, 1964) of
[1+0.3 (Re )1/2J
where
ReP = p (UP - V
For a falling 20-(j.m diameter water droplet (U - U ), U =1.2 cm/sec
_ 2 P 8 P
and Re = 1.6 x 10 , so that this ventilation factor is small. For
P 2
acoustical U = 5.4 x 10 cm/sec, and the Reynolds number becomes 7.3,
O
so the evaporation rate is increased by about a factor of 1.8. At 80
percent r.h. and 20 C, a 20-um diameter water droplet has a lifetime of
2.4 sec (Green and Lane, 1964); droplet lifetime is proportional to
droplet area for droplets larger than 1 p.m or so. The Braxton device
has residence times -^ 1 sec so that the droplets should not have evap-
orated completely before exiting the acoustic chamber.
An approximate calculation for the Braxton coagulation efficiency
follows:
V = 5,000 cfm = 200 ft3/sec ( . . _
„ > t = 1.2 sec
v = 100 ft3 ! res
c
f 2
r = 10 urn spray droplets = (r.. )
J = 5 hp
i~ = 10 un
m = 1 gal/min of spray
3 2
m = 4 liters/min = 4 x 10 g/60 sec = 0.66 x 10 g/sec
nl rl2 = m/^ *T Pw
48
-------
2
U = 5.4 x 10 cm/sec
g
.'. Q = 3.14 (2.8 x 10~2 cm"1) (5.4 x 102 cm/sec) (1.2 sec)
Q = 57
Q » 1 so n2/(n ) « 1 and thus Braxton device should be
efficient in reducing the small particles by droplet-particle
collision.
(This kind of calculation can only tell whether n/n « 1 or not and
should not be taken as a means for getting n/n = exp (-Q) more accurately
than to an order of magnitude in Q. There are many second-order effects.)
A somewhat more sophisticated calculation of the acoustic agglomera-
tion would take into account the tendency of the sub-micron fraction of
the aerosol to avoid capture by following the streamlines of flow around
the larger droplets as the flow oscillates. This can be added to the
model through an efficiency factor e-multiplying Q, acknowledging that
the geometric cross-section of the large particles, n r,^, is different
—2
from the effective collision cross-section, e it r, . This factor e will
be a function of the aerodynamic diameter of the small particles, or
e = e(r»). Theory and practice indicate that € «• 1 for T9 U /r » 1
^ 9 £ 1
and e -» 0 for T_ U /r.. « 1. For r = 10 \m, U = 6 x 10 cm/sec,
T» U /r = 1 at T- = 1.7 x 10 sec, which corresponds to d_ = 2 r_ =
0.7 urn. Thus we would expect the agglomerator to be significantly less
efficient for particles smaller than l/2-(am diameter.
49
-------
REFERENCES
Fuchs, N.A. (1964): Mechanics of Aerosols, Pergamon Press, New York.
Green, H.L. and Lane, W.R. (1964): Particulate Clouds; Dusts, Smokes
and Mists.
Mednikov, E.P (1963): Translator: Larrick, C.V., Acoustic Coagulation
and Precipitation of Aerosols, Consultants Bureau, New York, 1965.
50
-------
APPENDIX B
CALIBRATION OF SAMPLING EQUIPMENT
Cyclone Calibration;
The cyclones were calibrated in the GCA laboratory using a calibrated
orifice to determine the flow rate. The pressure drop across the cyclone
was plotted against the flow rate through the orifice on log-log paper and
the line of best fit drawn. The resulting calibration curves are presented
in Figures B-l through B-4.
Critical Orifice Flow Rate Determination;
The critical flow rates through the orifices were determined using a
Schutte and Koerting calibrated rotameter. The measured flow rates corrected
to standard conditions are presented in Table B-l.
Table B-l. Critical Orifice Flow Rates
Orifice
Flow Rate at Standard Condi-
tions, scfm
A
B
0.965
1.060
Cyclone Removal Efficiency by Particle Size
Field tests 15 and 17 were run to determine the cyclone removal
efficiency by particle size for fly ash and cupola dust. Andersen impactors
were located upstream and downstream of the sample train cyclone. The samples
were collected from the inlet sampling port of the Braxton AVP system.
System parameters are presented in Table B-2.
51
-------
IOO
9O
8O
TO
6O
50
40
30
fe M
oc
Q 20
15
10
I
I
I
I
I
I 2 3 456789 10
PRESSURE DROP ACROSS CYCLONE, INCHES OF WATER
Figure B-l. Cyclone #1 Curve
52
-------
IOO
90
80
TO
6O
5O
40
U.
O 30
LJ
b 25
o
20
15
10
I
I
I
I
I I
I 2 3 456789 10
PRESSURE DROP ACROSS CYCLONE, INCHES OF WATER
Figure B-2. Cyclone #2 Calibration Curve
53
-------
IOO
90
80
70
60
5O
2 40
u_
O
H 30|-
a
as
20
15
10
I
I
I
I I I
I 2 3 456789 10
PRESSURE DROP ACROSS CYCLONE, INCHES OF WATER
Figure B-3. Cyclone #3 Calibration Curve
54
-------
IOO
90
80
70
60
50
2
li_
O
h- 30
^
cr.
20
15
IO
I 2 3 456789 10
PRESSURE DROP ACROSS CYCLONE, INCHES OF WATER
Figure B-4. Cyclone #4 Calibration Curve
55
-------
Table B-2. System Parameters During Cyclone Removal Efficiency Tests
TEST #
15
17
DUST TYPE
fly ash
cupola
SYSTEM FLCW,
m3/s
2.5 (5000 cfm)
2.5 (5000 cfm)
CONCENTRATION,
g/m3
2.80 (1.22 gr/ft3)
3.06 (1.34 gr/ft3)
DUST
PRECONDITIONING
none
none
Results of cyclone removal efficiency measurements for fly ash and
cupola dust are presented in Figures B-5 and B-6. Difficulties encountered
necessitated combining the mass collected on certain impactor stages. In the
fly ash test, removal efficiencies were determined for the combined first and
second stage and for the combined fourth, fifth and sixth stages. Combining
was required for the first and second stage because of reintrainment which
resulted from the collection of too-large a sample. Combining of stages 4, 5
and six was used to compensate for the clogging of several of the jets which
impact on stage 5. In the cupola dust test, removal efficiencies were also
calculated for the combined fourth, fifth and sixth stages due to clogging of
the jets which impact on stage 5.
56
-------
100
90
80
O
Ul
O
£C
UJ
Q.
O
z
UJ
O
UJ
I
UJ
OL
70
60
50
40
50
20
10
I 23456
PARTICLE DIAMETER, \un
Figure B-5. Cyclone Removal. Efficiency by Particle Size of Fly Ash
-------
t_n
00
100
90
80
S 70
o
(T
UJ
Q. 60
O
o
UJ
50
40
30
UJ
CC 20
10
0
O
O
Figure B-6.
23456
PARTICLE DIAMETER, iun
Cyclone Removal Efficiency by Particle Size of Cupola Dust
-------
APPENDIX C
PARTICLE SIZE DISTRIBUTION DATA
The following table presents the particle size distribution data as
determined by six stage Andersen impactors. Before use, the collection
surfaces were coated with a thin, uniform layer of petroleum jelly to
minimize particle bounce and reintrainment.
The aerodynamic particle size cutoff for each stage is based upon an
assumed particle density of 2.0. Although this value is only an estimate,
any shift in the size cutoff due to a different density would be exactly
the same for both upstream and downstream impactors.
59
-------
Table C-l. PARTICLE SIZE DISTRIBUTION DATA
TEST 1
OUTLET
Size Range, urn
< 0.36
> 0.36 - < 0.7
> 0.7 - < 1.4
> 1.4 - < 2.3
> 2.3 - < 3.9
> 3.9 - < 6.5
> 6.5
Cumulative Mass,
Percent
0.0
17.7
40.8
56.9
57.7
75.4
100.0
TEST 2
INLET
Size Range, qm
< 0.36
> 0.36 - < 0.7
> 0.7 - < 1.4
> 1.4 - < 2.3
> 2.3 - < 3.9
> 3.9 - < 6.5
> 6.5
Cumulative Mass,
Percent
0.0
42.9
70.1
70.1
79.2
87.0
100.0
OUTLET
Size Range, urn
Cumulative Mass,
Percent
Filter
6
5
4
3
2
1
< 0.36
> 0.36 - <
> 0.7 - <
> 1.4 - <
> 2.3 - <
> 3.9 - <
> 6.5
0.7
1.4
2.3
3.9
6.5
0.8
34.2
40.8
46.7
55.8
89.2
100.0
-------
Table C-1 (continued). PARTICLE SIZE DISTRIBUTION DATA
TEST 3
TEST 4
Stage
Filter
6
5
4
3
2
1
OUTLET
Cumulative Mass,
Stage Size Range, urn Percent
Filter < 0.36 15.4
6 > 0.36 - < 0.7 48.1
5 > 0.7 - < 1.4 59.6
4 > 1.4 - < 2.3 63.5
3 > 2.3 - < 3.9 67.3
2 > 3.9 - < 6.5 69.2
1 > 6.5 100.0
INLET
Cumulative Mass,
Size Range, urn Percent Stage Size Rai
< 0.36 29.0 Filter < 0
> 0.36 - < 0.7 52.2 6 > 0.36
> 0.7 - < 1.4 68.0 5 > 0.7
> 1.4 - < 2.3 71.3 4 > 1.4
> 2.3 - < 3.9 72.4 3 > 2.3
> 3.9 - < 6.5 72.4 2 > 3.9
> 6.5 100.0 1 > 6
OUTLET
Cumulative Mass,
Percent
42.8
79.5
91.6
92.4
93.5
93.5
100.0
-------
Table C-l (continued). PARTICLE SIZE DISTRIBUTION DATA
TEST 5
M
S tage
Filter
6
5
4
3
2
1
TEST 6
Stage
Filter
6
5
4
3
2
1
INLET
Size Range, urn
< 0.36
> 0.36 - < 0.7
> 0.7 - < 1.4
> 1.4 - < 2.3
> 2.3 - < 3.9
> 3.9 - < 6.5
> 6.5
INLET
Size Range, um
< 0.36
> 0.36 - < 0.7
> 0.7 - < 1.4
> 1.4 - < 2.3
> 2.3 - < 3.9
> 3.9 - < 6.5
> 6,5
Cumulative Mass,
Percent
21.4
41.9
62.0
67.1
70.7
73.7
100.0
Cumulative Mass,
Percent
20.9
40.2
60.4
67.8
78.8
82.6
100.0
OUTLET
Size Range, urn
< 0.36
> 0.36 - < 0.7
> 0.7 - < 1.4
> 1.4 - < 2.3
> 2.3 - < 3.9
> 3.9 - < 6.5
> 6.5
OUTLET
Size Range, urn
< 0.36
> 0.36 - < 0.7
> 0.7 - < 1.4
> 1.4 - < 2.3
> 2.3 - < 3.9
> 3.9 - < 6.5
> 6.5
Cumulative Mass,
Percent
16.1
43.9
71.1
79.4
80.5
81.9
100.0
Cumulative Mass,
Percent
17.2
37.0
62.4
69.3
69.5
70.7
100.0
-------
Table C-l (continued). PARTICLE SIZE DISTRIBUTION DATA
TEST 7
INLET
OUTLET
Stage
Filter
6
5
4
3
2
1
TEST 8
Stage
Filter
6
5
4
3
2
1
Size Range, urn
< 0.36
> 0.36 - < 0.7
> 0.7 - < 1.4
> 1.4 - < 2.3
> 2.3 - < 3.9
> 3.9 - < 6.5
> 6.5
INLET
Size Range, urn
< 0.36
> 0.36 - < 0.7
> 0.7 - < 1.4
> 1.4 - < 2.3
> 2.3 - < 3.9
> 3.9 - < 6.5
> 6.5
Cumulative Mass,
Percent
25.6
50.4
68.4
71.8
71.8
71.8
100.0
Cumulative Mass,
Percent
43.6
81.6
95.1
95.1
95.1
95.1
100.0
Stage
Filter
6
5
4
3
2
1
Stage
Filter
6
5
4
3
2
1
Size Range, urn
< 0.36
> 0.36 - < 0.7
> 0.7 - < 1.4
> 1.4 - < 2.3
> 2.3 - < 3.9
> 3.9 - < 6.5
> 6.5
OUTLET
Size Range, urn
< 0.36
> 0.36 - < 0.7
> 0.7 - < 1.4
> 1.4 -~< 2.3
> 2.3 - < 3.9
> 3.9 - < 6.5
> 6.5
Cumulative Mass,
Percent
18.3
42.0
65.7
72.4
74.6
76.8
100.0
Cumulative Mass,
Percent
6.7
34.7
85.4
85.7
86.1
86.4
100.0
-------
Table C-l (continued). PARTICLE SIZE DISTRIBUTION DATA
TEST 9
Stage
Filter
6
5
4
3
2
1
TEST 10
Stage
Filter
6
5
4
3
2
1
INLET
Size Range, urn
< 0.
> 0.36 -
> 0.7 -
> 1.4 -
> 2.3 -
> 3.9 -
> 6.
36
< 0.7
< 1.4
< 2.3
< 3.9
< 6.5
5
INLET
Size Range, urn
< 0.
> 0.36 -
> 0.7 -
> 1.4 -
> 2.3 -
> 3.9 -
> 6.
36
< 0.7
< 1.4
< 2.3
< 3.9
< 6.5
5
Cumulative Mass,
Percent
27.6
55.1
88.2
88.2
88.2
88.2
100.0
Cumulative Mass,
Percent
18.3
50.6
76.6
78.6
78.6
78.6
100.0
Filter
6
5
4
3
2
1
OUTLET
Cumulative Mass,
Size Range, urn Percent
< 0.36
> 0.36 - < 0.7
> 0.7 - < 1.4
> 1.4 - < 2.3
> 2.3 - < 3.9
> 3.9 - < 6.5
> 6.5
OUTLET
Size Range, urn
< 0.36
> 0.36 - < 0.7
> 0.7 - < 1.4
> 1.4 - < 2.3
> 2.3 - < 3.9
> 3.9 - < 6.5
> 6.5
11.4
27.2
91.2
92.1
94.8
95.7
100.0
Cumulative Mass,
Percent
10.2
40.7
81.0
86.9
86.9
86.9
100.0
-------
Table C-l (continued). PARTICLE SIZE DISTRIBUTION DATA
TEST 12
INLET
OUTLET
Stage
Filter
6
5
4
3
2
1
TEST 13
Stage
Filter
6
5
4
3
2
1
Size Range, urn
< 0.36
> 0.36 - < 0.7
> 0.7 - < 1.4
> 1.4 - < 2.3
> 2.3 - < 3.9
> 3.9 - < 6.5
> 6.5
INLET
Size Range, urn
< 0.36
> 0.36 - < 0.7
> 0.7 - < 1.4
> 1.4 - < 2.3
> 2.3 - < 3.9
> 3.9 - < 6.5
> 6.5
Cumulative Mass,
Percent
10.7
27.5
48.9
55.2
57.2
58.8
100.0
Cumulative Mass,
Percent
4.1
16.9
51.1
57.2
60.3
63.2
100.0
Stage
Filter
6
5
4
3
2
1
Stage
Filter
6
5
4
3
2
1
Size Range, urn
< 0.36
> 0.36 - < 0.7
> 0.7 - < 1.4
> 1.4 - < 2.3
> 2.3 - < 3.9
> 3.9 - < 6.5
> 6.5
OUTLET
Size Range, urn
< 0.36
> 0.36 - < 0.7
> 0.7 - < 1.4
> 1.4 - < 2.3
> 2.3 - < 3.9
> 3.9 - < 6.5
> 6.5
Cumulative Mass,
Percent
3.1
14.0
47.0
69.3
71.5
71.9
100.0
Cumulative Mass,
Percent
1.6
17.2
39.2
51.8
54.4
56.3
100.0
-------
Table C-1 (continued). PARTICLE SIZE DISTRIBUTION DATA
TEST 14
Stage
Filter
6
5
4
3
2
1
TEST 15
Stage
Filter
6
5
4
3
2
1
INLET
Size Range, urn
< 0.36
> 0.36 - < 0.7
> 0.7 - < 1.4
> 1.4 - < 2.3
> 2.3 - < 3.9
> 3.9 - < 6.5
> 6.5
Cumulative Mass,
Percent
4.6
20.1
53.6
70.6
81.4
86.6
100.0
BEFORE CYCLONE
Size Range, urn
< 0.36
> 0.36 - < 0.7
> 0.7 - <.1..4
> 1.4 - < 2.3
> 2.3 - < 3.9
> 3.9 - < 6.5
> 6.5
Cumulative Mass,
Percent
0.0
1.4
3.6
6.5
13.5
53.5
100.0
OUTLET
Cumulative Mass,
Size Range, qm Percent
< 0.36 0.0
> 0.36 - < 0.7 18.6
> 0.7 - < 1.4 69.2
> 1.4 .- < 2.3 85.5
>2.3 -<3.9 88.4
> 3.9 - < 6.5 88.4
> 6.5 100.0
AFTER CYCLONE
Cumulative Mass,
Size Range, urn Percent
< 0.36
> 0.36 - < 0.7
> 0.7 - < 1.4
> 1.4 - < 2.3
> 2.3 - < 3.9
> 3.9 - < 6.5
> 6.5
1.2
17.2
59.3
66.0
68.6
70.9
100.0
-------
Table C-1 (continued). PARTICLE SIZE DISTRIBUTION DATA
TEST 16
Stage
Filter
6
5
4
3
2
1
TEST 17
S tage
Filter
6
5
4
3
2
1
INLET
Size Range, pun
< 0.36
> 0.36 - < 0.7
> 0.7 - < 1.4
> 1.4 - < 2.3
> 2.3 - < 3.9
> 3.9 - < 6.5
> 6.5
Cumulative Mass,
Percent
4.4
24.4
62.9
79.2
85.9
89.6
100.0
BEFORE CYCLONE
Size Range, um
< 0.36
> 0.36 - < 0.7
> 0.7 - < 1.4
> 1.4 - < 2.3
> 2.3 - < 3.9
> 3.9 - < 6.5
> 6.5
Cumulative Mass,
Percent
0.0
3.7
8.0
12.0
15.0
25.3
100.0
OUTLET
Cumulative Mass,
Size Range, qm Percent
< 0.36 0.0
> 0.36 - < 0.7 28.1
> 0.7 - < 1.4 65.6
> 1.4 - < 2.3 71.9
> 2.3 - < 3.9 71.9
> 3.9 - < 6.5 73.5
> 6.5 100.0
AFTER CYCLONE
Cumulative Mass,
Size Range, um Percent
< 0.36
> 0.36 - < 0.7
> 0.7 - < 1.4
> 1.4 - < 2.3
> 2.3 - < 3.9
> 3.9 - < 6.5
> 6.5
2.1
20.5
56.2
64.4
71.4
77.6
100.0
-------
Table C-l (continued). PARTICLE SIZE DISTRIBUTION DATA
TEST 18
INLET
Stage
Filter
6
5
4
3
2
1
Size Rang^e
< 0.36
> 0.36 - <
> 0.7 - <
> 1.4 - <
> 2.3 - <
> 3.9 - <
> 6.5
Cumulative Mass,
, urn Percent
0.7
1.4
2.3
3.9
6.5
2.5
13.1
44.9
55.5
59.4
61.4
100.0
OUTLET
Size Range, urn
< 0.36
> 0.36 - < 0.7
> 0.7 - < 1.4
> 1.4 - < 2.3
> 2.3 - < 3.9
> 3.9 - < 6.5
> 6.5
Cumulative Mass,
Percent
5.7
31.1
79.0
84.3
84.3
84.3
100.0
2
06
-------
APPENDIX D
LABORATORY SIZING OF TEST DUSTS
The test dusts used in the field program were first dispersed in the lab-
oratory and their particle size distribution determined with Andersen cascade
impactors. The results are shown in Tables D-l and D-2 and Figures D-l and
D-2.
The dusts were dispersed with an air ejector similar to the one used in
the Braxton dust feed system except the compressed air supplied to the
laboratory system was considerably higher (90 psi) then that supplied to the
field system (20 psi). This probably led to more nearly complete breakup of
the agglomerates in the bulk dust and hence an aerosol of smaller particle
size than was achieved in the Braxton dust dispersion system.
As expected the cupola dust displayed a much smaller particle size dis-
tribution than the fly ash. Under the laboratory dispersion conditions the
mass median diameter of the cupola dust was about 1.1 micrometers and the
mass median diameter of the fly ash was about 9 micrometers.
69
-------
Table D-l. Particle Size Distribution for Laboratory
Dispersed Cupola Dust
Stage
Filter
6
5
4
3
2
1
Size Range, urn
<0.36
>0.36 - <0.7
>0.7 - <1.4
>1.4 - <2.3
>2.3 - <3.9
>3.9 - <6.5
>6.5
Cumulative Mass,
Percent
16.2
32.3
54.5
54.5
54.5
54.5
100.0
Table D-2. Particle Size Distribution for Laboratory
Dispersed Fly Ash
Stage
Filter
6
5
4
3
2
1
Size Range, mn
<0.36
>0.36 - <0.7
>0.7 - <1.4
>1.4 - <2.3
>2.3 - <3.9
>3.9 - <6.5
>6.5
Cumulative Mass,
Percent
1.8
3.2
6.8
12.2
20.7
36.5
100.0
70
-------
10-
8-
7-
6
um
Pi
W
H
a
o
.7
.6
.3
001
O.I
10
10 40 SO 6O TO
BO
9O
93
90 99
9S.9
S9.99
PERCENT MASS < STATED PARTICLE DIAMETER
Figure D-l. Particle Size Distribution for Laboratory Dispersed Cupola Dust
-------
10—
9
8 ~
7 —
6 —
5 —
w
H
S '-
Q .9
a a
0 .7
M
§ '
f^ .3
001
O.I
5 10 2O 30 40 50 60 70 80 90 99
PERCENT MASS < STATED PARTICLE DIAMETER
98 99
99.9
99.99
Figure D-2. Particle Size Distribution for Laboratory Dispersed Fly Ash
-------
TECHNICAL REPORT DATA
(Please read liiinmclinn* on the n-rerse before completing)
I. HEPOHT NO.
EPA-650/2-74-036
2.
3. RECIF'ILNT'S ACCGSSIOIV NO.
4. - ITLE AND SUBTITLE
Braxton Sonic Agglomerator Evaluation
5. REPORT DATE
Mayl974
G. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
Richard Dennis, Robert Bradway, and.Reed Cass
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME ANO ADDRESS
GCA Corporation
GCA/Technology Division
Bedford, Massachusetts 01730
10. PROGRAM ELEMENT NO.
1AB012; ROAP 21ADL-04
11. CONTRACT/GRANT. NO.
68-02-1316 (Task 1)
12. SPONSORING AGENCY NAME ANO ADDRESS
EPA, Office of Research and Development
NERC-RTP, Control Systems Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT ANO PERIOD COVERED
Final
14. SPONSORING AGENCY CODE
IS. SUPPLEMENTARY NOTES
16. ABSTRACT,
The report'is an evaluation of a novel air pollution control system developed
by the Braxton Corporation. The alternating velocity precipitator, or sonic agglom-
erator, is designed to decrease the number and increase the size of particles in a
gas stream by agglomeration induced by a standing sound wave through which the
aerosol moves. A prototype alternating velocity precipitator was tested to.determine
its basic performance characteristics and to evaluate the effect of adding water and/
or steam to the systems performance.^ Tests indicate that the device decreases the
mass of fine particles but that the reduction is more highly correlated to the use of
water sprays than to the use of the sonic generator. The correlation coefficients
relating the use of sound and water sprays to fine particle reduction, however, were
not statistically significant. The particle size distributions of the fine particles at
both the inlet and outlet of the sonic agglomerator were determined with Andersen
cascade impactors. Although shifts between the inlet and outlet size distributions
were often observed, there was no clear trend to the changes and they could not be
correlated to system operating parameters.
7.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COSATI l-'icld/Croup
Air Pollution
Acoustics
Precipitators
Agglomeration .
Particle Size
Particle Size Distribution
"as Flow
Aerosols
Sound Waves
Air Pollution Control
Stationary Sources
Braxton Sonic Agglom-
erator
Alternating Velocity
Precipitator
Particulates
13B
20A
7A
7D
20D
i. uiurmuurioN STATEMENT
Unlimited
19. SE.CUHITY CLASS (Tills Krporl)
Unclassified
21. NO. OF CAGES
20. SECURITY CLASS (Thil pagt)
Unclassified
22. PRICE
E/-A Form 2220-1 (9-73)
------- |