EPA-R2-73-288
July 1973
RELATIONSHIP
AND
:¥:Wii:iiiS:H:SSi:i:S:K:S:v:vi«ii
:•:•:•:•:•:••.••:•:•:•:•:
I.*.*.* •.
m
•$x*
11
I
•.•.-.-.•.
•WS
•:•:•:•>:•;
•.*.•.•.•.•.
;$|fciji#;:;:S|::^^
3(|iii:§|::i|§iKii^S^
i!j;:|:|^^
-------
EPA-R2-73-288
RELATIONSHIP BETWEEN
FABRIC STRUCTURE
AND FILTRATION PERFORMANCE
IN DUST FILTRATION
by
Dean C. Draemel
Control Systems Laboratory
National Environmental Research Center
Research Triangle Park, North Carolina 27711
Project 21ADJ51
Program Element 1A2012
NATIONAL ENVIRONMENTAL RESEARCH CENTER
OFFICE OF RESEARCH AND MONITORING
U.S. ENVIRONMENTAL PROTECTION AGENCY
RESEARCH TRIANGLE PARK, NORTH CAROLINA 27711
July 1973
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Monitoring,
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 Office of Research and
Monitoring, EPA, and approved for publication. Approval does not
signify that the contents necessarily reflect the views and policies
of the Environmental Protection Agency, nor does mention of trade
names or commercial products constitute endorsement or recommendation
for use.
ii
-------
ABSTRACT
The report identifies a semi-empirical relationship between clean
cloth-fabric structural parameters, dust parameters, and filtration
performance. The criterion for high outlet concentration as a result
of bleeding or seepage of dust is a function of the pore size distribution
of the fabric versus size properties of the dust. The presence of a
significant number of pores with a characteristic dimension roughly 10
times the mass mean particle diameter of the dust being filtered leads
to bleeding and seepage of dust. This conclusion results from studies
with three dusts (fly ash, limestone, and silica), a number of fiber
types, and a range of fabric construction variables. Pressure-related
filtration performance can be correlated against clean fabric free area
if well defined yarn boundaries are present. Since filtration fabric
yarn boundaries are generally not well defined, pressure-related
filtration performance can be correlated against clean cloth. Frazier
permeability.
m
-------
CONTENTS
CONCLUSIONS
RECOMMENDATIONS 5
INTRODUCTION 7
Background 7
Reasons for Performing Work 7
Approach and Objectives 8
EQUIPMENT AND PROCEDURES 9
Single-Compartment Baghouses—Initial Study 9
Bench-Scale Filtration Apparatus—Supplemental Study 14
Single-Compartment Baghouses—Supplemental Study 14
RESULTS AND DISCUSSION 17
Single-Compartment Baghouse Study--Free-Area Correlation 17
with Pressure-Related Filtration Response
Group 1 Fabrics 18
Group 2 Fabrics 23
Group 3 Fabrics 23
Group 4 Fabrics 30
Bench-Scale Filtration Study—Microscopic Analysis of the 34
Filtration Process
Fabric 015 34
Fabric 038 37
Fabric 088 37
Fabric 120 38
Fabric 088 (3=1Oum Test Dust) 39
-------
Page
Single-Compartment Baghouse Study--Dust-Size/Pore-Size 41
Correlation with Filter Efficiency
Filtration Performance of Dust/Fabric Combinations 41
Filtration Performance of Commercial Fabrics 45
Fabric Parameters Generating Significant Dust/Fabric 50
Interactions
REFERENCES 55
NOMENCLATURE 57
CONVERSION FACTORS TO METRIC UNITS 59
APPENDIX A-DESCRIPTION OF 123 TEST FABRICS 61
I. Multifilament-Dacron Series 61
II. Staple-Dacron Series 63
APPENDIX B--STANDARD FABRIC AND YARN CHARACTERIZATION TESTS 65
I. Fabric Analysis 65
II. Yarn Analysis 67
APPENDIX C--EFFECT OF WEAVE ON FILTRATION PERFORMANCE 69
APPENDIX D—YARN AND FABRIC ANALYSIS 73
I. Fabric Analysis—Test Results 73
II. Yarn Analysis—Test Results (Denier, Bulk and Fiber 74
Density)
III. Yarn Analysis—Test Results (Twist, Width, and 75
Thickness)
APPENDIX E-RESULTS OF SUPPLEMENTAL CHARACTERIZATION TESTS AND
FABRIC FILTRATION PERFORMANCE DATA*
*Not included in this report, but available on request from author.
VI
-------
FIGURES
Figure No. Title Page
1 Single-Compartment Baghouse Test Equipment 10
2 Characteristic Resistance Curve 12
3 Bench-Scale Filtration Apparatus 15
4 AP, and S Versus Free Area ~ Group 1 Fabrics 20
5 K Versus Free Area--Group 1 Fabrics 21
6 CQ Versus Free Area--Group 1 Fabrics 22
7 APf and S Versus Free Area--Group 2 Fabrics 25
8 K Versus Free Area--Group 2 Fabrics 26
9 APf and S Versus Free Area -- Group 3 Fabrics 28
10 K Versus Free Area--Group 3 Fabrics 29
11 APf and Se Versus Free Area— Group 4 Fabrics 32
12 K Versus Free Area-- Group 4 Fabrics 33
13 The Four Interstice Types 36
14 Pore Dimensions 48
15 Non-1inearvAP Versus t Response 51
16 Schematic Representation of Dust/Fabric Interactions 53
vii
-------
TABLES
Table No. Title Page
1 Free Area and Performance Data—Group 1 Fabrics 19
(Dacron)
2 Free Area and Performance Data—Group 2 Fabrics 24
3 Free Area and Performance Data—Group 3 Fabrics 27
4 Free Area and Performance Data—Group 4 Fabrics 3'
5 Fabric Description Data for Microscopic Analysis 35
and Dust/Fabric Combination Studies
6 Minimum Pinhole Sizes During a Filter Cycle 40
7 Filtration Performance Data for Dust/Fabric 42
Combinations
8 Relative Yarn Dimensions for Continuous-Filament 44
Versus Staple Yarns Compared to Outlet Concentration
Data
9 Filtration Performance Data—Fabric Composition Study 46
10 Pore Size Distributions of Selected Fabrics 49
Cl Pore Type Effects on Fabric Behavior 69
C2 Continuous-Filament Fabrics—Pore Type Effects 71
C3 Staple Yarn Fabrics—Pore Type Effects 72
C4 Minutes of Filtration to Form Continuous Dust Cake 72
vm
-------
CONCLUSIONS
Fabric filtration performance is dependent on structural properties
of the fabric. Within the limits of this study (mainly fly ash at a
standard set of test conditions) the following conclusions can be made:
1. Efficiency or outlet concentration is a function of the pore size
distribution of a given fabric. Bleeding or leaking of dust is
a function of the number of pores above a critical size which
is related to size properties of the dust being filtered. This
study indicates that the presence of a significant number of
pores with a characteristic dimension roughly 10 times the
characteristic dimension (a mass mean particle diameter) of the
dust being filtered (fly ash, limestone, amorphous silica) leads
to bleeding or leaking of dust; i.e., increased penetration.
Projecting fibers within the pores and above the fabric's
surface greatly reduce the effective characteristic pore
dimension with respect to dust bridging. Such fibers do not
significantly affect the gas flow through the pore; i.e.,
although efficiency is a function of projecting fibers, pressure
drop is relatively unaffected. Although staple yarns provide
less yarn cover per unit weight of fiber, as opposed to con-
tinuous filament yarns, the added benefit of the projecting
fibers from staple yarns can be valuable.
2. K values (specific cake resistance) with a given dust are
dependent on the structure of the underlying fabric. The
deep channel-like pores, formed by more rounded yarns, can
lead to significant deposition of dust under velocity conditions
of an order of magnitude or more greater than the average face
velocity of the fabric. Deposition at local increased velocity
would tend to increase dust packing density and thus increase K.
-------
(Dusts subject to cake collapse phenomena imply pressure
and/or velocity dependence on dust packing density.) Very
shallow pores and a smooth fabric surface with no projecting
fibers can be very efficient in particle retention but lead
to a completely unsupported dust layer which has a
characteristically high K value and is subject to cake
collapse as pressure increases. Projecting fibers appear
to support a more porous dust cake (lower K values), less
subject to cake collapse. The dense projecting fibers
found with napped fabrics may tend to produce non-linear
AP vs t response, indicating a deviation from the cake law
type of filtration behavior normally seen with a woven
fabric. K values with a given dust may vary considerably
as a function of fabric even though efficiency remains
relatively constant for the same dust/fabric combinations.
3. Effective drag (flow resistance of the cleaned filter
fabric) can be correlated with free area if yarn boundaries
are well defined. If yarn boundaries are not well defined,
clean cloth permeability generally is a good indicator of
effective filter drag. Factors such as deep, channel-like
pores with no projecting fibers appear to increase effective
filter drag values for the aged filter fabrics just as the
K values are increased.
4. Structural properties of a fabric strongly affect the
filtration performance of a fabric and the fabric's inter-
action with a dust. The chemical composition of the fiber
does not appear to be a major variable although intrinsic
properties of certain fiber types, such as the curly
nature of cotton, may modify or determine significant
structural features of a fabric.
-------
5. Fabric parameters such as weave and yarn construction
also interact to influence significant structural properties
of a fabric; i.e., pore depth which in turn affects
performance.
-------
RECOMMENDATIONS
1. Bench-scale filtration tests and microscopic examination of
potential filters with the dust and system (temperature, relative humidity,
etc.) of interest will provide design data and insight into the probable
performance. This will allow determination of significant dust fabric
interactions such as bleeding criteria and relative K values of dust fabric
combinations for the application under consideration. Care must be taken
to perform such tests with the dust and conditions expected in the application
because performance may be affected by dust properties and system variables
such as relative humidity.
2. Additional work should be conducted to extend the limits of this
work to a greater range of operating conditions and more dust/fabric
combinations.
3. An attempt should be made to more exactly define the criteria for
bridging and bleeding as a function of dust and pore properties.
4. Performance of operating fabric filter systems should be observed
in the field to provide practical data useful in future design and specifications
of fabrics for industrial dust and fume control.
-------
INTRODUCTION
BACKGROUND
Much literature exists concerning flow and pressure drop through
porous media. Refinements and simplifying assumptions must be made when
considering dust filtration through porous media. When considering dust
filtration through heterogeneous (woven fabrics) versus homogeneous
porous media (mats, felts, papers, membranes, etc.), the situation
becomes increasingly complex.
Exact theoretical treatment of dust filtration through woven
fabrics is largely impossible without geometrical simplifications.
Actual industrial fabric filters seldom correspond to screens, mono-
filament fabrics, arrays of perfect orifices, or single fibers as
assumed in theoretical analyses.
General empirical correlations can be developed but, lacking any
fundamental backup, are often subject to catastrophic failures occurring
almost randomly. Improved relationships between fabric structure and
filtration performance should result from practical considerations of
theory and empiricism.
REASONS FOR PERFORMING WORK
Preliminary work relating fabric structure to filtration performance
was conducted by Spaite and Walsh. This work pointed out significant
changes in performance with what appeared to be relatively small changes
in fabric structure. Analysis of filtration literature by Borgwardt and
2
Durham cites the fact that equations normally used to predict filter-
house pressure drops "assume dust drag to be determined by intrinsic
properties of the dust only."
-------
The Kozeny equation, (1), relates pressure versus flow through
porous media for a bed of fixed configuration:
a3
This equation can be used to describe pressure drop through a filter
cake as a function of time if the dust feed rate and gas properties
are held constant. Under these conditions, cake law filtration
(linear AP vs t) results when L increases linearly with time and a is
constant. If a is not constant, L will increase non-1inearly with t
generating a non-linear AP vs t filtration response. Implications of
differing K (specific cake resistance proportional to AP\ values with
Atl
the same dust on different fabrics and the occasional non-linear
pressure response of some fabrics are that a is a function of the
fabric; i.e., K is not solely an intrinsic property of a given dust
alone.
4
Durham and Harrington showed a significant interaction of
fabrics and dusts as a function of relative humidity. Their research,
along with a number of other smaller studies, has led to the develop-
ment of the program reported here.
APPROACH AND OBJECTIVES
Slight variations in fabric structure have been shown to
significantly affect filter performance. No quantitative determi-
nation has been made of exactly what factors are responsible for the
observed changes in filtration performance. The objective of this
study was to relate performance to clean cloth fabric parameters,
and to obtain quantitative relationships with such parameters.
8
-------
EQUIPMENT AND PROCEDURES
SINGLE-COMPARTMENT BAGHOUSES—INITIAL STUDY
A group of 123 fabrics (listed in Appendix A) was custom woven and
sewn into filter bags. These fabrics were subjected to a number of
standard fabric characterization tests (listed in Appendix B). Results
of these standard tests were used to obtain an accurate quantitative
description of the structure of each fabric. The filter bags of each
fabric were tested on single-compartment, highly instrumented baghouses
to characterize the fabric's filtration performance. Results of the
supplemental characterization tests and the filtration performance data
for each fabric are listed in Appendix E. (Appendix E is not included
in this report. It is available on request.)
A single-compartment, instrumented baghouse is shown schematically
in Figure 1.
Dust-laden air, entering at the top of the baghouse, is deposited
on the inside of the bag. The air stream then passes out of the baghouse,
through a venturi flow-measuring device, flow control valve, efficiency
sampling section, and a positive-displacement rotary blower. The air
stream is conditioned in a controlled-humidity control led-temperature
chamber before entering the flow system. Dust is fed by a variable
speed screw feeder and a dispersion venturi. The baghouses are equipped
with adjustable timers to automatically control filtration time, shake
time, and delay time. Flow volume through the filter bags is automatically
controlled and both flow and pressure drop through the bags are continuously
recorded.
Test filters of each of the 123 fabrics were 5.56 in.* in diameter
and 71.5 in. in overall length with a 2-in. cuff on each end. Total
*
Although it is EPA's policy to use the metric system for quantitative
descriptions, the British system is used in this report because not to
do so would tend to confuse the reader. Readers who are more
accustomed to metric units may use the table of conversions on Page 59
to facilitate the translation.
-------
HUMIDITY CONTROL
CHAMBER
MILLIPORE FILTER
SAMPLING TRAIN
DUST FEEDER
COLLECTION HOPPER
Figure 1. Single-compartment bag house test equipment.
10
-------
2
filter area was 8.1 ft . Flow was controlled at 32.5 acfm to provide
a filter ratio of 4.0 ft/min. (4.0 fpm). Dust feed was set to give
3
an inlet dust concentration of 3.0 grains/ft . The test dust was
redispersed Detroit Edison fly ash which had been classified to remove
coarse particles; i.e., > 20 pm. The test dust was 50 percent by
weight less than 3.7 ym, and 90 percent by weight less than 11 ym by
Coulter Counter analysis. Relative humidity of the gas stream was
controlled at 30 .5% at 70 -10°F. Filter bags were installed with no
apparent slack under a very slight tension (estimated at ^ 1 Ib).
Each filter bag was run for 24 hours at the conditions specified to
break in the fabric and allow it to attain rough equilibrium. The
filter cycle consisted of 20 minutes of filtration, a 1-minute delay,
2 minutes of shaking, and another 1-minute delay before repeating
the cycle. The bottom of the bag was shaken with a horizontal shake
1-3/4 in. in length, at a frequency of 240 cycles per minute.
A sample of 4.0 acfm was withdrawn isokinetically through a
probe downstream of the filter bag. The sample was drawn through a
0.45 urn Mi Hi pore filter. A Coulter Counter analysis was run on
the dust sample collected. Efficiency was calculated by the weight
of dust collected versus the known dust feed rate. Fallout was not
measured but was assumed to be reasonably constant at an insignificant
level for the fixed dust, bag dimensions, and flow'conditions used.
Each filter bag was sampled for three consecutive cycles after the
24-hour equilibration period. Filtration performance data recorded
in Appendix E (available on request) represent the averages from the
three consecutive cycles.
A typical pressure drop versus areal dust loading curve for a
fabric filter is shown in Figure 2. Four responses describing fabric
performance are taken from this curve and the Mi Hi pore filter dust
sample. The linear portion of the curve, representing cake law
filtration, is described by:
11
-------
TERMINAL DRAG
SLOPE SPECIFIC CAKE RESISTANCE
END OF FILTER CYCLE
AREAL DUST LOADING, Ib/ft2
Figure 2. Characteristic resistance curve.
12
-------
np- = Kw + Se (2)
The three responses describing system resistance are:
effective drag, S /in> H2°\
e \ft/min |
specific cake resistance, K K H20/n/min
\ lb/ft2
terminal drag APf A /1n" H2°\
Q (ft/min /
2
where w is the areal dust loading in lb/ft . For the data reported in
this paper, final pressure drop, AP^ (in. HpO), is used in place of
terminal drag. For a given filtration cycle, only two of the three
responses are independent; however, performance is more easily
visualized with the aid of the three responses. The fourth response,
outlet concentration, represents an average over the filter cycle, and
is calculated from the Millipore sample by using:
C = weight on Millipore (3)
sample volume
Inlet dust concentration is defined as:
C. = R (4)
1 Q
The efficiency is thus:
r
Eff = 1 - _p_ weight percent. (5)
Ci
13
-------
BENCH-SCALE FILTRATION APPARATUS-SUPPLEMENTAL STUDY
Selected fabrics from the 123 fabrics tested on the pilot-scale
baghouses were also tested on a small bench-scale filter system. This
bench-scale apparatus, shown in Figure 3, consists of a grooved-disc
dust feeder, filter chamber, filter holder, blower, and rotameter.
Pressure drop across the filter was measured by an inclined manometer.
Dust was fed to the lower face of the horizontally mounted filter.
2
The 1-ft filter fabric was held in a removable frame, clamped between
the filter chamber sections. The fabric was tensioned slightly by
hand, to hold it firm and flat in the frame. The frame consisted of
?
two pieces of 1/4-in. plastic with a 1-ft section cut out of the
center. The fabric was placed between the frame sections and the
assembly was bolted together. A thin strip of soft rubber provided
an airtight gasket against both sides of the fabric and between the
filter frame and the filter chamber sections.
Dust was fed and the pressure increase was recorded as a function
of time. The flow was stopped and the filter fabric frame was removed
occasionally to allow microscopic examination and weighing. Weighings,
used to calculate dust/fabric resistance coefficients (K values), did
not represent efficiency because of the relatively large amount of dust
that settled in the filter chamber sections. Photomicrographs were
taken of the dust/fabric surface at various times throughout the filter
cycle. Since pertinent information from the limited number of runs on
the bench-scale filtration apparatus was fairly qualitative in nature,
it is not presented as a data summary in an appendix.
SINGLE-COMPARTMENT BAGHOUSES--SUPPLEMENTAL STUDY
Additional runs were conducted on 9 fabrics with structural
properties differing from those of the previously tested 123 fabrics
14
-------
INCLINED
MANOMETER
DUST
FEEDER
FILTER CHAMBER
ROTAN1ETER
Figure 3. Bench-scale filtration apparatus.
-------
mainly in the composition of the yarn fibers. These fabrics are
listed in Appendix D with supplemental fabric characterization test
data. The filter bags sewn from these fabrics were 3.58 in. in
diameter and roughly 71.0 in. in overall length, with a 2-in. cuff
on each end. Flow and dust feed were adjusted to maintain dust
loading and filter ratio at the same values used for the previously
tested 123 fabrics. The same four responses were used to describe
the performance of these fabrics.
16
-------
RESULTS AND DISCUSSION
SINGLE-COMPARTMENT BAGHOUSE STUDY--FREE-AREA CORRELATION WITH PRESSURE-
RELATED FILTRATION RESPONSE
An attempt was made to correlate single-compartment baghouse
filtration performance of statistically chosen blocks of fabrics against
structural features of the fabrics such as weave, warp count, fill count,
and yarn type (staple or continuous). The results of this type of
correlation were poor.
An attempt was then made to correlate filtration performance of
groups of fabrics against more fundamental structural features shown by
the supplemental test data in Appendix E (available on request). Clean
cloth Frazier permeability (CFM at 0.5 in. HpO) was found to correlate
with pressure performance (APf, S , K), but not with outlet concentration
(CQ). A "free area", calculated for each fabric, was found to provide a
better correlation with pressure performance and a weak correlation with
the outlet concentration of the filter fabrics.
The free area for each fabric was calculated from yarn count and
yarn width data as follows:
FA (free area) = 1 - (WCWW + FcFw - W^J^FJ = [1 - cover factor]
where
subscript c = count, threads per in.
subscript w = yarn width, in.
W = warp
F = fill
The free area is simply a measure of the total pore area of the
fabric, measured from a direction normal to the fabric's surface (i.e., a
projected pore area). A cover factor greater than 1 for either the warp
or the fill yarns alone implies a free area of 0.0.
17
-------
This section of the report presents results and a discussion of
filtration performance of four groups of fabrics. Each group was
selected to study the effects of specific fabric structural variables
on filtration performance. The four groups of fabrics and the
structural variables for each group are:
Group 1 - Warp count, fill count, and yarn type (continuous
filament or staple fiber).
Group 2 - Fiber (Dacron or nylon) and yarn type (continuous
filament or staple fiber).
Group 3 - Fill count and napping.
Group 4 - Weave and yarn type (continuous filament or staple
fiber).
Group 1 Fabrics
The first group of fabrics analyzed is listed in Table 1, along
with filtration performance data and calculated free area values. The
pressure-related performance data are plotted against calculated free
areas in Figures 4 and 5. Figure 6, showing outlet concentration,
indicates an obvious lack of a consistent relationship. This first
group of fabrics consists of a balanced set with a high and a low
warp and fill count for continuous-continuous, continuous-staple, and
staple-staple warp-fill yarn combinations, respectively (Table 1).
The effective drag, final pressure drop, and the resultant K
values are inverse functions of the free area. Very low free areas,
implying very small straight-through flow paths for the fabrics
tested, lead to more erratic results. Data for all fabrics except
fabric 34 appear to correlate well against free area. There is some
doubt of the accuracy of the yarn width measurements on fabric 34
(Frazier permeability of 113.0 CFM at 0.5 in. H20), because its
permeability and performance are very similar to those of fabric 33
(Frazier permeability of 84.4 CFM at 0.5 in. HpO); however, the
calculated free area is quite different.
18
-------
Table 1. FREE AREA AND PERFORMANCE DATA — GROUP 1 FABRICS (DACRON)
Fabric No.
(Refer to
Appendix A)
5a
8a
9a
15a
30b
33b
34b
> 41b
116C
118 + ngc
2
73C
88C
Thread
Count
W x F
(threads/in.)
64 x 54
64 x 82
76 x 54
76 x 82
64 x 54
64 x 82
76 x 54
76 x 82
64 x 54
64 x 82
76 x 54
76 x 82
Yarn Width (in.)
Warp
0.01543
0.01490
0.01303
0.01323
0.01380
0.01350
0.01350
0.01242
0.00886
0.00830
0.00839
0.00894
Fill
0.01490
0.01357
0.01436
0.01700
0.00945
0.00864
0.00972
0.00864
0.00890
0.01051
0.00945
0.01083
Calculated
free area
0.002
0.0
0.002
0.0
0.057
0.039
0.0
0.017
0.249
0.074
0.178
0.036
Outlet
concentration
(grains/1 Q3ft3)
88.7
1.92
50.3
8.10
175
268
273
33.5
275
140
154
106.5
A Pf
(in. H20)
3.12
4.31
4.23
4.34
0.70
1.69
1.46
3.21
0.25
0.60
0.52
1.08
aContinuous filament 250/50 warp and fill
in SH 0\
\ fpm* )
0.48
0.78
0.69
0.80
0.11
0.26
0.13
0.56
0.03
0.10
0.09
0.18
K
/in. H20/fpm*\
I lb/ft2 )
* /
9.14
8.60
10.97
8.50
2.06
5.00
7.03
7.20
1.10
1.50
1.80
3.15
Continuous filament 250/50 warp — staple 250 equivalent denier fi
cStaple 250 equivalent denier warp and fill
* fpm = ft/min.
ll
-------
ro
o
5.0
4.0
o
CM
I. 3.0
a.
g
2.0
1.0
0.02
i r
0.04
0.06
0.08 0.10 0.12
FREE AREA, in.2/in.2
0.14
i r
0.16
0.18
1.0
0.8
o
CNI
0.6 •£
C9
a
0.4 >
fe
U_
UJ
0.2
0.20
Figure 4. A P{ and Se versus free area-group 1 fabrics.
-------
IV)
o
CM
12
10
^ 8
o
I 4
O
LU
D
0.02
i i
D '
i r
0.04
0.06
0.08 0.10 0.12
FREE AREA, in.2/in.2
0.14
0.16
0.18
0.20
Figure 5. K versus free area—group 1 fabrics.
-------
ro
ro
"2. 300
CD
S,
-i 200
o
K
O
O
100
0.02
0.04
0.06
0.12
0.08 0.10
FREE AREA, in.2/m.2
Figure 6. CQ versus free area-group 1 fabrics.
0.14
0.16
0.18
-------
Group 2 Fabrics
A second group of fabrics was analyzed in the same manner as
Group 1. The second group of fabrics was not a balanced set: it
represented extreme limits in warp and fill counts and contained both
Dacron and Dacron-nylon fabrics with various yarn types. The pressure-
related performance data for these fabrics (Table 2) are plotted
against calculated free areas in Figures 7 and 8. The outlet concen-
tration data again showed a weak correlation and are not plotted.
Outlet concentration will be discussed in greater detail following
the section on microscopic analysis of the filtration process.
The effective drag, final pressure drop, and resultant K values
are again shown to be inverse functions of the free area with more
erratic results occurring as the free area approaches zero. Considering
the range of construction variables shown by these fabrics, the
correlation is fairly good. The presence of either staple or continuous
filament nylon fill yarns does not appear to cause a significant
shift in any of the performance responses.
Group 3 Fabrics
The third group of fabrics represented a wide range of fill counts
for staple Dacron fabrics, both unnapped and napped. The performance
data and supplemental test data for these fabrics are given in Table 3.
Pressure-related performance data for these fabrics are plotted against
calculated free areas in Figures 9 and 10.
The correlation between pressure-related performance and calculated
free area for these fabrics is very good. The data points representing
napped fabrics are identified. It can be seen that all of the pressure-
related responses are slightly lower for the napped fabrics. Except
for the fabric 91 data point, the trend is to lower the average overall
pressure drop for the napped fabrics by roughly 0.12 in. H^O without
23
-------
Table 2. FREE AREA AND PERFORMANCE DATA — GROUP 2 FABRICS
Fabric No.
(Refer to
Appendix A)
Dacron
Series
la
29a
38b
120C
115C
79c
Nylon fill
Series
45a
48a
50a
52a
54b
56b
Thread
Count
W x F
(threads/in.)
59 x 54
80 x 82
76 x 73
59 x 54
80 x 87
76 x 68
64 x 54
64 x 82
76 x 63
76 x 82
64 x 63
64 x 82
Yarn width fin.)
Warp
0.01622
0.01270
0.01269
0.00969
0.00862
0.00878
0.01593
0.01458
0.01296
0.01269
0.01269
0.01377
Fill
0.01739
0.01220
0.00945
0.01031
0.00933
0.00925
0.01701
0.01188
0.01458
0.01134
0.01026
0.00972
Calculated
free area
0.003
0.0
0.012
0.189
0.059
0.123
0.0
0.002
0.002
0.002
0.067
0.025
Outlet
concentration
(grains/103 ft.3)
42.8
2.30
84.1
261
35.8
185
100
30.3
32.5
17.6
269
100
A Pf
(in.H20)
2.88
5.40
3.30
0.17
1.86
0.67
2.37
4.98
5.27
5.01
0.90
1.53
Se
/in.H20\
fpm* |
0.42
1.12
0.51
0.01
0.30
0.10
0.33
0.84
0.95
0.96
0.11
0.25
K
/in.H20/fpm*\
( Ib /ft* j
9.04
6.09
10.90
1.12
5.05
2.01
7.70
12.00
10.70
8.00
3.25
3.80
ro
Continuous filament warp and fill
Continuous filament warp ~ staple fill
°Staple warp and fill
*fpm = ft/min.
-------
ro
in
0.0
0.04.
0.06
0.14
0.08 0.10 0.12
FREE AREA, in.2/in.2
Figure 7. APj and Se versus free area-group 2 fabrics.
0.16
0.18
0.20
-------
12
f—r
10
t>o
o
CM
UJ
O
ft
ae.
o
o
o
LU
a.
0.02
0.04 0.06 0.08 0.10 0.12
FREE AREA, m.2/in.2
Figure 8. K veisus free area—group 2 fabrics.
0.14
0.16
0.18
0.20
-------
Table 3. FREE AREA AND PERFORMANCE DATA ~ GROUP 3 FABRICS
Fabric No.a
(Refer to
Appendix A)
73
74b
76
77b
79
80b
82
3 83b
85
86b
88
89b
91
92b
Thread
Count
W x F
(threads/ in.)
76 x 54
76 x 54
76 x 59
76 x 59
76 x 68
76 x 68
76 x 73
76 x 73
76 x 78
76 x 78
76 x 82
76 x 82
76 x 87
76 x 87
Yarn width (in.)
Warp
0.00839
0.00858
0.00854
0.00772
0.00878
0.00874
0.00894
0.00933
0.00846
0.00909
0.00894
0.00913
0.00839
0.00898
Fill
0.00945
0.00941
0.00870
0.00937
0.00925
0.00980
0.00941
0.01000
0.00976
0.01004
0.01083
0.00988
0.00949
0.01039
Calculated
free area
0.177
0.171
0.170
0.185
0.124
0.112
0.100
0.079
0.085
0.067
0.036
0.058
0.063
0.031
Outlet
concentration
(grains/103ft3)
154.0
78.2
170.4
225.6
182.0
61.6
229.0
149.0
242.0
43.4
106.5
129.0
131.0
46.6
A Pf
(in.H20)
0.52
0.54
0.75
0.57
0.65
0.86
0.92
0.58
1.00
0.90
1.08
1.00
1.58
1.01
Se
fpm*
0.09
0.06
0.10
0.06
0.10
0.13
0.14
0.07
0.14
0.14
0.18
0.13
0.18
0.15
K
in.H20/fpm
lb/ftz
1.80
2.39
2.76
2.42
2.00
2.72
3.11
2.38
3.68
2.85
3.15
3.96
6.26
3.45
Nap
thickness
(In.)
—
0.0228
—
0.0209
—
0.0197
0.0144
0.0232
—
0.0244
___
0.0244
—
0.0195
All fabrics staple Dacron 250 equivalent denier warp and fill
Napped fabrics
* fpm = ft/min.
-------
ro
00
4.0
o
CM
j 3.0
a.
o
2.0
1.0
0.0
0.02
0.04
0.06
0.08 0.10 0.12
FREE AREA, in.2/in.2
0.14
O NAPPED APf
D UNMAPPED AP,
A NAPPED Se
• UNMAPPEDSe
0.16
0.18
0.20
0.15 §
CM
0.10
a:
o
LU
0.05 t
0.20
Figure 9. A Pf and Se versus free area—group 3 fabrics.
-------
VO
I 6
UJ
o
E *
^J *)
E Z
o
ut
a.
VI
0.02
0.04
,NAPPED
UNMAPPED
I I
I I I
0.06
0.08
0.14
0.10 0.12
FREE AREA, in.2/in.2
Figuie 10. K versus free area-group 3 fabrics
0.16
0.18
0.20
-------
causing any significant differences in the rate of pressure increase
as a function of time for napped versus unnapped fabrics. In other
words, nap does not appear to affect the K values. A possible
explanation may be that all the fabrics are fairly light (4.72 - 5.96
2 2
oz/yd versus industrial fabrics generally ranging 5-14 oz/yd ) with light
napping. Actual nap thickness averages 0.01 in.; e.g., the comparison
between the nap thickness for fabric 82 (unnapped) and that for fabric 83
(napped) in Table 3. A nap of 0.01 in. is really a light fuzz; it
would not be considered a nap on a commercial filter fabric. These light
naps result from using a yarn with a twist of 15-20 turns per in. which
does not lend itself to napping. The overall effect of a light nap was
to reduce average pressure drop slightly and increase efficiency
significantly. These effects are probably due to the slight decrease
in the number of yarn fibers as a result of the napping process and the
nap or fuzz providing deposition sites for the dust which allowed fewer
unobstructed large flow paths. This is discussed at greater length
following Figure 15.
Group 4 Fabrics
The fourth group of fabrics represented six different weaves in
both continuous filament and staple Dacron fabrics. The performance
data for these fabrics are given in Table 4 and shown graphically in
Figures 11 and 12. There appears to be a good correlation between
free area and pressure-related performance for these fabrics. There
do not appear to be any significant variations in performance solely
as a result of weave, except for the much higher pressure drops
produced by the plain weave fabric (not plotted). -Outlet concentration
data once again poorly correlate with free area and are not plotted;
however, they will be discussed in following sections. These data are
not adequate for a detailed examination of weave effects because the
fabrics tested lie at the extremes of free area values studied.
30
-------
Table 4. FREE AREA AND PERFORMANCE DATA — GROUP 4 FABRICS*
Fabric No.
(Refer to
Appendix A)
nb
17b
18b
19b
20b
21 b
94C
97c
100C
103C
106C
109C
Weave
3 x 1
3x2
2x2
Plain
Satin
Crowfoot
3 x 1
3x2
2x2
Plain
Satin
Crowfoot
Yarn width (in.)
Warp
0.01303
0.01404
0.01377
0.01296
0.01296
0.01620
0.00894
0.00878
0.00807
0.00882
0.00846
0.00795
Fill
0.01410
0.01620
0.01620
0.01350
0.01647
0.01458
0.00898
0.00965
0.00882
0.00917
0. 00898
0.00909
Free
Area
0.001
0.0
0.0
0.002
0.0
0.0
0.139
0.130
0.172
0.139
0.155
. 0.169
Outlet
concentration
(grains/103 ft3)
49.1
2.10
3.79
3.69
6.89
2.91
352
141
332
187
427
194
A Pr
f
(in. H20)
4.77
3.93
4.16
7.34
4.37
4.25
0.66
0.79
0.53
0.60
0.50
0.66
c
be
/in. H20\
( fpm* /
0.83
0.72
0.82
1.57
0.83
0.83
0.08
0.13
0.07
0.09
0.07
0.09
^in. H20/fpm*^
\ Ib2/ft2 )
10.60
7.46
6.45
6.79
7.51
6.82
2.60
2.73
2.17
2.18
1.81
2.37
All fabrics have 76 x 63 thread count
Continuous filament 250/50 warp and fill - Djicron
°Staple 250 equivalent denier warp and fill - Dacron
* fpm = ft/min.
-------
5.0
co
r\>
4.0
o
CXI
3.0
O
UJ
K 2.0
UJ
1.0
0.02
0.04 O.Ofi 0.08 0.10 0.12
FREE AREA, in.2/m.z
0.14
0.16
0.18
1.0
0.8
o
CM
0.6 .s
^
IS
0.2
0.20
Figure 11. A Pf and Se versus free area—group 4 fabrics.
-------
GJ
12
10
,
0.
CO
i r
i i
i i r
i i i
0.02 0.04 0.06 0.08 0.10 0.12
FREE AREA, in.2/tn.2
Figure 12. K versus free area—group 4 fabrics.
0.14
0.16
i i
0.18
0.20
-------
BENCH-SCALE FILTRATION STUDY-MICROSCOPIC ANALYSIS OF THE FILTRATION
PROCESS
A bench-scale test series was conducted to further develop an
understanding of fabric-structure filtration-performance relationships
with emphasis on outlet concentration. Four fabrics were chosen,
representing the entire range of free areas seen, for a detailed study
2
on the bench-scale 1-ft filtration apparatus. These four fabrics
had shown consistent performance and uniform structure in the test
studies discussed previously. A description of these fabrics, along
with some of the supplemental test data, is given in Table 5. Results
of the study, as well as its implications concerning the previously
presented data, follow.
A series of photomicrographs was taken of four different fabrics
(Table 5) from the group of 123 fabrics. These fabrics were all 3 x 1
twill weave, but differed in thread count and yarn type (continuous or
staple). These differences caused a range of free areas in the fabrics.
An analysis of the data has been made, partly on the basis of theoretical
interstice analysis of woven fabrics done by Backer. The four basic
interstice types are shown in Figure 13. In the following discussion,
the words "pore" and "interstice" are interchangeable.
2
A set of runs was made on a 1-ft bench-scale filtration apparatus
(Figure 3). Each clean piece of fabric was run at standard conditions
(4 ft/min, 3 grains/ft3, 30% R.H., and 70°F). No initial break-in was
used because attempts to manually clean fabric samples were not
reproducible. The test dust was a fine fly ash with d = 4.9pm by
Coulter Counter analysis. Each fabric was removed and photomicrographed
after 2, 4, 6, 8, 14, 22, and 30 minutes of filtration. A description
of the photomicrographs and implications of the data follow.
Fabric 015
Initial deposition occurred almost entirely at Type II pores. Deposits
continued to form at these interstices until joining together occurred. A
34
-------
Table 5. FABRIC DESCRIPTION DATA FOR MICROSCOPIC ANALYSIS AND DUST/FABRIC COMBINATION STUDIES
Fabric
No.
015
038
088
n 120
Yarn3
Warp
Continuous,
250 denier
Continuous,
250 denier
Staple, 210
equivalent
denier
Staple, 210
equivalent
denier
Fill
Continuous,
250 denier
Staple, 210
equivalent
denier
Staple, 210
equivalent
denier
Staple, 210
equivalent
denier
Yarnb
count
76 x 82
76 x 73
76 x 82
59 x 54
Yarn width (in.)
Warp
0.01323
0.01269
0.00894
0.00969
Fill
0.01701
0.00945
0.01083
0.01031
Calculated
free areac
0.001
0.0109
0.0370
0.1898
Permeability
(CFM at 0.5 in. H20)
15.0
55.0
89.0
432.0
Average pore
dimension
x (u)
0-10
17-35
50-60
^196
Dacron yarn
As woven
ft,
Free area = (1 - cover factor)
Yarn counts are not exact due to weaving procedures and handling. Average pore dimensions are calculated as a range
using the yarn counts "as woven" and the actual yarn counts measured later in supplemental tests.. These average pore
dimensions are calculated from yarn width and yarn count data assuming square pores of dimension x.
-------
Ill
IV
Figure 13. The four interstice types.
36
-------
uniform dust cake was formed at t = 14 minutes. Initial pressure drop
was high (APi = 1.27 in. H20), relative to the other fabrics tested in
this series. Pressure increase was gradual, giving APf = 1.62 in. HJO
and K = 2.98. Very limited deposition occurred at Type III pores due
to the spreading out of the continuous-filament yarns when unrestrained
by an adjacent cross yarn (refer to Appendix C on weave yarn interactions),
No deposition occurred at Type I pores. Weight gain was roughly 14 grams
out of 24.6 grams fed, or 56.9%.
Fabric 038
Initial deposition occurred almost entirely at Type II pores. Dust
continued to deposit at these pores, forming very dense deposits. The
pores appeared to have a significant depth relative to the total fabric
thickness. Dust deposition remained below the fabric surface until t = 8
minutes and, thus, no joining of deposits started until this point. The
deposits gradually joined together, aided by some deposition on the few
fibers projecting above the fabric's surface. A relatively uniform dust
cake did not appear to form until t = 22 minutes. Initial pressure drop
AP. = 0.59 in. H«0 and a steep rise in the AP versus t curve gave AP^ =
1 t. i\, T
1.06 in. HpO and K = 4.37. Weight gain was 11.8 grams out of 22.5 grams
fed, or 52.3%.
Fabric 088
Initial deposition occurred mainly at Type II and III pores, with
less uniform deposits forming at Type I pores. Deposits built up
gradually both in and around the pores and on fibers projecting above
the surface of the fabric. A rough, fairly continuous dust cake was
formed at t = 22 minutes. Pinhole leaks occurred through the fabric
and dust throughout the 30-minute cycle. Initial pressure drop, AP.,
was 0.18 in. H20. Pressure increase was gradual, giving APf = 0.4 in.
HpO and K = 2.94. Yarn spacing was not uniform, causing a wide range
of pore dimensions. Weight gain was 8.2 grams out of 22.47 grams fed,
or 36.5%.
37
-------
Fabric 120
Initial deposition occurred very slowly. Dust built up gradually,
bridging over the smaller pores first. Bridging and dust buildup were
significantly affected by individual fibers projecting above the fabric's
surface and across individual pores to which large deposits of dust
became attached. A continuous dust cake never formed due to the presence
of a number of extremely large pores (-\400um) which the dust could not
bridge. Pinhole leaks occurred throughout the 30-minute cycle. Deposi-
tion occurred without preference to pore type. Initial pressure drop
was very low, AP. = 0.03 in. H^O. Pressure rose slowly to AP- = 0.10 in.
H20, giving K = 1.22. Weight gain was 6.3 grams out of 22.5 grams fed,
or 28.0%.
During experimental runs with a standard d = 4.9um fly ash test dust,
the analysis showed that pinhole leaks occurred through fabrics 088 and
120, but not through fabrics 015 and 038. These fabrics had a range of
free areas of from ^ zero to 18.9%. Table 5 shows the free area and an
average pore dimension for each fabric, calculated from yarn width
measurements and yarn count. Yarn counts measured in the supplemental
fabric testing varied slightly from the fabric specifications as
ordered. The range for pore sizes given in Table 5 indicates the
difference in the calculated values using the "as ordered" count or
the measured yarn count. (Note: It should also be noted that both
dust size distributions and pore size distributions are polydisperse
and that an average or mean value does not completely describe such size
distributions.) Billings and Milder have postulated that a given dust
can bridge roughly 10 particle diameters. The test dust used for the
experimental work was fly ash with d = 4.9um by Coulter Counter analysis.
Fabrics 015 and 038 have typical average pore dimensions of less than 10
particle diameters for the d = 4.9um dust. Fabrics 088 and 120, on the
other hand, have typical average pore dimensions of more than 10 particle
38
-------
diameters. As mentioned earlier pinhole leaks were not observed through
fabrics 015 and 038, but were observed through fabrics 088 and 120. To
check the hypothesis on dust bridging, another test with d = 10pm was
prepared from a coarser blend of the same fly ash. This test dust was
run with fabric 088 in a manner identical to the other bench-scale
experimental runs. The following results were observed.
Fabric 088 (d = IQym Test Dust)
Initial deposition occurred mainly at Type II and III pores with
somewhat less heavy deposits at Type I pores. Dust built up fairly
rapidly on the surface of the fabric and on projecting fibers to form
a rough, fairly continuous dust cake at t = 10 minutes. Some pinhole
leaks were observed throughout the 30-minute cycle. Initial pressure
drop, AP., was 0.24 in. H?0. Pressure drop increased gradually, giving
APf = 0.66 in. HpO and K = 2.88. Weight gain was 17.35 grams out of
24.45 grams fed, or 71.0%.
A comparison was made between the photomicrographs of fabric 088
using d = 4.9pm and d = 10pm test dust. Measurements were made of the
minimum pinhole size over the entire 30-minute cycle by manually measuring
pinhole dimension shown in photomicrographs at lOOx.
Table 6 shows the smallest pinhole size measured for each dust at
each time interval, as well as the average over the cycle. The average
minimum pinhole dimension over the 30-minute cycle was 48pm with fly ash
of d = 4.9pm, and 100pm with fly ash of d = 10pm. These measurements,
although approximate, show strong support of a dust's ability to bridge
roughly 10 particle diameters under the aerodynamic conditions existing
in a fabric filter. All other factors being equal, the increased dust
retention with the coarser dust suggests reduced bleeding through the
fabric because of fewer pinhole leaks.
39
-------
Table 6. MINIMUM PINHOLE SIZES DURING A FILTER CYCLE
Time (min.)
2
4
6
8
14
22
30
Average
Dust size (microns)
4.9ym
40
40
50
50
60
—
—
48
10pm
70
no
70
50
150
120
130
100
40
-------
SINGLE-COMPARTMENT BAGHOUSE STUDY--DUST-SIZE/PORE-SIZE CORRELATION WITH
FILTER EFFICIENCY
Filtration Performance of Dust/Fabric Combinations
To add further support to the proposed relationship between dust
size, pore dimensions, and bleeding phenomena, additional runs were
conducted using the single-compartment baghouses to study the effect
of changing the dust type on the relation between pore dimensions and
outlet concentration. Fabrics 015, 038, 088, and 120 (Appendix A)
were tested using: fly ash dust with a mass mean particle diameter of
3.7ym; limestone dust with a mass mean particle diameter of ^18.5Mirr,
and amorphous silica dust with a mass mean particle diameter of i!7.0um.
Standard test conditions were used: filter ratio of 4.0 ft/min; inlet
dust concentration of 3.0 grains/ft ; relative humidity of 30% at ^70°F;
and a 20-minute filter cycle with 24 hours of break in. Filtration
performance results are shown in Table 7 for these three dusts.
Comparing the outlet concentration data for the three dusts tested
with the rough pore dimensions given in Table 7 shows that none of the
dusts tested produced significant outlet concentrations unless the
fabric on which they were being filtered had a rough average pore
dimension on the order of 10 times the mass mean particle diameter.
Pressure drops and K values differed drastically between these dusts.
A comparison between the outlet concentration data in Table 7 for
limestone and silica dust indicates that, once a threshold is reached
where significant bleeding may occur, the magnitude of the outlet
concentration may be partially dependent on the pressure drop during
the filtration cycle. In addition, it should be noted that K values
may vary considerably as a function of fabric while outlet concentration
remains relatively constant. This indicates that pressure-related
performance may be optimized while maintaining high efficiency.
41
-------
Table 7. FILTRATION PERFORMANCE DATA FOR DUST/FABRIC COMBINATIONS
Dust
Fly ash,
d = 3.7um
Limestone,
d = 18.5wm
Amorphous
silica
d = 17.0wm
Performance response
Outlet concentration
(grains/103ft3)
AP, (in. H,0)
Se in. H20
ft/mi n
v /in. H,0/(ft/min)\
I
1 lb/ft2 /
Outlet concentration
(grains/103ft3)
APf (in. H20)
S /in. H00\
e f 2
|ft/min |
„ /in. H,0/(ft/min)\
N f f. \
\ Ib/ft2 )
Outlet concentration
(grains/103ft3)
APf (in. H20)
S/in. H20
[ft/min
K /in. H?0/(ft/min)\
I Ib/ft2 /
Average pore dimensions
(microns)
Fabric No.
015
8.0
4.34
0.80
8.50
8.62
7.86
1.37
15.32
2.96
15.74
2.21
43.52
0-10
038
84.0
3.33
0.51
10.95
4.51
5.79
0.92
13.98
2.86
11.33
1.58
33.18
17-35
088
106
1.09
0.18
3.15
11.4
3.21
0.44
11.49
1.48
6.45
0.61
27.19
50-60
120
261
0.18
0.01
1.12
142
0.77
0.40
4.93
272
2.56
0.01
18.36
•x/196
42
-------
Another interesting observation concerns the pore types (Figure 13)
at which dust deposition occurred. Backer indicated that the four main
interstice types can be used to predict clean-cloth permeability if a
fabric is considered to have a close-packed warp; i.e., if adjacent warp
yarns touch. The relative minimum void cross-sectional areas from the
close-packed warp model indicate minimum flow through Type III pores with
greater flow through Type I, II, and IV pores, respectively. The fabrics
containing continuous-filament warp yarns approximate a close-packed warp
model because of the high warp yarn cover factor resulting from very low
yarn twist. The fabrics with continuous-filament warp yarns showed dust
deposition (indicating gas flow) at the various pore types in proportion
to the relative flows indicated by the close-packed warp model. The
fabrics containing staple warp yarns, on the other hand, represent a
cylindrical yarn model because of the lower yarn cover factors resulting
from high yarn twist. The fabrics with staple warp yarns showed less
uniform dust deposition (indicating gas flow) which did not correspond
well to the relative flows indicated by a close-packed warp model.
Theoretical analysis, such as that done by Backer , provides a
baseline for understanding flow through fabrics. Assumptions about
fabric structure, such as the close-packed warp model, are useful if
tempered with an understanding of the effects of weave and yarn type.
Constraints on yarn cross-section dimensions, such as the twist on staple
yarns, strongly affect structure with respect to flow and filtration.
A more detailed discussion of effects of weave and weave yarn interactions
is presented in Appendix C.
It was noted that the free areas and the pores of the 123 fabrics
studied were much smaller for the continuous-filament fabrics, than for
the staple yarn fabrics, due to the former's higher yarn cover factors.
Yarn width and thickness data, along with outlet concentration data, for
selected fabrics are presented in Table 8. Although the weight of fiber
per unit area is roughly the same for both the staple and continuous-
filament fabrics, the continuous-filament fabrics are much more efficient
43
-------
Table 8. RELATIVE YARN DIMENSIONS FOR CONTINUOUS-FILAMENT VERSUS
STAPLE YARNS COMPARED TO OUTLET CONCENTRATION DATA
Fabric
No.
on
019
020
094
103
106
Yarn
construction
Weave (Dacron)
3 x 1 Continuous-
filament,
Plain 250
denier
Satin
3 x ] Staple,
Pla1n equivalent
Satin den1er
Average
yarn
width
(in.)
0.01356
0.01296
0.01296
0.00894
0.00882
0.00846
Average
yarn
thickness
(in.)
0.00467
0.00456
0.00457
0.00728
0.00634
0.00705
Fabric
weight
o
(oz/yd1^)
4.66
4.42
4.60
5.13
5.16
4.83
Outlet
concentration
Q O
(grains/10*3 ft3)
49.1
3.69
6.89
352.0
187.0
427.0
44
-------
filters due to the more elliptical yarn cross sections which provide a
higher yarn cover factor and a more uniform positive barrier to dust.
This difference in yarn cross-sectional shape results from the high
twist on the staple yarns relative to the very low twist on the
continuous-filament yarns.
Cross-sectional dimensions of continuous-filament yarns are more
subject to change from adjoining cross yarns than are those of staple
yarns. This implies an interaction between yarn type and weave that
influences relative flow through the various pore types (refer to
Appendix C).
The previous discussion of bench-scale filtration tests reveals
fabric structural features responsible for filtration performance.
For the fabrics tested, flow and dust deposition are primarily at
Type II pores. The free area, a measure of the average minimum pore
area in a normal direction (calculated from yarn width measurements)
is a good indication of the pressure-related performance. The weave
interacts somewhat with the yarn to determine actual flow patterns.
Pores which are open to straight-through flow are subject to bleeding
as they increase in size; a critical pore size for bridging purposes
occurs at roughly 10 d of the dust being filtered. Thus, the pore
size distribution in a direction normal to the flow is largely
responsible for outlet concentration.
\
Filtration Performance of Commercial Fabrics
The previous paragraph does not consider fiber properties, such
as composition and surface characteristics. To check the effect of
fiber composition, a group of nine commercially available fabrics was
purchased and fabricated into filter bags of dimensions as stated in
the equipment and procedures section. These bags were then run on
the single-compartment baghouses, using the d = 4.9um fly ash test
dust and filtration conditions as specified previously. Performance
data are given in Table 9 for these fabrics.
45
-------
Table 9. FILTRATION PERFORMANCE DATA -- FABRIC COMPOSITION STUDY
Fabric
No.
1-39703
2-39707
3-39577
4-4589
5-33106
6-39704
7-889
8-4388
9-4400
Yarn
Warp
250 d. Dacron,
type 55
250 d. Dacron,
type 55
200 d. Oral on
Spun acrylic,
12.00/1
210 d. poly-
propylene
250 d. Dacron
type 55
Spun rayon,
21.00/1
Filament
polyester,
250 d./l
Spun Nomex
16.50/2
Fill
250 d. Dacron
type 55
250 d. Dacron
type 55
200 d. Oral on
Spun acrylic
12.00/1
210 d. poly-
propylene
250 d. Dacron,
type 55
Spun rayon,
14.00/1
Spun poly-
ester
16.50/1
Spun Nomex
16.50/2
Weave
3 x 1
3 x 1
3 x 1
3 x 1
3 x 1
3 x 1
Satin
3 x 1
Plain
Thread
count
W x F
(threads/in. )
78 x 65
68 x 54
78 x 70
76 x 51
67 x 53
67 x 58
96 x 86
77 x 77
46 x 38
Fiber type
Polyester
Polyester
Acrylic
Acrylic
Polypropylene
(olefin)
Polyester
Cellulose
Polyester
Polyamide
Outlet
concentration
(grains/103ft3)
6.71
52.0
3.92
3.45
422
89.0
38.9
2.92
6.05
*Pf
(in. H20)
6.11
2.35
4.02
2.75
2.33
3.22
1.94
3.82
3.08
Se
/in. H20\
( fpm* |
1.29
0.36
0.79
0.50
0.22
0.49
0.21
0.76
0.54
K
^in. H20/fpm*^
1 lb/ft2 /
6.68
6.55
5.95
5.00
9.48
8.27
7.58
5.74
6.50
*fpm = ft/min
-------
The four fabrics selected previously for a microscopic analysis
of the filtration process had well defined yarn boundaries and the
free areas calculated appeared to accurately describe the pore size
properties of the fabrics. Yarn width measurements conducted on the
nine commercial filter fabrics were inaccurate and did not appear
consistent with microscopic examination of the fabrics. Poorly defined
yarn boundaries and the more generally polydisperse pore size distri-
butions of the commercial fabrics resulted in a poor comparison between
calculated free areas and pore size properties of the fabrics. Filtration
performance of the nine commercial filter fabrics was therefore correlated
against microscopic measurements of fabric pore size properties, as well
as with permeability.
A straight-through pore size distribution was generated for each
fabric by measurements from photomicrographs. For each interstice, a
measurement was made of the minimum cross-section dimension for each
distinct flow channel. A projecting fiber, when present, was considered
a flow channel boundary. Figure 14 shows the minimum cross-section
measurements, A and B, for the two flow channels through the pore.
Dimension B is the maximum distance which a dust must bridge to seal
the pore and have cake law filtration.
The fallacy of using permeability as an indication of performance
can be seen by comparing fabrics 5 and 7 in Table 10. Even though fabric 7
has a higher clean cloth permeability, an examination of photomicrographs
indicates that a great number of fibers project across the pores. The
projecting fibers allow dust to deposit and bridge over the pores in
fabric 7, leading to cake law filtration and high efficiency. The
maximum size of the flow channel, as far as dust is concerned, is greatly
reduced by projecting fibers within the pore. The maximum size of the
flow channel, as far as the gas is concerned, is not drastically affected
by projecting fibers.
47
-------
YARNS
YARNS
Figure 14. Pore dimensions.
48
-------
Table 10. PORE SIZE DISTRIBUTIONS OF SELECTED FABRICS
(Number of pores of specified size for Mimm area.)
Fabric
No.
1
2
3
4
5
6
7
8
9
Pore size (microns)
50pm
2
1
—
—
3
6
—
—
2
60pm
1
—
2
70ym
3
2
80pm
5
?90pm
5+
Outlet concentration
(grains/103 ft3)
6.0
52.0
3.9
3.4
422.0
89.0
39.0
2.8
6.5
Permeability
(CFM at 0.5 in. H20)
18.5
58.8
11.5
33.0
130.0
47.1
146.0
19.6
38.2
49
-------
Pore size distributions were measured manually, using Polaroid photo-
micrographs at 100X for pores (considering projecting fibers as flow
channel boundaries) visible with low substage lighting. Care was taken
to avoid overexposure of photomicrographs which have led to erroneous
measurements.
The pore size distributions of these nine fabrics, for the pores
larger than roughly 10 times the mass mean particle diameter (d = 4.9ym),
are shown in Table 10. The outlet concentration data for these fabrics
show that the fabrics having significant numbers of pores larger than
roughly 10 average particle diameters are by far the least efficient
filters.
The significant factor, when considering outlet concentration, is a
function of pore size distribution over roughly 10 times the mass mean
particle size of the dust being filtered. It may be assumed that during
the course of filter usage a number of the larger pores may be eliminated
due to fiber or yarn rearrangement or that the dust does bridge over even
very large pores but that occasional bridge collapse and bleeding do
occur and are dependent on pore size.
FABRIC PARAMETERS GENERATING SIGNIFICANT DUST/FABRIC INTERACTIONS
The basic equation relating pressure drop to flow through porous
beds (Equation 1) indicates that non-linear AP response with respect
to time at constant dust and gas flow rate could be the result of
changes in o. The filtration taking place could still physically be
cake law filtration but the dust substrate may have changing properties.
At times, the existence of cake law filtration is questioned because
non-linearities are observed in AP vs t curves. Two examples are
discussed here and an explanation is offered.
Figure 15 shows the general form of AP vs t curves for two different
fabrics showing non-linear behavior.
50
-------
TIME (t) OR AREAL DUST LOADING
Figure 15. Non-lmearAP versus t response.
51
-------
Curve A is for fabric No. 038, a fairly light Dacron fabric with
deep pores. Microscopic examination showed that heavy deposits built
up in certain pores and that dust deposition took place mainly at
these pores for a significant portion of the filtration cycle. The
dust gradually formed a cake and a linear AP vs t response resulted.
A dust layer built up and pressure increased until a sudden discontinuity
was noted. This is known as cake collapse but is seen to occur during
the normal process of filtration. Cake collapse can be precipitated by
pressure increases caused by either increased gas velocity or_ by the
continued deposition of dust. Each successive linear portion of the
AP vs t response has a gradually lower K value, whereas the effective
overall K value increases with time because of the discontinuities
in the pressure increase.
Curve B is for a napped Dacron fabric where dust deposition is
intimately tied to the projecting nap fibers. The non-linear AP vs
t curve implies that the dust is being filtered by depth filtration
(i.e., in the nap) and that a, the packing density, is increasing as
more dust is deposited. It is expected that once the dust layer builds
up above the nap-dust layer, the deposition would result in a linear
AP vs t curve and a constant a for dust-on-dust at constant velocity
(barring cake collapse under pressure).
Figure 16 illustrates a possible physical relationship between
fabric structure and dust filtration. For a fabric of simple pores,
a two-dimensional structure is representative.
Given 5% free area and assuming little or no intrayarn flow (this
is confirmed by little deposition on yarns), it can be seen that the
velocity through the pores is roughly 20 times the average face velocity.
It would be unlikely to expect the same a for deposition at 4 ft/min as
at 80 or 100 ft/min. Once the dust builds up above the fabric surface,
filtration takes place either on previously deposited dust (region 2) or
52
-------
in
I
DUST
BUILDUP
I
FLOW
I
REGION 2'
REGION 2
FABRIC
REGION 3
Figure 16. Schematic representation of dust/fabric interactions.
-------
on a dust fiber matrix (region 3). If a dust fiber matrix exists, the
dust eventually builds up to dust-on-dust filtration (region 2'). It
can be seen that there may be significant interactions between the dust
and the filtering media. The dust fabric interactions are further
complicated by the non-uniform flow patterns resulting from a woven
fabric's heterogeneous physical structure. Since a number of physical
interactions indicate probable changes in a, linear AP vs t responses
should not necessarily be expected. Fabric structure can interact
considerably with a dust to lead to unexpected pressure-related
performance (Figure 16). This implies strongly that K values are not
just a function of participate structure. Care must be taken in fabric
selection to guard against unexpected filtration response with respect
to pressure.
An understanding of fabric structural features and their relationship
to filtration performance is useful in filter design, selection, and
application. Bench-scale filtration tests and simple microscopic exam-
ination of filters can be easily applied by potential makers or users
of fabric filter systems to determine boundary conditions for fabric
selection for the specific dust/gas system of interest. Standard
Frazier permeability gives a fair indication of pressure-related filtra-
tion performance for a general fabric. Microscopic examination must be
used to relate dust and fabric parameters both to ensure efficient
filtration and to guard against unexpected interactions.
54
-------
REFERENCES
1. Spaite, P. W. and G. W. Walsh. Effect of Fabric Structure on Filter
Performance. Ind. Hygiene J.: 357-365, July-August 1963.
2. Borgwardt, R. H. and J. F. Durham. Factors Affecting the Performance
of Fabric Filters. EPA. (Paper 29c, presented at 60th Annual Meeting
AIChE, New York, November 29, 1967.)
3. Stephan, D. G., G. W. Walsh, and R. A. Herrick. Concepts in Fabric Air
Filtration. Ind. Hygiene J. 21:1-14, February 1960.
4. Durham, J. F. and R. E. Harrington. Influence of Relative Humidity
on Filtration Resistance and Efficiency of Fabric Dust Filters.
Filtration and Separation: 389-392, July-August 1971.
5. Backer, S. The Relationship Between Structural Geometry of a Textile
Fabric and Its Physical Properties -- Part IV, Interstice Geometry and
Air Permeability. Text. Resj, October 1951.
6. Billings, C. E. and J. Wilder. Handbook of Fabric Filter Technology.
NAPCA Report CPA22-69-38 (NTIS No. PB200-648), Research Triangle Park,
N. C., December 1970.
55
-------
NOMENCLATURE
A filter area, ft2
2
A particle surface area, ft
2
C. coefficient in Kozeny equation, ft/mi n
33 3
C outlet concentration, grains/10 ft or grains/ft
d mass mean particle diameter, microns (pm)
E efficiency
FA free area
2
g conversion factor, ft Ib mass/mi n Ib force
2
g, local acceleration of gravity, ft/min
2
J conversion factor, in. H20/(lb force/ft )
2
K specific cake resistance, (in. ^O/ft/minJ/Ob/ft )
L cake thickness, ft.
AP pressure drop across filter, in. hLO
AP^ initial pressure drop across filter, in. H20
APf final pressure drop across filter, in. 1^0
Q air flow rate, ft /min
R dust feed rate,\grains/min
S effective drag, in. HJVft/min
uf fluid velocity, ft/min
V particle volume, ft
2
w areal dust loading, Ib/ft
fluid viscosity, II mass/ft min
57
-------
UNITS OF MEASURE - CONVERSIONS
Environmental Protection Agency policy is to express all measurements
in Agency documents in metric units. When implementing this practice will
result in undue costs or lack of clarity, conversion factors are provided
for the non-metric units used in a report. Generally, this report uses
British units of measure. For conversion to the metric system, use the
following conversions:
To convert from
°F
ft
ft2
ft3
ft/min (fpm)
ft3/min
in.
in.2
oz
oz/yd2
grains
grains/ft3
Ib. force
Ib. mass
Ib/ft2
in. H20/ft/min
in. HpO/ft/min
lb/ft2
To
Multiply by
°c
meters
2
meters
meters
centimeters/sec
centimeters /sec
centimeters
centimeters
grams
2
grams/meter
grams
grams/meter
dynes
kilograms
2
grams/centimeter
cm. ILO/cm/sec
f (°F-32)
0.304
0.0929
0.0283
0.508
471.9
2.54
6.45
28.34
33.89
0.0647
2.288
4.44 x 105
0.453
0.488
5.00
cm HpO/cm/sec
5
gm/cm
10.24
59
-------
APPENDIX A
DESCRIPTION OF 123 TEST FABRICS
I. Multifilament-Dacron Series (with 250/50 Dacron Warp Fiber)'
Fabric No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34 h r
35,36?,37^
38,39 ,40C
41 h r
42,43D,44C
Description of fill fiber
250/50 Dacron3
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
Staple Dacron (250 equivalent
denier)
M
II
II
3x2 Twill
2x2 Twill
Plain
Satin
Crowfoot
3 x 1 Twill
Thread
count
59 x 54
59 x 64
59 x 73
59 x 82
64 x 54
x 63
x 73
x
x
64
64
64
76
76 x
76 x
82
54
59
63
76 x 68
76 x 73
76 x 78
76 x 82
76 x 87
76 x 63
76 x 63
76 x 63
x 63
x
x
76
76
70
63
54
70 x 63
70 x 73
70 x 82
80 x 54
80 x
80
80
x
63
x 73
x 82
64 x 54
64 x 63
64 x 73
64 x 82
76 x 54f
76 x 63C
76 x 73°
76 x 82,
76 x 87£
61
-------
APPENDIX A (Continued)
Thread
Fabric No. Description of fill fiber Weave count
45 210/34 Nylon3 3 x 1 Twill 64 x 54
46 " " 64 x 63
47 " " 64 x 73
48 " " 64 x 82
49 " " 76 x 54
50 " " 76 x 63
51 " " 76 x 73
52 " " 76 x 82
53 Staple Nylon (210 equivalent " 64 x 54
denier)
54 " " 64 x 63
55 " " 64 x 73
56 " " 64 x 82
57 b " " 76 x 54.
58,59D,60C " " 76 x 63a
61 " " 76 x 73
62 " Plain 76 x 63
63 " 3x1 Twill 76 x 82
64 . 300/50 Rayon " 76 x 63H
65,66D,67C Staple Rayon (300 equivalent " 76 x 63°
denier)
68 ," Plain 76 x 63
69 250/50 Dacron. Fine metal 3 x 1 Twill 76 x 54
wire every 10th yarn in warp.
70 " " 76 x 63
71 " " 76 x 73
72 " " 76 x 82
62
-------
APPENDIX A (Continued)
II. Staple-Dacron Series (250 Equivalent Denier, Dacron Fiber)
Fabric No.
73
76
79
82
85
88
91
94
97
100
103
106
109
112
113
114
115
116
117
118
119
120
121
122
123
74b
K
77
oUi.
83?
86h
89h
92h
Q
98b
ioib
104h
107h
110b
75c
78
81 r
84
87c
90^
93c
96
ggC
102^
105^
108^
nr
Weave Pattern
3 x 1 Twill
Count
3 x 2 Twi11
2 x 2 Twill
Plain
Satin
Crowfoot
3 x 1 Twill
76
76
76
54,
59
68
X
X
X
76 x 73
76 x 78
76 x 82
76 x 87
76 x 63'
76 x
76 x
76 x 63
x
x
x
x
x
80 x 87
64 x 54
64 x 63
64 x 73
64 x 87
59 x 54
59 x 63
59 x 73
59 x 87
76
76
80
80
80
63
63
63C
54
63
73
Conti nuous-fi1ament
Light nap
°Heavy nap
Triple order—enough fabric to produce 30 bags:
nap, and 10 heavy nap
10 regular, 10 light
63
-------
APPENDIX B
STANDARD FABRIC AND YARN CHARACTERIZATION TESTS
I. Fabric Analysis
1. Type of Weave Pattern - 3x1 twill, plain, satin, etc./ASTM D 1910-64
2. Thread Count (warp and fill) - threads/in. /ASTM D 1910-64
3. Fabric Weight - oz/yd2 fabric/ASTM D 1910-64
4. Crimp (warp and fill) - percent/ ASTM D 1910-64
p
6. Air Permeability - cfm air/ft of fabric at 0.50 in. H90 pressure
5. Fabric Thickness - in. /ASTM D 1777-64
Air Permeability -
drop/ASTM D 737.69
7. Bulk Density - grams fabric/ cm fabric
8. Type of Yarn (warp and fill) - continuous filament, staple, or
combination
9. Abrasion Resistance of Textile Fabrics-ASTM D 1175-64T (tested
warp and fill)
a. Flexing and Abrasion Method - number and average number of
cycles required to rupture a specimen, tension and pressure
used, and condition of specimens.
b. Flexing and Abrasion Method - average percentage loss of
breaking strength obtained after abrasion for one or more
specified number of cycles, tension and pressure used, and
condition of specimens.
c. Oscillatory Cylinder Method - percentage loss in breaking
strength.
10. Stiffness of Fabrics (Cantilever Test) - flexural rigidity of
the warp and the filling separately and an overall average/ASTM
D 1388-64
11. Tear Resistance of Woven Fabrics by Falling-Pendulum (Elmendorf)
Apparatus - individual values and average tearing force in grams
for each direction of tear, capacity of tester, puckering, number
of tests rejected/ASTM D 1424-63
65
-------
12. Breaking Load and Elongation of Textile Fabrics (Grab Test) •
average breaking load, etc./ASTM D 1682-64
13. Tearing Strength of Woven Fabrics by the Tongue (Single Rip)
Method (Constant-Rate-of-Traverse Tensile Testing Machines) •
average tearing strength etc./ASTM D 2262-64T
66
-------
APPENDIX B (Continued)
II. Yarn Analysis
14. Yarn Twist - amount and direction, turns/in./ASTM 1244-69T
and (where applicable) ASTM D 1422 and D 1423 (fabrics)
15. Yarn Number - cotton count, worsted count, tex, and denier/ASTM
1244-69T
16. Number of Filaments per Yarn (continuous filament only) - number/
yarn
17. Filament Diameter - in.(measured from logitudinal sections at
500x magnification)
18. Diameter of Staple Fiber - in./for cotton/ASTM D 1444-63
19. Yarn Width and Thickness - in. Width -- microscopic measurement
with occular filar micrometer at 12x, average of 10 measurements.
Thickness - microscopic measurements from embedded sample sections
at 500x.
20. Bulk Density of Yarn - grams fiber/cm yarn (calculated from
linear density, yarn width, and thickness)
o
21. Fiber Density - grams fiber/cm fiber (yarns) (calculated from
vibrascopic and filament diameter measurements)
22. Thickness of Nap - in. (difference between Shirley Thickness
gauge measurements at 0.01 and 1.0 psi)
2
23. Weight of Nap - grams nap/cm cloth area (difference between
cloth weights with and without nap)
24. Bulk Density of Nap - grams nap/cm nap (could not be determined)
67
-------
APPENDIX C
EFFECT OF WEAVE ON FILTRATION PERFORMANCE
Filtration tests were made on the bench-scale filtration apparatus
(Figure 3) with each of three selected fabrics using fly ash test dust.
The three fabrics had different weaves and, thus, different pore type
distributions (Figure 13).
Table Cl. PORE TYPE EFFECTS ON FABRIC BEHAVIOR
Fabri c
No.
015
019
020
Weave
3 x 1
Plain
Satin
Pore types, %
I II III IV
25 50 25 --
100
-- 80 20 —
APf
(in. H20)
1.62
4.75
2.45
APi
(in. H20)
1.27
4.1
1.75
K
in. HpO/fpm
lb/ft2
2.76
4.30
4.18
A description of the photomicrographs for each fabric will be made and
a number of inferences will be drawn.
FABRIC 015
Deposition occurs almost entirely at Type II pores. Deposits pile up
at these points and join together to form a fairly uniform dust cake at
t = 14 minutes. Initial pressure drop was low (&P. = 1.27 in. hLO) and
pressure increase was gradual giving AP, = 1.62 in. hLO and K = 2.76. Very
limited deposition was observed at Type III pores due to the spreading out
of the continuous-filament yarns when unrestrained by an adjacent cross yarn,
69
-------
FABRIC 019
Plain weave fabric is composed entirely of Type I pores. Deposition
occurred uniformly at all Type I pores. Initial pressure drop was high
(AP. =4.1 in. FLO) due to both the nature of the pores and tightness of
the weave. Uniform deposition led to joining of the individual deposits
and formation of a relatively continuous dust cake at t = 8 minutes.
Pressure increase was gradual, giving APf = 4.75 in. HJQ and K = 4.30.
FABRIC 020
This fabric was a four-harness satin; i.e. yarns go over four and
under one as compared to the 3 x 1 twill where a yarn goes over three
and under one. The weave was not a twill and the pore type distribution
was thus 80% Type II and 20% Type III (refer to Figure 13). Cross sections
between cross yarns were very broad (roughly 400 versus 300 urn for the
plain weave). Deposition was fairly uniform along the floats with a
slight decrease at the Type III pores. Initial pressure drop was low
(APi = 1.75 in. H20) with a final pressure drop (APf) of 2.45 in. H20
and K = 4.18. A fairly uniform dust cake was formed at t = 8 minutes.
DISCUSSION
The K values for these fabrics were not significantly different,
implying that construction effects due to weave probably are not important
when considering dust cake characteristics and cake properties. In
previous photomicrographs, a change in K values was noted mainly due to
yarn type and pore depth.
The previous photomicrographs of fabrics with a 3 x 1 twill weave
had shown deposition occurring almost entirely at Type II pores. The
satin weave fabric had a greater fraction of Type II pores and it was
expected that its filtration behavior would reflect this. Actually,
the continuous-filament yarns spread out considerably between cross
yarns and reduced effective pore area (refer to Table 8). The effect
70
-------
of a more elliptical cross section with the continuous-filament yarns
is shown in the clean cloth Frazier permeabilities (Table C2) for these
similar fabrics.
Table C2. CONTINUOUS-FILAMENT FABRICS-PORE TYPE EFFECTS
Fabric
No.
on
019
020
Frazier
permeabi 1 i ty
35.6
2.17
10.8
Yarn count
(threads/in.)
76 x 63
76 x 63
76 x 63
Weave
3 x 1
Plain
Satin
5
Theoretical interstice analysis done by Backer identifies two
theoretical conceptual models of the different pore types (Figure 13).
If the yarns are considered to be cylindrical, the relative minimum
cross-sectional areas for the different pore types indicate increasing
flow through Type I, II, IV, and III pores, respectively. If a close-
packed warp model is assumed, increasing flow would be expected through
Type III, I, II, and IV pores, respectively. Microscopic examination of
fabrics Oil, 019, and 020 show their structure to be somewhat similar to
that of the close-packed warp model; fabric 020 deviates from the model
because of weave/yarn interactions (Table 8); i.e., the continuous-
filament yarn cross sections vary considerably as a function of position
along the yarn. The permeabilities shown in Table C2 for these fabrics
do not reflect behavior predicted by either of the conceptual models.
Microscopic examination of fabrics 094, 103, and 109 shows their structure
to be quite similar to that of the pore models, assuming cylindrical yarns.
The yarns are reasonably cylindrical (Table 8) because of the high yarn
twist (^ 15 turns per in.), and the fabric permeabilities reflect the
relative values indicated by the conceptual model assuming cylindrical
yarns (Table C3).
71
-------
Table C3. STAPLE YARN FABRICS — PORE TYPE EFFECTS
Fabric
No.
094
103
106
Yarn count
(threads/in. )
76 x 63
76 x 63
76 x 63
Weave
3 x 1
Plain
Satin
Permeability
(CFM at 0.5 in. H20)
201
98.4
301
A significant effect of weave and pore type is indicated by the time
needed to establish a relatively uniform dust cake. The 3x1 twill
fabric (Fabric 015) showed deposition almost exclusively at Type II pores.
Type II pores constitute only 50% of the total number of pores through
fabric 015. Fabric 019 (plain weave) showed deposition at all pores
because of its having 100% Type I pores. Fabric 020 (satin weave) showed
deposition mainly at Type II pores (80% of its pores). The time needed
for the deposits at the individual pores to join together and form a
continuous cake is reduced when more deposition sites are available. This
is shown by the approximate time needed to form a continuous cake on the
three fabrics.
Table C4. MINUTES OF FILTRATION TO FORM CONTINUOUS DUST CAKE
Fabric No.
015
019
020
Time to form cake (min.)
14
8
8
72
-------
APPENDIX D
YARN AND FABRIC ANALYSIS
I. FABRIC ANALYSIS-TEST RESULTS
Fabric
style
1-39703
2-39707
3-39577
4-4589
5-33106
6-39704
7-884
8-4388
9-4400
Test la
weight
(oz/yd2)
5.19
4.48
3.73
7.59
3.89
4.37
6.63
6.70
7.41
Test 2a
yarn crimp
(%)
Warp
1.73
0.61
0.96
3.08
3.75
1.13
4.59
1.87
5.95
Fill
4.86
6.15
8.21
11.26
4.82
6.67
3.69
10.18
9.16
Test 3b
Thickness
(in.)
0.0114
0.0114
0.0114
0.0205
0.0118
0.0114
0.0185
0.0165
0.0197
Test 4C
permeability
of air
(ft3/min/ft2)
18.5
58.8
11.5
33.0
130.0
47.1
146.0
19.6
38.2
'ASTM D 1910-64
JASTM D 1777-64
-ASTM D 737-69
73
-------
APPENDIX D (Continued). YARN AND FABRIC ANALYSIS
II. YARN ANALYSIS-TEST RESULTS (DENIER, BULK AND FIBER DENSITY)
Fabric
style
1-39703
2-39707
3-39577
4-4589
5-33106
6-39704
7-884
8-4388
9-4400
Test 9a
denier
Warp
276
269
193
408
251
269
284
276
651
Fill
299
294
209
472
245
294
362
369
646
Test 11
bulk density
(gm fiber/cm yarn)
Warp
0.729
0.729
0.713
0.819
0.585
0.756
0.745
0.706
0.894
Fill
0.847
0.847
0.848
0.851
0.641
0.876
1.330
1.180
0.725
Test 12
fiber density
2
(gm fiber/cm fiber)
Warp
1.43
1.43
1.40
0.96
1.32
1.35
Fill
1.38
1.38
1.39
0.95
1.37
*ASTM D 1244-69T
74
-------
APPENDIX D (Continued). YARN AND FABRIC ANALYSIS
III. YARN ANALYSIS—TEST RESULTS (TWIST, WIDTH, AND THICKNESS)
Fabric
style
1-39703
2-39707
3-39577
4-4589
5-33106
6-39704
7-884
8-4388
9-4400
Test 8a
yarn twist
(tpi-Z)
Warp
3.8
3.8
3.7
12.3
7.3
3.8
12.5
4.6
7.7
10.9b
Fill
4.2
4.3
4.2
14.4
7.6
4.5
14.0
18.6
8.2
10.7b
Test 10
Warp
Width
(in.)
0.0145
0.0169
0.0155
0.0124
0.0143
0.0170
0.0116
0.0121
0.0186
Thickness
(in.)
0.0052
0.0048
0.0038
0.0088
0.0046
0.0046
0.0072
0.0071
0.0087
Filling
Width
(in.)
0.0131
0.0125
0.0123
0.0143
0.0133
0.0125
0.0089
0.0118
0.0176
Thickness
(in.)
0.0056
0.0061
0.0044
0.0085
0.0063
0.0059
0.0067
0.0058
0.0111
*ASTM D 1244-69T, D 1422, and D 1423
J2-ply S-twist
75
-------
BIBLIOGRAPHIC DATA
SHEET
Report No.
EPA-R2-73-288
3. Recipient's Accession No.
4. Title and Subtitle
Relationship Between Fabric Structure and Filtration
Performance in Dust Filtration
5' Report Date
July 1973
6.
7. Author(s)
Dean C.
Draemel
8. Performing Organization Rept.
No.
9. Performing Organization Name and Address
EPA, Office of Research and Development
NERC-RTP, Control Systems Laboratory
Research Triangle Park, North Carolina 27711
10. Projeci/Taslc/Work Unit No.
21 ADJ 51
11. Contract/Grant No.
In-house Report
12. Sponsoring Organization Name and Address
IX Type of Report & Period
Covered
Final
14.
15. Supplementaty Notes
16. Abstracts The report identifies a semi-empirical relationship between clean cloth
fabric structural parameters, dust parameters, and filtration performance. High
outlet concentration caused by bleeding or seepage of dust is a function of the pore
size distribution of the fabric vs. size properties of the dust. A significant number
of pores with a characteristic dimension roughly 10 times the mass mean particle
diameter of the dust being filtered leads to bleeding and seepage of dust. This
conclusion results from studies with three dusts (fly ash, limestone, and silica), a
number of fiber types, and a range of fabric construction variables. Pressure-
related filtration performance can be correlated with clean fabric free area if yarn
boundaries are well defined. Since many yarn boundaries are not well defined, clean
cloth Frazier permeability may be used as an alternative method of correlating
pressure-related filtration performance.
17. Key Words and Document Analysis. 17o. Descriptors
Air Pollution Interstices
Fabrics Woven Fabric
Dust Seepage
Filtration
Fly Ash
Limestone
Silicon Dioxide
Fibers
Yarns
17b. Identifiers/Open-Ended Terms
Air Pollution Control
Stationary Sources
Yarn Boundaries
Frazier Permeability
Baghouse
17e. COSATI Field/Group 13B, 7A, HE, 13H, 14B
18. Availability Statement
Unlimited
19. Security Class (This
Report)
UNCL/
uASSIFIED
Class (Thi!
20. Security Class (This
Pane
[CLASSIFIED
21. No. of Pages
84
22. Price
FORM NTIS-35 (REV. 3-72)
76
USCOMM-DC M932-P7Z
-------
INSTRUCTIONS FOR COMPLETING FORM NTIS-35 (10-70) (Bibliographic Data Sheet based on COSAT1
Guidelines to Format Standards for Scientific and Technical Reports Prepared by or for die Federal Government,
PB-180 600).
1. Report Dumber. Each individually bound report shall carry a unique alphanumeric designation selected by the performing
organization or provided by the sponsoring organization. Use uppercase letters and Arabic numerals only. Examples
FASEB-NS-87 and FAA-RD-68-09.
2. Leave blank.
3. Recipient's Accession Number. Reserved for use by each report recipient.
4. Title and Subtitle. Title should indicate clearly and briefly the subject coverage of the report, and be displayed promi-
nently. Set subtitle, if used, in smaller type or otherwise subordinate it to main title. When a report is prepared in more
than one volume, repeat the primary title, add volume number and include subtitle for the specific volume.
& Report Dote. I .ich repnrt shall carry a date indicating at least month and year. Indicate the basis on which it was selected
(e.g., date of issue, date of approval, date of preparation.
6- Performing Organisation Cede. Leave blank.
7. Authors). Give name(s) in conventional order (e.g., John K. Doc, or J.Robert Dor). List author's affiliation if it differs
from the performing organization.
8* Performing Organ!zotion Report Number. Invert if performing organisation wishes to a.s&ign this number.
9. Performing Organisation Name and Address, dive name, si reel, « iiy, state-, and rip c ode. List no more than two levrls of
an organizational hierarchy. Display the name of the organisation exactly as it should appear in Government indexes such
as USCRDR-I.
10. Pro|ect/Tosk/Work Unit Number. Use the proien, task and work unit numbers under which the report was prepared.
11. Contract/Grant Number. Insirl contract or grant number under whuh report wa.s pri pared.
12. Sponsoring Agency Nome and Address. Include zip code.
13. Type of Report and Period Covered. Indicate interim, final, etc., and, if applicable, dates coveted.
14. Sponsoring Agency Code. Leave blank.
15. Supplementary Notes. Kntcr information not included elsewhere but useful, such as: Prepared in cooperation with . . .
Translation of ... Presented at conference- of ... To hi published in ... Supersedes . . . Supplements .
16. Abstract. Include a brief (200 words or less) factual summary of the most significant information contained in the report.
If the report coniam.s a significant bibliography or literature survey, mention it here.
17. Key Words and Document Analysis, (a). Descriptors. Select from the Thesaurus of l-.ngineering and Scientific Terms the
proper authorized terms that identify the ma|or concept of the research and are sufficiently specific and precise to he used
as index entries for cataloging.
(b). Identifiers and Open-Ended Terms. Use identifiers for project names, code names, equipment designators, etc. Use-
open-ended terms written in descriptor form for those subjects for which no descriptor exists.
(c). COSATI Field/Croup. Field and Group assignments are to be taken from the 1965 COSATI Subject Category List.
Since the majority of documents are mult {disciplinary in nature, the primary Field/Group assignmem(s) will be the specific
discipline, area of human endeavor, or type of physical object. The application^) will be cross-referenced with secondary
Field/Group assignments that will follow the primary posnng(s).
18. Distribution Statement. Denote releasabihty to the public or limitation for reasons other than security for example "Re-
lease unlimited". Cite any availability to the public, with address and price.
19 & 20. Security Classification. Do not submit classified reports to the National Technical
21. Number of Pages. Insert the total number of pages, including this one and unnumbered pages, but excluding distribution
list, if any.
22. Price. Insert the price set by the National Technical Information Service or the Government Printing Office, if known.
FORM NTIS-39 (REV. 3-72) USCOMM-OC I49S2-P72
------- |