U.S. Environmental Protection Agency Industrial Environmental Research EPA~600/7'*7T~!lfl5fl
Office of Research and Development Laboratory
Research Triangle Park, North Carolina 27711 NCWfWIHWf l9f7
EPA FABRIC FILTRATION STUDIES:
5. Bag Cleaning Technology
(High Temperature Tests)
Interagency
Energy-Environment
Research and Development
Program Report
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into seven series. These seven 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 seven series
are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
This report has been assigned to the INTERAGENCY ENERGY-ENVIRONMENT
RESEARCH AND DEVELOPMENT series. Reports in this series result from the effort
funded under the 17-agency Federal Energy/Environment Research and Development
Program. These studies relate to EPA's mission to protect the public health and welfare
from adverse effects of pollutants associated with energy systems. The goal of the
Program is to assure the rapid development of domestic energy supplies in an environ-
mentally-compatible manner by providing the necessary environmental data and
control technology. Investigations include analyses of the transport of energy-related
pollutants and their health and ecological effects; assessments of, and development
of, control technologies for energy systems; and integrated assessments of a wide
range of energy-related environmental issues.
REVIEW NOTICE
This report has been reviewed by the participating Federal Agencies, and approved
for publication. Approval does not signify that the contents necessarily reflect the
views and policies of the Government, 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 Information
Service, Springfield, Virginia 22161.
-------
EPA-600/7-77-095b
November 1977
EPA FABRIC FILTRATION STUDIES:
5. Bag Cleaning Technology
(High Temperature Tests)
by
B.E. Daniel, R.P. Donovan (RTI),
and J.H. Turner
Environmental Protection Agency
Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, North Carolina 27711
Program Element No. EHE624
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Washington, D.C. 20460
-------
PREFACE
This report is the fifth in a series of-reports, entitled EPA
Fabric Filtration Studies, which summarize the results of EPA laboratory
testing of new baghouse fabric materials and present the conclusions of
specialized research studies in fabric filtration. These tests have
been carried out over the past 5 years by the Industrial Environmental
Research Laboratory, Research Triangle Park, N. C., and previously by
predecessor agencies. The purpose of these investigations was to
evaluate the potential of various new fabrics as baghouse filters and to
obtain data for use by the fabric filtration community. The testing
consisted of simulating baghouse operation in a carefully controlled
laboratory setting that allowed measurement and comparison of bag per-
formance and endurance.
The work reported in this paper was based on a laboratory simulation
of high temperature baghouse operation, the only work in this series to
use this apparatus. Cement dust was the only dust used here, whereas
flyash was previously the only test dust used. Inlet dust loading was
not measured and was not precisely controlled, since no performance
parameter was monitored other than pressure drop across the bag. The
primary purpose of the high temperature facility was to detect tempera-
ture induced bag failure or phenomena.
As in all previous reports, British units are used primarily.
Their widespread use in the existing literature makes them the preferred
choice in spite of EPA's policy to use metric units. Use of metric
units would seriously inconvenience the majority of the intended reading
audience. For those readers more familiar with the metric system a
conversion table for changing the British units used in the report to
their metric equivalents appears in Appendix B.
ii
-------
The projected EPA Fabric Filtration Studies series consists of the
following reports:
1) "Performance of Non-Woven Nylon Filter Bags," 0. H. Turner,
EPA-600/2-76-168a (NTIS No. PB 266271/AS), December 1976.
2) "Performance of Non-Woven Polyester Filter Bags," G. H. Ramsey
et al., EPA-600/2-76-168b (NTIS No. PB 258025/AS), June 1976
3) "Performance of Filter Bags made from Expanded PTFE Laminate,"
R. P. Donovan et al., EPA-600/2-76-168c (NTIS No. PB 263132/AS),
December 1976.
4) "Bag Aging Effects," R. P- Donovan et al., EPA-600/7-77-095a,
(NTIS No. PB 271966/AS), August 1977.
5) "Bag Cleaning Technology (High Temperature Tests)," (this
report).
6) "Analysis of Collection Efficiency by Particle Size."
iii
-------
ABSTRACT
The influence of high temperature operation (operation in an air
flow whose temperature has been adjusted to the maximum continuous
operating temperature recommended by the fabric filter manufacturer)
on the selection of shake-cleaning parameters is the subject of this
work. Two bags each of cotton and Dacron were operated in a laboratory
baghouse using heated air passed through cement dust as the source of
dirty air. The bags cleaned at high "g" forces (-5 g's) showed more
deterioration in strength properties than those cleaned at 1.9 g's. The
observations generally confirm the Dennis/Wilder analysis of mechanical
cleaning and suggest that temperature is not a first order variable in
the analysis of mechanical shake-cleaning. The cursory tests conducted
here do not conclusively rule out temperature as an important parameter;
they merely report that in this limited investigation it was not.
iv
-------
TABLE OF CONTENTS
Rage
Preface ii
Abstract iv
List of Figures vi
List of Tables vi
List of Abbreviations and Symbols vii
Acknowledgment viii
Secti on
1 Introduction 1
2 Conclusions 2
3 Background 3
4 Experimental Procedures 12
5 Results 15
References 28
Appendix A - Test Procedures for Filter Bag Properties (Ref. 3) . 29
Appendix B - Conversion Factors 32
-------
LIST OF FIGURES
Number Page
1 Residual fabric loading versus average bag
acceleration (from Dennis and Wilder [Ref. 1]) .... 6
2 Cloth loading and filter drag characteristics for
typical shaking regimes—composite curve (from
Dennis and Wilder [Ref. 1]) 7
3 Residual fabric loadings for various fabrics with
flyash aerosol (8 cps, 1 in. amplitude shaking)
(from Dennis and Wilder [Ref. 1]) 9
4 Tensile properties for a 10-foot by 6-inch sateen
bag (from Dennis and Wilder [Ref. 1]) 10
5 Laboratory baghouse for high temperature tests .... 13
6 Pressure drops of cotton bags filtering cement
dust at 180°F 17
7 Pressure drops of Dacron bags filtering cement
dust at 275°F 18
8 Replot of Figure 7 using total operating time as
the abscissa 20
LIST OF TABLES
Number
1 Summary of Runs 16
2 Bag Weights 23
3 Fabric Properties of Cotton Bags [Ref. 3] 25
4 Fabric Properties of Dacron Bags [Ref. 3] 26
VI
-------
LIST OF ABBREVIATIONS AND SYMBOLS
A = filtration area of fabric (sq ft)
CQ = mass outlet concentration (grains/1000 cu ft)
E = mass collection efficiency (percent)
KU = true value of specific cake resistance (in. H20/fpm)/(lb/sq ft)
l<2 = measured value of specific cake resistance (in. H20/fpm)/(lb/sq ft)
APE = pressure drop across bag at time zero of filtration cycle (in. H20)
APT = pressure drop across bag at end of filtration cycle (in. H20)
SE = effective drag (in. H20/fpm)
ST = terminal drag (in. H20/fpm)
A/C = air/cloth ratio (fpm)
2
Wn = cloth loading after cleaning (Ib/ft )
K
2
W = cloth loading just prior to cleaning (Ib/ft )
vn
-------
ACKNOWLEDGEMENT
The authors acknowledge, with pleasure, the comments and suggestions
made by Richard Dennis of 6CA Technology Division to improve this report.
viii
-------
SECTION 1
INTRODUCTION
Optimum parameters for shake-cleaning fabric filters have pre-
viously been studied, both theoretically and experimentally, by Dennis
and Wilder [Ref. 1]. They showed that the residual dust remaining on a
fabric filter after a shake-cleaning correlated with the reciprocal of
the square root of the average bag acceleration during the shake-cleaning.
The exact relationship varies with varying dust/fabric systems and also
depends on other variables such as humidity, electrostatic charge and
bag age.
The amount of residual dust, however, seems not to be related to
the initial dust loading on the bag prior to the shake-cleaning.
Dennis and Wilder supported their theoretical models with measure-
ments made while filtering flyash at room temperature. The fabrics they
used included cotton sateen and Dacron. In the cursory work to be
reported here the cleaning cycles recommended by Dennis and Wilder were
repeated on cotton and Dacron in a test facility that allowed the fabric
filters to be operated at their maximum recommended temperatures (180°F
for the cotton; 275°F for the Dacron). The purpose of-the work was to
determine if the analysis of the shake-cleaning cycle previously confirmed
for room temperature operation remained valid at the temperature maximums
of each of the fabrics.
-------
SECTION 2
CONCLUSIONS
The high temperature measurements reported in this paper are
qualitatively consistent with the shake-cleaning analyses and room
temperature measurements previously reported by Dennis and Wilder (D&W).
Stronger conclusions in support of the D&W work are not justified be-
cause the dust used in the experiments reported here differed from that
used by Dennis and Wilder (cement dust here vs. flyash) and the instru-
mentation was not as complete: bag tension, an important D&W parameter,
was controlled only crudely; and the only performance parameter monitored
here was pressure drop across the bag, a parameter treated only sketchily
by D&W. Within these limitations, however, the importance of bag
_a£celeration during cleaning was demonstrated and, as in the D&W model,
shown to be a variable of first order importance in the shake-cleaning
of fabric filters. No new, temperature-dependent phenomenon was identified
to modify or upset the D&W analysis. Both bag performance, as measured
by the pressure drop, and bag life, as measured by the physical properties
of the fabric, depended more on the shake-cleaning action than on time
at temperature. As in the D&W work, the dust loading of the cotton bags
greatly exceeded that of the Dacron bags. Measurements of the absolute
values of various properties of the used fabric, especially abrasion
resistance, suggest that the Dacron bags would last longer. For both
the cotton and the Dacron bags, shake cleaning at high "g" forces reduces
bag strength (and presumably ultimate bag life) more rapidly than does
low "g" cleaning. No direct measurements of bag life were made, however.
-------
SECTION 3
BACKGROUND
The analysis of bag shake-cleaning carried out for EPA by Dennis
and Wilder [Ref. 1] develops a theory of bag motion in terms of shake-
frequency, stroke length and various bag properties including dimensions,
elastic modulus and mounting tension. The inertial forces transmitted
to the bag by the shaking force applied to one end of the bag must
exceed the forces holding the dust at any specific region in order to
effectively remove the dust. While the cleaning efficiency of a specific
shake cycle depends upon the magnitude of the dust trapping forces as
well as the motion of the bag, the Dennis/Wilder analysis concentrates
primarily on the latter. The assumption is that tensile stress between
the dust cake and the fabric is the only effective removal mechanism—
the inertia! forces perpendicular to the fabric surface during accelera-
tion and deceleration separate the dust from the fabric, although shear
force may assist in breaking adhesive bonds between dust and fabric.
In analyzing the bag motion Dennis and Wilder treat the bag like a
vibrating string, oscillating in dampened harmonic motion. A displace-
ment introduced by the shaker mechanism propagates along the bag to the
end where it is reflected. At certain frequencies the reflected wave
reinforces the applied displacements—these frequencies"constitute
resonant frequencies.
At all frequencies some dampening occurs and a minimal requirement
for cleaning is that the applied shaking energy be sufficient to intro-
duce a traveling wave that is not completely dampened before reaching
the end of the bag. Otherwise no shake-cleaning would occur at the
motionless end remote from the shaker mechanism.
-------
Bag tension is an important variable in determining wave propa-
gation and dampening. It varies along the length of the vertically
suspended bag because of gravity, increases with time of filtration
because of dust loading and varies with applied forces and bag motion
during the shake-cleaning. Dennis and Wilder derived the following
expression for relating the average bag amplitude, 7, to bag tensions
during shake-cleaning:
V •
VT. - Ti,n> "'
where Y = the average amplitude of bag displacement (a)
f = the shaker frequency (t )
o
M = the elastic modulus of the bag filter (m/t )
L = the bag length (£) (between clamps)
p = the mass per unit length of the bag (m/£)
T_ = the dynamic bag tension averaged at its midpoint
in o
Ti m = tne initial, average midpoint tension (static)
I ,111 n
The value of Y calculated from Equation 1 underestimated the
photographically measured* displacement amplitudes by about 30 percent
[Ref. 1], Equation 1 predicts that the average amplitude decreases with
*The procedure was to measure a maximum amplitude at a node and a minimum
amplitude at an anti-node and average the two amplitudes to obtain an
average amplitude.
-------
increasing shake frequency. This relationship is not simple to confirm,
however, because the tensions also vary with shake frequency, peaking in
the vicinity of a resonance. Away from the resonances the bag tensions
generally increase with increasing frequency partially cancelling the
frequency-dependence of amplitude explicitly contained in Equation 1.
Once knowing the displacement at any point on the bag, the maximum
acceleration, am, at that point is [Ref. 1]:
am = 4 iT2f2Y (2)
All points on the bag are assumed to move at the same frequency as
the shaker arm.
Dennis and Wilder [Ref. 1] further showed that the residual dust
loading of the fabric filter varied inversely as the square root of the
average bag acceleration (Figure 1). The residual dust loading of the
fabric is the dust remaining on the fabric after a specific shake-
clean cycle as characterized by an average acceleration—the average of
the maximum acceleration at all points of the bag. The residual dust
loading is independent of the initial dust loading prior to the shake
cleaning. To the first order the residual dust loading depends only on
the average bag amplitude, (Y), and frequency of the shaker, (f),
assuming uniform bag tension at rest.
Figure 2 is a composite curve from Dennis and Wilder that summarizes
this behavior. At the end1 of the filtration cycle the terminal drag is
ST and the cloth loading, Wy. The values of drag and cloth loading
following a shake-cleaning are plotted for four different sets of shake-
cleaning parameters (A to D). Although the inverse square root relation-
ship between average acceleration and residual dust loading is not
strictly followed, the residual dust loading clearly decreases with in-
creasing bag acceleration.
-------
1000,.
(O
s_
500,
X-
-O
(O
200
Unnapped Sateen
Weave Cotton 3
flyash at 3.5 grains/ft
Filter Velocity, 3 ft/min
360 Shakes
100
I
Figure 1
2 5
Average Bag Acceleration, g's
Residual fabric loading versus average bag
acceleration (from Dennis and Wilder [Ref.l]).
10
-------
o>
4->
fl
CD
O
03
+->
l/l
01
cc
01
•fj
5.0
4.0
3.0
A
B
C
D
2.0
1.0
10 ft x 6 in. Cotton Sateen
Shajdin£ Regime (350 Shakes)
- AMPL. 1 in., FREQ. 10.1 cps
- AMPL.3/4 in., FREQ. 10.7 cps
- AMPL. 2 in., FREQ. 4.0 cps
- AMPL. 1 in., FREQ. 4.9 cps
0.02
0.04
0.06
OTTJ8"
Total Cloth Loading, lbs/ft2
OTlT
1.6
1.2
0.8
0.4
Figure 2.
Cloth loading and filter drag characteristics for typical shaking regimes--
composite curve (from Dennis and Uilder [Ref. 1]).
OJ
4->
«3
en
to
Q
L-
0)
-------
Total number of shakes is also a factor. Dennis and Wilder specify
a minimum of 100 shakes for the observed relationships to be reproducible.
By 200 shakes the residual dust has attained 80 to 95 percent of its 360
shake value and the optimum number becomes a compromise between the
diminishing contribution to the cleaning and the linear increase in
mechanical wear on the fabric.
Bag age also influences the observed relationship between the
residual dust loading and the average acceleration during the shake-
cleaning. The curves shown in Figure 3 compare residual dust as a
function of total number of shakes for various new fabric filters and
used bags of the same fabric (the "old", 0, designation in Figure 3).
For all fabrics the residual dust loading decreased with bag age, per-
haps because of "irreversible stretching in the [fabric] media" (Figure
4), and/or shedding of fibers that project across pores.
In summary, the general recommendations for shake-cleaning developed
by Dennis and Wilder include:
1) Shaker parameters (amplitude and frequency of shake) selected
so as to produce an average bag acceleration in the range 1.5
to 7 g's.
2) Total number of shakes between 200 and 400.
3) Control (and monitoring) of bag tension as a parameter in
achieving No.l; in particular, adequate tension to ensure
propagation of the oscillating motion along the entire length
of the bag.
Other variables, such as dust type, fabric type, and bag age,
influence the specific relationship between cleaning efficiency and
shake-cleaning technique. Hence, the optimum shaker parameters cannot
be specified a priori with complete confidence. Some trial and error
will be necessary. The purpose of the work reported here is to observe
the high temperature behavior of fabric filter bags, shake-cleaned in
accordance with the general recommendations listed above. High tem-
perature means the maximum temperature for continuous operation specified
8
-------
1000
500
LO
sz
03
S-
O)
CD
c
•r-
-o
(0
o
o
S-
-Q
03
CO
3
T3
•r—
V)
OJ
o:
200
100
D-
--- 0 -- Q-2
O-o^^
Curye_
1,2
3,4
5,6
Fabric
Napped Sateen Weave Cotton
Plain Weave Dacron
Crowfoot Dacron 7
N - New, <104 Shakes, 0 Old, 2 x 107
Shakes
1
50 100 200
Total Number of Shakes
500
Figure 3. Residual fabric loadings for various fabrics with flyash
aerosol (8 cps, 1 in. amplitude shaking) (from Dennis and
Wilder [Rof. 1]).
-------
0.6
c
o
O
LU
Cn
to
cn
0.4
0.2
New
New-16.5 1bs./1n.
Used- 45 Ibs./lfh
Used (2 x 10' shakes)
4 6
AppHed Weight Load, Ibs.
10
Figure 4.
Tensile properties for a 10-foot by 6-Inch sateen bag (from Dennis and
Wilder [Ref. ij).
-------
by the fabric manufacturer- The investigations carried out attempted
to determine if the general Dennis/Wilder recommendations apply also for
high temperature operations or whether new forces and interactions
dominate the problem.
11
-------
SECTION 4
EXPERIMENTAL PROCEDURES
The apparatus used to carry out the high temperature evaluation
consisted of a custom-assembled chamber sized to hold four bags (Figure
5). Cement dust, repeatedly entrained in the metered hot air flow
entering the teg, was used for all tests. The entrained dust was then
removed from the air flow by the fabric filter during the filtration
cycle and shake-cleaned into the dust pot at the bottom of the bag
during the cleaning cycle. During the next filtration cycle, the dust
became re-entrained once more as the heated air entered the bag through
various ports in the dust pot. The dust was thus continuously trans-
ferred from the dust pot to the fabric (the entrainment/filtration
portion of the cycle) and then from the fabric back to the dust pot
(the shake-cleaning portion of the cycle). The filtration period was
always 75 sees; the shake-cleaning, 35 sees. No time delay separated
these periods. Filtration ended and shake-cleaning began simultaneously;
conversely, the shake-cleaning ended and the air flow for the next
filtration period began at the same time.
Temperature of operation was controlled by passing the inlet air
through a furnace heater, preset to the desired operating temperature.
The actual temperature inside the baghouse was monitored by thermocouples
located at various positions in the clean air side of the baghouse. The
shaker arm was fabricated from hollow tubing which allowed pressure
measurements to be made across the bag; that is, access to the inside of
the bag for pressure measurements was through the shaker arm and the bag
mount at the top. Since the inside of the bag is the dirty side of the
air flow, the tubing became clogged with dust periodically.
12
-------
Inlet
Blower
f
Thermocouples
i
\
To Shaker Assembly
Ball Bushing
Support
Frame
Ball •
Bushing
Shaker^
Arm
CT,
fO
CQ
J_
O)
Bypass
•Rotameter
4—Flow
Control
Valve
Solenoid
Valves
Heater
Dust
Pot
Figure 5. Laboratory baghouse for high temperature tests.
-------
The variables of the shake-cleaning were controlled by a standard
motor/cam arrangement not shown in Figure 5.
No performance characteristics (efficiency, outlet concentration,
etc.) were measured other than pressure drop. The purpose of the test
was to detect any major departure from room temperature behavior that
high temperature operation would introduce.
All bags were 31.75 in. long and 5.5 in. in diamter, with a total
2
bag area of 3,81 ft . Unlike the Dennis and Wilder work iRef. 1] bag
tension was not monitored continously; rather, it was adjusted initially
by measuring the bag slack. After mounting the bag with zero slack, the
tension was tightened or loosened by a fixed length to achieve the de-
sired tension. This crude control of tension was deemed adequate for
the confirmation of qualitative bag behavior.
Total air flow through the bag was determined by a rotameter In the
feed line upstream of the furnace. This air flow was held constant
throughout any given run at some value between 7-5 and 8.4 cfm, yielding
an air/cloth ratio of about 2 fpm for all the tests reported here.
Inlet dust loading was not Treasured (nor were outlet concentration
or bag efficiency). The dust feed mechanism, relying totally on air
flow through settled dust, probabably produced non-uniformities in the
inlet loading, as discussed in Section 5.
14
-------
SECTION 5
RESULTS
Four runs, two with cotton bags and two with Dacron bags, were
carried out using cement dust as the test dust. The independent
variables of these runs are summarized in Table 1 along with the
calculated total number of shakes and the maximum bag acceleration
during the shake cycle. The total operating time was adjusted to
achieve over 3 million shakes during each run regardless of the shaker
rpm. The stroke is the total distance moved by the shaken end of the
bag and is therefore twice the amplitude of the sinsusoidal wave motion
of the i>ag—the amplitude, Y, used in Equation 2 to calculate the
maximum acceleration, was taken to be half the stroke. The operating
temperatures were the maximum recommended by the manufacturers for the
specific fabrics.
Unlike the Dennis and Wilder work iRef. 1], bag dust loading was
•not measured in situ. Hence, the Dennis/Wilder correlation between
residual dust loading and the inverse square of average bag acceleration
during shake-cleaning (Figure 1) could not be confirmed directly. What
was observed was the pressure drop across the bags at the time the
shake-cleaning cycling commenced. This variable (actually the drag,
AP/EA/C]) has been shown previously to correlate qualitatively with the
dust loadings, both residual and terminal, of a shake-cleaned bag
operating on a fixed time sequence [Ref. 2] (a fixed time sequence is
one in which the durations of the filtration period, the cleaning period
and all other intervals of the operating cycle are constant in time).
It was used in this work to investigate the role of acceleration during
cleaning upon the dust loading of the bags.
Figure 6 is a plot of pressure drop for the two cotton bags; Figure
7, for the Dacron bags. The ordinate is pressure droo rather than drag,
15
-------
TABLE 1. SUMMARY OF RUNS
CTl
Run
No.
37
38
39
40
Bag No.
6034-2
6034-3
6031-1
6031-2
6031-3
6031-4
6031-5
6031-6
6031-7
6034-4
6034-5
6034-6
Fabric
Cotton
Cotton
Dacron
Dacron
Dacron
Dacron
Dacron
Dacron
Dacron
Cotton
Cotton
Cotton
Operati ng
Time
(hours)
703
703
704
704
704
462
450
450
488
460
460
460
Temp
(°F)
180
180
275
275
275
275
275
275
275
180
180
180
Shake
(rpro)
240
240
240
240
240
390
390
390
390
370
370
370
Stroke
(in.)
2.36
2.36
2.36
2.36
2.36
2.36
2.36
2.36
2.36
2.36
2.36
2.36
Total
Shakes
(x lo6)
3.22
3.22
3.23
3.23
3.23
3.44
3.35
3.35
3.63
3.25
3.25
3.25
am (from Equation 2)
[x 106 in./min2 (g)]
2.68 (1.9)
2.68
2.68
2.68
2.68
7.08 (5.1)
7.08
7.08
7.08
6.38 (4.6)
6.38
6.38
-------
1ft-
81.
45*.
Q.
<
KEY
Dag No.
X 6034-21
-3' (Run flo. 37)'
1-2i
,f —240 mdtn shake
-31 fbi.H fin 171
a -51 —?70 rpm shake
1 (Run No. 40)
I
total number of shakes,xl
Figure 6. Pressure drops of cotton bags filtering cement dust at 180°F.
-------
KEY
7_
Baq Mo.
O 6031-1
X -2
a -3
-- — 240 rpm shake
(Run No. 38)
---390 rpm shake
(Run No. 39)
oo
X
Q
O
X
o
TT
O
Total number of shakes
Figure 7. Pressure drops of Dacron bags filtering cement dust at 275°F.
-------
since the gas flow was held constant for all these measurements. On a
few measurements the air/cloth ratio varied—because of obstructions or
other bag problems—but these were exceptions and were remedied
immediately.
The data plotted in Figures 6 through 8 represent averaged pressure
drops of 8 to 15 readings each. The curves, drawn by eye to fit the
data, assume a simple linear behavior within two or three sequential
time intervals and attempt to draw only first order distinctions between
the compared curves.
The immediate conclusion from the plots of Figures 6 and 7 is that
the runs carried out under higher "g" cleaning conditions operated at
lower pressure drop, corresponding to a bag of lower dust loading. This
conclusion is qualitatively consistent with the predictions of Dennis
and Wilder; a stronger supporting statement is not made because the dust
used here is different and the instrumentation was not as complete as
theirs.
In Figure 7 there is a hint of a decrease in pressure drop for the
tlata of run No.38, the Dacron run shake-cleaned at low "g". Thus turn-
down in "the curve suggests the onset of bag wearout £Ref. f\ after about
2.6 x 10 shakes. No such "wearout" suggestion is contained in the
curve for run No.39 for which the AP data do not reflect any fall off to
over 3.6 x 10 shakes. If total number of shakes is a valid measure of
bag life, then the two curves represent different behavior. If, however,
because of the elevated temperature of operation, operating time alone
is a better measure of bag life, the two curves compare as shown in
Figure 8.. In Figure 8 the abscissa has been changed to operating time
and the two curves may be consistent, since the operating time of the
390 rpni run is much less than the 240 rpm run—it simply may not have
had sufficient running time to reach the wearout period. If the wearout
mechanism is more temperature-dependent than shake-dependent, the display
in Figure 8 is the more realistic presentation. Figures 6 and 7 assume
that the number of mechanical shakes is the dominant variable by which
to measure bag life.
19
-------
KEY
rv>
o
/ m
x
E
2 24
2
4
I 4
3
4
4 I
0
X
E
i j 4 4
I
Bag No.
O 6031-1
X
m
i
2
3
4
#
-3)
240 rpm
shake (Run
No. 38)
-390 rpm
shake (Run
No.39)
I
100
200
500
COO
300 400
Operating time,hrs
Figure 8. Replot of Figure 7 using total operating time as the abscissa.
700
-------
The fact that the Figure 8 plot removes the minor inconsistency
from the Figure 7 plot is the only evidence found in these investigations
to suggest that time at temperature may be a significant variable. If
valid this dependence does not conflict with the Dennis/Wilder observations;
at most it adds another variable to consider in formulating a shake-
cleaning schedule.
The measurements of pressure drop reflect large scatter. One major
cause of variation in the measured pressure drops was the long, thin
line through which the inside bag pressure was detected. Uncontrolled
pressure drop along this line, because of obstruction and dust buildup,
caused measurement errors that were characterized by gradual drifts to
lower and lower values of measured pressure. When the line would sub-
sequently be cleaned or blown clear, the indicated pressure drop would
jump to a new, high value, introducing severe, unreal discontinuities
into the record. This plugging problem was never solved but increased
alertness for incipient blocks reduced its severity toward the end of
the experiments.
An additional source of error arose because of non-uniformity in
the re-entrainment of dust from the dust pot. The re-entrainment depended
upon high velocity jets of incoming hot gas blowing through the dust.
These jets could also become plugged, shifting the air flow to a higher
positioned or less obstructed jet and pathway which then rapidly shifted
the dust from its vicinity to that of the plugged jet port or ports. In
any event the dust loading delivered to the bag would vary and become
either erratic or dramatically reduced. Failure to spot this occurrence
introduced additional error into the data.
The fabrics themselves are quite different in what appears to be
the steady-state value of pressure drop. The higher of the two curves
in Figure 7 (the Dacron fabrics) is less than the lower of the two
curves in Figure 6 (the cotton fabrics). Residual dust loading following
the shake-cleaning was not measured directly. The weight of the dust
loaded bags was determined at the end of the test period by removing
21
-------
them from the baghouse and weighing them. They were then vigorously
hand-shaken and reweighed. Table 2 summarizes these measurements.
The striking observation is the large difference~in dust loading
between the cotton bags and the Dacron bags. The weight of a cotton
bag plus its.dust load was at least twice that of the new cotton bag.
The Dacron bags gained only a small additional load when weighed dirty.
These differences, although observed and noted by Dennis and Uilder,
were not as..pronounced for them. Because absolute values of dust
loading are not predictable from the Dennis/Wilder work and must be
determined-independently for each new system, a quantitative comparison
cannot .be made. In any event, the cement dust/cotton bag data of Table
2
2 yields a value of terminal dust loading of 0.10-0.13 Ib/ft , a range
not too different from that given by Dennis and Wilder for the flyash/
cotton system (Figure 2). The terminal dust loading of the cement
dust/Dacron system, on the other hand, is on the order of only 0.003-
2
0.005 Ib/ft . Dennis and Wilder do not give any terminal dust loading
for the flyash/Dacron system but their published residual dust loadings
are an order of magnitude lower for the flyash/Dacron than for the
flyash/cptton system.
The "after run" data listed in Table 2 cannot be classified as
either the W,. or WR (see Figure 2) values of Dennis and Wilder. These
"after run" weights are those of the bags after removal from the bag-
house at the completion of the test runs. The runs ended at some
arbitrary time during a cleaning cycle and hence are more likely to be
nearer their WR value than their WT value. Little difference in weight
is evident between the two Dacron runs except for the anomalous no-
weight gain of 6031-3. The cotton bag cleaned at low "g" does have a
significantly higher dust loading than those cleaned at high "g"—in
agreement with the predictions of Dennis and Wilder, if one chooses to
interpret the "after run" weights as a basis for calculating residual
dust loadings.
22
-------
TABLE 2. BAG WEIGHTS
Bag No.
6034-2
6034-3
6031-1
6031-2
6031-3
6031-4
6031-5
6031-6
6031-7
6034-4
6034-5
6034-6
Run No.
(fabric)
37
(cotton)
38
(Dacron)
39
(Dacron)
40
(cotton)
New, gm
164
165
177
177
177
177
177
177
177
1 164
| 165
1 165
1
After Run
(dust loaded)
gm
396
185
185
177
184
183
183
184
348
348
354
After hand-
shaking.* gm
196
202
203
i 202
1
23
-------
The physical properties of the fabrics making up the filter bags
were measured before and after the shake-clean test runs. The pro-
perties, determined by the procedures described in Appendix A, were
carried out on fabric samples cut from the bags by the School of
Engineering and Textiles, North Carolina State University [Ref. 3].
Their results are summarized in Tables 3 (cotton) and 4 (Dacron). The
used fabrics measured include one sample from each run so that the six
fabrics evaluated consisted of:
1) an unused sample of both the cotton and the Dacron fabrics;
2) one sample of each used fabric (6034-2 and 6031-2), operated
for over 700 hrs at maximum rated temperature but shake-
>
cleaned at the relatively mild maximum acceleration of 1.9
g's; and
3) one sample of each used fabric, operated for only 450+ hrs at
maximum rated temperature but shake-cleaned with a maximum
acceleration on the order of 5 g's.
The total number of shakes on all used fabrics was approximately
the same (> million shakes, Table T).
The major differences between the new fabric and the used fabrics
of the same type were:
1) the used fabric is heavier (presumably because of residual
dust);
2) it is less permeable to air; and
3) it exhibits reduced tongue tear strength.
In addition to the above differences the used cotton fabrics showed
dramatically reduced abrasion resistance; the used Dacron fabric did
not.
The shake-clean cycles themselves produced some differences, the
high "g" cleaning action invariably proving more detrimental:
1) the ravel strip tensile strength (No.3, Tables 3 and 4) of
both the cotton and the Dacron was significantly lower for the
fabric cleaned with the high "g" cycle; and
24
-------
TABLE 3. FABRIC PROPERTIES OF COTTON BAGS [Ref, 3]
ro
01
Property
Height, oz/yd2
Construction: *
Filling, pp1
Warp, epl
Strength-Ravel Strip Tensile:
Filling, Ib
Warp, Ib
Elongation-Ravel Strip Tensile:
Filling, %
Warp, %
Strength-Tongue Tear:
Filling, Ib
Warp, Ib
Within Specimen Variability-
Tongue Tear:
Filling, Ib
Warp, Ib
Strength-Ball Burst, Ib:
Air Permeability, ft3/m1n/ft2:
Abrasion Resistance, cycles:
Residual Dust, % of 1nU1al wt:
Typical New
Average
9.86
60.3
95.4
110.6
91.6
14.8
17.0
12.8
10.9
—
186. 0
12.5
43,642
35,250**
Standard
Deviation
—
0.45
0.55
4.Z5
25.74
0.54
2.35
0.41
0.37
1.53
0.37
14.07
0.37
—
6034-2
(703 hr, 1.9 g)
Average
10.3
60.7
98.7
115.4
92.9
15.8
18.2
6.67
5.32
198
7.16
3,654
3,141**
29.1
Standard
Deviation
__.
0.45
0.55
17.25
21.74
0.31
2.42
0.66
0.54
0.65
0.22
22.8
0.515
2,611
6034-4
(460 hr, 4.6 g)
Average
10.6
60.3
97.2
51.3
46.6
15.7
19.2
5.74
5.27
...
201.8
7.35
4,374
4,336**
21.7
Standard
Deviation
...
0.45
1.10
4.11
3.14
1.74
3.01
0.02
0.22
0.69
0.31
5.63
1.60
811
s per inch and ends per Inch."
**Geometr1c average,
-------
TABLE 4. FABRIC PROPERTIES OF DACRON BAGS [Ref. 3]
ro
Property
Weight, oz/yd2
Construction:*
Filling, ppi
Warp, epi
Strength-Ravel Strip Tensile:
Filling, Ib
Warp, Ib
Elongation-Ravel Strip Tensile:
Filling, %
Warp, %
Strength-Tongue Tear:
Filling, Ib
Warp, Ib
Within Specimen Variability-
Tongue Tear:
Filling, Ib
Warp, Ib
Strength-Ball Burst, Ib:
Air Permeability, ft3/min/ft2:
Abrasion Resistance, cycles:
Residual Dust, % of initial wt:
Typical New
Average
10.1
47.9
74.4
131.6
305.1
41.9
51.6
21.1
34.1
___
438.2
29.44
86,205
84,120**
Standard
Deviation
...
0.10
0.55
21.33
23.82
4.68
1.37
2.44
3.46
1.21
0.82
20.87
5.48
21 ,854
6031-2
(704 hr, 1.9 g)
Average
10.98
49.3
74.6
162.1
231.9
39.3
42.2
12.6
18.1
403
17.6
70,688
64,420**
2.6
Standard
Deviation
—
0.447
0.548
7.72
33.08
1.49
5.23
0.55
3.66
0.52
0.31
11.0
2.32
30,682
.
6031-7
(488 hr, 5.1 g)
Average
11.1
48.9
74.8
109.1
237.5
26.6
35.4
10.7
16.6
285.2
12.32
277,068
259,900**
11.9
Standard
Deviation
...
0.74
0.45
23.74
10.61
5.12
0.92
0.54
2.45
0.28
0.17
40.92
1.76
100,299
*Picks per inch and ends per inch.
**Geometric average.
-------
2) the elongation (No.4, Table 4) and the ball burst strength
(No.7, Table 4) of the Dacron cleaned at high "g" were
significantly lower than those of the Dacron cleaned at low
"g" forces.
The abrasion resistance of the high "g" Dacron sample was anomalously
high (No.9, Table 4) and may reflect a major physical change in the
fabric surface, caused, perhaps, by heat generation during abrading.
For whatever reason, the fabric surface of this sample became extremely
smooth and polished during the abrasion test, the only sample of those
tested to do so and, hence, the only sample to exhibit an increase in
abrasion resistance [Ref. 3].
None of the bags was tested to failure and all appeared to be in
good condition following the test cycle—at least to the eye. The
physical properties of the fabric, however, do not rule out a correlation
between shaker parameters and bag life. All fabric properties that
deteriorated did so more rapidly when the "g" forces increased during
the cleaning cycle. The samples that were operated at high temperature
for a longer time, but at lower "g" cleaning conditions, retained more
of their new fabric properties. The data justify only this qualitative
statement, however.
27
-------
REFERENCES
1. Dennis, R. and J. Wilder, "Fabric Filter Cleaning Studies," EPA-
650/2-75-009, NITS No,..PB 24.0372/AS, .January 1975, GCA Technology
Division.
2. Donovan, R. P., B. E. Daniel and J. H. Turner, "EPA Fabric Filtration
Studies: 4. Bag Aging Effects," EPA-600/7-77-095a, NTIS No. PB
271966/AS, August 1977, EPA/Industrial Environmental Research
Laboratory, Research Triangle Park.
3. Letter Reports, W. C. Stuckey, North Carolina State University, to
J. H. Turner, EPA/Industrial Environmental Research Laboratory-
Research Triangle Park, January to July 1973.
28
-------
APPENDIX A
TEST PROCEDURES FOR FILTER BAG PROPERTIES [Ref. 3]
A. Weight Per Square Yard--(oz/sq yd)
1) Rip seams of bag to obtain rectangular piece.
2) Measure full length and full width of three separate pieces to
nearest sixteenth of an inch. Determine average of each
dimension.
3) Weigh full piece of fabric to nearest 0.1 gram.
4) Calculate weight in ounces per square yard.
B. Construction—Thread Count
1) Count warp yarns (ends) in a 3-inch length at five different
places.
2) Count filling yarns (picks) in a 3-inch length at five different
places.
3) Calculate average warp ends/inch and average filling picks/inch.
C. Tensile Strength and Elongation—Ravel Strip Method^
1) Mark five 1-1/2 x 6 inch specimens on the fabric in both the
warp and filling directions so that no two warp specimens
contain the same warp yarns, nor any two filling specimens
contain the same filling yarns. Mark the longer dimension of
each specimen parallel to the yarn component to be tested for
strength or elongation.
2) Cut all specimens from the base fabric and ravel equally on
both sides frqm the 1-1/2 to a 1-inch dimension.
3) Break the specimens using the Instron tester with the following
test conditions:
a) "D" cell—200 Ib Full Scale Load (FSL)
b) Clamp surfaces--!-1/2 x 1-1/2 inches
c) Gage length: 3 inches
d) Crosshead speed: 0.6 inches/minute
e) Chart speed: 3 inches/minute
29
-------
4) For each specimen, record the breaking load in pounds and
elongation in inches.
5) For both warp and filling yarn components, calculate average
breaking load in pounds and average elongation in percent.
Tongue Tear Strength
1) Mark five 3x8 inch specimens in both the warp and filling
directions. Mark the 3-inch dimension parallel to the yarn
component to be tested for tear resistance. Mark so that no
two specimens contain the same yarn component to be tested.
2) Cut all specimens from the base fabric. Cut into the 3-inch
side of each specimen, 1-1/2 inches from each end (i.e., in
the center of the 3-inches). Extend the cut into the body of
the specimen 3-inches, to make two strips or tongues on the
specimen.
3) Tear each specimen on the Instron, mounting one tongue in one
clamp and the other tongue in a second. (The specimen tears
when the two clamps are separated.)
4) Operate the Instron so that the clamps separate 3 inches
greater than the initial gage, resulting in a 1-1/2 inch tear
in the specimen. Use the following test conditions:
a) "C" cell—20 Ib FSL
b) Clamp surfaces--!-1/2 x 3 inches
c) Gage length: 3 inches
d) Crosshead speed: 2 inches/minute
e) Chart speed: 2 inches/minute
5) Determine tearing strength for each specimen by dividing the
chart for the 1-1/2 inch tear into five equal sections and
reading the highest peak in each section. The average of the
five peaks is the tearing strength of that particular specimen.
6) Calculate average warp and filling tearing resistance.
30
-------
E. Ball Burst Strength
1) Mark and cut from fabric five 4-inch diameter specimens so
that no two specimens include the same warp and filling yarns.
2) Use Scott Model J pendulum tester with 300 pound capacity for
burst tests.
3) Calculate and report average strength in pounds.
F- Air Permeability
1) Use Frazier instrument and make five tests by randomly positioning
the fabric over the chamber opening. (No cutting of specimens
is necessary.)
2) Use No. 4 nozzle (3 mm) or whatever is necessary, and adjust
surface pressure on fabric to 0.5 inch on the inclined manometer
for each determination prior to reading the vertical manometer.
3) Calculate average air flow in cu ft per sq ft of fabric per
minute.
G. Abrasion Resistance
1) Cut five 3-3/4 inch diameter specimens so that no two specimens
include the same warp and filling components.
2) Abrade until failure, using a Schiefer abrasion tester with
a square cut tungsten abradent blade, a 5-lb head weight,
and a 1-inch diameter sample pedestal.
3) Calculate and report geometric mean of number of cycles to
failure.
31
-------
APPENDIX B
CONVERSION FACTORS
To Convert From:
foot?
yard
Ib (force)
foot
inch
mil
yard
grain
Ib (mass)
inch ofpwater (60°F)
lb/inchp (psi)
lb/foor
foot/mi n (fpm)
foot?
inch,
yard
oz/yd2
3
grains/ft ^
grains/1000 ft"5
To:
meter2
meter2
meter
newton
meter
meter
meter
meter
kilogram
kilogram
2
newton/meter2
newton/meter^
newton/meter
meter/sec
meter3
meter.,
meter
kg/m2
kg/m3 •
g/m3
°K
Multiply By:
9.29
6.45
8.36
4.45
3.05
2.54
2.54
9.14
6.48
4.54
2.49
6.89
4.79
5.08
2.83
1.64
7.65
3.39
2.29
2.29
°K =
x 10~2
x 10~T
x 10 '
«K]
x 10 J
x 10_ =
x 10 '
«">:?
x 10 '
X 10*o
x 10 7
x 10 '
xio'2
x 10"2
x 10"?
x 10"'
xlO-2
x 10.-3
x 10 *
— (°F -
9 v r
459.67)
32
-------
TECHNICAL REPORT DATA
(Please read /uuructions on the reverse before completing
1. REPORT NO.
EPA-800/7-77-095b
2.
3. RECIPIENT'S ACCESSION NO.
4.Tm.E ANDSUBT,TLE EPA FABRIC FILTRATION STUDIES:
5. Bag Cleaning Technology (High Temperature Tests)
5. REPORT DATE
November 1977
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
B.E. Daniel, R. P. Donovan (RTI), and J.H. Turner
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
10. PROGRAM ELEMENT NO.
See Block 12.
EHE624
11. CONTRACT/GRANT NO.
NA—Inhouse Report
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Task final: 6/74-1/77
14. SPONSORING AGENCY CODE
EPA/600/13
is. SUPPLEMENTARY NOTES IERL-RTP project officer for this inhouse report is J.H.Turner,
Mail Drop 61, 919/541-2925. Other reports in this series are in the EPA-600/2-
76-168 series.
16. ABSTRACT
repOrt gjves results of a laboratory study to determine the influence of
high temperature operation (operation in an air flow whose temperature has been
adjusted to the maximum continuous operating temperature recommended by the
manufacturer) on the selection of fabric filter shake- cleaning parameters. Two cotton
and two Dacron bags were operated in a laboratory baghouse, using heated air pas-
sed through cement dust as the source of dirty air. The bags cleaned at high 'g'
forces (about 5 g's) showed more deterioration in strength properties than those
cleaned at 1. 9 g's. The observations generally confirm the Dennis /Wilder analysis
of mechanical cleaning and suggest that temperature is not a first order variable it
the analysis of mechanical shake- cleaning. The cursory tests conducted here do net
conclusively rule out temperature as an important parameter; they merely report
that, in this limited investigation, it was not.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENIIFIERS/OPEN ENDED TERMS
c. COSATI 1 icicl/Group
Air Pollution
Filtration
Air Filters
Fabrics
Cotton Fabrics
Polyester Fibers
Cleaning
High Temperature
Tests
Cements
Air Pollution Control
Stationary Sources
Fabric Filtration
Bag Fitters
Baghouses
Shake Cleaning
13B
07D
13K
11E
13H
14B
13C.11BJ
13. DISTRIBUTION STATEMENT
Unlimited
19. SECURITY CLASS (ThisReport)
Unclassified
} NO. Of FAGbS
41
20. SECURITY CLASS (Tills page)
Unclassified
?2. PRICG
EPA Form 2220-1 (9-73)
33
------- |