EPA-600/2-76-168C
December 1976
Environmental Protection Technology Series
EPA FABRIC FILTRATION STUDIES:
3. Performance of Filter Bags
Made From Expanded PTFE Laminate
SB.
111
CD
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-600/2-76-168c
December 1976
EPA FABRIC FILTRATION STUDIES:
3. PERFORMANCE OF FILTER BAGS
MADE FROM EXPANDED PTFE LAMINATE
by
Robert P. Donovan (Research Triangle Institute),
Bobby E. Daniel, and James H. Turner
Industrial Environmental Research Laboratory
Office of Energy, Minerals, and Industry
Research Triangle Park, NC 27711
Program Element No. E HE 624
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Washington, DC 20460
-------
PREFACE
This report is the third 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 4 years by EPA's Industrial Environmental
Research Laboratory, Research Triangle Park, N. C. Related work by
predecessor agencies predates the present series. 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 a baghouse operation in
a carefully controlled laboratory setting that allowed measurement and
comparison of bag performance and endurance. The simulation discussed
in this paper covered only a very narrow range of operating conditions:
1) Redispersed, classified flyash (mass median diameter
between 5 and 6 urn) entrained in air was the only dust
used.
2) All filtering was done at room temperature.
3) Humidity was varied from about 30 to 70 percent.
4) The air to cloth ratio was varied between 4:1 and
10:1.
5) The inlet dust loading was held in the vicinity of
3 grains/ft3 (6.9 g/m3)*.
*EPA policy is to use metric units only or to list both the common
British unit and its metric equivalent. For convenience and clarity, to
the anticipated reading audience. British units are used in this
report. Readers more familiar with the metric system may use the
factors in the Appendix to convert to that system.
iii
-------
Extreme caution should be used in extrapolating the results reported
here to the substantially different conditions that occur in all field
applications. The usefulness of the present results is primarily as an
initial screen of candidate fabrics for baghouse applications.
The projected EPA Fabric Filtration Studies series-will consist of
the following reports:
1) "Performance of Non-Woven Nylon Filter Bags," J. H. Turner
(in press)
2) "Performance of Non-Woven Polyester Filter Bags," G. H. Ramsey
et a!.. EPA-600/2-76-168b, June 1976
3) "Performance of Expanded PTFE Laminate Filter Bags" (this
report)
4) "Bag Aging Effects"
5) "Bag Cleaning Technology"
6) "Analysis of Collection Efficiency by Particle Size"
-------
CONTENTS
Preface iii
Figures vi
List of Abbreviations and Symbols vii
Acknowledgments viii
1. Introduction 1
2. Conclusions 6
3. Experimental Methods 8
4. Results n
Modified Cleaning Cycles 22
Endurance Test 29
5. Discussion 34
References 41
Appendix, Conversion Factors .... 42
-------
FIGURES
Number
1 Gore Tex film (2150X) 3
2 Cross-section of Gore Tex/Nomex fabric (215X) 4
3 Test apparatus 9
4 Outlet concentration as a function of air/cloth ratio . . 12
5 Measured cake resistances 13
6 Pressure drops and calculated drags as a function of air/
cloth ratio 14
7 Humidity dependence of filtration efficiency and outlet
concentration 15
8 "Standard" time dependence of particle concentration
during filtration cycle (Polyester spunbonded fabric) . . 17
9 Time dependence of flyash particles penetrating Gore Tex/
Nomex bags 18
10 Concentration of penetrating particles as a function of
air/cloth ratios 20
11 Size distribution of penetrating particles at various
air/cloth ratios 21
12 Pressure drops during modified cycle (4 sec cleaning,
variable delay time) 24
13 Initial pressure drops for modified sequences 26
14 Final pressure drops for modified sequences 27
15 Average total number density of particles during last
filtering cycle of modified sequence 28
16 Size distribution of penetrating particles at various
delay times 30
17 Pressure drop as a function of filtration time 31
18 Endurance data for Gore Tex/Nomex bag filter 33
19 Close-up of fibrillated film of Gore Tex/Nomex (6450X) • • 35
20 Spunbonded polyester (Reemay*) (645X) 37
21 Gore Tex/Nomex at 108X 38
vi
-------
LIST OF ABBREVIATIONS AND SYMBOLS
A = filtration area of fabric (sq ft)
C = mass outlet concentration (grains/100 cu ft)
E = mass collection efficiency (percent)
l<2 = true value of specific cake resistance (in. ^O/fpmVOb/sq ft)
Kp = measured value of specific cake resistance (in. ^O/fpnO/Ob/sq ft)
APE = pressure drop across bag at time zero of filtration cycle (in. H20)
AP-p = pressure drop across bag at end of filtration cycle (in. H20)
SE = effective drag (in. H^O/fpm)
ST = terminal drag (in. H^O/fpm)
A/C = air/cloth ratio (fpm)
-------
ACKNOWLEDGMENTS
All fabric filters used in this study were donated by W. L. Gore
and Associates, Inc., Rt. 213, North Elkton, MD 21921. It is also a
pleasure to acknowledge the constructive criticism and advice provided
by Mr. Edward De Garbolewski of that organization.
vi ii
-------
SECTION 1
INTRODUCTION
This paper summarizes a laboratory evaluation of filter bags made
of Gore Tex.* Gore Tex is expanded polytetrafluoroethylene (PTFE)
deposited as a thin, fibrillated film (shown in Figure 1). The filters
evaluated here for baghouse applications consisted of this fibrillated
film deposited on a 2/1 twill fabric woven from spun staple Nomex**
fiber. A cross-section through the composite fabric is shown in Figure
2. The coarse woven fibers are the Nomex substrate, one side of which
is covered with the thin PTFE film. This Gore Tex/Nomex composite is
but one of a family of Gore Tex laminates. Other readily available
backing materials include woven polyester and Gore Tex expanded PTFE
itself. All backing scrims are highly porous. Their chief function is
to provide strength for the composite fabric.
Measured properties of the Gore Tex/Nomex fabric are listed in
Table 1. The high tensile and burst strengths are characteristic of a
woven fabric. Permeability is low which might suggest pressure drops
and drags somewhat higher than desired for bags made from this material.
As will become evident, this potential shortcoming does not, in fact,
exist; the dust/fabric system investigated here allowed economical
baghouse operation, more like the best of the previously investigated
systems than the poorest [Refs. 2,3].
Among the properties specified by the manufacturer as being of
particular interest for baghouse applications are the chemical inactivity
of the PTFE ("resists acids, alkalies, weathers well; and is non-flammable")
*Tradename of W. L. Gore and Associates, Inc.
Elkton, MD 21921
**Tradename of E. I. DuPont Company, Inc.
-------
and its temperature stability ("very stable from -350 to 500°F"). It
is a good thermal and electrical insulator and its surface has a low
affinity for water. To fully realize these advantages, however, the
fabric must be an all Gore Tex laminate—Gore Tex PTFE film on a Gore
Tex PTFE backing. All the laminates evaluated in this study were of
the Gore Tex/Nomex type, whose chemical and physical properties are
governed primarily by the Nomex. The film side of the Gore Tex/Nomex
fabric feels slick and smooth--!ike any PTFE coated surface; the reverse
however, shows no influence of the PTFE film.
-------
£r;r- ".fs-jiisj
**»>• .vtmrA'imSS* '>*
^••Pr */JP*- '*^ *A ^m1
• ^%^S
^c :"v^
" -..mlj*"'t ""*,»,<*"* ";*'
'«•.**;•
.
IHH V
>.
- ^ J^*1!
:^2^S* :^>7r5i»-.*a|
,^'^nSS^^-' W.*1*-''* ;-'f
«* f jJP ^ V* « j., «* dflB ' • -^--
_ —^~*^ - "^•^ •' •? ^^ rt»» -
— *"' »«*>
S*». :•**?« ™*
Figure 1. Gore Tex film (2150X) [Ref.ll.
3
-------
Figure 2. Cross-section of Gore Tex/Nomex fabric (215X) [Ref.l]
-------
TABLE 1. MEASURED PROPERTIES OF GORE TEX/NOMEX*
2
Weight/unit area [oz/yd ] 5.90
Thickness [mils] 12.3
Grab Tensile [lbs/in.]
Warp 217
Filling 121
Grab Elongation [%]
Warp 31.7
Filling 27.8
Tongue Tear [Ibs]
Warp 6.73
Filling 6.30
Mullen Burst [psi] 233
Frazier permeability at 0.5 in. H90 press, diff. 10.6
[fWmin/ft2] *
*ASTM test methods carried out by FRL, An Albany International Company,
Route 128 at Route 1, Dedham, MA 02026 [Ref.l].
-------
SECTION 2
CONCLUSIONS
Laboratory evaluation of bag filters made from Gore Tex/Nomex (a
fibrillated film of polytetrafluorethylene supported by a woven lattice
of Nomex) shows that: "^
1) The Gore Tex/Nomex bag filters exhibit very high filtration J5
efficiency (~99+ percent), low outlet concentration (less than ££-
4 grains/1000 ft ), acceptable effective and terminal drags <
and acceptable cake resistance. Their filtration efficiency
was the highest of any bag filters evaluated in this series.
2) Other than showing an increase in pressure drop across the
bag, performance is insensitive to variations in
air/cloth ratio between 4 fpm and 10 fpm. Outlet mass
concentration and cake resistance did not vary significantly
over this range of air/cloth ratios.
3} Conclusions based on data acquired with a particle counter
disagreed with the conclusions based on mass weighings, as
stated in 2, in that the total number concentration of
outlet particles increased with air/cloth ratio (it was almost
twice as great at an air/cloth ratio of 10 fpm as it was at an
air/cloth ratio of 4 fpm). In addition the size distribution
of the outlet particles shifted slightly toward the larger
particles with increasing air/cloth ratio.
4) Bag efficiency and the other performance parameters did not
vary significantly with relative humidity over the 30 percent
to 70 percent range.
-------
5) Optical counter data suggest that the dominant dust/
fabric interaction is sieving. This conclusion is
also supported by scanning electron micrographs of the fabric
surface, revealing structure and dimensions that are comparable
in size to or smaller than the median diameter of the test
flyash.
6) Modifications in the test cycle showed that acceptable
cleaning could be obtained by short (4 sec) periods of shaking.
Reducing the shake cleaning time by a factor of 30 (from 2 min
to 4 sec) produced increases of about 50 percent in the pressure
drops (APr and APy). Reducing the delay times between the
filtration and the shake cleaning increased both the pressure
drops and the number density of particles in the outlet (optical
counter data). A delay time of 5 sees was clearly inadequate;
the optimum delay is a tradeoff between downtime (no filtering)
and energy costs of filtering at high average pressure drop.
7) The one sample endurance-tested failed after 11 million
shakes, making it comparable in endurance to the spunbonded
bags but less durable than the woven bags. The failure time
in this test depended strongly upon mounting technique; the
bag failure occurred in the vicinity of the cuff clamps.
Bag filters made from Gore Tex/Nomex appear particularly well
suited for applications requiring high efficiency filtration. The
chemical inactivity and broad temperature range of operation characteristic
of PTFE (both properties cited by the manufacturer) make Gore Tex/Gore
Tex and, to a lesser degree, Gore Tex/Nomex filter bags suitable over a
wide spectrum of applications and perhaps the optimum solution for
certain filtering problems such as the control of fine hazardous or
toxic particulate emissions. The dimensions and structure of Gore Tex
laminate fabric suggest that it would be a particularly effective fabric
for filtering subrricron particles because many of the PTFE fibers are of
submicron diameter as would be required for efficient filtration of
submicron dusts according to classical filtration theory.
-------
SECTION 3
EXPERIMENTAL METHODS
The series of evaluating tests is similar to those previously
carried out on Cerex* [Ref. 2] and Reemay** [Ref. 3] filter bags. As
before,the tests were conducted in a single compartment baghouse con-
taining only one bag (Figure 3). The Gore Tex/Nomex bags were sewn so
as to present the PTFE coated side of the fabric to the dirty gas flow.
The nominal bag area was 8.5 sq ft for operation at an air/cloth ratio
of 4 fpm. To achieve air/cloth ratios higher than 5 fpm, the bottom of
the test bag was coated with a thin film of Silastic***, thereby reducing
the filtering, length of the bag and the bag area. At a constant total
flow through the baghouse, the air/cloth ratio varies inversely with the
bag area. The higher air/cloth ratios (up to 10 fpm) were achieved
primarily by reducing the bag area and only secondarily by altering the
total gas flow.
The performance parameters measured were filtration efficiency,
outlet concentration, effective and terminal drags and specific cake
resistance. The only dust used was powerplant flyash, classified to
remove oversized particles. By Coulter counter analysis, 10 percent by
weight of the flyash was less than 4 pm in diameter; 90 percent was less
than 16 ym; and the median diameter was between 5 and 6 urn.
The dust loading was held constant throughout all testing, the feed
rate varying with the total flow to produce a constant dust loading at
all air/cloth ratios. Outlet concentration and filtration efficiency
were determined by isokinetically sampling the outlet and weighing the
sample collected on a Millipore filter. The size distribution of the
particles in the outlet gas was also determined using a Climet particle
*Tradename of Monsanto.
**Tradename of DuPont.
***Tradename of Dow Corning, Inc.
-------
HUMIDITY
CONTROL CHAMBER
TEST
FILTER
BAG
^FILTER
CHAMBER
MECHANICAL
SHAKER DISPERSION
OPTICAL
PARTICLE
COUNTER
VARIABLE SPEED
DUST FEEDER (FLYASH)
MILL!PORE FILTER
SAMPLING TRAIN
ROTARY BLOWER
Figure 3. Test apparatus.
9
-------
counter, a forward light scattering instrument. Climet data, taken at
the limit of the counter's capability, were used as secondary measures
of performance only; efficiency and outlet concentration were based
exclusively upon the Millipore sampling as described in Reference 3.
The effect of relative humidity upon the performance parameters was
•checked by varying the relative humidity in a random manner between 30
percent and 70 percent. The equilibrating time between humidity changes
was 48 hours. The relative humidity of the ,gas being cleaned was measured
both before filtration and after filtration, using wet-bulb/dry-bulb
thermometers. All the standard performance parameters were measured as
a function of humidity; the size distribution of particles in the outlet
was also determined by the Climet counter as a function of filtration
time at each value of relative humidity.
Bag endurance was measured on one sample using the number of the
shake cleaning cycles to failure as a measure. The operating cycle for
the endurance test consisted of 15 minutes of shake cleaning (240 rpm,
1-7/8 in. stroke) followed by 1 minute delay, then 2 minutes of dust
feed followed by a 1 minute delay before resuming the 15 minute shake.
Periodically the cycle was switched to the standard 20 minute filtration
cycle for measuring bag performance (20 minutes filtration, 1 minute
delay, 2 minutes shake, 1 minute delay). In both cases the maximum
6 2
harmonic acceleration during shaking was about 2.1 x 10 in./min
(~1.5 g's).
10
-------
SECTION 4
RESULTS
The filtration efficiency of all Gore Tex/Nomex bags tested was
very high, regardless of the air/cloth ratio. Similarly, the outlet
concentrations were very low. Figure 4 shows outlet concentration as a
function of air/cloth ratio. Over the range of 4 fpm to 10 fpm, no
dependence on the air/cloth ratio exists, the scatter in the values
being typical of these low concentrations.
Other performance parameters are plotted in Figures 5 and 6 as a
function of air/cloth ratio. Comparative data points (at 4 fpm only) are
shown for a woven polyester fabric, a spunbonded nylon fabric and a
spunbonded polyester fabric. The specific cake resistance, the effective
drag and the terminal drag (not shown) of the Gore Tex/Nomex fabric are
either similar to or lower than those of woven polyester and spunbonded
.nylon bags but higher than the values of the best spunbonded polyester
bags [Refs. 2,3]. This means that fan operating power is greater for
the Gore Tex/Nomex bags than for the spunbonded polyester bags but is
comparable to or less than the power requirements of woven polyester
bags or spunbonded nylon bags.
When used to filter flyash, Gore Tex/Nomex bags show very little
performance dependence upon humidity (Figure 7). This lack of humidity
dependence is different from that of many fabric/dust systems in which
the filtration efficiency normally improves at high humidity [Ref. 4].
A second striking difference in the interaction between Gore Tex/Nomex
and flyash is the relative stability of mass filtration efficiency.^ .._.
during the filtration cycle. Most fabrics exhibit low efficiency
immediately after cleaning [Ref. 5]. As the filter cake builds up,
efficiency improves. Most of the particles passing the filter do so in
the first half of the filtration cycle. The number of particles in the
11
-------
ro
o
O
o
(O
S-
CT
o
•p-
•M
fO
O)
u
•M
OJ
J 1
6 7
Air/Cloth Ratio (fpm)
8
10
Figure 4. Outlet concentration as a function of air/cloth ratio.
-------
10
CM
O
CM
-CM
0)
Q.
OO
x Spun-
bonded
nylon, woven
polyester O
[Refs.2,3]
Spunbonded polyester [Ref.3]
8
10
Air/Cloth Ratio (fpm)
Figure 5. Measured cake resistances.
-------
O
O
E
3
T
A Sr
8
o
OJ
(O
co
in
o
u
£ 4
O)
to
in
O)
£ 3
1 -
X Woven Polyester
X Spunbonded
Nylon
X Spunbonded
Polyester
567
Air/Cloth Ratio (fpm)
8
.1
10
Figure 6. Pressure drops and calculated drags as a function of
air/cloth ratio.
14
-------
100.00
£ 99.50
c
•f—
(J
99.00
t-t
n n
_
OT- "^
j
D D
O
n
'
a
-
-
-
-
v
*
f ; • • t , .
n _ -
CO
10 "~ -
0
o
o
9 ^ -
1/1
c.
8 i -
cr-
1 0°
-
r U
3 C
O
O
"OJ
1 o
3
2
I -
.1
.1
.9
.8
.7
.6
.5
.4
.3
.2
.1
30
Figure 7.
40
50
60
70
% Relative Humidity
Humidity dependence of filtration efficiency and outlet concentration.
o
CM
(O
C
1
O)
-------
filter outlet varies with time (measured from the end of the previous
cleaning cycle) as shown in Figure 8. This plot is for a spunbonded
polyester fabric filter [Ref. 3].
Similar measurements made on Gore Tex/Nomex filter bags reveal a
very different behavior (Figure 9). At an air/cloth ratio of 4 fpm the
concentration of particles passing through the filter remains essentially
constant throughout the filtration cycle.
This result suggests that the Gore Tex layer itself rather than the
filter cake dominates the dust/fabric interaction.
Filter cake buildup still occurs with the Gore Tex/Nomex, as evidenced
by the increase in pressure drop during the filtration cycle. Since the
increase in pressure drop is not accompanied by an increase in filtration
efficiency, the role of the filter cake may be to plug the Gore Tex
layer rather than to serve as a filtering medium. In any event the
peaks in the concentration of penetrating particles, characteristic of
previous bags tested [Ref. 3 and Figure 8], are greatly reduced for
particles filtered through Gore Tex/Nomex bags.
A pronounced pulsing effect occurs at high air/cloth ratio. Figure
9 compares plots of particle concentrations in the outlet versus filtration
times for two air/cloth ratios. At an air/cloth ratio of 9 fpm numerous peaks
in particle outlet concentration appear at various times during the
filtration cycle. One hypothesis is that these peaks correlate with
filter cake sloughing off the bag. This sloughing action consists of
substantial portions of the filter cake falling off or becoming dislodged
by vibration of the air flow, resulting 1n a sudden release of particles
through the bag. At the lower air/cloth ratio of 4 fpm no pronounced
peaks occur and presumably the sloughing is less abrupt.
Two observations conflict with this model of dust cake "sloughing
action:"
1) According to this model the dust cake plays a measurable role
in the filtration process (the loss or sloughing of dust
cake is hypothesized to produce the bursts of penetrating
particles). The data of Figure 9a suggest the opposite—
that the dust cake has little effect on filtration efficiency.
16
-------
0
t 2-4 inn
+ 1-2 v*
O 0.5-1.0 MM
O 0.3 - 0.5 tiff
a. Relative Humidity
8 10
12
14 16 18 20
Filtration Time (min)
b. Relative Humidity 28X
6 8 10 12 14 16
Filtration Time (min)
18 20
Figure 8. "Standard" time dependence of particle concentration during
filtration cycle (Polyester spunbonded fabric).
17
-------
CO
4r
U
c
O
OJ
O
•£ 1
S-
(O
O.
•o
a. A/C 4 fpm
a — tr
4 6 8 10 12
Filtration Time (min)
14 16
A 1-2 ym
D 0.5-1 ym
O 0.3-0.5 ym
b. A/C 9 fpm
ro
-P
U
c:
O
CD
M 2
S-
10
0
16 18
2 4 6 8 10 12 14
Filtration Time (min)
Figure 9. Time dependence of flyash particles penetrating Gore Tex/Nomex
bags.
18
-------
2) If the dust cake is continually sloughed, the increase in
pressure drop across the bag should be limited to a fixed
value representative of the "effective" dust cake layer that
never gets removed. A special clear.ing cycle should be
unnecessary because the sloughing action constitutes a
naturally occurring cleaning step itself. Neither of these
conclusions is true—pressure drop builds up continuously
across the bag and the bag is not self cleaning.
Dust cake accumulation is, admittedly, far less evident on these
bags than other bags tested in this series. Its absence, however,
implies not so much sloughing as simple inability of the dust to adhere
to the PTFE film. This lack of dust adherence ("good release properties")
has been noted by others [Ref. 6] and proclaimed by the manufacturer as
an important advantage for using Gore Tex laminates as fabric filters
[Ref. 7].
The increased concentration of penetrating particles at high air/
cloth ratio evident in Figure 9 is amplified in Figure 10. Each datum
point of Figure 10 represents the average of the 18 or 19 individual
counts taken at 1 minute intervals throughout the 20 minute filtration
cycle—for example, the 9 fpm data of Figure 10 are simply the average
of all the individual measurements shown in Figure 9b. As the air/cloth
ratio increases, the number of penetrating particles in all size ranges
increases. Those in the larger two sizes Increase a little more than
the smallest size range (Figure 11). The 0.3 to 0.5 ym class makes up
66 percent of the total number at an A/C of 4 fpm but only 57 percent of
the total at an A/C of 10 fpm; conversely the 0.5 to 1.0 urn particles
contribute 29 percent of the total at an A/C of 4 fpm but 35 percent
when A/C « 10 fpm. The breakdown of the total penetrating particle
population, classified into three size ranges, is given as a function of
air/cloth ratio in Figure 11. The small shift to larger-sized penetrating
particles at high A/C agrees qualitatively with Impactor data previously
reported by McKenna et al. [Ref. 6] using Gore Tex/Nomex bags on an
industrial boiler.
19
-------
ro
CO
-I-J
c
o
-------
ro
70
60
50
o
o
o
0)
U -in
s- 30
(U
20
10
O
.-o-
o
-o-
•O
O 0.3 vm
O 0.5 ym
O 1.0 urn
O
-O-
6 7
Air/Cloth Ratio (fpm)
8 9 10
Figure 11. Size distribution of penetrating particles at various air/cloth ratios.
-------
MODIFIED CLEANING CYCLES
Because of the unusual interaction between the flyash and the Gore
Tex/Nomex bags and to further investigate the dust release properties of
this system, two modifications in the standard cleaning cycle were
investigated:
1) Reduction of the shake cleaning period to 4 sees instead
of the standard 2 minutes.
2) Reduction of the delay between the end of the feed cycle
and the beginning of the shake cleaning cycle to various
times less than the standard 1 minute.
Successful operation at reduced shake cleaning and/or delay times
implies more economical field operation in that less time and energy are
required for bag cleaning.
Variations in Delay Time and Shake Cleaning Time
Modifications in the delay time were studied using the 4 second
shake cleaning period exclusively. Two time delays are built into the
standard sequence: one between the end of the filtration cycle and the
beginning of the shake cleaning; and the other between the end of the
shake cleaning and the resumption of the filtration that starts the next
sequence. Both of these delays originate from the same timer so that a
change in one delay always means a similar change in the other.
The standard test cycle is a fixed time test cycle consisting of 20
minutes filtering, 1 minute delay, 2 minutes shake cleaning (at 4 cycles/sec
for a total of 480 complete shakes) and finally another 1 minute delay
before beginning the next 20 minute filtering cycle. The sequence of
steps Is:
FILTRATION
ADJUSTABLE
TIME
DELAY
SHAKE
CLEANING
Standard Cycle:
Modified Cycle:
20 minutes
20 minutes
1 minute
1 minute to 5 seconds
2 minutes
4 seconds
22
-------
Figure 12 shows the effect of reducing shake cleaning and delay
times at an A/C of 4 fpm. The initial and final pressure drops character-
istic of the standard cycle are plotted at the extreme left. The bars
just to right—those labelled with delay time equal to 1 minute—show
the effect of reducing the shake cleaning cycle to 4 sees (16 shake
cycles total). Cleaning is not as effective with the 4 sec shake cycle
in that both AP£ and APy are nearly 50 percent larger but, considering
that the shake cleaning time period has been reduced by a factor of 30,
this magnitude of increase in the pressure drops may be surprisingly
small. The conclusion is that the 480 cleaning shakes of the standard
cycle remove only fractionally more dust cake than do the 16 shakes of
the modified cycle. This effect is analogous to that found by Walsh and
Spaite [Ref. 8] for cotton sateen and flyash, but the mim'umum number of
cleaning strokes required 1s reduced by a factor of 4 over that reported
by Walsh and Spaite.
The remaining data graphed in Figure 12 show the effect of shortening
the delay times. During the delay periods, gas flow through the bag
stops and the system simply rests, allowing suspended dust to settle.
An identical delay period separates the change from filtering to cleaning
and from cleaning back to filtering.
In general, as the delay time 1s shortened, both the Initial and
the final pressure drops Increase, suggesting that, at shortened delay
times, either the dust 1s not as efficiently dislodged from the bag
surface or that more and more of the dust dislodged by the shaking
becomes retrapped on the bags rather than deposited in the hopper or
both. Any forward air flow during the shake cleaning dramatically
reduces the ability of the shake cleaning action to dislodge the trapped
surface dust. If the delay time between the cessation of the filtration
period and the beginning of the shake cleaning 1s not sufficient to
allow the pressure differential across the bag to decay to essentially
zero, the dust dislodgement effectiveness of the short, 4 sec shake
cleaning may be greatly reduced.
The effectiveness of the shake cleaning can also be compromised by
insufficient delay between the shake cleaning and the resumption of
filtration. Ideally the cleaning action shakes the dust off the bags to
23
-------
o
-------
form a dense cloud of agglomerated particles which gradually settle into
the collection hopper. If, however, the gas flow of the filtration
cycle is resumed before the dust cloud has completely or even mostly
settled, those particles and agglomerates still in suspension are rapidly
swept back onto the bag.
Regardless of which delay action dominates, the data of Figure 12
show that with only a 5 sec delay, the effective cleaning of the modified
cycle is greatly reduced.
The data graphed in Figure 12 represent average values of five
consecutive cycles at each delay time. As evident 1n Figures 13 and 14,
which plot sequential values at each delay time, the pressure drops at 5
sees delay are actually worse than indicated in Figure 12 1n that they
have not reached an equilibrium value even after the sixth consecutive
cycle. The pressure drops at 10 sees delay and longer appear to have
reached a steady state value (or nearly so) after five sequential cycles.
Clearly, 5 sees delay time is too short and thwarts the cleaning
action. There is little difference in pressure drops between the 60 sec
delay and the 32 sec delay. Twenty-two sec delay causes increased
pressure drops and 10 sec delay even greater.
The Climet data generally support the conclusion that shortening
the time delay results in higher dust loads on/in the fabric filter and
greater penetration through the fabric. Typical penetration data, as
gathered by the Climet counter, are shown in Figure 15. The bars represent
the average total number density of particles measured during the last
cycle of each modified sequence carried out at the delay time Indicated.
A particle measurement was generally made every minute during the 20
minute filtration cycle. The value graphed 1n Figure 15 is the average
of the 18 to 20 measurements of total particles made at 1 minute intervals
during the filtration step. The values of total particles did not
change much with time--the values recorded 1, 2, or 3 minutes after the
filtration cycle began were essentially the same as those measured 17,
18 or 19 minutes into the 20 minute filtration period. This lack of
time dependence for the Gore Tex/Nomex fabric filters was described
previously and depicted by the data of Figure 9a for operation at an A/C
of 4 fpm.
25
-------
All modified sequences consist of 20 min feed, 4 sec shake clean, and delays as follows:
5 sec 0
10 sec A
22 sec <$>
32 sec Q
1 min El
A/C: 4 fpm
to
0)
CQ
D
Cycles
Figure 13. Initial pressure drops for modified sequences.
-------
Ml modified sequences consist of 20 minute feed, 4 second shake clean and delays as follows;
ro
o
CM
v>
-------
4
o
u
OJ
-M
O
ro
CO
CO
ID
O
to
c
(U
O
-------
Climet data were taken during each of the five or six consecutive
cycles run at a given delay time. The numbers graphed 1n Figure 15
represent the average of those measured during the last cycle only. For
the 5 sec delay, the average value during the last cycle differed sub-
stantially from the average of all cycles, increasing in time somewhat
similarly to the pressure drops at 5 sec delay, previously plotted in
Figures 13 and 14. Far more particles penetrated through the fabric
during the 6th cycle of the 5 sec time delay sequence than during the
1st or 2nd. This increase of particle penetration with time was not
true for cycles using a 10 sec or longer delay time, again being similar
in behavior to the pressure drop data presented in Figures 13 and 14.
The size distribution of the penetrating particles changed only
slightly with delay time, shifting toward larger particle penetration
during the 5 sec delay cycle (Figure 16). The magnitude of the shift is
a 5 percent effect.
Variations in Filtration Time
Variations in the length of the filtration cycle were also in-
vestigated. Holding the delay time at 1 minute (the delay time of the
standard cycle) and the shake cleaning step at 4 sec (the cleaning time
of the modified cycle), the length of the filtration cycle was varied
between 5 and 40 minutes. Figure 17 plots the measured pressure drops
over this range. The major effect of increasing the length of the
filtration cycle is to increase the final pressure drop. The cleaning/
delay cycle restores the initial pressure drop for all filtration cycle
times, as is evident from the unchanging values of AP^. These measure-
ments show that the 4 sec shake cleaning with 1 minute delay would lend
itself nicely to an operating sequence based on a fixed pressure drop to
initiate cleaning.
ENDURANCE TEST
For the endurance test the air/cloth ratio was held at 4 fpm through-
out all filtration. The outlet concentration measured during the periodic
checks of performance parameters 1s plotted in Figure 18. The first
29
-------
80
70
60
50
3
O
43
o
40
CO
o
c
-------
8
«*-
o
in
4
CO
at o
3
CQ
At
20 min.
30 min.
40 min.
o.
o
•o-
_L
Slope KQ'
10.865
9.976
10.827
-O
10
Figure 17.
15
30
20 25
Filtration Time (min)
Pressure drop as a function of filtration time.
35
40
45
I
-------
evidence of performance degradation appeared after 4.5 million shakes.
Two small tears formed on the bag where it was clamped to the bottom
support cone. The outlet concentration increased but leveled off at 12
3
to 14 grains/1000 ft until two new tears developed at the top of the
bag, causing the outlet concentration to jump significantly. These new
tears were also in the vicinity of the cuff, this time the top cuff.
From Figure 18, the bag failure can be seen to occur between 10 and 11
million shakes.
The endurance measured in this test depended primarily on mounting,
similar to the conclusion reported by McKenna et al. [Ref. 6] when using
Gore Tex/Nomex in a pilot scale baghouse on an industrial boiler.
The bag failure here stemmed from fabric abrasion against the support
cones at either end. The bag fabric remote from the mounts was in good
condition with only slight evidence of wear or thinning. Superior
mounting methods could greatly increase the measured life of the bag.
The similarly measured, shake endurance of woven polyester bags
2
was 54 million; of 3 oz/yd Reemay bags, 3.5 million shakes; and of 6
P
oz/yd Reemay bags, 22 million shakes [Ref. 3]. The heaviest weight
Cerex bag tested exceeded 6 million shakes without catastrophic failure
[Ref. 2], Translating the number of mechanical-shakes-to-failure into
bag life is complicated by the influence of other bag properties upon
the frequency of cleaning. While the spunbonded fabrics [Refs. 2,3] and
the Gore Tex/Nomex fabric do not tolerate as many shakes as woven
polyester bags, their lower values of drag and specific cake resistance
mean that they will not have to be shaken as often so that they could
conceivably last longer in a given field application.
32
-------
1 2 34 5 6 78 9 10 11 12 13 14 15 16
No. of Shakes x 106
Figure 18. Endurance data for Gore Tex/Nomex bag filter.
33
-------
SECTION 5
DISCUSSION
The particle counter data reveal some distinctive properties of the
interaction between flyash and fabric filters made of Gore Tex/Nomex.
Particle concentrations according to three optical sizes, as shown in
Figure 9 for two air-to-cloth ratios, do not vary in time in the same
manner as they do for a fabric filter that utilizes the dust cake for
additional particle capture and removal. As previously discussed, when
the dust cake plays the dominant role in the filtration process, filtration
efficiency is lowest immediately after the cleaning cycle and increases
as the dust cake rebuilds during the subsequent filtration cycle, being
the highest immediately before the next cleaning cyle. Figure 9 shows
that the Gore Tex/Nomex fabric does not behave this way. At an A/C of
4 fpm, the particle concentrations are roughly constant throughout the
filtration cycle; at A/C = 9 fpm (Figure 9b), the particle concentrations
actually increase as filtration proceeds, exhibiting peaks of penetrating
particles superimposed upon a slowly rising background.
The hypothesis of dust cake sloughing was rejected as an explanation
because the dust cake appears not to dominate the filtration process
even though its buildup causes an increase in pressure drop across the
bag.
The results presented in the previous section are qualitatively
consistent with a model in which the fibrillated film of PTFE (Figures 1
and 19) removes most of the flyash by sieving. Referring to the scanning
electron micrograph of Figure 19 (printed at 6450X, meaning a 1-in. length
in the micrograph represents about 4.0 ym in reality), the structure and
dimensions of the fibrillated film makes sieving of 5 to 6 ym median
diameter flyash appear highly likely, since few passageways 1n the film
34
-------
Figure 19. Close-up of fibriHated film of Gore Tex/Nomex
(6450X) [Ref.l].
35
-------
are as large as 6 ym. These dimensions contrast markedly with those of
the Reemay fabric, for example, shown in Figure 20 at 645X, a full order
of magnitude lower in magnificiation than the Gore Tex/Nomex micrograph
in Figure 19. The gross differences in fiber dimensions and spacings
make differences in the dust/fabric interaction not only plausible but
highly likely.
Accepting sieving as the primary dust removal mechanism then leads
to the following consequences:
1) The particle trapping depends primarily on the properties of
the fibriHated Gore Tex film deposited, like a fine spider
web, across the woven Nomex lattice (Figure 21).
2) The dust cake that collects does not improve filtration
efficiency over the 0.3 ym to 1 ym particle size range
although it increases the pressure drop across the fabric
by blocking spaces and passageways—by plugging the sieve.
3) The shake cleaning removes the dust cake, thereby unplugging
the filter and reducing the pressure drop across the filter.
The efficiency of filtration, depending only on the Gore
Tex film, remains unaffected by the cleaning.
4) At high gas flow—high air-to-cloth ratios—the dust cake
increases in quantity, plugging more of the film sieve and
giving rise to high pressure drops across the fibrillated
film. These high pressure drops stretch the film, causing
it to open its pores and perhaps rupture at discrete points.
As the film is stretched and opened up, additional particles
pass through the sieve. The peaks in penetrating particles
(Figure 9b) are characteristic of pore stretching to pass
a pulse of particles, followed by partial or complete
relaxation to the original pore size. Ruptured filaments
represent an increased pore size and could account for part
of the increase of penetrating particles with time as depicted
in Figure 9b. Rupture is not a necessary feature of the
model, however, since increased pressure drop alone means
increased pore size in the film and pressure drop has been
36
-------
Figure 20. Spunbonded polyester (Reemay*) (645X) [Ref.l],
*Registered tradename of E. I. Dupont Co., Inc.
37
-------
Figure 21. Gore Tex/Nomex at 108X [Ref.l]
38
-------
shown to build up throughout the filtration cycle
(Figure 17) and with increasing A/C (Figure 6).
The sieving model just described is based almost exclusively on
data from the optical counter. These data are not always supported by
the measurements of outlet concentration based on total mass collected
on a Millipore filter (Figure 4). In Figure 4 outlet concentration (C )
appears insensitive to A/C; if anything it decreases with increasing
A/C. The Figure 4 data, however, reflect very low values of outlet
concentration—all are below 4 grains/1000 ft . Errors in this measurement
3
become very large at C values below 10 grains/1000 ft . Therefore
these data may not be significant.*
The sieving model, including pressure-drop-induced stretching of
PTFE, explains most of the major features of both the optical and the
mass data. Lack of sensitivity to relative humidity, as portrayed in the
mass data of Figure 7, is consistent with sieving as the primary filtration
mechanism. Humidity affects particle agglomeration and sticking coefficient,
neither of which is critical for sieving. The increase in the number
of penetrating particles with A/C, as shown in Figure 10, is consistent
with pore stretching because of the higher pressure drop occcurring at
high A/C (Figure 6). Pore stretching can also account for the slight
shift to larger sized penetrating particles at high A/C (Figure 11). In
all modified cycle testing (Figures 12-17) anything that produced increased
pressure drop across the bag also caused an increase in the number
density of penetrating particles and a shift to larger sized penetrating
particles (compare Figures 13 through 16).
*In general quantitative correlation of optical counter number data with
Millipore mass data has not been good throughout these experiments.
Sampling ports are located differently for the two measurements (Figure
3). Isokinetic flow is used for the mass data while a fixed total flow,
adjusted to be near isokinetic by varying the sampling nozzle dimensions,
is used for the optical measurements. The measurements depend on different
particle populations (the Millipore data measures total mass; the optical
counter, number density over a narrow size range) and particle properties
so that quantitative correlation would not necessarily be direct and
simple. Qualitative correlation should be expected but does not exist
between measurements of C. vs A/C (Figure 4) and particle penetration vs
A/C (Figures 9, 10). °
39
-------
While the sieving explanation seems to account for the present
observations, it is still tentative and speculative. Further testing
using a finer dust (a dust composed primarily of submicron particles)
would provide valuable additional data. With these finer particles,
sieving should become less important and the role of the dust cake more
similar to that in the previous dust/fabric systems studied in this
series [Refs. 2,3].
40
-------
REFERENCES
[1]. FRL Client Report G 74321, "Characterization of Filter Fabrics,"
12 Nov. 1974, FRL, An Albany International Company, Route 128 at
U.S.I, Dedham, Mass. 02026.
[2]. Turner, J. H., "EPA Fabric Filtration Studies: Performance of
Non-Woven Nylon Filter Bags," (in press).
[3]. Ramsey, G., R. P. Donovan, B. E. Daniel and J. H. Turner, "EPA
Fabric Filtration Studies: 2. Performance of Non-Woven Polyester
Filter Bags," EPA-600/2-76-168b, (NTIS No. PB 258-025/AS), June 1976
Research Triangle Park, N. C. 27711.
[4]. Durham, J. E. and R. E. Harrington, "Influence of Relative Humidity
on Filtration Resistance and Efficiency of Fabric Dust Filters,"
Filtration and Separation 8. July/August 1971, pp. 389-393.
[5]. Dennis, R., "Collection Efficiency as a Function of Particle Size,
Shape and Density: Theory and Experience," J. Air Poll. Control
Assoc. 24., Dec. 1974, pp. 1156-1163.
[6]. McKenna, J. D., J. C. Mycock and W. 0. Lipscomb, "Applying Fabric
Filtration to Coal-Fired Industrial Boilers, A Pilot Scale
Investigation," EPA-650/2-74-058-a, (NTIS No. PB 245-186/AS), August
1975, Enviro-Systerns and Research Inc., P. 0. Box 658, Roanoke,
VA. 24004.
[7]. "Gore TexR Filter Bags for Dust and Product Collection in Baghouses,"
Product Information, Feb. 1976, W. L. Gore and Associates, Inc.,
P. 0. Box 1220, Rt. 213 North, Elkton, Md. 21921.
[8]. Walsh, G. W. and P. W. Spaite, "An Analysis of Mechanical Shaking in
Air Filtration," J. Air Poll. Control Assoc. 12., Feb. 1962, pp.57-61.
41
-------
APPENDIX
CONVERSION FACTORS
To Convert From:
footo
To
Multiply By:
yard
grains/foot *
grains/1000 ft
Ib (force)
foot
inch
mil
yard
grain
Ib (mass)
inch of2water (60°F)
lb/inch? (psi)
lb/foor
foot/min (fpm)
foot3
inch.,
yardj
oz/yd2
meter
kg/m3
g/m3
newton
meter
meter
meter
meter
kilogram
kilogram
2
newton/meter2
newton/meter2
newton/meter
meter/sec
meter3
meter,
meter
kg/m3
9.29 x
6.45 x
8.36 x
2.29 x
2.29 x
4.49
3.05 x
2.54 x
2.54 x
9.14 x
6.48 x
4.54 x
2.49 x
6.89 x
4.79 x
5.08 x
2.83 x
1.64 x
7.65 x
o **n ^
10-4
10-1
10 '
10"?
10"J
10"!
10l5
IS-'
'"I
10 '
io:|
lo-i
10 l
ao"3
10"?
10-1
10 J
-.n-2
42
-------
TECHNICAL REPORT DATA
(Please read lauructions on the reverse before completing)
1. REPORT NO.
EPA-600/2-76-168C
2.
3. RECIPIENT'S ACCESSION-NO.
4. TITLE AND SUBTITLE
EPA Fabric Filtration Studies: 3. Performance of
Filter Bags Made From Expanded PTFE Laminate
i. REPORT DATE
December 1976
6. PERFORMING ORGANIZATION CODE
AUTHOR(S) Robert P. Donovan (Research Triangle Insti-
tute), Bobby E. Daniel, and James H. Turner
8. PERFORMING ORGANIZATION REPORT NO.
IERL-RTP-233
. PERFORMING ORGANIZATION NAME AND ADDRESS
See Block 12.
10. PROGRAM ELEMENT NO.
HE 62 4
11. CONTRACT
NA—Inhouse Report
2. 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
[nhouse Final: 10/74-10/75
14. SPONSORING AGENCY CODE
EPA-ORD
5.SUPPLEMENTARY NOTEsjERL-RTP project officer for this inhouse report is J. H. Turner,
919/549-8411 Ext 2925, Mail Drop 61.
6.ABSTRACTTne report? third in an EPA Fabric Filtration series, gives results of an
evaluation of fabric filters made of an expanded polytetrafluoroethylene (PTFE) film
supported on a woven Nomex scrim--the Gore Tex/Nomex fabric. Filtration efficiency
ivas very high and other performance parameters (drag and effective cake resistance),
acceptable. The one fabric bag tested for endurance failed prematurely near the bag
cuff; even so, it gave evidence of acceptable bag life. Because of the small fiber
dimensions and spacings of the PTFE film, the dominant mechanism for particle re-
moval appears to be sieving. This mechanism is not usually the dominant filtering
mechanism for fabric filters; consequently, the Gore Tex/Nomex fabric exhibits some
properties that are different from those of other fabrics evaluated in this series. The
most important difference is in the role of the dust cake which, for the system repor-
ed here, is jiot a major factor in determining efficiency. Filtration efficiency is as
;ood or better with little or no dust cake on the filter (such as at the beginning of a
iltration cycle) than it is after a cake has had a chance to form (such as at the end of
he filtration cycle). Thick dust cakes were simply not seen on this fabric, however.
This conclusion applies only to the flyash used in these experiments. Finer dusts may
behave differently.
7.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS C. COSATI Field/Group
Air Pollution Fly Ash
Dust Caking
filtration
Dust Filters
Fabrics
Tetrafluoroethylene Resins
Air Pollution Control
Stationary Sources
Polytetrafluoroethylene
Particulate
Fabric Filters
Baghouses
Sieving
13 B
11G
11D
13K
HE
HE
2 IB
07A,13H
ttg_
JRVI
18. DISTRIBUTION STATEMENT
Unlimited
19. SECURITY CLASS (This Report}
Unclassified
21. NO. OF PAGES
51
20. SECURITY CLASS (This page)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
43
------- |