EPA-600/2-76-164
June 1976
Environmental Protection Technology Series
EVALUATION OF REXNORD GRAVEL BED FILTER
Industrial Environmental Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
-------
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 reviewed 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.
-------
EPA-600/2-76-164
June 1976
EVALUATION
OF REXNORD
GRAVEL BED FILTER
by
Joseph D. McCain
Southern Research Institute
2000 Ninth Avenue South
Birmingham, Alabama 35202
Contract No. 68-02-1480
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
-------
ABSTRACT
This report presents the results of fractional and overall
mass efficiency tests of a Rexnord, Inc., gravel bed filter system.
The tests were performed on a full scale system used for controlling
particulate emissions from a Portland cement plant clinker cooler.
Total flue gas particulate mass concentrations and emission rates
were determined at the inlet and outlet of the gravel bed system
by conventional (Method 5) techniques. Inlet and outlet emission
rates as a function of size were determined on a mass basis using
cascade impactors for sizes from about 0.5 ym to 5 ym, and on a
number basis for sizes smaller than about 1 ym using optical and
diffusional methods.
The text of this report includes brief descriptions of the
Portland cement process, the Rexnord gravel bed filter system, the
measurement methods, inlet and outlet particle size distribution
data, and fractional efficiencies.
This report was submitted in partial fulfillment of contract
number 68-02-1480 to Southern Research Institute under the
sponsorship of the Environmental Protection Agency. The work
reported here was completed January 31, 1976.
-------
TABLE OF CONTENTS
Contents Page
SECTIONS
I - Conclusions 1
II - Introduction 2
III - Discussion 4
Description of the Gravel Bed Filter System 5
Method 5 Results and Overall Collection Efficien-
cies 5
Cascade Impactor Results 9
Ultrafine Particulate Data 14
Optical Particle Counter Results 18
APPENDICES
A - Cascade Impactor Data By Run 33
B - Plant Production Data 39
C - Method of Operation of Gravel Bed Filters and Design
Specifications of the System Tested 40
FIGURES
la - Portland cement kiln and clinker cooler layout.
(One of two kilns shown) 3
Ib - Gravel bed system layout and sampling locations.
The system tested was comprised of eight modules in
two parallel sets of four rather than the illustra-
ted 3
2 - Inlet and outlet size distributions on a cumulative
percentage by weight basis 11
3 - Inlet and outlet size distributions on a mass basis.. 12
4 - Fractional efficiencies as determined using Cascade
impactors. August data 15
5 - Fractional collection efficiencies as determined
using Cascade impactors. November data 16
111
-------
TABLE OF CONTENTS (Cont'd.)
FIGURES (Cont'd.) Page
6 - Fractional efficiencies as determined using
diffusional, optical and inertial (Cascade
impactor) methods. August data 20
7 - Temporal variations of particulate concentrations
in three size bands at the outlet of the gravel bed
filter system. Concentration units are arbitrary
and not to the same scale for the three size bands.... 23
8 - Relative emission rates in three particle size
intervals during explorations of the effects of back-
wash operations on emissions. For explanation of
symbols see text 26
9 - Representative cut diameters as a function of pres-
sure drop for several scrubber types, after Calvert
(1974), J. APCA 24:929 31
Cl - Operation of gravel bed filter 41
TABLES
1 - Design specifications of the system as tested.. 6
2 - Inlet 7
3 - Outlet 8
4 - Gravel bed filter fractional efficiencies as measured
with Andersen Impactors 13
5 - Ultrafine particulate data - August test 19
6 - Ultrafine particulate data - November test 19
7 - Summary of small particle optical data 22
8 - Effect of backflush interval on particulate emission
rate 24
9 - Percentage of emissions in puffs 28
10 - Fractional efficiencies from August and November data
with and without contribution from cleaning cycle 29
IV
-------
TABLE OF CONTENTS (Cont'd.)
TABLES (Cont'd.) Page
Al - Inlet mass loading by size interval from
Andersen impactor data. Mass loading in
indicated size interval, mg/DNM3 34
A2 - Outlet mass loadings by size interval for August
test series. Mass loading in indicated size inter-
val, mg/DNM3 35
A3 - Outlet mass loadings by size interval for November
test series. Mass loading in indicated size inter-
val , mg/DNM3 36
A4 - Inlet mass loading by size interval from Brink
Impactor Data. Mass loading in indicated size
interval, mg/DNM3 37
A5 - Andersen impactor blank weight gains 38
v
-------
SECTION I
CONCLUSIONS
This evaluation was one of a series of studies being conducted
by the Industrial Environmental Research Laboratory of the Environ-
mental Protection Agency to identify and test novel devices which
are capable of high efficiency collection of fine particulates.
The test methods used may not be consistent with compliance-type
methods, but were state-of-the-art techniques for measuring mass
and fractional efficiency using the standard mass train and iner-
tial, electrical, optical and diffusional techniques.
The collection efficiency of the Rexnord gravel bed filter,
determined by conventional (Method 5) techniques on a source
producing particulate having a mass median diameter of about 200
ym ranged from 95% to 98% during three days of testing throughout
which the collector was not operating in an optimum mode. Over-
all efficiencies determined from cascade impactor data during
a second two-day test series were found to be 99.3% and 99.7%.
The system pressure drops in the first test series ranged from 11.6
to 17.6 cm w.c. while during the second test series the system
pressure drop ranged from 9.6 to 14 cm w.c. Measured fractional
efficiencies were about 50% at 0.04 ym, zero or negative over the
size interval from about 0.08 ym to 1.0 ym, approximately zero at
1 ym, 30% at 2 ym, and about 97.5% at 5 ym. The system energy
usage during the tests was approximately 1780 joules /SCM (47.7 BTU/
1000 SCF) at a pressure drop of 11.8 cm (4.7 inches) w.c.
Most of the devices tested to date under the novel device
test program have been scrubbers. For this reason it has been
convenient to compare their performance to a conventional venturi
scrubber. The Rexnord gravel bed, while not a scrubber, has also
been compared on the same basis as shown in Figure 9. It was
-------
determined that the power consumption of the Rexnord unit was
somewhat higher but not substantially different from that of a
well-designed venturi scrubber giving the same particulate col-
lection efficiency.
SECTION II
INTRODUCTION
This report presents results of tests conducted by Southern
Research to determine the capability of the Rexnord gravel bed
filter to collect fine particulates. The goals of the tests
were to determine the overall mass efficiency and the fractional
efficiency of the filter when operating under normal conditions
in controlling the effluent from a Portland cement plant clinker
cooler.
Figure 1 is a schematic of the basic Portland cement kiln,
clinker cooler, and gravel bed systems showing the inlet and out-
let sampling locations. The tests were conducted on a gravel bed
filter system controlling the emissions from clinker coolers on
two 500-tons/day Portland cement kilns. Both kilns were operating
at full capacity during the majority of the tests. In the process
a slurry of water and powdered raw materials is introduced to a
kiln in which they are calcined, forming clinker which drops onto
a moving bed clinker cooler. A series of fans beneath the moving
bed blow cool, ambient air through it to reduce the temperature
of the product. The bulk of the air from the clinker cooler is
used as combustion and secondary air for the kiln. Scavenge air
from the system (that air not required by the kiln) is drawn off
through the gravel bed filter. Because the air to the gravel bed
is scavenge air, the airflow through the filter is subject to
considerable variation in both temperature and volumetric flow
rate. Inlet gas temperatures ranged from 70°C (170°F) to 230°C
-------
Burner
Clinker Cooler
Product
Discharge
Figure la. Portland cement kiln and clinker cooler layout.
Arrows indicate directions of air flow.
Inlet sampling
j § s §
Outlet sampling
points
X
12".
h"--1
Gravel Bed Filter
System
Stack
Figure Ib. Gravel bed filter system layout and sampling locations,
-------
(450°F). Outlet temperatures ranged from 71°C (160°F) to 129°C
(265°F). The temperature drop between the inlet and outlet
results primarily from the addition of approximately 283 m3/mi.n
(10,000 cfm) of ambient air which is used to periodically clean
the gravel beds. During these tests, the gravel bed system pressure
drop was approximately 12 cm water column (4.7 in). The system
flow rate ranged from about 600 to 1000 DNM3/™in at the inlet.
(The latter figures exclude the backflush air). Testing took place
during the months of August and November 1975. The second series
of tests were done as a result of a determination by the manufacturer
that the system was not operating under optimal conditions during
the August test. The results of the November tests did show sub-
stantial improvement in system performance as compared with the
August data.
SECTION III
DISCUSSION
A total of five measurement techniques were used during the
tests. These were: (1) diffusional techniques using condensation
nuclei counters and diffusion batteries for determining concentration
and size distribution on a number basis for particles having dia-
meters less than approximately 0.2 ym, (2) optical techniques to
determine concentrations and size distribution for particles
having diameters between approximately. 0.3 ym and 1.5 ym, (3) op-
tical techniques for monitoring outlet concentration variations
over the size range from 0.6 to 50 ym, (4) inertial techniques
using cascade impactors for determining concentrations and size
distributions on a mass basis for particles having diameters be-
tween approximately 0.5 ym, and 5 ym, and (5) standard mass train
(Method 5) measurements for determining total inlet and outlet mass
loadings.
-------
Description of the Gravel Bed Filter System
The gravel bed filter is constructed on a modular basis, with
eight modules making up the system in this instance. Each module
consists of a cyclonic inlet section followed by two gravel beds
operating in parallel.
At any one time, during normal operations, seven (7) of the
eight (8) modules are on line in the forward flow direction with
one being cleaned and renewed by backflushing with heated ambient
air. In the installation tested, the modules were cleaned in se-
quence with a backflush time of 6 to 20 minutes. During the first
45 seconds of the backflush period, a mechanical raking system is
actuated to stir up the dirty gravel. The particulate laden air
from the module being backflushed is exhausted into the inlet plenum
of the remaining modules. This backflush and raking operation and
the frequency with which it takes place has a very pronounced in-
fluence on the collection efficiency achieved by the system as will
be discussed later in this report. The tests of this unit were
done under long term steady-state conditions with backwash periods
of 12 and 6 minutes. In addition, limited short term tests under
transient conditions were made using 6 and 20 minute backwash inter-
vals during a period when the system was normally being operated
with a 12-minute backwash interval. A more complete description
of the gravel bed system, together with illustrations, are given
in Appendix A. Design values for the system tested are given in
Table 1.
Method 5 Results and Overall Collection Efficiencies
Because of economic considerations and limited working space,
Method 5 testing was done only during the first three days of the
first (August) test series.
The data obtained by Method 5 technique is summarized in
Tables 2 and 3. The overall collection efficiency for each of
these tests is also given in Table 3, together with the gas flow
per module (at flue conditions). It would appear from these data
that the collection efficiency of the system is sensitive to the
gas flow per module and improved markedly with decreasing gas flow
-------
Table 1
Design Specifications Of The System As Tested
Inlet Volume Flow: 2266 ACM/min at 204°C
(80,000 ACFM at 400°F)
Backflush Volume Flow: 317 ACM/min at 66°C
' (11,200 ACFM at 150°F)
Pressure Drop: 25.3 cm w.c. (10 inches w.c.)
Gravel Size: 4 mm (5/32 inch) x No. 6 mesh
Bed Depth: 11.4 cm (4% inch)
Bed Area: 3.72 m2/Bed (40 ft2/Bed)
2 2
(For a total of 59.5 m of bed area with 52 m actively
filtering in normal operation.)
-------
Table 2
Mass Emission Tests - Method" 5
Inlet
Run #
Date
Time
%, Moisture
Velocity, m/s (f/s)
ACM/min (ACFM)
SDCM/min (SDCFM)
Grams/ACM (Grains/ACF)
Grams/SDCM (Grains/SDCF) 3.192(1.395)
Kg/hr. (Lbs/hr.) 179.46(395.65)
8-25-75
1350
2.33
10.44(34.24)
1467.3(51812)
937.1(33089)
2.039(0.891)
8-26-75
1015
1.65
9.70(31.48)
1364.5(48140)
1088.4(38431)
1.144(0.500)
1.435(0.627)
93.68(206.54)
8-26-75
1435
2.60
7.42(24.35)
1043.5(36847)
761.6(26891)
1.602(0.700)
2.197(0.960)
100.37(221.27)
8-27-75
1050
1.54
8.50 (27.89)
1195.2(42203)
875.3(30906)
2.130(0.931)
2.911(1.272)
152.84(336.96)
8-27-75
1515
1.80
7.97(26.15)
1120.6(39570)
739;9(26128)
2.078(0.908)
3.146(1.375
139.68(307.94)
-------
Table 3
Mass Emission Tests - Method. 5
Outlet
Run #
Date
Time
Velocity, m/s(f/s)
%, Moisture
00 ACM/min (ACFM)
SDCM/min(SCDFM)
Gr ams/ACM(Grains/ACF)
Grams/SDCM(Grains/SDCF)
Kg/hr'. (Lbs/hr.)
No. of Active
Modules
Average Flow per
Active Module
in ACM/min(ACFM)*
Efficiency
8-25-75
1400
8.82(28.94)
2.29
1631.7(57619)
1239.3(43759)
0.094(0.041)
0.121(0.053)
9.02(19.88)
7
233.1(8230)
95.00
8-26-75
1015
8.76(28.73)
1.83
1619.9(57201)
1326.4(46837)
0.030(0.013)
0.037(0.016)
2.91(6.42)
7
231.4(8170)
96.9
8-26-75
1445
6.79(22.29)
1.86
1256.8(44379)
1017.5(35927)
0.064(0.028)
0.080(0.035)
4.89(10.78)
4
8-27-75
1100
7.98(26.18)
1.64
1476.2(52124)
1174.2(41461)
0.043(0.019)
0.055(0.024)
3.87 (8.53)
7
314.2(11095) 210.8(7445)
95.1
97.5
8-27-75
1515
7.26(23.81)
1.38
1342.5(47405)
1049.7(37067)
0.034(0.015)
0.043(0.019)
2.74 (6.04)
7
191.7 (6770)
98.0
*Includes backflush air.
-------
over the range of values that were obtained during this series of
tests. The differences in the inlet and outlet gas flows in Tables
2 and 3 are due to the addition of the backflush air. During these
tests, the backwash duration was 12 minutes.
On a service-inspection trip by Rexnord personnel subsequent
to this test series, it was found that the rakes were being
activated a fraction of a second to a few seconds before the back-
wash valves were actuated, which would tend to reduce the average
collection efficiency of the device. In addition, accumulated ex-
perience with the gravel bed system in this application indicated
that the beds were not being adequately cleaned with the 84 minute
forward-flow/12 minutes backwash cycle, which results in a buildup
of dust within the beds. This increases the energy requirements
and decreases the collection efficiency of the system. Resetting
the rake timers and changing the backwash intervals to provide a
42-minute forward flow/6 minute backwash cycle resulted in a sub-
stantial improvement in performance as reported by plant personnel.
Consequently, a second series of tests were run during the month
of November, 1975. Method 5 measurements were not made during the
retests. The outlet impactor data during the second test series,
indicated a reduction 'in the outlet particulate loading by approxi-
mately a factor of 3.5 as compared to the impactor outlet data
obtained during the first test series, while inlet data from cas-
cade impactors indicated no discernable differences in loading
and size distribution between the two test series. Thus, in steady
state operation under a more nearly optimum operating cycle, the
overall efficiency in this application is estimated to be approxi-
mately 99.5%.
Cascade Impactor Results
Inertial sizing was accomplished using Brink and Andersen
Cascade impactors for inlet measurements and Andersen Impactors
only for outlet measurements. Sampling was done at near isokinetic
rates. Errors due to deviations from isokinetic sampling should
be of little consequence for particles having aerodynamic diameters
-------
smaller than 5 ym. Inlet sampling times of three hours produced
catches of substantially less than one milligram on all stages
collecting particles smaller than 3 ym with the Brink impactors.
These catches were considered too low to give reliable loadings
so only the Andersen inlet data was used for fractional efficiency
calculations.
Cyclone precollectors were required on the inlet impactors
because of the very high concentrations of +10 ym particles
(approximately 98% by weight of the influent particulate was in
particles larger than 10 ym). The cyclones which were available
for use with the Andersen impactors in this application had parti-
cle collection characteristics such that no information on the
inlet size distribution could be obtained for sizes larger than
about 5 ym. Because this program was concerned primarily with
fine particles, this limitation was of little consequence.
The impactor data are summarized in Figures 2 and 3. Figure
2 shows averaged inlet and outlet mass size distributions on a
cumulative percentage versus aerodynamic diameter basis. Figure 3
shows the same size distributions on a cumulative mass concen-
tration basis. The high concentrations of large particles in the
gas streams coupled with the non-ideal particle separation and
collection characteristics of the impactors tend to make the im-
pactor filter catches difficult to interpret. In sampling parti-
culate having the properties and size distribution of the type
encountered here the impactor back-up filter catches can be dom-
inated by oversize particles, a small fraction of which are not
retained by the stages which should collect them. As a result,
no fractional efficiencies were based on the impactor back-up
filters and the size distribtuions obtained with the impactors
as shown in Figures 2 and 3 are given both with and without the
back-up filter catches.
The fractional efficiencies as calculated from the impactor
data are given in Table 4 together with the gravel bed operating
10
-------
0
•H
N
•H
CO
XJ
EH
01
CO
0)
tn
•H
0)
PQ
C
0)
O
M
a>
Andersen with filter]
0 Andersen less filter!
• Brink with filter
O Brink less filter
1.0
August with filter
V August less filter
November with filter
November less filter
2 5 10 20
Aerodynamic Diameter, yM
Figure 2. Inlet and outlet size distributions on a cumulative
percentage by weight basis.
-------
1000
0)
N
13
•H
4J
rt
100
to 2700 mg/DSCM
Asymptotic
August outlet with
V August outlet less filter
November outlet with filter
A November outlet less filter
^Average inlet with filter (Andersen data)
^Average inlet less filter (Andersen data)
) Average inlet (Brink data)
Hi
10
Aerodynamic Diameter, yM
100
Figure 3. Inlet and outlet size distributions on a mass basis.
12
-------
Table 4
Gravel Bed Filter Fractional Efficiencies As
Measured With Andersen Impactors
u>
Date
8/27
8/28
8/28
8/29
8/29
11/4
11/5
Inlet
Gas Flow
DSCM/min.
739
917
606
1039
682
732
856
Inlet
Temp.
°C
149
181
227
193
149
166
152
Typical
Bed Pres.
Drop
cm/H20
11.0
13.6
10.2
13.26
9.75
6.4
9.1
System
Pres.
Drop
cm/H20
12.39
17.3
12.8
17.58
11.58
9.6
14.0
Backf lush
Period
12 min.
12 min.
12 min.
12 min.
12 min.
6 min.
6 min.
.72
-37
-146
4
13
-229
-104
-31
Collection Efficiencies At
Indicated Aerodynamic Diameters
1.09 2.04 3.70 5.45 8.26
-68
-56
-32
-43
-56
-29
0
32
2
-42
11
25
66
78
60
34
19
37
67
93
94.1
91 97.7
>71
>40
>67
>57
98.1
96.3
-------
pressure loss and. are shown in Figures 4 and 5, Also shown in
Figures 4 and 5 are fractional efficiency curves derived from
overall averages of the respective data obtained during the two
test series. Because the outlet loadings appear to he decoupled
from the inlet insofar as short term behavior is concerned, these
overall averages may better illustrate the system performance than
do the individual tests. The negative efficiencies shown for some
particle sizes are discussed in a later section of the report.
Data from all impactor runs including blank runs which were made
to determine the possible level of any interferences on the im-
pactor data are given in Appendix B.
The sizes reported here for the inertial data are given in
two forms, "aerodynamic" and "physical", or Stokes' diameters.
The "physical" diameters are based on a particle density of 2.7
g/cm3, which was determined with a helium pycnometer from a bulk
sample of the particulate. If the true particle densities are
lower than this value, the sizes as given should.be increased by
a factor equal to the square root of ratio of the assumed density
of true density. Aerodynamic diameters are diameters based on a
particle density of 1 g/cm3.
Ultrafine Particulate Data
Data on the concentration and size distribution of ultrafine
particulates were determined during the first test series using
diffusional sizing techniques with General Electric Condensation
nuclei counters for determining the various ultrafine particulate
concentrations. Attempts were made to use a Thermosystems Model
3030 Electrical Aerosol Analyzer; however, rapid random fluctua-
tions in particulate concentrations and size distribution in this
size range rendered the data from this method almost totally un-
interpretable. In addition, a Gardner small particle detector
(a manually operated CNC) was used throughout the tests as a crude
monitor of the exit concentration and size distribution (using a
variable supersaturatibn method) of ultrafine particles. Some
inlet data were also obtained with the variable supersaturation
method during the November tests. This instrument was used without
14
-------
o
0)
•r\
O
•r\
<4-t
M-l
W
C
O
•H
-M
O
0)
O
o
PRES. DROP
(cm w.c.)
Typical
Bed System
8/27
8/28
8/28
8/29
8/29
Average
-180
2 5
Aerodynamic Diameter, yM
Figure 4. Fractional efficiencies as determined using Cascade
impactors. August data.
15
-------
§
•H
O
•H
M-l
IH
W
C
O
•M
+J
U
0)
O
u
-140
-160
-180
Aerodynamic Diameter, yM
Figure 5. Fractional collection efficiencies as determined using
Cascade impactors. November data.
16
-------
dilution and much of the time the particulate concentration ex-
ceeded its range, however, it showed qualitative agreement with
the data obtained using the more elaborate system in those instances
during which they were at a common location. During the second
(November) tests series, only the Gardner small particle detector
was used to obtain data on ultrafine particulates.
The useful concentration ranges of the electrical aerosol
analyzer and the condensation nuclei counter are such that exten-
sive dilution of the gas streams being sampled was required in
order to obtain reliable data. Dilution factors of about 50:1 were
used for both inlet and outlet 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 diffu-
sional driers were utilized in the lines carrying the diluted
samples to the various instruments.
Because only one set of electrical and diffusional sizing
equipment was available, it was not possible to obtain simultaneous
inlet and outlet data with these methods. The system was first
installed at the inlet sampling location, and all the inlet data
were obtained. For the purposes of calculating the efficiency of
the filter, the assumption was made that the process was suffi-
ciently 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..
The ultrafine particulate data were confused by the nature of
the particulate source which produced large, sporadic concentration
changes. Because the clinker cooler is not a combustion process,
it is likely that the ultrafine particles were introduced to the
flue gas for the most part via the ambient air supplied to the
clinker cooler fans. The concentration and size distribution of
this component of the ultrafines would be expected to be highly var-
iable. An additional complicating factor was the ultrafine parti-
culate brought into the system by the backwash air, which was also
17
-------
subject to large fluctuations in concentration. Accuracy in the
measurements was limited by the rapid and frequent concentration
variations and the efficiencies derived from these data are un-
certain. However, the trends in the fractional efficiencies derived
from the data are probably real and the fractions of the influent
materials that penetrate the scrubber are believed to be generally
correct to within a factor of two. The data obtained during the
August test series are summarized in Table 5. The fractional
efficiencies derived from the data are shown in Figure 6. Also
shown on Figure 6 are fractional efficiencies obtained from simul^-
taneous measurements with the small particle optical counter and
cascade impactors. The limited comparison data obtained during
the November tests are given in Table 6. The efficiency for the
+0.2 ym diameter particles in the November data appears to be
substantially higher than was the case in August. It is unlikely
that the collection efficiency in the .02 to .2 range was increased
to a value as high as 80 to 90%, however values of about 50% cannot
be excluded.
Optical Particle Counter Results
Two optical particle counters (Royco Model 225) were used
during these tests. The first was used in conjunction with the
ultrafine measurement system to obtain data on fractional effi^-
ciencies over the size range from 0.35 ym to 2.0 ym. The second
was used only at the outlet of the gravel bed for obtaining real
time information on relative changes in concentration of particles
over the size range from 0.6 to 50 ym.
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, but also serve as rough checks
on the data obtained with the cascade impactors.
18
-------
Table 5
Ultrafine Particulate Data—August Test
Cumulative Concentration Larger Than Indicated Size
Diffusional Method, Variable Supersaturation Method
Inlet
8/26
Inlet
8/27
Outlet
8/28
Date:
Dia., urn
.01
.02 2.0xl06 l.SxlO6 1.7x10
.063 5.5xl05 4xl05 9.6xl05
2xl05 4.4xl05
.10
.20
3.5xl06 2.3xl06 4.2xl06
6
2.9xl05
7x10"
3x10" 1.4xl05
Outlet
8/26
Outlet
8/27
Outlet
8/28
1.3xl06 2.5xl06 >2.8xl06
7.9xl05 2xl06 >2.8xl06
6.9x10" 2.1xl05 2.2xl05
Table 6
Ultrafine Particulate Data—November Test
Date:
Dia., ym
.01
.02
.20
Cumulative Concentration Larger Than Indicated Size*
Inlet Outlet
11/4
11/5
>2.8xl06 >2.8xl06
>2.8xl06 >2.8xl06
2.1xl05 3.0xl05
11/4
11/5
>2.8xl06 >2.8xl06
>2.8xl06 >2.8xlOb
l.SxlO5 3.5xl05
*A11 by variable supersaturation method.
19
-------
to
o
•H
u
M-l
100
80
60
40
20
0
-20
-40
-60
-80
-100
-120
-140
-160
-180
0 Diffusion
O Optical 1
A Impactor
__,.. i
-.^.t .;:. .)•-
lal Method
Method
s (Aerodynami
. ' ' ' ' '
>!,,•'
c
*\
?
I
Di<
•
j>
V
ai
^
ne
s
,t
m '1-
•
ters) 1 --I
T |
- ±T..t.
a. jj.) _
T T it
T Tt
" t T ~~
l"
i.i
1
I'" "^
| i j
i j 1
1
_ /
k
i
i
! i i
i,!i
1 1
I fl M
•" +•• • i
i !
"T
1 [
m
i
(
i
i
)
i
_
^__ _p__
.
,\
I
_!_. 1
1
-T- -1 i 1
;•:•••
"*" T
, T
i |llj
""I r
in HI
rtj .n
ftt t
t"r tt
ii..n
I If., n}
ti. jijn
_L 4i ^
T
U.Ol
0.02
0.05 0.5
Particle Diameter, yM
Figure 6. Fractional efficiencies as determined using diffusional,
optical and inertial (Cascade impactor) methods. August
data.
-------
Inlet and outlet data with the small particle unit CO.35 to
2 ym) were obtained at the same times as were the ultrafine data.
These data are thus not simultaneous sets and are subject to some
of the same uncertainties as were the diffusional data. The inlet
concentrations of particles in this size range was much more
stable than the ultrafine concentrations, although still subject
to large concentration swings, and were probably dominated by
entrained particulate from the clinker cooler. Concentrations in
this size range were believed to be relatively unaffected by the
ambient airborne particulate. The results of the measurements are
given in Table 7. Fractional efficiencies determined from the
optical data were shown in Figure 6, from which it can be seen
that the negative fine particle collection efficiencies which were
found from the impactor data were verified by the particle counter
data.
The outlet data from both the small particle (0.3 to 2 ym)
and large particle (0.6 to 50 ym) systems showed pronounced effects
from the backflush cycle on emission rates. Typical examples of
the time variations in the outlet particulate concentrations are
shown in Figure 7, which represents an approximate 30«-minute data
segment obtained with the continuous monitoring large particle
system during the August tests. A large "puff" followed by a
slow decline in concentration occurs each time a clean module is
put back on line (once every 12 minutes in this case). During
the August test series, this 12-minute interval was varied for
two special one-hour duration tests (one with 6-minute intervals,
the other at 20 minute intervals) in order to allow some estimate
to be made of the sensitivity of the gravel bed system to changes
in operating conditions. The results of these tests are given
in Table 8.
It appears from the data shown in Table 8 that the effects on
emissions resulting from changing the insertion intervals were
small and only slightly size dependent, at least during these
brief tests. Because of the brevity of these special condition
21
-------
Table 7
Summary of Small Particle Optical Data
Concentration In Size Range (number/cc)
Inlet
Date:
Averaging Time
Dia . Range , ym*
.35-. 43
.43-. 58
.58-1.2
1.2-1.4
1.4-2.0
8/26
(2.5 hrs)
S.OxlO2
3.4xl02
2.4xl02
1.4xl02
3.1xl02
8/27
(6 hrs)
7.46xl02
4.7xl02
3.1xl02
l.SxlO2
4.1xl02
Outlet
8/28 8/29
(0.75 hrs) (5 hrs)
14.3xl02 9.0xl02
14.0xl02 IS.OxlO2
9.6xl02 7.9xl02
5.2xl02 3.6xl02
6.0xl02 3.9xl02
*Polystyrene equivalent diameter
22
-------
to
CO
T)
C
0)
N
•H
CO
0)
-P —
(0 W
U -P
•H -H
-d c
fi D
CH M
•r-\ ($
en 4-1
C -H
O XJ
-H S-4
-P
-------
Table 8
Effect of Backflush Interval On Particulate Emission Rate
Clean Module Insertion Relative Particulate Emission Rate In Size
Interval (Minutes) Bands Indicated (Arbitrary Units', 1 Hr. Avg.)
.6-1.8 ym dia. 1.8-3.6 ym dia. 3.6-7.2 ym dia.
6 36 42.5 37
12 31 39.2 40
20 25 38.5 52
24
-------
tests, the system did not have time to come to a steady state condi-
tion, thus the results obtained during the one^hour test periods
do not well represent what might be obtained if sufficient time
had been available for testing under steady-state conditions. HOWT
ever, these results do provide some indications of the system sensi-^
tivity to changes in operating conditions.
The November tests, which were made with the gravel bed system
operating in a steady state conditions with a 6-minute backflush
time, showed a substantial improvement in performance as evidenced
by the impactor data, in contrast to the rather small performance
change observed during the one hour of operation with a 6^-minute
backflush cycle during the August tests.
The combined particle counter and cascade impactor data
suggest that a large portion of the outlet fine particle emissions
result from the breaking up of agglomerated fine particles during
the raking operations in the cleaning of the off-line beds. The
nature of the clinker cooler process is such that a substantial
portion of the particulates lofted from the clinker bed by the
cooling air could be agglomerates. A limited number of tests were
run in November in order to explore the hypothesis that the negative
efficiencies resulted from the breakup of agglomerates. These
tests were made using only the realtime large particle system for
monitoring the particulate concentrations in the gravel bed effluent
gas stream. During these tests, clean beds were put on line with-
out simultaneously backflushing another bed and dirty beds were
backflushed with and without raking with no clean bed being put on
line.
The effects of these operations are illustrated in Figure 8
which shows strip chart recordings of the particle counter output
during these tests. Events marked as "N" are the puffs normally
observed when a new bed is placed in the backflush mode at which
time a clean bed is simultaneously placed on line. Events marked
"C" are the result of placing a clean bed on line without initiatina
25
-------
0.6-1.8 yM
HP tilW^
•tfpW trif ti^fittt ^itMTO^W^ fa
iiitai i-Hi- WJM iMiKMMittM
Uu^iL
]v)/p^}:.:|i M Ljj
4V^iiJiU.LuJ iiU-iiiiUivi^A!
r];-[:':"i:|"::|::':;T 1" :::
].U.li li Ulii jiiiiil lii • 1{ Lii j 0 i.L! Liiiili. i.i iJ i ii!.U L[
u :
Time
Figure 8. Relative emission rates in three particle size intervals during explorations
of the effects of backwash operations on emissions. For explanation of sym-
bols see text.
-------
a backflush cycle on another module. Events marked "R" are the re-
sult of raking a dirty bed which is being backflushed without
simultaneously placing a clean bed on line. The event marked "B"
resulted from backflushing a bed without raking it and without
simultaneously placing a clean bed on line.
It appears that when a clean bed is placed on line, some
residual particulate which has been loosened by the backwash opera-
tion is carried out directly to the stack causing a short duration,
low amplitude puff. The valving operations also appear to produce
a very short duration, concentration spike designed "V" in Figure
8. The initial burst of backflush air to a bed that is entering
the cleaning cycle appears to create a puff of lesser amplitude.
The raking process during the backwash process produces moderate
to large amplitude concentration increases which last slightly
longer than the time period during which the rakes are activated.
The puffs labeled "B" and "R" result from particulate that was
entrained in the backwash air and passed in the normal, forward
flow direction through the on line, active, modules. It was there-
fore concluded that the backwash and raking cycle does in fact
result in a large amount of deagglomeration of the particulate
which has been collected by a bed, and that in some aerodynamic
size range, the emissions resulting from the portion of this reen-
trained material which is not recollected by the active beds ex-
ceeds the rate at which particulate in those size ranges is carried
into the system from the clinker cooler. Estimates of the percentages
of total emissions in several size intervals resulting from the
puffs were made from the data obtained with realtime monitors. These
estimates are presented in Table 9 for the August and November tests.
Estimates of the fundamental collection efficiencies of the
gravel beds were made by subtraction of relative contributions
due to the puffs from the cascade impactor outlet data and recal-
culating fractional efficiencies based on the revised data. The
results of these calculations are given in Table 10, together with
27
-------
Table 9
Percentage of Emissions In Puffs
Monitor #1
Monitor #2
.35-. 6
29
N.A.
.6-1.2
51
N.A.
.6-1.8
55
(29)
1.8-3.6
70
67
3.6-7.2
62
(41)
Date: Size Interval:
August (12 min Backwash)
November (6 min Backwash)
N.A.: Not Available - Monitor #1 not used during November tests.
Values in parentheses are uncertain.
28
-------
Table 10
Fractional Efficiencies From August and November
Data With and Without Contribution From Cleaning Cycle
Collection Efficiencies
Date
8/27
8/28
8/28
8/29
8/29
11/4
11/5
.72
with
-37
-146
4.4
13
-229
-104
-116
Vim
w/o
42
-3
71
63
-38
(-53)
(-62)
with
-68
-56
-32
-43
-56
-29
15.5
At Indicated Aerodynamic Diameters
1.09
w/o
13
19
31
26
19
(33)
(55)
2.04
with
32.1
2.2
-42
11
24.5
66.2
73.8
w/o
69
56
36
60
66
90
92
3.
with
60
33.8
19
37.2
67
93
94.8
7
w/o
72
54
43
56
77
97
98
Value in parentheses uncertain.
29
-------
the efficiencies calculated from the original data. These esti-
mates indicate that the collection efficiency of the gravel bed
under the conditions of these tests was low for particles smaller
than about 2 ym even without the effect of the puffs,
The performance obtained with the Rexnord gravel bed filter
during these tests can be compared with the expected performance
figures for scrubbers using the "cut diameter" method described by
Calvert (1974) J APCA, 24: 929. This method is based on the idea
that the most significant single parameter to define both the
difficulty of separating particles from gas and the performance of
a scurbber is the particle diameter for which the collection effi-
ciency is 0.5 (50%). Figure 9 is adapted from one presented by
Calvert (op cit). It shows control device aerodynamic cut diameter
graphed against power per unit flow rate Chp/1000 acfm). Also
shown is the equivalent air pressure drop if all power went into
moving the volume of air through a flow resistance. The lines
shown are theoretical and not experimental. Lines la and Ib are
for sieve plate-type scrubbers with froth density, F=0.4, and hole
diameters, d,=0.5 and 0.3 cm, respectively. Line 3 is for impinge-
ment plates and line 4 is for a packed column from 1 to 3 M high
and packing of nominal 2.5 cm diameter. Line 2a for venturi scrub-
bers (f=0.25) represents the performance of venturi scrubbers in
collecting hydrophobic particles while line 2b (f=0.5) represents
the same class of scrubbers in collecting hydrophyllic particles.
From the gravel bed operating conditions and the measured perform-
ance data as previously given, the gravel bed produced the aero-
dynamic cut diameters shown in Figure 9 when operating at respective
power levels. The estimated cut diameters with the effects of the
puffs removed are also shown in Figure 9.
The difference between the August test results and the November
results is attributed primarily to the more frequent backwashing
resulting in less pluggage of the beds. This results in lower
operating pressure drops for the same gas flows, At the same time,
30
-------
5.0
4.0
3.0
2.0
1.5
PRESSURE DROP , i nches
6 7 8 9 10 15
*
20
30
40
I
50 60
80
I
100
o
a.
TJ
a:
UJ
o
o
o
cr
UJ
1.0
0.8
0.6
0.5
o
I °-4
<
0.3
0.2
O.I
I I I
Backflush
Time 12 min.
7 Active beds
O-18/27
D 8/28
A} 8/28
V\8/29
OJ8/29
0] H/4
OJ n/5
Backflush time
6 min.
7 Active beds
Filled symbols represent Gravel Bed
cut diameter versus system pressure drop; open symbols
represent gravel bed cut diameter versus typical bed
pressure drops; half-filled symbols represent gravel bed
cut diameters with the effects of the puffs removed
versus bed pressure drops.
0.25 0.5
i I I i I I
0.8 1.0 2.0
POWER , hp/IOOO actm
3.0
5.0
8.0 10
_L
_L
I l I l i
5 6 7 8 9 10
20 30
PRESSURE DROP , cm
50
70 90 KX)
200
300
Figure 9.
Representative cut diameters as a function of pressure
drop for several scrubber types, after Calvert (1974),
J. APCA 24:929.
-------
as a result of the better cleaning of the beds, the residual
dust concentration in the beds is reduced which in turn reduces
the loading surge to the active beds when a bed is backwashed.
The lower bed loadings also reduce the available dust which can
be carried directly out to the stack when a bed is put back on
line. It is also possible that with the beds in the highly plugged
condition that existed during the August tests they may have be-
come partially fluidized during the backwash cycle. This would
enhance the deagglomeration process and perhaps result in the
production of particulate from coarser particles by mechanical
actions.
32
-------
APPENDIX A
CASCADE IMPACTOR DATA BY RUN
33
-------
Table A-l
Inlet Mass Loading By Size Interval From Andersen Impactor Data
Mass Loading In Indicated Size Interval, mg/DNM3
Date
Start
8/27
8/27
8/27
8/27
8/28
8/28
8/28
8/29
8/29
1320
1620
1730
1800
0950
1105
1440
1015
1400
11/4
11/4
11/4
11/4
11/5
11/5
11/5
11/5
1100
1130
1430
1435
0935
0930
1415
1415
Duration
(Minutes)
20
20
20
20
45
120
120
120
120
65
60
135
120
120
120
120
120
Dia., um
Dia., ym
p=1.0 >4.70
3459
3411
3131
581 l
2737
4252
2098
2711
2131
2896
3569
2223
' 2474
' 6612
1941
1791
2384
2453
\ 4607
\4232
^-- •
2840
3650
2031
p=2.7 >2.8
3.1-4.7
28.6
20.9
14.5
25.8
9.2
11.1
9.3
12.2
14.4
16.22
20.69
11.74
18.8
19.3
14.9
12.4
9.11
14.6
17.5
19.7
15.79
18.31
13.27
1.8-2.8
1.4-3.1
20.7
13.8
11.4
14.8
5.20
7.82
6.44
7.44
9.34
10.77
13.85
7.69
9.5
12.2
10.7
8.53
14.5
7.51
7.97
10.6
10.19
11.72
8.61
0.83-1.8
.87-1.4
4.84
3.59
2.43
3.5
1.11
2.66
3.26
2.31
2.42
2.90
3.55
2.25
2.4
0.51
3.7
2.53
1.12
1.32
3.31
4.02
— •
2.36
3.23
1.50
.48-0.83
.63-. 87
2.42
1.00
1.42
1.39
0.48
0.52
2.64
2.04
0.42
1.37
1.90
0.84
0.12
1.03
0.53
0.67
0.23
0.03
4.37
3.42
- .
1.30
2.41
0.19
.34-0.48
<.63*^y
\
13.5
4.99
2.23
2.52
4.65
3.84
13.5
6.89
1.43
^/
5.94
9.70
3.09
4.60
1.05
8.14
<0.34
*Filter catches - May be
dominated by oversize
particles.
'Nozzles turned downstream
to avoid overloading upper
stages.
Average 8/27-8/29
90% Upper Confidence Limit
90% Lower Confidence Limit
2Nozzle pointed wrong
direction.
Average 11/4-11/5
90% Upper Confidence Limit
90% Lower Confidence Limit
-------
Table A-2
Start
Date Time Duration
CO
8/25
8/25
8/26
8/26
8/261
8/261
8/27
8/27
8/27
8/27
8/28
8/28
8/28
8/28
8/29
8/29
8/29
8/29
1440
1440
1119
1124
1515
1515
1100
1150
1515
1515
1045
1045
1415
1415
1000
1000
1400
1400
120
120
84
84
120
120
120
120
120
120
120
120
120
120
120
120
120
120
Outlet Mass Loadings By Size Interval For August Test Series
Mass Loading* In Indicated Size Interval, mg/DNM3
Dilution
Correct-
ion Pres.
Factor Drop
1.32
1.32
1.22
1.22
1.34
1.34
1.34
1.34
1.42
1.42
1.31
1.31
1.47
1.47
1.27
1.27
1.41
1.41
16.5
16.5
17.0
17.0
15.8
15.8
15.8
15.8
12.9
12.9
17.3
17.3
12.8
12.8
17.6
17.6
11.6
11.6
Dia. i__
p=1.0 >14.3 10.1-14;3 6.2-10.1
2.57
0.45
2.19
3.06
2.86
3.42
3.14
3.93
2.89
2.48
2.20
2.25
2.60
2.26
3.11
1.76
1.17
2.56
3.24
3.73
2.76
4.4-6.2 2.9-4.4 1.35-2.9
2.89
2.79
2.39
3.81
9.09
8.87
12.0
7.98
7.40
11.4
7.16
7.74
7.72
5.69
10.2
5.79
2.69
6.92
8.85
10.71
6.99
2.77
2.74
1.57
2.24
2.34
2.54
3.31
5.30
2.83
2.26
2.22
1.97
2.14
2.96
2.13
1.60
1.51
1.68
3.31
3.87
2.74
2.90
0.90
3.55
4.47
4.75
3.90
4.37
5.21
4.02
3.12
3.17
4.35
3.25
3.68
4.14
3.05
15
3.35
4.58
5.23
3.93
4.83
1.76
5.97
6.45
64
21
80
31
56
11
23
40
26
88
77
27
55
5.19
7.15
8.13
6.17
4.41
3.07
4.21
5.46
6.09
6.14
7.16
7.50
7.43
6.99
5.17
5.35
5.57
6.67
5.13
5.21
3.93
6.10
7.55
8.43
6.67
Dia.
p=2.7
>8.6 6.1-8.6 3.7-6.1
2.6-3.7 1.7-2.6 .78-1.7
.82-1.35
2.13
1.56
3.36
2.33
2.38
3.11
3.48
4.14
4.36
4.01
2.71
2.58
3.00
2.70
2.53
2.57
2.30
3.07 _
3.77
4.34
3.19
.46-.78
.58-.82 <.58*
0.96
0.15
0.94
0.71
0.97
1.52
1.54
1.41
1.77
1.47
1.15
0.88
1.92
1.53
1.70
1.23
1.28
0.93
1.49
1.32
0.21
0.33
10.2
1.10
0.93
1.43
1.54
0.90
1.61
1.79
5.73
4.67
3.98
3.82
1.27
1.03
1.66 2.74**
1.95 3.76***
1.37 1.73****
.32-.46 <-32
*Because of the dilution that results from the addition of the backwash air, the results from the outlet impactors must be
adjusted to compensate for the difference in inlet and outlet gas flows before comparisons among the various operating
conditions can be made and before fractional collection efficiencies can be calculated. The correction factor by which
the measured outlet concentrations must be multiplied in order to effect this adjustment are given for each impactor run
in the table.
**Average after correcting for dilution.
***Upper 90% Confidence Limit.
****Lower 90% Confidence Limit.
-------
Table A-3
Outlet Mass Loadings By Size Interval For November Test Series
Mass Loading* In Indicated Size Interval, mg/DNM3
Date Start Duration
Dia. , vim
p=1.0 >13.9
9.8-13.9 6.0-9.8 4.2-6.0 2.8-4.2 1.3-2.8 .79-1.3 .57-.79 <.57*
U)
11/3
11/3
11/4
11/4
11/5
11/5
1545
1545
1130
1130
0945
0945
120
120
240
240
240
240
1.35
1.35
1.39
1.39
1.32
1.32
NA
NA
9.6
9.6
14.0
14.0
0.66
18. 41
0.23
0.00
0.99
1.35
.86
1.55
.17
Dia., urn
p=2.7 >8.4
0.43
0.71
0.13
0.14
0.49
0.45
.53
.77
.28
5.9-8.4
0.40
0.64
0.19
0.16
0.51
0.38
0.51
0.71
0.31
3.6-5.9
0.43
0.61
0.18
0.34
0.61
0.56
.61
.80
.43
2.6-3.6
8.012
1.08
0.64
0.92
1.16
0.20
1.08
1.59
.58
1.7-2.6
9.452
3.80
2.63
3.36
3.78
0.65
3.60
5.19
2.00
.76-1.7
2.56
2.21
2.58
2.22
2.51
1.16
2.99
3.61
2.37
.44-. 76
0.96
0.66
1.02
0.68
0.86
1.03
1.18
1.36
0.99
.31-. 44
1.15
0.63
0.57
0. 38
0.97
1.12
1.18**
1.51***
.85****
<.31
* (Same as for Table A-2) .
**Average after correcting for dilution.
***Upper 89% Confidence Level.
****Lower 90% Confidence Level.
'Nozzle scrapped port on entry
2Omitted from average
-------
Table A-4
Inlet Mass Loading By Size Interval From Brink Impactor Data
Mass Loading In Indicated Size Interval, mg/DNM3
Date Time Duration Dia., p=l >17.8 12.6-17.8 7.2-12.6 4.3-7.2 3.0-4.3 1.6-3.0 1.14-1.6
U)
8/26
8/26
8/27
8/27
8/28
8/28
8/29
8/29
1010
1700
1045
1605
0945
1310
1000
1400
20
40
180
180
180
180
180
180
Dia., p=2.7
*Average
**Upper 90% Confidence Limit
***Lower 90% Confidence Limit
2020
2160
963
824
1720
597
3640
1150
1634
2301
967
36.6
38.9
17.7
65.9
50.8
48.2
49.5
116
53.0
72.5
33.4
28.1
15.2
7.7
22.4
17.8
12.3
84.3
19.2
25.9
42.3
9.4
22.1
17.2
8.2
29.3
11.4
14.7
22.5
15.4
17.6
22.2
13.0
13.6
6.1
7.0
9.4
4.3
3.3
6.9
6.1
7.09
9.24
4.93
14.5
3.0
2.6
2.6
1.9
1.9
2.5
2.6
4.06
6.92
1.20
7.66
3.5
1.2
1.8
1.1
1.6
2.0
.6
2.43
3.97
0.90
7.6-10.8 4.3-7.6
2.6-4.3
1.8-2.6
.92-l.i
.65-92
0.67-1.14
.36-.65
<.67
20.4
8.6
1.0
2.2
0.8
1.6
2.0
.6
4.65
9.28
0.02
5.96
12.6
2.83
4.00
2.8
6.4
6.9
11.8
6.66*
9.19**
4.13***
<.36
-------
Date
Weight Gain/
Stage
Table A-5
Andersen Impactor Blank Weight Gains
Weight Gain of Substrate, mg.
8/25 8/26 8/26 8/27 8/28 8/29
11/4 11/5
1
2
3
4
5
6
7
8
F
Average
Std. Dev.
-.02
-.40
.04
-.15
.27
-.18
-.41
-.58
-.08
-.25
.21
.20
.18
.22
.25
.20
.19
.23
.17
.14 t
.21
.03
.32
.34
.21
.24
.28
.18
.19
.15
.45
.24
.07
.16
.25
.18
.32
.21
.18
.12
.21
.13
.20
.06
.04
.03
.01
.07
.09
.00
.00
.10
.20
.04
.04
.08
.10
.09
.01
.03
.08
.11
.06
.25
.07
.03
.33
.02
.07
-.11
.17
.03
.10
.26
.40
.11
.14
.01
-.27
.00
-.06
-.01
-.23
-.09
-.07
.29
-.09
.11
38
-------
APPENDIX B
PLANT PRODUCTION DATA
Date 8/25 8/26 8/27 2/28 8/29 11/4 11/5
Production
Tons/Day 742 533* 961 1031 1064 1063 995
*Single kiln
39
-------
Appendix C
Method of Operation of Gravel Bed Filters and Design
Specifications of the System Tested
The gravel bed filter system is a system comprised of particulate
filter beds used in conjunction with mechanical or other particulate
separators in different configurations and combinations for the pur-
pose of collecting dry particulates from dust-laden gas streams.
The gravel bed filter system process consists basically of pre-
separating large coarse particles from the dust-laden gas stream
by means of mechanical collectors, although other means can be
employed. The mechanical collector is usually the least expensive
method for achieving this precleaning operation. The mechanical
collectors can be in series with the filter beds, an integral part
of each module, or they can be separate pieces of equipment connect-
ed in different arrangements by ductwork.
The multichamber cyclone-gravel bed filter combination consists
of several units of equal size, connected by common raw gas and
clean gas ducts. The operation of the system is illustrated in
Figure Cl. The raw gas is led into the filter via a
parallel-connected raw gas plenum (1). An immediate separation
of very coarse materials, by settling, takes place in this plenum
chamber. From there, the gas enters the cyclone type preliminary
separator (2), where the entrained medium and coarse dust is separa-
ted and removed via the discharge airlock (3) at the outlet. The
precleaned raw gas now rises from the cyclone through the vortex
tube (4) and enters the filter chambers (5). It passes from the
top of the horizontal filter beds to the bottom, so that the re-
maining fine dust is deposited on the quartz grains and in the
interstices of this bed (6). The cleaned gas now flows through
the clean gas collection chamber (8), and passes via the 3-way
40
-------
OPERATING PHASE
BACKFLUSH PHASE
14
12
BACKFLUSH
AIR
1. INLET CHAMBER
2. PRIMARY COLLECTOR (CYCLONE)
3. DOUBLE TIPPING GATE (DUST DISCHARGE)
4. VORTEX TUBE
5. FILTER CHAMBER
6. GRAVEL BED
7. SCREEN SUPPORT FOR BED
8. CLEAN GAS COLLECTION CHAMBER
9. EXHAUST PORT
10. BACKWASH CONTROL VALVE
11. BACKWASH AIR INLET
12. VALVE CYLINDER
13. STIRRING RAKE
14. STIRRING RAKE MOTOR/REDUCERS
Figure Cl. Operation of gravel bed filter.
41
-------
valve (10) into the clean gas duct (9). Cleaning of a filter unit
can be initiated by means of a preset sequencer or by a pressure
differential across the filter bed. During the cleaning cycle,
the unit is isolated from the gas stream by the 3-way valve.
Then, backwash air is admitted to the filter chamber in a reverse
flow direction. It is either forced in by using a backwash air
blower, or sucked in by negative pressure (11). The backwash air
loosens the filter bed. During the cleaning process the rake-
shaped double arm stirring device (13) is rotated by the geared
motor (14). Thus the dust is thoroughly removed from the gravel
and entrained by the backwash air, and the fluidized filter gravel
is again settled after the cleaning operation. The large agglom-
erated dust particles are carried by the backwash air via the
vortex tube (4) into the preseparator (2), where the velocity is
reduced and the gas stream deflected so that a large percentage
of the dust is settled out. The backwash air, containing the re-
maining dust, mixes with the dust-laden air in the raw gas duct
and is then subjected to cleaning in the remaining units of the
filter. The filter, as has been described, consists of the filter
top section, containing the filter bed, and the upstream cyclone
located underneath. The flow rate of the filter bed is, of course,
important to the dimensioning of the cyclone, which generally has
a smaller diameter than the filter section proper. Attainment of
optimum flow rate for both filter bed and cyclone would result in
a very large filter bed relative to the cyclone.
To efficiently use the space required for the overall unit,
it is desirable to increase the bed area without increasing the
bed diameter to a value much greater than that of the cyclone.
As can be seen from Figure C2, this objective was obtained by
placing two parallel-connected filter beds one on top of the other.
The cyclone-gravel bed filter combination is automatically
controlled via a program transmitter. Following the signals of
this continuously operating program transmitter, the clean gas
port for a chamber to be cleaned is shut and the backwash air port
42
-------
opened. The drive of the raking device is started after a time
lag. The adjustment of the duration of the cleaning process and
of the intervals presents no difficulty. When the cleaning pro-
cess is over, the raking device is stopped first. Then, after
another time lag, the clean gas port is opened; and, at the
same time, the backwash air flow is stopped so that filtering
can be resumed in this chamber.
The design of the double bed filter element and its stirring
(leveling) device and drive is also simple.
The rake-shaped double arm of the raking device is firmly
connected to the vertically arranged geared motor via a rigid
coupling. This coupling is provided with an asbestos washer
serving as a thermal insulation. The bearing of the geared motor
is reinforced so that the forces, which result when the rake is
rotating, can be absorbed without any further intermediate bear-
ing. The filter bed is spread on a woven supporting screen made
of spring or stainless steel. Thanks to the special method applied
for dust removal, this supporting screen is not subject to signifi-
cant wear.
As has already been mentioned, the gases enter the cyclone at
relatively low velocities, so that generally no significant wear
occurs in the cylindrical section. The outlet cone of the presepara-
tor is protected against wear by means of a ceramic lining. A
weighted double-acting valve assembly (tipping valves) is arranged
beneath the outlet cones, where it serves as a dust discharge device
and as an air seal. In many cases, this valve assembly is mechan-
ically actuated. All moving parts required for the cleaning process
will be in operation during that period only. Normally, the dura-
tion of the cleaning phase is about three to five minutes.
Once the large dense particulates have been removed in the
first pass of the raw gas through the precleaning stages of the
system, the gas containing the residual fines passes through the
43
-------
packed solids filter beds. In this passage, the impingement of
the small particulates on the discrete filter media causes an
agglomeration to take place in which the fine particles are
joined with other fine particles to form larger agglomerates.
Periodically that particular filter is taken out of the system
and back-washed by reversing a gas flow upwardly through the
bed while it is being mechanically stirred by a rake. The bed
is both mechanically and pneumatically disturbed and this re-
sults in a semifluidized state in which the agglomerates (which
are lighter than the filter media and have a greater surface
area/weight ratio) are blown backwards from the bed and are
then collected in the common precleaning cyclones or recollected
in the remaining gravel bed.
44
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/2-76-164
2.
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
Evaluation of Rexnord Gravel Bed Filter
5. REPORT DATE
June 1976
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
Joseph D. McCain
8. PERFORMING ORGANIZATION REPORT NO.
SORI-EAS-76-299
9. PERFORMING OROANIZATION NAME AND ADDRESS
Southern Research Institute
2000 Ninth Avenue South
Birmingham, Alabama 35202
10. PROGRAM ELEMENT NO.
1AB012; ROAP 21ADL-004
11. CONTRACT/GRANT NO.
68-02-1480
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
Final; Through 1/31/76
14. SPONSORING AGENCY CODE
EPA-ORD
is.SUPPLEMENTARY NOTESEPA-650/2-74-129a was earlier report in this series. IERL-RTP
project officer for this report is D.L.Harmon, Mail Drop 61, 919/549-8411, Ext 2925.
16. ABSTRACT
repOrt gives results of fractional and overall mass efficiency tests of a
full-scale Rexnord gravel bed filter system used to control particulate emissions from
a Portland cement plant clinker cooler. Total flue gas particulate mass concentra-
tions and emission rates were determined at the inlet and outlet of the gravel bed sys-
tem by conventional (Method 5) techniques. Inlet and outlet emission rates as a func-
tion of size were determined on a mass basis using cascade impactors for sizes from
about 0. 5 to 5 micrometers , and on a number basis for sizes smaller than about 1
micrometer using optical and diffusional methods. The report briefly describes the
Portland cement process , the Rexnord gravel bed filter system , the measurement
methods , inlet and outlet particle size distribution data, and fractional efficiencies.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COSATI Held/Group
Air Pollution
Filters
Dust Control
Portland Cement
Industrial Processes
Clinker
Flue Gases
Measurement
Tests
Evaluation
Air Pollution Control
Stationary Sources
Rexnord Filter
Gravel Bed Filter
Particulate
Mass Efficiency
13B
11B,13C
13H
21B
14B
18. DISTRIBUTION STATEMENT
Unlimited
19. SECURITY CLASS (ThisReport)
Unclassified
21. NO. OF PAGES
50
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
45
------- |