EPA-600/7-91-001
January 1991
A SELF-CONSISTENT DEUTSCHIAN ESP MODEL
by
M. G. Faulkner
J. L. DuBard
Southern Research Institute
Birmingham, Alabaina 35255
Cooperative Agreement No. CR 812811
EPA Project Officer
L. S. Hovis
Air and Energy Engineering Research Laboratory
U. S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Prepared for:
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, D.C. 20460
-------
- - TECHNICAL REPORT DATA
(riease read Instructions on the reverse before complet" '
1. REPORT NO. 2.
EPA-600/7-91-001
3.
4. TITLC AND SUBTITLE
A Self-consistent Deutschian. ESP Model
5. REPORT DATE
January 1991
6. PERFORMING ORGANIZATION CODF
7. AUTHOR(S)
M. G. Faulkner and J. L. DuBard
a. PERFORMING ORGANIZATION REPORT NO.
S. PERFORMING ORGANIZATION NAME ANO ADDRESS
Southern Research Institute
P.O. Box 55305
Birmingham, Alabama 35255
10. PROGRAM CLEMENT NO.
11. CON 1 HACT/GHANT NO.
CR 812811
12. SPONSORING AGENCY NAME ANO ADDRESS
EPA, Office of Research and Development
Air and Energy Engineering Research Laboratory
Research Triangle Park, North Carolina 27711
13. TYPfc OF REPORT AND PERIOD COVERED
Final; 10/86 - 12/87
14. SPONSORING AGCNCY CODE
EPA/600/13
is.supplemfntary notes project officer is Louis S. Hovis, Mail Drop 4, 919/541-
3374.
i6.abstract Tjie rep0rt presents a new version of the EPA/Southern Research Institute-
electrostatic precipitator (ESP) model. The primary difference between this and the
standard (Revision 3) versions is in the treatment of the particulate space charge.
Both models apply the Deutsch equation to narrow particle size bands over short ESP
lengths to determine collection efficiency. The standard version estimates space
charge by a formula which predicts an effective mobility for ions and particles and a
reduced ion density for particle charging. The estimated values are used to calculate
the electric field at the plate and the particle charge, both required for the Deutsch
equafcieanHewever-, in the new version the particulate space charge is treated expli-
citly, allowing -the^.interrelation of the particle charge and electric field calculations.
The charge and field calculations are alternated until they become self-consistent
within each length increment of the ESP. Self-consistency occurs when the charge
used for the space charge in the field calculation is the same as that calculated using
the results of the field calculation. The explicit treatment of the space charge direc-
tly relates the particle charge and electric field calculations, and therefore the col-
lection efficiency calculation, to the dust load present in the gas stream. The report
gives operating instructions for the new model.
17. KEY WORDS AND DOCUMENT ANALYSIS
a. DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. cosati Field/Group
Pollution
Electrostatic Precipitators
Mathematical Models
Particles
Space Charge
Collection
Efficiency
Pollution Control
Stationary Sources
Deutsch Equation
Particulate
Collection Efficiency
13 B
131
12 A
14 G
20C
18. DISTRIBUTION STATEMFNT
Release to Public
19. SfcCURITY CLASS (This Report}
U nclas sif Icq.
21. NO. OF PAGES
92
20. SfcCURITY CLASS {Thispage)
Unclassified
22. PHICE
EPA Form Z22C-1 (S-73) B~lO
1
-------
NOTICE
This document has been reviewed in accordance with
U.S. Environmental Protection Agency policy and
approved for publication. Mention of trade names or
commercial products docs not constitute endorsement
or recommendation for use.
ii
-------
ABSTRACT
The electrostatic precipitator (ESP) model developed by Southern Research
Institute (SRI) for EPA provides an acceptable simulation of the performance of
cold-side utility fly ash ESPs with typical values of inlet mass loading. To
increase the accuracy of model predictions in unusual situations, such as high
inlet mass loading or abnormally low current, a revised version of the model has
been developed. The revised model is unique in that it rigorously calculates the
effects of particulate space charge on the interelectrode electric field and on
subsequent particle charging.
This report provides operating instructions for the revised model, with
examples of the input and output data. A discussion of the theory of the model
is also included. Although the revised ESP model was developed on a main-frame
computer, it can be run on an IBM PC-compatible microcomputer.
-------
CONTENTS
Section Page
Abstract -j j -j
Figures vi
Tables vii
1. Introduction 1
2. Conclusions and Recommendations 3
3. Operating Instructions 5
Input Data 5
Video Display 10
Output Data 10
4. Theory 14
5. Comparison of Results 22
6. Evaluation 33
References 39
Appendices
A. Description of Input Data A-l
General Description A-l
Construction of the Basic Data Set A-l
Construction of Shortened Data Sets A-28
Dynamic Rapping Data Set A-30
Opacity Data Set A-33
Unknown Operating Conditions A-34
Reference A-37
B. Output Data File ESPREV.OUT Corresponding to the Input Data
File Shown in Figure 1 B-l
V
-------
FIGURES
Number Pap:e
1 Input data set ESPREV.DAT 6
2 Video display during running of data set ESPREV.DAT 11
3 Logic diagram for the standard version of the ESP
model (Revision 3) 19
4 Logic diagram for the revised version of the ESP model 20
5 Particulate charge calculated by the standard and
revised versions of the ESP model 23
6 Calculated electric fields for the standard and
revised versions of the ESP model 24
7 Total space charge density calculated by the standard
and revised versions of the ESP model 25
8 Particulate and ionic components of the space charge
calculated by the revised version of the ESP model 27
9 Ionic and particulate current densities predicted by
the standard and revised versions of the ESP model 28
10 Kass penetration calculated by the standard and
revised versions of the ESP model 29
11 Mass penetration calculated by the revised version
of the ESP model with 5 percent gas sneakage 30
12 Measured and calculated percent penetration for the
Edgewater Unit 4 ESP 31
13 Particle charging rates predicted by the revised ESP
model for three input dust loads 34
14 ESP percent penetration as a function of inlet dust load .... 35
15 Ratio of measured migration velocity to computed
migration velocity for 10 plants (revised model
without the ramp function) 37
16 Ratio of measured migration velocity to computed
migration velocity for 10 plants (revised model
with revised ramp) 38
A-l Flow chart for the input data logic A-2
vi
-------
TABLES
Table Page
1 Revised ESP Model Input Data 7
2 Glossary of Variables 15
A-l Opacity Related Input Data A-10
A-2 Reduced Effective Negative Ion Mobilities for Various
Gas Compositions A-18
A-3 Values of Viscosity for Air at Various Temperatures
and Water Contents A-26
A-4 User Options When the Operating Voltage and Current are
Not Known A-35
vi 1
-------
SECTION 1
INTRODUCTION
A more general arid more powerful mathematical model of electrostatic
precipitation (ESP) has been developed by Southern Research Institute (SRI). The
standard version of the ESP model developed by SRI for EPA, now in its third
revision (1) , provides an acceptable simulation of the performance of cold-side
utility fly ash precipitators with typical values of inlet mass loading. To
increase the responsiveness of the ESP model to situations having high inlet mass
loading and/or abnormally low corona current, a revised version of the model has
been prepared (2).
Both versions of the model apply the Deutsch equation to narrow particle
size bands over short ESP length increments to determine particle collection
efficiency. In the standard version, the effects of particulate space charge are
estimated by a formula that predicts an effective mobility for combined ions and
particles and a reduced ion density for particle charging. These estimated
values are then used to separately calculate the electric field at the plate,
EeL, and the particle charge, Q, which are required for the Deutsch equation.
The revised version differs from the standard version in that the
particulate space charge is treated explicitly, which allows the interrelation
of the particle charge and electric field calculations. Hie charge and field
calculations are alternated until they become self-consistent within each length
increment throughout the entire ESP. Self-consistency occurs when the charge
used for the space charge in the field calculation is the same as the charge
calculated using the results of the field calculation. The explicit treatment
of the space charge directly relates the particle charge and electric field
calculations, and therefore the collection efficiency calculation, to the dust
load present in the gas stream.
This report describes the revised model. The report covers the following
topics:
operating instructions for the revised model,
a description of the input data, with an example,
a description of the display on the computer screen during the
operation of the revised model,
a description of the output data, with an example,
a discussion of the underlying theory of the revised model,
a comparison of the logic in the revised and standard (revision 3,
published by NTIS) models, and
comparisons of the output data of the revised and standard models
based on data from a utility ESP.
1
-------
Appendix A is the input data chapter from the user manual for revision 3
of the ESP model. This appendix provides in-depth information about each item
of the input data and describes all of the options available to the user of the
ESP model.
Appendix B contains a revised model output data set. The data correspond
to the input data set that is used for an example in this report.
Although the revised ESP model described in this report was developed on
a main-frame computer, it can be run oil an IBM PC-compatible microcomputer.
Because the model performs a large number of mathematical calculateons, it is
recommended that the microcomputer be equipped with a math coprocessor to
minimize the time required for running the model.
2
-------
SECTION 2
CONCLUSIONS AND RECOMMENDATIONS
The revised ESP model represents an improvement over the standard model in
that the calculations of particle charge and collecting electric field required
for the Deutsch equation are not separate but are interrelated such that the
charging and field calculations are made self-consistent in each length increment
of the ESP. This is done by treating the particulate space charge explicitly and
by alternating the calculations of particle charge and field until both become
constant. The algorithms used in the revised model are conceptually rigorous,
with the exception of the continued use of an average interelectrode electric
field in the charging calculation. However, this field is not the voltage
divided by wire-to-plate distance, as in the standard version of the model, but
rather the spatial average of the field values obtained throughout the inter-
electrode space. Considering the random motion of fine particles through regions
of different field strength and the computational problem of keeping track of
particle charge accumulation, this may be the best way to perform the charge
calculation. The revised algorithm contains no estimated effects such as the
space charge adjustment or combined mobility of ions and particles that are found
in the standard model. These features result in a model that is responsive to
changes in dust load as well as electrical conditions. Since there are no
estimates based on full-sized precipitators used in the calculation, the revised
model can be used to calculate efficiencies for unusual cases such as pilot
precipitators with very low currents.
Following the theoretical efficiency calculation, there are two non-
rigorous corrections that are applied to model predictions. These are the
calculation of rapping reentrainment and the correction of small-particle
migration velocities. The rapping correction must be empirical in nature because
no applicable theory exists. That a correction factor is necessary for a
rigorously Deutschian model to match the small-particle migration velocities in
full-size ESPs indicates a shortcoming in the Deutsch theory, possibly due to an
oversimplification in the underlying assumptions of the theory. At this time,
however, there are no competing theories of ESP particle collection that do not
also require empirically derived constants.
The revised model has been tested against the SRI data base of conventional
utility fly ash precipitators to verify that the same answers are obtained as
from Revision 3. However, the revised ESP model has never been thoroughly tested
for the high load conditions for which it was developed. To date, the only ESP
that has operated long enough with LIMB for steady state testing and for which
there exists reliable data is the Edgewater Unit 4 ESP. Neither the standard ESP
model nor the revised model performance predictions were very close to the
measured performance of the Edgewater ESP, for reasons discussed in Section 5 of
this report. A better model test would be a smaller ESP, which would have lower
efficiency than the Edgewater unit. This would ensure that there would be enough
mass penetrating the ESP to minimize uncertainties in the impactor measurements.
It would also reduce the effects of non-ideal conditions such as spontaneous
reentrainment. Therefore, in order to develop a high level of confidence in the
3
-------
revised model, it is recommended that a measurement program be performed using
a smaller ESP.
4
-------
SECTION 3
OPERATING INSTRUCTIONS
The revised ESP model, ESPREV.FOR, is written in Microsoft compatible
FORTRAN and occupies 67,574 bytes of memory. The executable file, ESPREV.EXE,
has been compiled for use with a math coprocessor. For use in systems lacking
a math coprocessor, the program will need to be recompiled with the appropriate
options. ESPREV.EXE occupies 264,298 bytes of memory. It is recommended that
these files be copied onto a hard disk before use.
To run the model, type ESPREV and press enter. The program will prompt the
user for the names of a file containing input data and a file into which to write
the output data. Definitions of the input data are given in Table 1. Appendix
A gives detailed descriptions of the input data. Descriptions of the video
display generated by the model and the output data are given below. Due to the
large amount of data generated by this model, the output data are written to a
file. To obtain a hard copy of the data, it is necessary to print, this file using
a PRINT command. An option for shortening the amount of output data generated
is provided.
INPUT DATA
The revised ESP model reads the same input data format as the standard
version of the model. The instructions for creating a data set, excerpted from
the standard model instruction manual, are given in Appendix A. New data sets
may be created by writing the data into the prescribed data format using an
editor or word processor (store in ASCII format) , or by using the input data file
creation program described in a separate topical report. The file creation
program contains default values for some of the variables, prompts the user for
the data, and writes the data into the correct format. The data set shown in
Figure 1 is used as an example data set for this report. The data are for a
hypothetical four-section ESP.
Although the input data format is the same as for the standard version of
the model, there are changes in how some of that data are used and in what
numbers are required. Table 1 briefly describes the data by line number. For
each line, the data are listed sequentially by variable name. This list is
included to facilitate understanding of Figure 1 and to indicate which data have
been changed from the way they are used in the standard ESP model. Table 1 is
not intended to fully explain the input data. For complete details of the
meaning of each variable, refer to Appendix A. In Table 1, changes in the data
are indicated by N/A (the variable does not apply to the revised model) , NEW
(this datum does not appear in previous versions of the model), or REV (this
variable is used differently in the revised model). A blank in the space used
for N/A, REV or NEW means that there is no change in the datum.
5
-------
19 1 0 0 015.00 0 0 0 8.00 3.50
TEST PLANT 2: 4 SECTION ESP WITH 320 SCA, BITUMINOUS COAL SIZE DISTRIBUTION
1 2 11511 2 0 1113
10
2.000 36.000 99.900 2400.00 100.00 1.200.000270 1.000.15E+070.20E+11
0.050.15 4.00.100.25 4.00.000.00 4.0
0.010 0.070 0.100 0.140 0.220 0.340 0.500 0.700 1.000 1.400
2.200 3.400 5.000 7.000 10.000 14.000 22.000 34.000 100.000
16.300 3.400
410101010
2.70001-+04 4 .3100E+04
3.0000E+00 3.3750E+05
2. 7000E+04 4.2000E+04
3.0000E+00 3.3750E+05
2.7000E+04 3.9800E+04
3 .0000E+00 3.3750E+05
2.7000E+04 3.8200E+04
3.0000E+00 3.3750E+05
99
4.9000E-01
5.0000E+00
7.3700E-01
5.0000E+00
7.4800K-0I
5.OOOOE+OO
9.8100E-01
5.OOOOE+OO
7000E+04
0000E+02
7000E+04
OOOOE+02
7000E+04
0000E+02
7000E+04
0000E+02
7.5000E-02
1.OOOOE+OO
7.5000E-02
l.OOOOElOO 2.
7.5000E-02 4.
1.OOOOE+OO 2.
7.5000E-02 4.
1.OOOOE+OO 2.
4.
2.
4.
5000E+00
3000E-05
5000E+00
3000E-05
5000E+00
3000E-05
5000E+00
3000E-05
1.8000E+01
9.0000E-01
1.8000E+01
9.0000E-01
1.8000E+01
9. OOOOE-Oi
1.8000E+01
9.OOOOE-OI
Figure 1.
Input data set ESPREV.DAT.
6
-------
TABLE 1. REVISED ESP MODEL INPUT DATA
Line 1:
NENDPT
NDATA
NRAP
MODL
NOPRNT
PATHL
NCOMPS
NWAVES
NLAMDA
PGH20
PC02
NOSNK
N/A -
REV -
N/A -
N/A -
N/A -
NEW -
NEW -
NEW -
number of size band endpoints, 21 max, 99 to end program
type of data set, 1 = full, 2-4 shortened
type of rapping calculation (normal rapping only)
type of input data, 0 = normal, 1 = metric, other values N/A
shorter output if 0, expanded if 1
optical path length for opacity, 0 no opacity calculation
index of refraction data are built in
% H20 for calculation of gr/dscf and lb/106Btu. Format F5.2
in columns 31-35.
% 02 to go with % H20. Format F5.2 in columns 36-40.
allows program to skip sneakage calculation in selected ESP
sections for inclusion of prechargers, etc., where no
sneakage is encountered. These data are given as 10 one
place integers 0 or 1 for a maximum of 10 sections. If
NOSNK(I) = 0 or blank, sneakage is calculated for section I.
If NOSNK(I) = 1, sneakage is not calculated for section I.
E.g. 0100 : no sneakage occurs in the second section.
Using _1 has the same effect. Format 1011 in columns
41-50. These data are blank in Figure 1.
Line 2: Title - no change
Line 3:
NEST - N/A - type of E field calculation (always rigorous)
NDIST - - type of PSD data, cumulative mass or mmd
NVI - N/A - type of V-I data (measured data must be used)
NX - REV - maximum number of E-field grid points in X direction (wire-
to-plate) is now 25
number of grid points in Y (gas flow) direction
number of iterations to converge
calculations are for NGALC = 0 regardless of value given in
input data. NCALC = 0 type of data must be used (i.e.,
Line 4, NN, must appear)
if greater than 1, additional data will be read in but
not used
NEFF - N/A - source of rapping dust = last section
NTEMP - N/A - cold-side only
NONID - REV - if greater than 1, additional data will be read in but
not used
NY
NITER - N/A -
NCALC - REV -
NRAPD - REV -
Line 4:
NN
number of increments in the integration of the particulate
charging rate (typically 5)
7
-------
TABLE 1. (continued)
Line. 5:
DL - - dust loading
PL total ESP length
ETAO - N/A - estimated efficiency is not used in this model
DD particle density
EPS - - dielectric ratio
VRATIO - - peak-to-average voltage ratio
US ion mobility
FPATH - - ionic path factor
EBD - N/A - electrical breakdown field strength
RHO - N/A - particulate resistivity
Line 6:
ASNUCK - REV - only the first set of non-ideal condition data (sneakage,
AZIGGY - REV sigma g, and number of gas baffles) are used. If NONID is
AZNUMS - REV greater than 1, additional data will be read but will not be
used.
Lines 7 - 8:
ENDPT - - particle size band endpoints (21 maximum)
Line 9:
D50 - - particle size distribution mass median diameter
SIGMAP - - particle size distribution standard deviation
Line 10:
NUMSEC - - number of ESP sections
LSECT - - number of length increments per section
Lines 11 - 12 (data for section 1):
AS plate area
VOS - - voltage
TCS - - current
WLS - - total wire length
ACS - - corona wire radius
BS - - wire-to-plate spacing
NFS - - number of wires per lane
SYS - - 1/2 wire-to-wire spacing
VGS - - gas volume flow rate
VGASS - - gas velocity
TEMPS - - gas temperature
PS gas pressure
VISS - - gas viscosity
LINCS - - incremental length
8
-------
TABLE 1. (continued)
Lines 13 - 14: data for section 2
Lines 15 - 16: data for section 3
Lines 17 - 18: data for section A
Line 19:
Line 19 is the first line of the next data set. In this case, since only one
data set is used, the only entry in line 19 is 99, the code to indicate that
no more data sets are to be read.
9
-------
VIDEO DISPLAY
While the revised ESP model is running, certain information is requested
from the user and other information is displayed on the computer screen to advise
the user of the progress of the calculation. Figure 2 shows a portion of the
screen display which was created during execution of the example data set. A
line-by-line explanation of Figure 2 is given below.
Line 1. The program is started by entering the name of the executable
version of the program (ESPREV).
Lines 2 - 3. The program does not have the name of a file from which to
read input data. The file ESPREV.DAT has been specified.
Lines 4-5. The program does not have the name of a file into which to
write the output data. The file ESPREV.OUT has been specified.
The program now has the information it needs to complete its calculations. No
further action by the user is required until the simulation is over.
Line 6 - end. The rest of the data displayed on the screen are to reassure
the user that the calculation is progressing and to indicate the current
status of the calculation. The following data are displayed:
INCR - identifies the ESP length increment being calculated,
FIELD - shows how many times the electric field has been
calculated in this increment,
Q - shows how many times the particulate charge has been
calculated in this increment.
Last line. The final line will display the code 11111 to indicate that the
program has terminated normally. At this point, control will return to the
keyboard.
OUTPUT DATA
The ESP model output data are written into the file whose name was supplied
at the beginning of the model run. To obtain a written copy, this file must then
be printed. The following listing describes the output data. Items marked with
an asterisk are not printed if the input variable NOPRNT is zero. The output data
corresponding to the input data in Figure 1 are shown in Appendix B.
1. Input data. A tabulation of the input data is provided for reference.
2. Incremental data. For each length increment in the ESP, the following data
are printed:
Charge suppression factor (if applicable). Tf the normal charging
calculation would result in too much space charge for the convergence
10
-------
C> ESPREV
Please enter an input data file name:
ESPREV.DAT
Please enter an output data file name:
ESPREV.OUT
ITERATION
COUNTERS:
INCR
=
1
FIELD
=
1
Q
=
l
ITERATION
COUNTERS:
INCR
=
I
FIELD
-
3
Q
-
2
ITERATION
COUNTERS:
INCR
-
1
FIELD
=
5
Q
=
3
ITERATION
COUNTERS:
INCR
1
FIELD
=
7
Q
4
ITERATION
COUNTERS:
INCR
-
1
FIELD
-
9
Q
-
5
ITERATION
COUNTERS:
INCR
=
1
FIELD
=
11
Q
=
6
ITERATION
COUNTERS:
INCR
=
2
FIELD
=
1
Q
=
1
ITERATION
COUNTERS:
INCR
-
2
FIELD
-
2
Q
-
2
ITERATION
COUNTERS:
INCR
-
3
FIELD
=
1
Q
=
1
ITERATION
COUNTERS:
INCR
3
FIELD
2
Q
2
ITERATION
COUNTERS:
INCR
37
FIELD
1
Q
1
ITERATION
COUNTERS:
INCR
=
37
FIELD
=
2
Q
=
2
ITERATION
COUNTERS:
INCR
=
38
FIELD
=
1
Q
=
1
ITERATION
COUNTERS:
INCR
=
38
FIELD
=
2
Q
=
2
ITERATION
COUNTERS:
INCR
-
39
FIELD
-
1
Q
-
1
ITERATION
COUNTERS:
INCR
=
39
FIELD
=
2
Q
=
2
ITERATION
COUNTERS:
INCR
=-
40
FIELD
=
1
Q
=
1
ITERATION
COUNTERS:
INCR
-
40
FIELD
2
Q
=
2
Return code 11111
Figure 2. Video display during running of data set ESPREV.DAT.
11
-------
of the electric field calculation, the program reduces the charge
until convergence can occur. If charge suppression occurs, the factor
by which the charge is reduced is printed with a message.
Current density (J)." The final values of the particulate, ionic,
total (particulate + ionic) , and input current densities on the plate
are printed.
Electric field (E).* The final values of the electric field at the
plate and the average interelectrode electric field are printed.
Space charge (Rho).* The particulate (which is uniform) and the
average ionic space charge densities are printed.
Mass removal. The mass of particles per cubic meter of flue gas that
was collected in this increment and up to this point in the ESP are
printed.
Penetration. The percent of the original mass that has penetrated
this increment of the ESP is printed.
3. Sectional data. At the end of each ESP electrical section, mass and
percent penetration data corrected for the effects of gas sneakage are
printed.
4. Charge accumulation data.* The charges accumulated on each particle size
are tabulated for each ESP length increment. There will be places in the
tabulation where the charge is reduced from the previous increment due to
the averaging of charges when the sneakage gas is remixed.
5. ESP Outlet Statistics. In this section, size-dependent values of
efficiency, penetration, and migration velocity; and overall values of
efficiency and penetration are printed. These values include the
corrections for gas sneakage (calculated inside the incremental computation
rather than afterward) , non-uniform gas velocity (sigma g) , and small
particle migration velocity error. (If NOPRNT = 1, the statistics are
printed after each correction is made.) The data are then corrected for
rapping emissions and reprinted.
6. Summary. The final section summarizes the modeling results by printing the
following information:
Data title,
SCA,
Sneakage fraction,
Sigma g (gas flow standard deviation),
Rapping puff size distribution,
Total efficiency and penetration without rapping,
Total efficiency and penetration with rapping,
Dust loadings in gr/dscf and Ib/MBtu (if % H20 and 02 were given),
12
-------
Plume opacity, total and part due to rapping (if PATHL is not 0),
Stack diameter used in opacity calculation, and
Extinction coefficient for calculating opacity for a different stack.
13
-------
SECTION 4
THEORY
The underlying assumptions for the revised version of the model are listed
below. A glossary of the variables used in this section is presented in Table
2.
1. The space charge due to charged particles is constant in a given length
increment and is uniformly distributed in the gas stream. This allows the
development of a rigorously Deutschian model, as these were the conditions
for which the Deutsch equation was derived. The assumption of uniformity
is particularly good for fine particles in a turbulent gas flow. The fine
particles are especially important in ESP modeling as these are the most
difficult particles for an ESP to collect.
2. The space charge due to ions is neither uniformly distributed nor constant
because the ions follow the electric field lines, which are non-uniform in
the interelectrode space.
3. The total space charge density is the sum of the particulate and ionic
space charge densities (p pp + pL). In the revised model, pp and are
treated separately and explicitly, in contrast to the estimated treatment
of a combined ionic and particulate space charge found in the standard
version of the model.
4. The current is ionic except in the laminar boundary layer at the collection
plate. This is due to the assumption that the particles are stationary in
a given length increment. This is a good approximation since the
particulate mobility is several orders of magnitude less than the mobility
of the ions. Since the particles are stationary, only the ionic current
density appears in the current continuity equation (v-^ =- 0) .
5. The ionic mobility, bt, is used only in the calculation of the ionic current
density on the plate (Jp = E^p^) . The mobility drops out of the equations
in the remainder of the interelectrode space.
6. The particulate current density Jp is included when determining the total
current density on the plate (J = Ji + Jp). Jp is computed from the
calculated charges and Deutsch migration velocities of the different size
particles (Jp - E^Q^) .
14
-------
TABLE 2. GLOSSARY OF VARIABLES
E Electric field in the interelectrode space (V/m)
EpL Electric field at the collection plate (V/m)
Eavg Average electric field in the interelectrode space (V/m)
Q Particle charge (C)
p Total space charge density (C/m3)
Particulate component of p
pL Ionic component of p
J Total current density on the collection plate (A/mz)
Jp Particulate component of J
Jt Ionic component of J
Nk Number density of particles in the kth size band (number/m3)
Qk Charge on the particles in the kth size band (C)
u>g Deutsch migration velocity of particles in the kth size band (m/s)
I Measured plate current (amps)
A Plate area (m2)
V Electric potential (volts)
pf Free ion density used for charge calculation (C/m3)
e Electron charge (1.6 x 10"19 C)
B Scale factor to increase or decrease charging rate
bt Ionic mobility (m2/V-s)
e0 Permittivity of free space (C2/N-m2)
15
-------
7. Overall E field convergence is tested using the measured avei-age plate
current density (J = I/A).
8. The calculations of the electric field and the particle charge are
alternated until self-consistency is obtained in each incremental length
before proceeding to the next incremental length in the precipitator. The
self-consistency is determined by comparing the changes in Eavg between
successive field-charge iterations. When the change in the field is
sufficiently small, the calculation is assumed to have converged.
The basic equations of the E field routine have been redefined, starting
with Poisson's equation
(v2V = -p/e0)
and the current continuity equation
(0 = vj,
= v-(p^E).
The equations are solved for V and p on a calculation grid, called the V/p grid,
which stretches from wire to mid-wire to plate. The following notation is used
on the V/p grid:
wire
mid-wire
2
plate
t
ax
i
«- ay -
16
-------
The equation for the potential at point 0 is
2(^2 4 ay2)V0 = a/(V2 + V4) + ^(Vj. + V3) -t \ a/a/p,,,
which is unchanged from previous versions of the model, except that pQ = pi + pp
explicitly. The space charge density equation is
Pa ¦ -K ± / (K2 + e0Expz /s^. + e0Eyp3 /ay) ,
where K - (e0Ex/ax + c0Ey/ay - pp) / 2.
The equation for p is similar to that of the standard version of the model with
the following exceptions:
1. In the revised version, the particulate and ionic space charge densities
are treated explicitly and summed to find the total space charge (p pp
+ Pi)-
2. Mobility does not appear in the revised equations. The only mobility which
could occur in a stationary cloud of charged particles is the ionic
mobility, and it cancels algebraically in the interelectrode space.
3. In the standard version of the model, K does not contain the -pp term that
arises from the explicit treatment of the space charge. The sign before the
square root is taken as negative for negatively charged particles. The
close similarity in the equations allows the use of the existing E field
subroutine structure with only minor changes in the FORTRAN code.
Following the solution of the V/p grid and the subsequent calculation of
Eavg and Epl, the particulate charge is calculated. Charging is calculated by the
rigorous charging procedure used in the standard version of the model. To
obtain correct charging in an iterative charging procedure, the charges on the
particles are reduced to the values from the previous length increment in each
iteration before recharging using the most recent values of Eavg and free ion
density (pf). The value of the free ion density is obtained from
Pf - -Bpi/e,
in which pt is the average ionic space charge on the V/p grid, e is the elemental
charge and B is a factor < 1 that may be used to reduce the charge in a given E
field/particulate charge (E/Q) iteration to avoid nonconvergence of the V/p
calculation. Steps have been taken to ensure that the amount of charge reduction
in a given E/Q iteration is less than in the previous iteration. This results in
slow and steady growth of the particulate charge until full charge is reached
without causing nonconvergence of the V/p calculation. Full charge is defined as
that charge that would be achieved if all of the available free ions were used
for charging.
Once the particles are charged, the program returns to the E field
calculation, where the most recent charge values are used in the space charge
calculation. The calculations of ionic and particulate space charge density,
17
-------
average field and collecting field (E field at the plate), plate current density,
and particle charge are all iterated to self-consistency within each
computational length increment. After the charge calculation in each iteration,
if full charge has been reached (i.e., B 1.0), the average E field is tested
for convergence. If the value of E averaged across the calculation grid is
within 5% of its previous value, the E/Q calculation is assumed to have
converged. In modeling some ESPs, it may prove impossible to build up full
particulate charge without causing V/p nonconvergence. In this case, the
particulate charge is made as large as possible and the E/Q calculation is
assumed to have converged with reduced charging. This usually occurs only in the
first few increments of the ESP. Following the termination of the E/Q
calculation, the collection efficiency is calculated for each particle size using
the Deutsch equation. The number of particles in each size band remaining in the
gas stream is determined, and the efficiency calculation proceeds to the next
length increment in the ESP.
A comparison of the logic used in the standard and revised versions of the
ESP model may be obtained from Figures 3 and 4. Both versions of the model use
iteration to achieve self-consistency in the calculation. In the standard model,
an estimated value of the overall collection efficiency must be furnished as part
of the input data. This value is used to estimate the space charge effects, as
shown in Figure 3. The particle charge and collecting electric field are then
calculated independently of each other in each incremental length. The test of
convergence is made when the overall collection efficiency is compared to the
estimated value. If the two do not agree, the calculated value is taken as the
new estimate and the entire process is repeated until the efficiency values do
agree or the maximum number of iterations is reached.
By comparison, in the revised model, each increment is made fully self-
consistent before the calculation proceeds to the next length increment (Figure
4). Since the space charge effects are developed within each increment, the need
for an estimated efficiency is eliminated. Further, since the E field and
particulate charge calculations are interrelated, the fields and charges in each
increment are consistent with each other. This feature is unique to the revised
ESP model and represents a significant improvement in the state of precipitator
modeling.
The algorithm in the revised ESP model includes corrections for the non-
ideal effects of gas sneakage, non-uniform gas flow, and rapping reentrainment.
Gas sneakage occurs when a portion of the particle-laden gas passes through the
ESP in a non-electrified path, such as along the roof or dust hopper. These
particles are not subjected to charging and collection. The effects of gas
sneakage have been incorporated into the revised model in the following way. At
the beginning of each section, a specified fraction of the particles in each size
band is set aside while the remaining particles are subjected to the
precipitation process. At the end of each section, the set-aside particles are
remixed with the other particles. To account for the reduction in charging due
to sneakage, the particle charge is averaged downward at the end of each section.
18
-------
REPEAT FOR EACH LENGTH INCREMENT
REPEAT
END
START
CALCULATE EFFICIENCY
CALCULATE CHARGE ON
EACH SIZE PARTICLE
ESTIMATE COLLECTION EFFICIENCY
CALCULATE ELECTRIC
FIELD AT PLATE
CORRECT FOB SNEAKAGE. ag. AND RAPPING
ESTIMATE SPACE CHARGE EFFECTS
BASED ON ESTIMATED EFFICIENCY
CHECK IF CALCULATED EFFICIENCY = ESTIMATED
IF NOT. REPEAT EVERYTHING WITH NEW ESTIMATE
Figure 3. Logic diagram for the standard, version of the ESP model (Revision 3).
19
-------
START
!
i
REPEAT UNTIL SELF-CONSISTENT
CALCULATE EFFICIENCY
CALCULATE NEW PARTICULATE SPACE CHARGE
FROM PARTICLE CHARGE
CALCULATE PARTICLE CHARGE FROM ELECTRIC
FIELD AND IONIC SPACE CHARGE
CALCULATE ELECTRIC FIELD AND TOTAL SPACE CHARGE
USING SPECIFIED PARTICULATE SPACE CHARGE
REPEAT FOR EACH LENGTH INCREMENT
CORRECT FOR SNEAKAGE AT END OF SECTIONS
END
CORRECT FOR og AND RAPPING
Figure 4. Logic diagram for the revised version of the ESP model.
20
-------
The effect of a non-uniform gas velocity in an ESP is to reduce the
collection efficiency. In the standard version of the ESP model, this is done
by applying a correction factor to the exponent of the Deutsch equation for each
size band. The correction factor is obtained from an equation which describes
a family of curves that were generated by calculating the efficiency of a pilot
ESP for a range of gas velocities and then summing the efficiencies according to
the distribution of the gas velocities measured in the ESP. For correctness,
such a study should be made for each ESP being studied. However, this is not
practical and the correction factor based on the pilot study has been used with
good results for a wide range of ESPs. Since no better method for performing
this correction is available, the correction is made in the same way in the
revised ESP model.
To correct the efficiency calculation for the effects of rapping
reentrainment, the rapping routine from the standard ESP model has been adapted
to the revised model. This routine was based on a study of utility fly-ash
precipitators. The algorithm adds mass to the ESP outlet emissions, in an amount
proportional to the mass collected in the final ESP section. This additional
mass is distributed among the particle size bands according to a 6-^m-mmd log-
normal distribution.
21
-------
SECTION 5
COMPARISON OF RESULTS
The revised version of the ESP model has been compared to the standard
version on the basis of computed changes along the length of the precipitator in
values of the following parameters:
accumulated particle charge,
average and collecting electric field,
total, ionic and particulate space charge density,
ionic and particulate current density at the plate, and
total particulate mass penetration.
The results of the comparison for a pulverized-coal-fired power plant
precipitator are shown in Figures 5 through 11. The ESP chosen for this
comparison features a 25-cm (10-in.) plate spacing, good electrical power input,
and a moderate dust loading of 3.7 g/m3 (1.6 gr/acf) . The ash has a moderate
resistivity (2 x 1010 ohm-cm) and a particle size distribution that is typical
for fly ashes from eastern bituminous coal. Figure 12 shows a percent
penetration comparison based on measured data for the Edgewater Unit 4 LIMB test.
Figure 5 shows the rate of charge accumulation for three particle sizes.
Although the predicted charging rate is reduced for the revised model, comparison
with experimental data shows that the fit to the data is equally good with either
model. The reduction in the charge can be attributed to the reduced average
electric field used in the charging calculation (shown in Figure 6). The reason
for the field reduction in the revised model is that the average field is
calculated by averaging the electric field over the V/p grid in each length
increment. Figure 6 shows that the average electric field is dependent on the
particulate space charge and decreases as particles are collected. In the
standard version, the value of the average field is obtained by dividing the
applied voltage in each field by the wire-to-plate distance. The steps in the
field curve for the standard version are due to changes in the applied voltage
between fields.
Figure 6 also shows the enhancement of the collecting electric field (the
field at the plate) due to charged particles. This effect is observable because
the revised version of the model calculates the effects of space charge on the
electric field explicitly. The standard version, which does not directly relate
the field calculation to particle charge, does not show this phenomenon. This
is an important effect because the Deutsch migration velocity is directly
proportional to the field at the plate.
Figure 7 compares the total space charge density at the plate (directly
opposite the wire) for the standard and revised versions of the model. In the
standard model, no specific requirements are made on the space charge density
regardlng the number of charged particles in the gas stream. The calculation is
made as if only Ions were present, except that an effective mobility for combined
motions of ions and particles is used instead of the ionic mobility. In the
22
-------
10
,-15
10
-16
ui
U
EC
<
I
o
10
-17
10
r18
6.0 fjtm DIAMETER
-o<
>o
£
,«¦o ^
1.2 Aim DIAMETER
0.6 fim DIAMETER
STANDARD MODEL
OOO REVISED MODEL
10
20 30 40
INCREMENT NO.
5676-108
Figure 5. Particulate charge calculated by the standard and revised versions
of the ESP model.
23
-------
IS
o
E
X
>
U)
EE
a
cc
a
a
HI
IS
o
*-
*
E
>
D
_i
LJJ
Li.
U
CC
h
O
HI
_i
LU
"N>-
E AVERAGE
+ REVISED MODEL
STANDARD MODEL
!~
~
E PLATE
10 20 30
INCREMENT IMO.
56 76-1 1 5
Figure 6. Calculated electric fields for the standard and revised versions of
the ESP model.
24
-------
+ REVISED MODEL
STANDARD MODEL
10"
10"
10 20
INCREMENT NO.
30
56 76-113
Figure 7. Total space charge density calculated by the standard and revised
versions of the ESP model. The calculation is at the plate directly
opposite the wire.
25
-------
revised model, where the particulate and ionic space charge are treated
explicitly, the total space charge density at every point in the interelectrode
space is constrained to be equal to or greater than the minimum charge density
provided by a uniformly-distributed, stationary cloud of charged particles. As
shown in Figure 7 the space charge density starts at a high value due to the
rapid charging of entrained particles and then decreases as particle collection
takes place.
In Figure 8, the total space charge density for the revised model shown in
Figure 7 has been broken down into its ionic and particulate components.
Comparison of the figures shows that in the first half of the ESP, the
particulate space charge density is the major component of the total space charge
density. Only in the last few increments of this example does the ionic space
charge exceed the particulate space charge. These data are not available in the
combined-ions-and-particles treatment used in the standard version of the model,
but may be easily obtained from the revised model which keeps track of ions and
charged particles separately.
In contrast to the dominance of the particulate space charge density, the
current density on the plate is overwhelmingly ionic in nature. Figure 9 shows
this for both the standard and revised versions of the model. Note that the
particulate and ionic current densities are plotted on the same scale,
demonstrating the difference in magnitude between the two components. The
minimum difference is a factor of 10, with the average difference exceeding 2
orders of magnitude.
Figure 10 shows the incremental mass penetration calculated for the standard
and revised versions of the model. The greater collecting electric field in the
revised model produces a greater efficiency (lower penetration) even though the
ultimate particle charge is less than in the standard version. These curves
represent the ideal efficiency only and do not contain any non-ideal effects.
Figure 11 shows the effects of 5% gas sneakage. The remixing of the gas
which had bypassed the electrified portion of the ESP causes a sudden increase
in the number of particles present at the end of each ESP section. This is
clearly shown in Figure 11 by the vertical lines at increments 10, 20, 30, and
40.
Figure 12 shows the results of modeling the data taken at Edgewater Unit
4 during the May 1989 humidified LIMB tests. The data for this modeling were
taken on May 24 (5 fields on line) and May 26 (3 fields on line). The data were
modeled for a single gas condition, 5 percent sneakage with
-------
10"
PARTiCULATE SPACE CHARGE DENSITY
IONIC SPACE CHARGE DENSITY
10
i-5
pj
E
o
CB
Z
LU
G
LU
O
tc
<
''>ir
,!!¦> *fa
10
r6
10
20 30
INCREMENT NO.
40
66 76-112
Figure 8.
Particulate and ionic components of the space charge calculated by
the revised version of the ESP model. The calculation point is at
the plate directly opposite the wire.
27
-------
IONIC
STANDARD MODEL
+ + + REVISED MODEL
PARTICULATE
ui
3 10
10
10
20
30
0
INCREMENT NUMEER
Figure 9. Ionic and particulate current densities predicted by the standard
and revised versions of the FSP model.
28
-------
10"
10-
P)
E
a
2
o
H
<
cc
t
10"
10-5
STANDARD MODEL
~ ~ ~ REVISED MODEL
0 10 20 30 40
INCREMENT NO. 6676-no
Figure 10. Mass penetration calculated by the standard and revised versions
the ESP model.
29
-------
A WITHOUT GAS SNEAKAGE
WITH GAS SNEAKAGE
P5
05
<
tr
H
UJ
LJ
0.
20 30
INCREMENT NO.
5868-34
Figure 11. Mass penetration calculated by the revised version of the ESP model
with 5 percent gas sneakage.
30
-------
<
cc
LU
o.
T
¦ STANDARD MODEL
A REVISED MODEL
O MEASURED
.01
.001
400
500
600
700
800
SCA, f^/kacfm
Figure 12. Measured and calculated percent penetration for the Edgewater Unit
h ESP.
31
-------
Figure 12 shows that for 3 fields, the ESP performed better than predicted
by either model. However, for 5 fields the measured performance was only
slightly better than for 3 fields. Both models indicate that the 5-field
performance should have been about an order of magnitude better than it was.
This may be the result of abnormally high reentrainment of the humidified ash.
The cause of the better-than-predicted penetration with 3 fields is not known.
Some factors which may have contributed to the high efficiency arc listed below.
The ESP had been switched from 5 fields (first field = field 2) to 3
fields (first field = field 4) at the end of the previous day's
testing. There may not have been enough time for a 3-field
configuration to come to equilibrium, resulting in lighter than normal
reentrainment from the third field.
The V-I data for the first field show an unusually low current
density, averaging 0.13 nA/cm2, compared to 4.7 nA/cm2 for the 5-field
configuration. Space charge corona suppression may have contributed
to this. However, no similar reduction was observed in the 5-field
configuration.
The net result is that this was not a good test for model validation. A
similar measurement program on a smaller ESP with less efficiency (so that more
mass could be collected in the impactors), that has been running long enough to
equilibrate, and that does not show abnormal V-I characteristics or high
reentrainment, would provide a better test.
32
-------
SECTION 6
EVALUATION
The primary reason for the development of the revised ESP model was to
provide a precipitator performance model that is responsive to changes in dust
loading. This goal has been met. The preceding data comparisons show that the
revised model clearly demonstrates the effects of its explicit space charge
calculation. The sensitivity of the revised model to changes in dust load is
shown in Figure 13, where the calculated particle charge for a 1.2 diameter
particle is plotted as a function of distance down the precipitator. The plant
which provided the model input data has an abnormally high inlet dust load of
22 g/mJ (9.4 gr/acf). In Figure 13, the dust: load has been arbitrarily reduced
for comparison. The lowest dust load, a value typical of many ESPs, shows high
initial charging followed by a slow increase due to diffusion charging. The
middle value shows a decreased charging rate for the first two length increments,
followed by slow charge increase at about the same level as the curve for low
dust load. For the highest dust load (the actual dust load for this ESP) the
charging rate is greatly reduced in the first three length increments. In the
next five increments, the charging is totally suppressed until the space charge
is reduced by particle collection, after which charging resumes. Similar total
suppression of charging due to high mass loading has been measured on a pilot ESP
at SRI (3) . Note that these are not real data, as the dust load cannot be
changed without changing the electrical operating conditions of the ESP.
However, these hypothetical data illustrate the point that the revised ESP model
reacts to changes in dust load. The standard ESP model gave the same charge for
all dust loads, yielding a curve which closely resembles the curve for the lowest
dust load in Figure 13.
Anolher example of the sensitivity of the revised ESP model Lo changes in
dust loading is shown in Figure 14, where ESP penetrations calculated by the
standard and revised ESP models arc compared. It is expected that as the space
charge decreases, the collecting electric field, and consequently the collection
efficiency, will also decrease. This effect is shown in the figure for both
models. However, changing the dust loading produces much greater variation using
the revised model, where space charge is used explicitly in the electric field
calculation. Note that although the percent penetration will be lower with
higher inlet mass loading, it is usually riot enough to overcome the extra mass,
resulting in higher mass emissions. As in the previous example, these data were
generated by varying the dust load and nothing else in the data sets. The curves
might have different shapes if measured operating data for varied dust loads were
used.
The second reason for revising the ESP model was to eliminate the three
deficiencies that have been identified in the standard model:
1. The space charge effects are not explicitly calculated but are estimated
based on an effective mobility which accounts for fast moving ions and slow
moving particles. The effective mobility is not a composite of mobilities
but is given by an equation that applies only to small particles near the
collection plate.
33
-------
V
V A
i r
x *
~ ~
H i 1 T
~ ~ ~ ~
©
X
0
lil"
<3
2
<
1
o
-------
REVISED MODEL
* STANDARD MODEL
Inlet Dust Load, gr/acf
Figure 14. ESP percent penetration as a function of inlet dust load.
3b
-------
2. The electric field and particle charge calculations are not mathematically
connected.
3. An empirical correction factor must be applied to the average migration
velocities of small particles to make their calculated efficiencies match
measured data.
The first two deficiencies were eliminated by the structure of the revised
model. It was hoped that making the revised model rigorously Deutschian would
remove the need for an empirical correction factor for small-particle migration
velocities (the third deficiency). That this was not the case can be seen in
Figure 15, which shows the ratios of measured migration velocity to migration
velocity calculated by the revised model without the correction factor, for 10
cold-side utility ESPs. These plants were chosen from the SRI F.SP data base (4).
The measured migration velocities were determined from the fractional penetration
data using the Deutsch equation. The calculated data were generated using the
values for gas sneakage and non-uniform gas velocity standard deviation that have
proven to be most typical of the power plants in the data base (0.05 and 0.15
respectively), and include the effects of rapping reentrainment. The curve in
Figure 15 represents a second-order least-squares fit to the data. This curve
is very similar to the curve produced by the same data using the standard ESP
model. By including an empirical correction factor similar to the one in the
standard model, the performance projections can be corrected, as shown in Figure
16.
36
-------
3.0
2.5
2.0
<
a.
0.2 0.3 0.4 0.6 0.8 1 2
PARTICLE SIZE, n
Figure lb.
Ratio of measured migration velocity to computed migration velocity
for 10 plants (revised model without the ramp function). The curve
is a second-order 1 east - squares fit. to the data.
37
-------
<
c
0.5
0.0
0.2 0.3 0.4 0.6 0.8 1 2
PARTICLE SIZE, nm
Figure 16. Ratio of measured migration velocity to computed migration velocity
for 10 plants (revised model with revised ramp).
38
-------
REFERENCES
1. Faulkner, M. G., and DuBard, J. L. A Mathematical Model of Electrostatic
Precipitation (Revision 3). EPA-600/7-84-069a,b ,c. (Volume I, Modeling
and Programming: NTIS PB84-212-679; Volume II, User Manual: NTIS PB84-212-
687; FORTRAN Source Code Tape: NTIS PBS/)-2.32-990) . U. S. Environmental
Protection Agency, Research Triangle Park, North Carolina. 1984.
2. Faulkner, M. C-. , DuBard, J. L. , and Hovis, L. S. A Self-Consistent
Deutschian ESP Model. In: Proceedings: Seventh Symposium on the Transfer
and Utilization of Particulate Control Technology, Volume 1. EPA-600/9-89-
046a (NTIS PB89-194-039). U.S. Environmental Protection Agency, Research
Triangle Park, North Carolina. 1989.
3. Faulkner, M. G., DuBard, J. L., and Sparks L. E. The Effect of High Mass
Loading on Fly Ash Precipitators. In: Proceedings: Sixth Symposium on
the Transfer and Utilization of Particulate Control Technology, Volume 2.
EPA-600/9-86-03Ib (NTIS PB87-147-625). U.S. Environmental Protection
Agency, Research Triangle Park, North Carolina. 1986.
4. DuBard, J. L and Dahlin, R. S. Precipitator Performance Estimation
Procedure. EPRI CS-5040. Electric Power Research Institute, Palo Alto,
California, 1987.
39
-------
Appendix A
DESCRIPTION OF INPUT DATA8
GENERAL DESCRIPTION
The format of the original computer program which performs the
calculations in the model for electrostatic precipitation has been
re-structured to make the inputting of data less cumbersome. The
number of cards which is necessary to input data has been reduced
significantly by allowing different operating conditions to be
analyzed from one basic set of input data. Due to the fact that
several options are available in using the model, the number of
cards and type of information in the input data may vary from one
set of data to the next. Thus, it is necessary for the user to be
familiar with the logic associated with the input data in order to
ensure that the desired operations will be performed.
The input variables may be read into the program entirely in
metric units or in mixed British and metric units depending on the
value of the variable MODL described below. Non-metric data are
internally converted to metric units prior to performing the
calculations. The input variables and format specifications are
discussed in detail in the following section. Where both British
and metric units are given, the British units are to be used when
mixed units are selected and the metric when all-metric input is
selected. Figure A-l contains a flow chart which shows the logic
involved in entering the data.
CONSTRUCTION OF THE BASIC DATA SET
The following is a sequential listing of the variables in the
first group of data which is read in, along with the descriptions
of the variables and the format specifications.
(1) NENDPT is the number of discrete points on the cumulative
percent versus particle diameter curve. NENDPT is
specified by the user and must have a value of at
least 1 but not greater than 21. After the
collection efficiency has been calculated for one
data set, the program will begin reading data for
another efficiency calculation starting with this
data group. If NENDPT has a value of 99, the
program terminates. Thus this data group will be
aThis appendix was taken from the EPA/SRI ESP model report.1
References to Volume 1 in this appendix refer to Volume 1 of the
ESP model report.
A-l
-------
STARTPROGRAM
READ; NENDPT, NOATA, NRAP, MOOL. NOPRNT,
PATHL. NCOMPS. MWAVES, NLAMOA >
YES
MOOL- 2
NO
READ: RHO,HDUCT
YES
TERMINATE
PROGRAM
NO
READ: ITL
NOATA = 1
NDATA - 2
READ: NEST. NDIST. NVI, NX
NY. NITER. NCALC. NRAPO,
NEFF. NTEMP. NONID
NO ^ -s YES
NDATA = 3
MOOL
READ: RHO. HDUCT
ZREAD: fVOSd). TCSfl). /
1 = 1. NUMSEC! /
READ: (VGS(l), VGASS(I)
I - t. NUMSEC)
/REAP: NN./
NVI - 2
READ: tFINAL, Jl 1, JI2
VISKIP. VISAME
4102-2S
Figure A-l.
Flow chart for the input data logic (Sheet 1 of 4)
A-2
-------
YES
NDATA>1
f READ: DL. PL, ETAQ, 00, EPS.
' VRATI0, US, FPATH. EBO, RHO
NO
READ: NUMSEC. (LSECTII
I = 1, NUMSEC!
NO
YES
START 00 LOOP OVER THE NUMSSR
OP ELECTRICAL SECTIONS
READ: !AR050(l!, AfiSIGMII!
I - 2. NRAPOt
READ: AS, VOS. TCS, WIS, ACS, BS. NWS
SYS, VGS, VGASS. TEMPS, PS. VISS, UNCS
READ: (ASNUCK(I), AZiGGY(l)
AZNUMSm, ( = 1, NONID)
YES
READ; (ENDPTIII. 1 = 1, NENDPT
NVI = 1
NO
NO
MOIST - 2
READ: RFS, START!. START2.
STARTS, VSTAR
YES
READ: 0«Q, SIGMAP
END 00 LOOP OVER THE NUMBER
OF ELECTRICAL SECTIONS
NO
MOIST = 1
YES
READ: (PRCU(I). I * 1. NENOPT) /
Figure A-l. Flow chart for the input data logic
A-3
(Sheet 2 of 4).
-------
YES
NRAP= 0
NO
YES
PATHL = 0
NO
READ: TIMRAP, MAPRAP. REPCT,
NUMBOL. NOAT
YES
TiMRAP = 0
NO
NO
MAPRAP(I) « 0
NO
NOAT = 1
YES
YES
READ: BOLFRC
READ: REMMD, RES1G
READ: BQLT1M, DDRAP
YES
PATHL = 0
NO
4 X 0 2-3 2
Figure A-l. Flow chart for the input data logic (Sheet 3 of 4)
-------
YES
NCOMPS- 0
YES
NLAMOA = 0
NO
READ: CM1.CMR
READ: CMI,CMR.UPTOU!/1
YES
NRAP-Q
NO
Figure A-l. Flow chart for the input data logic (Sheet 4 of 4).
A-5
-------
the first group of each data set and will also be
the final card in the input data. If 21
-------
input data. If the program terminates normally,
the calculations may or may not be correct,
depending on the input data and the action taken by
the computer.
(3) NR&P is an indicator which determines the type of
rapping reentrainment correction which will be
performed. If NRAP - 0, the correction is
performed by the same empirical process used in the
previous versions of this model. If NRAP = 1,2, or
3, a different process is used in which additional
data are read into the program. From these data,
the particle statistics of the reentrained dust are
determined. The program then repeats the
efficiency calculation for the input dust plus the
reentrained dust, which is introduced into the gas
flow at that point in the ESP at which it was
rapped loose. The additional data required for
dynamic rapping will be described later in this
section. The values 1,2, or 3 determine the amount
of ESP outlet information which is printed with 1
giving extensive details on individual particle
statistics and 3 giving summaries only. These
effects are described fully in the next section of
this report. NRAP is read in with an 12 format and
must be right justified in columns 5-6.
(4) MODL is an indicator which specifies the type of data to
be read into the program and may have values from 0
to 3. If MODL = 0, mixed British and metric data
are to be used. For MODL = 1, all data must be in
metric units. If MODL = 2, an internal data set is
used and the only other data required from the user
is the resistivity of the input particles and the
wire to plate spacing. This data set is described
later in this appendix. For MODL =3, the program
terminates after printing the voltage-current curves
for the ESP. The options and data sets required when
MODL = 3 are described in the section on unknown
operating conditions. MODL is read in with an 12
format and must be right justified in columns 7-8.
(5) NOPRNT is an indicator which may have the values of 0 or
1. If NOPRNT = 0, the printed output includes data
on particle charging and, if opacity is calculated,
the differential extinction coefficient for each
particle size band. If NOPRNT = 1, particle
charging data is not printed and only total opacity
figures are printed. NOPRNT is read in with an 12
format and must be right justified in columns 9-10.
A-7
-------
is the path length over which the opacity
calculation will be made and is in units of ft or
m. PATHL is read in with an F5.2 format in columns
11-15. If PATHL = 0.0, no opacity calculation is
performed.
is the number of indices of refraction for which
the opacity is to be calculated. NCOMPS may have
values from 0 to 10. If NCOMPS 3 0 the two default
values of the index of refraction are used. These
are 1.50 + O.Oi and 1.50 - 0.10i. If NCOMPS = 1 to
10, the calculation is performed for the values
which are read into the program. NCOMPS is read in
with an 12 format and must be right justified in
columns 19-20.
is the number of wavelengths at which opacity will
be calculated and may have values of 0 or 1. If
NWAVES = 0, the calculation is performed at 10
different points in the spectrum. If NWAVES = 1,
the calculation will be performed at a wavelength
of 0.55 urn only. NWAVES is read in with an 12
format and must be right justified in columns
24-25.
is the number of indices of refraction which must
be read in to perform the opacity calculation using
an index of refraction that varies with wavelength.
NLAMDA may have values from 0 to 10. If NLAMDA =
0, the index of refraction is assumed to be
independent of wavelength and no extra data must be
read in. If NLAMDA >0, then NLAMDA indices of
refraction and their associated wavelengths are
read into the program. If NLAMDA > 0, NCOMPS must
have a value of 1 or 2. If NCOMPS ¦ 0, the value
of NLAMDA is ignored. If NCOMPS > 2, and NLAMDA >
0, NCOMPS is set equal to 2. NLAMDA is read with
an 12 format and must be right justified in columns
29-30.
The overall format for this group is £512, F5.2, 3(3X,I2)].
The data contained in this group is on the first card and this
card must be the first card in each new data set. However, only
the first two parameters, NENDPT and NDATA, are required. The
remaining seven parameters are optional. If only the first two
parameters are given, the operation of the program is the same as
for Revisions I and II except that execution is much faster.
Consequently, data sets for the previous version of this model are
still usable and will result in the same printed output as before.
The use of the optional parameters will result in the inclusion
(6) PATHL
(7) NCOMPS
(8) NWAVES
(9) NLAMDA
A-8
-------
of the extra processes described in the preceding parameter
definitions. The last four parameters (PATHL, NCOMPS, NWAVES, and
NLAMDA) define the opacity calculation and are interrelated as
shown in Table A-l.
Data group 2 is for specifying information which will
identify the data set which is under consideration. All or part
of columns 1-80 on data card 2 can be used in identifying the data
set. The overall format for this card is (40A2). This data group
must be the second card in each new data set.
At this point, the third and successive data groups depend on
the choice of the value of NDATA. The basic data set must be read
into the program before shortened data sets can be used. For
NDATA=1, the program reads in the data groups in the basic data
set in the sequence discussed below.
The following is a sequential listing of the variables in
data group 3, along with the descriptions of the variables and the
format specifications.
(1) NEST is an indicator which can have the values of 1 and
2. If NEST = 1, the program will perform
extensive, detailed calculations in order to
determine precipitator performance. If NEST = 2,
estimation procedures are used to determine
precipitator performance. Both of these options
have been discussed in detail. Use of the
estimation procedure will result in considerable
savings in computer time and can be used to
establish trends or to establish ranges over which
to apply the more rigorous calculations. NEST is
read in with an 12 format and must be right
justified in columns 1-2.
(2) NDIST is an indicator which can have the values of 1 and
2. If NDIST » 1, the user must supply the inlet
particle size distribution corresponding to the
size bands specified by the variable ENDPT. If
NDIST = 2, the program will construct a log-normal
particle size distribution for these size bands
based on the parameters D50 and SIGMAP. The
technique used to construct the log-normal size
distribution is described in Volume 1. NDIST is
read in with an 12 format and must be right
justified in columns 3-4.
(3) an indicator which can have the value 1 or 2.
If NVI -1r the user must supply known or measured
values of the operating applied voltage and current.
A-9
-------
TABLE A-l.
>
i
i'
o
VALUE OF PARAMETER
PATHL NCOMPS NWAVES NLAMDA
0.
>0 0 0
>0
>0 .1,2 0 0
>0 1 to 10
>0 1,2 0 ltolO
OPACITY RELATED INPUT DATA
CALCULATIONS PERFORMED
No opacity calculation performed
Opacity calculated at 10 wavelengths using
default values of index of refraction
Opacity calculated at 1 wavelength using
default values of index of refraction
Opacity calculated at 10 wavelengths using
NCOMPS user supplied values of index of
refraction
Opacity calculated at 1 wavelength using
NCOMPS user supplied values of index of
refraction
Opacity calculated at 10 wavelengths using
NLAMDA user supplied values of frequency
dependent index of refraction
-------
If NVI = 2, the program will construct a voltage-
current curve (or curves) for a specified wire-
plate geometry up to a voltage which is specified
by the user. Refer to the section on unknown oper-
ating conditions (p. A-34) to determine when this
option should be used. NVI is read in with an 12
format and must be right justified in columns 5-6.
(4) NX is the number of grid points in the x-direction
(perpendicular to the gas flow) which is used in
the numerical techniques that determine the
electrical conditions. NX cannot exceed a value of
15. If NVI = 1, sufficient accuracy can normally
be obtained with NX > 1i. If NVI = 2, NX should be
set equal to 15. NX is read in with an 12 format
ana must be right justified in columns 7-8.
(5) NY is the number of grid points in the y-direction
(parallel to the gas flow) which is usee in the
numerical techniques that determine the electrical
conditions. If NVI * 1, sufficient accuracy can
normally be obtained with NY > 9. If NVI = 2, NY
should be set equal to 15. NY is read in with an
12 format and must be right justified in columns 9-
10.
is an indicator which serves two different
purposes. If NVI » 1, the value of NITER
determines the maximum number of iterations the
program will make on a loop which converges on
overall mass collection efficiency. If the overall
mass collection efficiency converges within 0.05%
before NITER iterations, the calculation of
collection efficiencies is completed at this point.
NITER serves the purpose of cutting the calculation
off in a reasonable amount of time when convergence
requires more iterations And computer time than is
warranted. For normal inlet mass loadings and
particle size distributions a value of NITER = 2 is
sufficient. For high inlet mass loadings or very
fine particle size distributions a value of NITER ¦
3 or 4 may be necessary to provide sufficient
accuracy. If NVI = 2, the value of NITER
determines the number of iterations which will be
performed over each incremental length of the
precipitator in order to obtain self-consistent
solutions for the electrical conditions. In its
present stage of development, the calculation
procedure yields the same results for all values of
NITER. Thus, in this case, set NITER = 1. The
A.-11
-------
calculation procedure is discussed in Appendix A of
Volume 1. NITER is read in with an 12 format and
must be right justified in columns 11-12.
(7) NCALC is an indicator which can have the values of 0 and
1. If NCALC = 0, particle charge is determined by
using equation (12) in Volume 1. Due to the number
of times particle charge must be calculated and the
use of numerical techniques in order to solve the
charging equation, the particle charging
calculations for NCALC «= 0 take a considerable
amount of computer time. If NCALC = 1, particle
charge is estimated empirically by using the sum
of the charges predicted from classical field and
thermal charging theories [see equation (15) in
Volume 1]. In this case, particle charge can be
determined very rapidly from analytical
expressions. Thus, in those cases where a
significantly shorter run time is more important
than the best accuracy possible, NCALC should be
set equal to 1. If NEST = 2, particle charge will
be performed as if NCALC s 1 regardless of the
value of NCALC. NCALC is read in with an 12
format and must be right justified in columns
13-14.
(8) NRAPD is an indicator which specifies the number of
rapping puff particle size distributions which will
be utilized by the program in predicting the effect
of rapping reentrainment on overall mass collection
efficiency. NRAPD must have a value'of at least 1
and can not exceed a value of 10. If NRAPD = 1,
the program will determine the rapping puff
particle size distribution based on the average of
data obtained from several field tests on full-
scale precipitators. These tests yield an average
rapping puff particle size distribution with a mass
median diameter (HMD) of 6.0 um and a geometric
standard deviation (
-------
(9) NEFF is an indicator which can have the values of 1 and
2. If NEFF = 1, the total mass reentrained at the
outlet due to rapping is determined from the mass
collected in the last field under adjusted no-rap
conditions. If NEFF = 2, the total mass
reentrained at the outlet due to rapping is
determined from the mass which would be collected
in the last field under unadjusted ideal
conditions. NEFF should normally be taken to be 1
since this case is physically meaningful. A value
of NEFF = 2 will result in rapping losses which are
significantly greater than for NEFF = 1. Thus, a
value of NEFF = 2 should only be used when a
precipitator design which is conservative with
respect to rapping losses is desired. NEFF is read
in with an 12 format and must be right justified in
columns 17-18.
(10) NTSMF
is an indicator which can have the values of 1 and
2. The mass reentrained due to rapping will differ
for cold-side ana hot-side precipitators. If NTEMP
= 1, the mass reentrained due to rapping is
estimated based on an equation for cold-side
precipitators. If NTEMP = 2, the mass reentrained
due to rapping is estimated based on equation for
hot-side precipitators. NTEMP is read in with an
12 format and must be right justified in columns
19-20.
(11) NONID is an indicator which specifies the number of
combinations of normalized gas velocity standard
deviation (
-------
variables in the next data group which is read in, along with the
descriptions of the variables and the format specifications.
(1) NN is the number of increments in the Runge-Kutta
integration of equation (12) in Volume 1. A value
of NN = 5 will provide sufficient accuracy for
normal use. Greater values found in data sets
created for previous versions of the model may be
used but will result in an increase in computer
time for no appreciable increase in accuracy. NN
is read in with an 12 format and must be right
justified in columns 1-2.
(2) NUMINC was previously the number of increments in the
Simpson's Rule integration over 6 in equation (12)
in Volume 1. Since this integration is now done by
the Gaussian Quadrature method, this variable is no
longer used. Data sets containing this variable
may still be used, however, as the space allowed to
NUMINC, columns 3-4, is not used for any other
purpose.
The overall format for this data group is (12) and all the
data are contained on a single card. If NCALC » 1, the above data
group is not read into the program.
If NVI = 2, the model must calculate a voltage-current curve.
In this case, the following is a sequential listing of the
variables in the next data group which is read in, along with the
descriptions of the variables and the format specifications.
(1) IFINAL is an indicator which causes the calculation of
successive points on the voltage-current curve to
cease after IFINAL points. This indicator allows
the user to have the voltage-current calculation
terminated at a point before the specified oper-
ating voltage is reached whenever it is taking an
excessive number of points to reach the specified
operating voltage. IFINAL is read in with an 12
format and must be right justified in columns 1-2.
(2) JH is an indicator which allows the initial increment
size on current density in the calculation of the
voltage-current curve to be changed after JI1-1
points are determined on the curve. Since the
voltage-current calculation finds the applied
voltage corresponding to a specified value of
current density, this indicator allows the user to
cover a large range of current densities without
using an excessive number of points. JI1 is read
in with an 12 format and must be right justified in
columns 3-4.
A-14
-------
(3) JI2 is an indicator which allows the second increment
size on current density in the calculation of the
voltage-current curve to be changed after JI2-1
points are determined on the curve. JI2 serves the
same function as JI1 and JI2 must have a value
greater than JI1 for proper usage. JI2 is read in
with an 12 format and must be right justified in
columns 5-6.
(4) VISKIP is an indicator which may have the values of 0 and
1. If VISKIP = 0, a voltage-current curve will be
calculated up to a specified operating voltage for
each successive length increment of the
precipitator. If VISKIP = 1, only the operating
current density which corresponds to a specified
operating voltage will be calculated based on an
estimation procedure discussed in Volume 1. In
most cases, the user will want to set VISKIP = 1
since this will result in a prediction of the
operating current density in each increment of
length of the precipitator without using the large
amounts of computer time required by the
calculation of a voltage-current curve. Only
extremely detailed analysis would warrant setting
VISKIP = 0. VISKIP is read in with an 12 format
and must be right justified in columns 7-8.
(5) VISAME is an indicator which may have the values of 1 and
2. The proper use of VISAME can result in
significant savings in computer time whenever the
applied voltage is the same in each electrical
section. If the applied voltage is the same in
each electrical section, set VISAMS = 1 and only
one "clean" voltage-current curve-will be
calculated. If VISAME = 1, as many data sets as
desired can be read into the program and all
calculations will be based on the one "clean"
voltage-current calculation. The use of VISAME * 1
is especially beneficial in studying hypothetical
cases due to the large savings in computer time.
If the applied voltage differs from one electrical
section to the next, the user must set VISAME = 2.
Whenever the operating voltage and current are
unknown and the user must specify the use of the
voltage-current calculations (NVI = 2), the
quickest run time will occur when VISKIP = 1 and
VISAME = 1. The longest run time will occur when
VISKIP = 0 and VISAME ¦ 2. VISAME is read in with
an 12 format and must be right justified in columns
9-10.
A-15
-------
The overall format for this data group is (512) and all data
are contained on single card. If NVI = 1, the above data group
is not read into the program.
The following is a sequential listing of the nest, data group
which is read in, along with the descriptions of the variables anc
the format specifications.
{1) DL^ is the inlet particulate mass loading in units of
grains/ft3 or kg/m3. DL is read in with a F8.0
format and must be left justified in columns 1-8.
(2) PL is the total electrical length of the precipitator
in units of ft or m. PL is read in with a F8.0
format and must be left justified in columns 9-16.
(3) ETAO is the overall mass collection efficiency in units
of percent and it has two different interpretations
depending upon the value of NVI. If NVI = 1, ETAO
represents the measured or estimated overall mass
collection efficiency and is used as a test for
convergence in an iteration loop on overall mass
collection efficiency. If NVI = 2, ETAO simply
represents the desired design efficiency and is not
used in the calculations. ETAO is read in with a
F8.0 format and must be left justified in columns
17-24.
(4) DD is the density of the particles in units of kg/m3.
DD is read in with a F8.0 format and must be left
justified in columns 25-32.
(5) EPS is the dielectric constant of the particles for use
in the particle charging calculations and is
dimensionless. Values of EPS must be equal to or
greater than 1. In most industrial applications,
the flue gas is sufficiently humidified that the
particle surface "becomes conductive and a value of
EPS = 100 can be used to simulate a conductor. EPS
is read in with a F8.0 format and must be left
justified in columns 33-40.
(6) VRATIO is the ratio of the peak voltage to the average
voltage and is dimensionless. In the calculation
of particle charge, it is assumed that the
particles will charge to an extent determined by
the peak voltage rather than the average voltage.
For industrial applications, VRATIO has a value
around 1.2. VRATIO is read in with a F8.0 format
and roust be left justified in columns 41-48.
A-16
-------
(7) US is the ionic mobility at standard temperature
(273 K) and standard pressure (1 atm) and is in
units of m2/(V-sec). This mobility is referred to
as the "reduced mobility". Values to use for
reduced ionic mobilities for flue gas compositions
are not well-established at the present time. The
reduced ionic mobility for air is in the range 1.2-
2.1 x 10-l+m2/(V-sec). Reduced ionic mobilities for
flue gas compositions have been reported that are
considerably larger than those reported for air.
These values cover the range of 2.2-5.4 x
10~'+m2/(V-sec). Some reported values of reduced
ionic mobility for various gas compositions are
given in Table A-2. Since the ionic mobility has a
strong influence on the electrical conditions
through the current and electric field
distributions, this is an important parameter in
determining precipitator performance. A value of
2.7 x 10"~^m2/(V-sec) should provide a
representative value to use for flue gases
emanating from coal-fired boiler applications. US
is read in with a F8.0 format and must be left
justified in columns 49-56.
(8) FPATH is a parameter which is used in the field charging
equation and is dimensionless. FPATH represents
the number of ionic mean free paths over which the
momentum of the ions will persist and allow the
ions to reach the surface of the particle even
though the saturation charge has been reached. The
effect of this parameter is to increase the
saturation charge. FPATH normally should have a
value in the range 0-2. It is recommended that
FPATH be assigned a value of 1. FPATH is read in
with a F8.0 format and must be left justified in
columns 57-64.
(9) EBP is the electrical breakdown strength of the gas or
the particulate layer in the region near the plate
and is in units of V/m. The value of this
parameter is a strong function of the resistivity
of the collected particulate layer and the
condition of the collection plates. At present,
mathematical techniques which are based on physical
principles do not exist for predicting the value of
BBD under differing conditions. Thus, experimental
data and prior experience must be used to choose
appropriate values of EBD. In practical
applications, EBD falls in the range of 2-15 kV/cm.
A value of 2 kV/cm should provide a conservative
estimate of EBD whereas a value of 15 kV/cm would
A-17
-------
TABLE A-2
REDUCED EFFECTIVE NEGATIVE
ION MOBILITIES
FOR VARIOUS GAS COMPOSITIONS
Reduced Effective
Gas Composition
Ion Mobility
(Volume Percent)
{cnr/V
-sec)
Mi
COz 02_ §Oz_
h2o
100.0
0-67 +
0.17a
100.0
2.46 +
0.06b
100.0
1.08 +
0.03b
100.0
0.35C
(Laboratory Air)
1.03d
(Laboratory Air)
1.26 -
1.96e
79.4
14.7 4.6 0.2
0.6
5.39f
73.5
13.6 4.2 0.2
8.4
2.93f
65.9
12.2 3.8 0.2
17.8
2.23f
71.0
11.2 3.7 0.0
14.0
2.35f
75. 7
11.6 3.2 0.0
9.4
3.02f
75.1
11.5 3.2 0.1
9.9
f
2.74
78.5
10.9 3.6 0.0
7.0
3.36f
78.3
19.8 3.6 0.1
7.0
2.67f
77.9
10.8 3.6 0-3
7,0
2.Z0f
77.6
10.7 3.7 0.7
7.0
2.43f
. tl.
J. Lowke and J. A.
Rees,
Australian J. Phys. 16, 447 (1963).
b. E.
W. McDaniel and H.
R. Crane, Rev.
Sci. Instru. 28, 684 (1959).
c. E.
W. McDaniel and M.
R. C.
McDowell,
Phvs.
Rev. 114, 1028 (1959).
d. B.
Y.E. Liu, K. T. Whitby, and H.H.S.
Yu, J-
Appl. Phys. 38,
1592 (1967).
e J.
Bricard, M. Cabane,
G. Modelaine,
and D.
Vigla, Aerosols
and Atmospheric Chemistry. Edited by G. M. Hidy, New York#
New York, 27 (1972).
f. H. W. Spencer, III, "Experimental Determination of the Effective
Ion Mobility of Simulated Flue Gas." In Proceedings of 1975
IEEE-IAS Conference, September 26, 1975, Atlanta, Georgia.
A-18
-------
in most cases provide the most optimistic value.
The value of EBD is used whenever NVI = 2 and a
voltage-current curve is generated. If the field
at the plate exceeds the value of EBD at any point
on the curve, a message to this effect is printed
out with the voltage-current calculation terminat-
ing at the corresponding applied voltage and cur-
rent density. These values of voltage and current
are then used in the projection of precipitator
performance. EBD is read in with a F8.0 format
and must be right justified in columns 65-72.
(10) RHO is the resistivity of the collected particulate
layer and is in units of ohm-cm. The resistivity
to be used must be determined experimentally by
either in situ or laboratory methods. RHO is used
in the model only if the operating conditions are
unknown. For this use refer to the section on
unknown operating conditions (p. A-34). RHO is
read in with an 8.2 format and must be right
justified in columns 73-80.
When the operating voltages and current are
specified, the value of RHO has no effect on
the operating conditions. To estimate the
effects of a change in RHO, the procedures
listed under unknown operating conditions
must be used.
The above data group has an overall format of (9F8.0, E8.2)
ana is contained on a single data card. This data set must be
read in with each basic data set, i.e. when NDA.TA ¦ 1.
The next data group which is read in depends on the value of
NRAPD. If NRAPD is greater than 1, the following is a sequential
listing of the variables in the next data group, along with the
descriptions of the variables and the format specifications.
(1) ARD50(I)
(2) ARSIGM(I)
is an array containing the mass median diameters in
um for log-normal particle size distributions of
the different rapping puff distributions which will
be utilized in the model. The values of this
variable are read in with a F4.0 format and must be
left justified in columns 1-4, 9-12, 17-20, 25-28,
33-36, 41-44, 49-52, 57-60, 65-68, and 73-76.
is an array containing the geometric standard
deviations for log-normal particle size
distributions of the different rapping puff
distributions which will be utilized in the model.
A-19
-------
Values of ARD50(I) and ARSIGM(I) with the same
index are used together to construct a log-normal
particle size distribution. The values of this
variable are read in with a F4.0 format and must be
left justified in columns 5-8, 13-16, 21-14, 29-32,
37-40, 45-48, 53-56, 61-64, 69-72, and 77-80.
ARSIGM(I) can not have a value less than 1.
The above variables must be read in for 1=2 up to I=NRAPD
where NRAPD can not exceed a value of 10. The overall format for
this data group is [10(2F4.0)] and is contained on a single data
card. If NRAPD=1. this data group is not read in. In this case,
only one rapping puff particle size distribution will be considered
where ARD50(1) = 6.0 in and ARSIGK(l) = 2.5. This case is built
into the program and relates to experimental data discussed in
Volume 1.
The following is a sequential listing of the variables in the
next data group which is read in, along with the descriptions of
the variables and the format specifications.
is an array containing different fractions of gas
flow which bypass the electrified region in each
baffled stage of the precipitator and/or different
fractions of the mass collected in each stage of
the precipitator which are reentrainea due to
factors other than rapping. The values of this
variable are read in with a F4.0 format and must be
left justified in columns 1-4, 13-16, 25-28, 37-40,
49-52, and 61-64 of the first two data cards in the
group and in columns 1-4, 13-16, and 25-28 of the
third data card in the group. ASNUCK(I) must lie
in the range 0.0 to 1.0.
is an array containing different normalized
standard deviations for the inlet velocity
distribution of the gas flow. The values of this
variable are read in with a F4.0 format and must be
left justified in columns 5-8, 17-20, 29-32, 41-44,
53-56, and 65-68 of the first two data cards in the
group and in columns 5-8, 17-20, and 29-32 of the
third data card in the group. AZIGGY(I) must be
equal to or greater than 0.0.
is an array containing the number of baffled stages
in the precipitator. The values of this variable
are read in with a F4.0 format and must be left
justified in columns 9-12, 21-24, 33-36, 45-48, 57-
60, and 69-72 of the first two data cards in the
group and in columns 9-12, 21-24, and 33-36 of the
third data card in the group. The values of
AZNUMS(I) must be whole numbers.
(1) ASNOCK(I)
(2) AZIGGY(I)
(3) AZNUMS(I)
A-20
I
-------
The values of ASNUCK(I), AZIGGY(I), and AZNUMS(I) with the
same index are used together to simulate one set of nonideal
parameters and to produce one set of no-rap efficiencies. The
values of I are determined by NONID which must have a value of at
least 1 and can not exceed a value of 15. Thus, at least one set
of these parameters must be read in. It is recommended that the
user take the first set of these variables to be ASNUCK(I) = 0.00/
AZIGGY(I) = 0.00, and AZNOMS(I) = actual number of stages so that
efficiencies under ideal conditions will be obtained. In
practical situations, a well-operating precipitator will have
values of ASNUCK and AZIGGY of approximately 0.1 and 0.25,
respectively.
The overall format for this data group is [6{3F4.0)] and the
data group is contained on 3 or less cards. For NONID < 6,
6
-------
The next data group which is read in depends on the value of
ND1ST. If NDIST =2, the following is a sequential listing of the
variables in the next data group, along with the descriptions of
the variables and the format specifications.
{1) D50 is the mass median diameter of a log-normal inlet
particle size distribution and is in units of urn.
The value of D50 must lie between 0.01 and 1,000
um. The value of D50 is read in with a F8.0 format
and must be left justified in columns 1-8.
(2) SIGMAP is the geometric standard deviation of a log-normal
inlet particle size distribution and is
dimensionless. The value of SIGMAP must be equal
to or greater than 1. The value of SIGMAP is read
in with a F8.0 format and must be left justified in
columns 9-16.
The program uses the values of D50 and SIGMAP to construct a
log-normal particle size distribution over the range and size
bands determined by the values of ENDPT(I). Any mass which is not
in the size range .determined by ENDPT(I) will be put into the size
band with the largest midpoint. This must be done to ensure that
the sum over all size bands of the percentage of total mass in
each size band will equal 100%.
The above data group has an overall format of (2F8.0) and is
contained on a single data card. This data set is not read in if
NDIST - 1.
If NDIST =1, the next data group which is read in consists
of a single array. The description of this array and its format
specification are given below.
(1) PRCU(I) is an array containing values of cumulative
percents corresponding to points on a curve of
inlet mass cumulative percent versus particle
diameter. The number of cumulative percents read-
in must equal the value of NEtiDPT which can not
exceed 21. The cumulative percents must match the
particle diameters specified in the array ENDPT(I).
The cumulative percents are entered in units of %.
The first value of PRCU(I) must be 0% and the last
value must be 100%. The program determines the
percentage by mass in each particle size band from
the values contained in ENDPT(I) and PRCC(I). The
user must supply values of PRCU(I) based on
measured or known particle size information for the
particular application under consideration. The
values of PRCU(I) are read in with a F8.0 format
and must be left justified.
a-22
-------
The overall format for this data group is (10F8.0) and the
data group is contained on 3 or less data cards. For NENDPT < 10,
1CKNENDPT < 20, and NENDPT = 21, the number of data cards
necessary are 1, 2, and 3, respectively. This data group is not
read in if NDIST = 2.
The following is a sequential listing of the variables in the
next data group which is read in, along with the descriptions of
the variables and the format specifications.
(1) NUMSEC is the number of electrical sections in the
direction of gas flow. The maximum number of
electrical sections which can be accounted for in
the program listed in Volume 1 is 10. The value of
this variable is read in with an 12 format and must
be right justified in columns 1-2.
(2) LSECT(I) is an array containing values of the number of
incremental lengths to be taken in each electrical
section in the direction of gas flow. These values
are determined by the user with increasing values
of I corresponding to electrical sections moving
from the inlet to the outlet of the precipitator.
. The total number of incremental lengths in all
sections must not exceed 99. It is not necessary
to have the same number in each section. The
values of this variable are read in with an 12
format and must be right justified.
The overall format for this data group is (12, 1012) and the
data group is contained on a single data card. This data group
must be read in with each basic data set.
The following is a sequential listing of the variables in the
next data group which is read in, along with the descriptions of
the variables and the format specifications.
{1} AS(NSECT) is the total collection plate area for a given
electrical section and is in units of ft2 or m2.
The values of this variable are read in with an
El 1.4 format and must be right justified in columns
1-11.
(2) VPS(NSECT) is the applied voltage in a given electrical
section and is in units of volts. If NVI = 1, the
value of VOS(NSECT) corresponds to a measured or
known value. If NVI - 2, V0S{NSECT) corresponds to
an applied voltage up to which a voltage-current
curve will be calculated. Then, this applied
voltage along with the corresponding current will
A-23
-------
be used in the calculation of precipitator
performance. The values of this variable are read
in with an E11.4 format and must be right
justified in columns 12-22.
(3) TCS(NSECT) is the total current in a given electrical section
and is in units of amperes. If NVI = 1, the value
of TCS(NSECT) corresponds to a measured or known
value. If NVI = 2, TCS(NSECT) has no meaning in
terms of input data since it will be calculated in
the program. In this case, the appropriate columns
on the data card can be left blank or any arbitrary
number can be entered. The values of this variable
are read in with an El 1.4 format and must be right
justified in columns 23-33.
(4) WLS(NSECT) is the total effective wire length in a given
electrical section and is in units of ft or m. The
values of this variable are read in with an E11.4
format and must be right justified in columns
34-44.
(5) ACS(NSECT) is the corona wire radius in a given electrical
section and is in units of in. or mm. The values
of this variable are read in with an £11.4 format
and must be right justified in columns 45-55.
(6) BS(NSECT) is the wire-to-plate spacing in a given electrical
section and is in units of in. or m. The values of
this variable are read in with an E11.4 format and
must be right justified in columns 56-66.
(7) NWS(NSECT) is the number of discharge electrodes per given
electrical section per gas passage and is
dimensionless. The values of this variable
normally should not exceed 20. If the values do
exceed 20, use 20 in the program. These values are
used to determine the number of terms in a series-
summation which determines the electrostatic
electric field distribution and 20 terms are more
than sufficient to reach convergence. The values
of this variable are read in with an El 1.4 format
and must be right justified in columns 67-77.
(8) SYS(NSECT) is one-half of the wire-to-wire spacing in a given
electrical section and is in units of in. or m.
The values of this variable are read in with an
El 1.4 format and must be right justified in columns
1-11.
A-24
-------
(9) VGS(NSECT) is the total gas volume flow rate in a given
electrical section and is in units of actual
ft3/min or m3/sec. The values of this variable
are read in with an El 1.4 format and must be right
justified in columns 12-22.
(10) VGASS(NSECT) is the gas velocity in a given electrical
section and is in units of ft/sec or m/sec. The
values of this variable are read in with an El 1.4
format and must be right justified in columns 23-
33. The gas velocity must be consistent with the
cjas volume flow rate and the precipitator geometry
in each electrical section"? These are related by
VGASS(I) = [LINCS(I)*LSECT(I)*VGS(I)]/[AS(I)*BS(I)*Constant],
where the constant = 1 for metric units and 5 for
British units.
(11) TEMPS(NSECT) is the gas temperature in a given electrical
section and is in units of °F or °C. The values of
this variable are read in with an E11.4 format and
must be right justified in columns 34-44.
(12) PS(NECT) is the gas pressure in a given electrical section
and is in units of atmospheres. The values of this
variable are read in with an El 1.4 format and must
be right justified in columns 45-55.
(13) VISS(NSSCT) is the gas viscosity in a given electrical
section and is in units of kg/(m-sec). Table A-3
gives values of viscosity for different
temperatures and water contents for a gas
composition whose components are those of air.
This table provides values of viscosity which cover
most cases found in practice although some
extrapolation is necessary for those cases
involving hot precipitators where temperatures are
greater than 3QQ°C. The values of this variable
are read in with an E11.4 format and must be right
justified in columns 56-66.
(14) LINCS(NSECT) is the incremental length size which will be
taken in a given electrical section and is in units
of ft or m. If NVI * 1, LINCS(NSEGT) should be
given a value of approximately 1 ft although
larger values can be used with, some loss in
accuracy in order to save computer run time. If
NVI » 2, LINCS(NSECT) must be given a value equal
as near as possible to the wire-to-wire spacing in
order for the numerical procedure to be valid. In
any case, the product of LSECT(NSECT) and
A-2 5
-------
TABLE A-3. VALUES OF VISCOSITY FOR AIR AT VARIOUS TEMPERATURES AND WATER CONTENTS*
Percent HzO
fC
0
1
2
3
4
5
6
I
8
9
10
10
1.767
1.758
1.748
1.739
1.730
1.721
1.712
1.702
1.693
1.684
1.675
20
1.81.0
1.801
1.792
1.783
1.774
1.765
1.755
1.746
1.737
1.728
1.719
30
1.854
1.844
1.835
1.826
1.817
1.808
1.799
1.790
1.780
1.771
1.762
40
1.900
1.887
1.978
1.869
1.860
1.850
1.841
1.832
1.823
1.814
1.805
50
1.938
1.929
1.920
1.911
1.902
1.892
1.083
1.874
1.065
1.856
1.B47
60
1.979
1.970
1.961
1.952
1.943
1.934
1.925
1.916
1.907
1.898
1.888
70
2.020
2.011
2.002
1.993
1.904
1.975
1.966
1.957
1.94B
1.939
1.930
80
2.059
2.050
2.042
2.033
2.024
2.015
2.006
1.997
1.988
1.979
1.970
90
2.099
2.090
2.081
2.072
2.063
2.054
2.046
2.037
2.028
2.019
2.010
100
2.137
2.129
2.120
2.111
2.102
2.093
2.085
2.076
2.067
2.058
2.049
110
2.175
2.167
2.158
2.149
2.140
2.132
2.123
2.1.14
2.105
2.097
2.088
120
2.213
2.204
2.195
2.189
2.178
2.169
2.161
2.152
2.143
2.135
2.126
130
2.250
2.241
2.232
2.224
2.215
2.207
2.198
2.189
2.181
2.172
2.164
140
2. 206
2.277
2.269
2.260
2.252
2.243
2.235
2.226
2.218
2.209
2.201
150
2.321
2.313
2.304
2.296
2.288
2.279
2.271
2.262
2.254
2.245
2.237
160
2.356
2.348
2.339
2.331
2.323
2.315
2.306
2.298
2.289
2.281
2.273
170
2.390
2.382
2.374
2.366
2.358
2.349
2.341
2.333
2.325
2.316
2.308
180
2.424
2.416
2.408
2.400
2.392
2.383
2.375
2.367
2.359
2.351
2.343
190
2.457
2.449
2.441
2.433
2.425
2.417
2.409
2.401
2.393
2.385
2.377
200
2.489
2.482
2.474
2.466
2.458
2.450
2.442
2.434
2.426
2.418
2.410
210
2.521
2.513
2.506
2.498
2.490
2.482
2.475
2.4 67
2.459
2.451
2.443
220
2.552
2.545
2.537
2.530
2.522
2.514
2.507
2.499
2.491
2.483
2.476
230
2.583
2.575
2.568
2.560
2.553
2.545
2.538
2.530
2.523
2.515
2.507
240
2.613
2.606
2.598
2.591
2.583
2.576
2.569
2.561
2.554
2.546
2.539
250
2.642
2.635
2.628
2.621
2.613
2.606
2.599
2.592
2.584
2.577
2.570
260
2.671
2.664
2.657
2.650
2.643
2.636
2.628
2.621
2.614
2.607
2.600
270
2.699
2.692
2.665
2.678
2.671
2.664
2.657
2.650
2.643
2.636
2.629
280
2.727
2.720
2.713
2.706
2.700
2.693
2.686
2.679
2.672
2.665
2.658
290
2.754
2.747
2.740
2.734
2.727
2.720
2.714
2.707
2.700
2.694
2.687
300
2.780
2.773
2.767
2.761
2.754
2.748
2.741
2.734
2.728
2.721
2.715
X 10"s kg/(m-sec)
fl Calculations according to:
C.R. Wilke. A Viscosity Equation for Gas Mixtures. J. Chera. Phy., 18^(4)s517-519 (April, 1950).
-------
LINCS(NSECT) must equal the total length of the
given electrical section. The values of this
variable are read in with an E11.4 format and must
be right justified in columns 67-77.
The overall format for this data group is [7(Ell.4)] and the
data group is contained on two data cards. This data group must
be read in with each basic data set.
The next data group which is read in depends on the value of
NVI. If NVI = 2, the following is a sequential listing of the
variables in the next data group which is read in, along with the
descriptions of the variables and the format specifications.
(1) RFS(NSECT? is the roughness factor for the wires in a given
electrical section and is dimensionless. In
precipitation practice/ if the wires are scratched
or dirty but not completely coated with air, then
the values of RFS(NSECT) lie in the range 0.5-1.0.
A value of 1.0 corresponds to wires which are in
perfect condition. The effect of decreasing the
roughness factor is one of increasing the current
that can be achieved at a given voltage level. If
the wires are completely covered with dirt, then
the effect may be one of increased wire diameter
with a roughness superimposed. This situation
would lead to compensating effects. The values of
this variable are read in with an E11.4 format and
must be right justified in columns 1-11.
(2) START!(NSECT) is the chosen initial current density at which
the calculation of a voltage-current curve starts
in a given electrical section and is in units of
A/m2. In generating the voltage-current curve, the
current density increments in steps of START1
(NSECT) until a change is specified. The values of
this variable are read in with an E11.4 format and
must be right justified in columns 12-22.
(3) START2(NSECT) is a chosen increment in current density which
is used in place of START1(NSECT) when the Jll-th
point on the voltaae-current curve is reached and
is in units of A/m*. The values of this variable
are read in with an E11.4 format and must be right
justified in columns 23-33.
(4) START3(NSECT) is a chosen increment in current density which
is used in place of START2(NSECT) when the JI2-th
point on the voltage-current curve is reached and
is in units of A/m . The values of this variable
A-27
-------
are read in with an E11.4 format and must be right
justified in columns 34-44.
(5) VSTAR(NSECT) is an estimate of the applied voltage
corresponding to the first point on the voltage-
current curve as defined by START1(NSECT) and is in
units of volts. If VSTAR(NSECT) is close to the
actual applied voltage, the calculation will be
performed faster. However, whatever the choice of
VSTAR(NSECT), it will not affect the accuracy of
the calculation. The values of this variable are
read in with an'E11.4 format and must be right
justified in columns 45-55.
The overall format for this data group is [7(Ell.4)] and the
data group is contained on a single data card. If NVI « 1r this
data group is not read in.
The data input starting with AS(NSECT) above must be repeated
for each electrical section of the precipitator, proceeding from
the inlet to the outlet of the precipitator. Thus, the data group
containing AS(NSECT) and possibly the data group containing
RFS(NSECT) must be read in NUMSEC different times.
At this point, the basic data set has been entered into the
program and precipitator performance will be projected based on
this data. The last card in the data section must have a 99 in
columns 1-2. This causes the program to terminate normally.
CONSTRUCTION OF SHORTENED DATA SETS
Once the basic data set is processed, then all the parameters
which are needed by the program to calculate precipitator
performance are stored in memory. By using values of NDATA equal
to 2, 3, or 4, shortened data sets can be entered after the basic
data set in order to analyze the effects of particle size
distribution, specific collection area, and electrical conditions
on precipitator performance. In the shortened data sets, the
values of a small number of variables which are stored in memory
are changed to new values in order to produce a new set of data.
In each shortened data set, the first two data groups and
data cards which are read in are the same as those discussed "for
the basic data set. The value of NDATA. on the first data card
determines the variables in memory that will be changed. The
effects of particle size distribution on precipitator performance
can be analyzed by setting NDATA - 2. In this case, the third
data group which is read in depends upon the value of NDIST which
is stored in memory. If NDIST = 2, an inlet mass median diameter
(D50) and geometric standard deviation (SIGMAP) must be read in
A-28
-------
according to the same specifications discussed for the basic data
set. If NDIST = 1, cumulative percents [PRCU(I)] corresponding to
the particle sizes [ENDPT(I)] stored in memory must be read in
according to the same specifications discussed for the basic data
set. After the third data group is read in, the shortened data
set is complete. By repeating this type of shortened data with
different choices of D50 and SIGMAP or PRCU(I), the effects of
particle size distribution can be analyzed with the use of only a
few data cards.
The effects of specific collection area (SCA) on precipitator
performance can be analyzed by setting NDATA =3. In this case,
the following is a sequential listing of the variables which are
inputted in the third data group, along with the descriptions of
the variables and the format specifications.
is the total gas volume flow rate in a given
electrical section and is in units of actual
ft3/min. The values of this variable are read in
with an E11.4 format and must be right justified in
columns 1-11, 23-33, and 45-55.
is the gas velocity in a given electrical section
and is in units of ft/sec. The values of this
variable are read in with an S11 -4 format and must
be right justified in columns 12-22, 34-44, and
56-66.
There must be a value of VGS(I) and VGASS(I) entered for each
electrical section. The overall format for this data group is
[3(2E11.4)] and the data group is contained on 4 or less cards
depending on the value of NUMSEC which is stored in memory. For
NUMSEC < 3, 3
-------
corresponds to a measured or known value. If the
value of NVI stored in memory is 2, VOS(I)
corresponds to an applied voltage up to which a
voltage-current curve will be calculated. Then,
this applied voltage along with the corresponding
current will be used in the calculation of
precipitator performance. The values of this
variable are read in with an E11.4 format and must
be right justified in columns 1-11r 23-33, and
45-55.
(2) TCS(X) is the total current in a given electrical section
and is in units of amperes. If the value of NVI
stored in memory is 1, TCS(I) corresponds to a
measured or known value. If the value of NVI
stored in memory is 2, TCS(I) has no meaning in
terms of input data since it will be calculated in
the program. In this case, the appropriate columns
on the data cards can be left blank or any
arbitrary number can be eneterea. The values of
this variable are read in with an El 1.4 format and
must be right justified in columns 12-22, 34-44,
and 56-66.
There must be a value of VPS(I) and TCS(I) entered for each
electrical section. The overall format for this data group is
[3(2E11.4)] and the data group is contained on 4 or less cards
depending on the value of NUMSEC which is stored in memory. For
NUMSEC < 3, 3
-------
defined by the variable LSECT in the basic data
set.) Since the maximum number of increments is
45, MAPRAP is read in with a 4511 format. If
TIMRAP > 0 and MAPRAP (I) > 0, then the Ith
increment of the precipitator is defined as having
been rapped. The program sets the dust thickness
on the collection plate at increment I equal to 0
and stores a portion of the collected dust for
reintroduction into the gas flow at that point. If
TIMRAP > 0 and MAPRAP(I) = 0, then the Ith
increment was not rapped. If TIMRAP = 0 and
MAPRAP(1) * 0, the program will treat the next card
as the beginning of an all new rapping scheme. In
this case, the next card read in will be the first
TIMRAP card of the new scheme. If TIMRAP = 0 and
MAPRAP(1) = 0, the rapping calculation is
terminated and the program resumes normal
operation. The next card that will be read in this
case is'card 1 of the basic data set. MAPRAP is
located in columns 11-55.
(3) REPCT is the fraction of the dust collected in one
electrical section that is reentrained during the
time required for the gas to flow past one
computational length increment, LINCS. This is
defined by
F Duration of gas flow past one increment
REPCT = * ,
NLANES Total duration of each rapping puff
where F is the total fraction of the collected dust
that will be reentrained during a rap and NLANES is
the number of parallel lanes in the section which
are rapped at different times.
(4) NUMBOL is the number of data points in the reentrainment
profile. NUMBOL is read in with an 12 format and
must be right justified in columns 61-65.
(5) NDAT is an indicator which can assume the values of 0 or
1. If NDAT = 1, the subsequent cards will define
the rapping profile. If NDAT = 0, the existing
rapping profile will be used in this rap. NDAT is
read in with an 15 format and must be right
justified in columns 66-70.
A-31
-------
The overall format for this card is (F10.5, 4511, F5.3, 215).
This card is used to define each rap of the precipitator. If two
or more sections of the precipitator are rapped at the same time,
a single card can contain all of the necessary information.
However, any rap which occurs at a different time will require
another TIMRAP card.
If NDAT = 1 on the TIMRAP card, the following cards must be
inserted following that card to define the rapping profile:
(1) BOLFRC(I,J) are the data points used to construct the
normalized histogram of fractions of dust
reentrained in a given time interval. There must
be NUMBOL values of BOLFRC on the card. BOLFRC is
read into the program with an F6.3 format.
The overal format of this card is (13F6.3). The program
expects the number of BOLFRC cards to equal the number of
precipitator sections {NUMSEC) so that a different reentrainment
histogram may be read in for each section. After NUMSEC cards are
read, the following data is expected:
(1) REMMD(I,J) are the mass median diameters for a log-normal size
distribution associated with each step of the
histrogram defined by BOLFRC(I,J). There must be
NUMBOL values of REMMD on the card. REMMD is read
with an F6.3 format and must be in units of um.
(2) RESIG{J) is the geometric standard deviation of the log-
normal size distributions defined by REMMD. The
value of RESIG must be equal to or greater than 1.
RESIG is read in an F6.3 format immediately
following the last value of R£MMD(I,J).
The overall format of this card is (13F6.3). The program
expects to read NUMSEC REMMD cards to match the NUMSEC BOLFRC
cards. After these cards are used, the following card is read
in:
(1} BOLTIM(I) are the durations in seconds of each of the steps
of the rapping histogram. The sum of the values of
BOLTIM(I) is the total duration of the rapping
emissions and must obey the following relation:
I Boltim(I) = NLANES*Total duration of each rapping puff
The total duration of the rapping emissions must
not exceed the time between raps as the model
cannot correct for overlapping rapping puffs.
There must be NUMBOL values of BOLTIM on the card
in an F6.3 format.
A-32
-------
(2) SDRAP is the particle density of the reentrained dust.
DD rap is in units of kg/m3 and is read in with an
F6.3 format.
The overall format of this card is (13F6.3). Only 1 BOLTIM
card will be read in as the durations of the steps in the
histogram are assumed to be the same for all sections of the
precipitator. The rapping profile must be read in after the first
TIMRAP card. Thereafter, it may be read in if it is changed.
After the BOLTIM card is read in, the next card expected by the
program is a new TIMRAP card.
OPACITY DATA SET
In the calculation of precipitator outlet opacity, additional
data is required if NCOMPS is greater than 0. If NCOMPS > 0 and
NLAMDA = 0, user supplied values of the read and imaginary parts
of the index of refraction must be read into the program:
(1) CMR(I) is the real part of the user supplied index of
refraction. This quantity is read in with an F5.3
format.
CMI(I) is the imaginery part of the user supplied index of
refraction. CMI(I) is read with an F5.3 format.
The NCOMPS values of CMR{I) and CMI(I) are read in as data
pairs with an overall format of [6(2F5.3)].
If NCOMPS > 0 and NLAMDA > 0, then the index of refraction
changes with wavelength. In this case the following data are read
in:
(1) CMR(J)
(2] CMI(J)
(3) OPTOLM(J)
CMR(J) and CMI(J) are the correct indices of
refraction. UPTOLM is in units of urn and is read
in with an F5.3 format. The final value of
UPTOLM(J) must be equal to or greater than 0.70.
is the real part of the user supplied index of
refraction for wavelengths ranging from the value
of UPTOLM(J-1) (assumed to be 0.43 um if J = 1) to
the value of UPTOLM(J). CMR is read in with an
F5.3 format.
is the imaginary part of the user supplied index of
refraction for wavelengths ranging from the value
of UPTOLM (J-1) to the value of UPTOLM (J). CMI is
read in with an F5.3 format.
is the upper limit of the wavelength band for which
A-33
-------
The NLAMDA values of CMR(J), CMI(J), and OPTOLM(J) are read
in the form of data groups, each complete for a given value of J.
The overall format for this data is [4(3F5.2)j.
UNKNOWN OPERATING CONDITIONS
The model may be used to calculate voltage-current curves and
collection efficiencies for precipitators with unknown operating
voltages and currents. Table A-4 summaries the options available
to the user and makes recommendations on the applicability of each.
If collection efficiency data for a utility fly ash precipitator is
desired, the following data set should be used.
The first data card has the same format as the first card in
the basic data set. MODL must be set to 2 in column 8. Since the
remainder of the data in this data group is redefined in
subroutine INTERN, the rest of the card may be left blank if
desired.
Data group 2 contains the following 2 data.
(1) RHO is the resistivity of the collected particulate
layer expressed in ohm-cm. REO is entered with an
E8.2 format and must be right justified in columns
I-8.
(2) HDUCT is the wire-to-plate spacing in units of inches.
HDUCT is read with an F5.2 format in columns
II-15.
If the ESP being studied is similar to a utility fly ash
precipitator, the above procedure may be used. However, if the
ESP is much different from a utility fly ash precipitator, a
different procedure should be used. This procedure is enabled
when NVI is set to 2 in data group 3. The description of the data
required for this case is contained in the construction of the
basic data set section of this Appendix (p. A-i).
If only voltage-current curves are desired, the required
data set again depends on whether the ESP is similar to a utility
fly ash precipitator. If the ESP is similar, then the first 3
data groups are the same as for the basic data set described
earlier in this Appendix. To get the voltage-current calculation,
the following values must appear:
(1) Card 1: MODL = 3 in column 8 (the rest of the card
may be blank), and
(2) Card 3: NVI * 1 in column 6 (the rest of the card
may be blank).
A-34
-------
TABLE A-4. USER OPTIONS WHEN THE OPERATING VOLTAGE AND CURRENT ARE NOT KNOWN
Variables
Procedure
Desired Calculation
MODL
NVI
VISKIP
VISAME
Recommended
Comments
Collection efficien-
cy of a utility fly
ash precipitator.
2
1
n/a
n/a
Yes
Uses internal data set based on 17
field tests of cold-side utility fly
ash precipitators to generate effi-
ciencies for a range of SCA values.
Fast.
0
or
1
2
1
1 or 2
No
Calculates approximate V-I curves;
then calculates efficiency for a
derived operating point. Slow.
VISAME2 (different voltages in each
section) ia slower than VISAME=1
(same voltage).
Collection efficien-
cy for an ESP that
is much different
from a utility fly
ash ESP.
2
1
n/a
n/a
No
Precipitator test data used in
internal data set are not similar
to this application. Accuracy may
be low. Fast.
0
or
1
2
1
1
Yea
Calculates approximate voltage-
current curves; then calculates
efficiency for a derived operating
point. Uses the same operating
point in each ESP section. Slow.
0
or
1
2
1
2
No
Same as above except calculates a
new operating point for each ESP
section. Slower.
0
or
1
2
0
2
No
Same as above except calculates
electric fields and potentials in
each ESP length increment. Very
slow.
(continued)
-------
TABLE A4
Variables
Desired Calculation MODL NVI VISKIP VISAME
Voltage-current curves 3 1 n/a n/a
only for a utility
fly ash precipitator.
3 2 1 1 or 2
Voltage-current curves 3 1 n/a n/a
only for an ESP that
is much different from
a utility fly ash ESP.
3
2 1 1 or 2
(CONTINUED)
Procedure
Recommended Comments
Vea Uses internal data set to generate
voltage-current curves. No
efficiency calculated. Fast.
N° Calculates voltage-current curve
for first section only using
approximate procedures. No
efficiency calculated. Slow.
No Uses internal data set to generate
voltage-current curves. Precipitator
test data used in internal data set
are not similar to this application.
Accuracy may be low. Fast.
YB8
Calculates voltage-current curve
for first section only using
approximate procedures. Mo
efficiency calculated. Slower.
-------
Card 4 will contain the following data:
(1) RHO is the resistivity of the collected particulate
layer expressed in ohm-cm. RHO is entered with an
E8.2 format and must be right justified in columns
I-8.
(2) HDUCT is the wire-to-plate spacing in units of inches,
HDUCT is read with an F5.2 format in columns
II-15.
Voltage-current curves and operating points for a six section ESP
will be printed. If the ESP is not similar to a utility fly ash
precipitator, a complete data set appropriate for an efficiency
calculation with NVI set to 2 is required. If MODL = 3 on card 1/
the calculation will terminate after the calculation and printing
of the voltage-current curve data for the first electrical
section.
REFERENCE
1. Faulkner, M. G., and DuBard, J. L. A Mathematical Model of
Electrostatic Precipitation (Revision 3). EPA-600/7-84-069a,b,c.
(Volume I, Modeling and Programming: NTIS PB84-212-679; Volume II,
User's Manual: NTIS PB84-212-687; FORTRAN Source Code Tape: NTIS
PB84-232-990). U. S. Environmental Protection Agency, Research
Triangle Park, North Carolina. 1984.
A-37
-------
Appendix B
OUTPUT DATA FILE ESPREV.OUT CORRESPONDING TO THE
INPUT DATA FILE SHOWN IN FIGURE 1
* *
* EPA ESP MODEL *
* *
* AEERL AND SRI *
¦A *
* REVISED ESP MODEL - DECEMBER 1990 *
~k a
icicic'k'k'k-^-k'k'kic'k'k'k'k'k'kic'k-k'k-k'k'k'-k'k'-k'k'k'k'k-k'k-k-k'k'Jc'k'kic'k
INPUT DATA FOR DATA SET NUMBER 1
TITLE: TEST PLANT 2: 4 SECTION ESP WITH 320 SCA, BITUMINOUS COAL SIZE
DISTRIBUTION
CALCULATION PARAMETERS
AAAAAAAAA A A AA A AAA A o A A A A A A A A A AAA A A A A A A A A A A A A A A A A A A A A A AAA A A A A A A A A A A A A A A A A A A A A^ A A A A
TYPE OF DATA SET (1 - 4) NDATA 1
MODEL TYPE (0-3) MODL 0
REDUCED PRINTING (0-3) NOPRNT 0
RIGOROUS OR ESTIMATED FIELD CALCULATION (1 OR 2) NEST 1
RIGOROUS OR ESTIMATED CHARGE CALCULATION (0 OR 1) NCALC 0
VI CURVES KNOWN OR CALCULATED (1 OR 2) NVI 1
DIMENSION OF X GRID (15 MAX) NX 15
DIMENSION OF Y GRID (15 MAX) NY 11
MAX NUMBER OF ITERATIONS TO CONVERGENCE NITER 2
NUMBER OF INTEGRATION INCREMENTS IN CHARGE CALC NN 10
ESTIMATED EFFICIENCY (%) ETAO 99.90
PHYSICAL PARAMETERS
INLET MASS LOADING DL 2.000 GR/ACF
PARTICLE DENSITY DD 2400. KG/M3
DIELECTRIC RATIO EPS 100.0
ION MOBILITY US .270E-03 M2/V/S
RESISTIVITY RH0 .200E+11 OHM-CM
B-l
-------
TOTAL ESP LENGTH
COLD SIDE OR HOT SIDE (1 OR 2)
PEAK-TO-AVERAGE VOLTAGE RATIO
ELECTRICAL BREAKDOWN STRENGTH
PL
NTEMP
VRATIO
EBD
36.00 FT
1
1.20
,150E+07 V/M
PERCENT WATER
PERCENT 02
PCH20
PC02
8.00 %
3.50 %
PARTICLE SIZE DATA
NUMBER OF SIZE BAND END POINTS NENDPT 19
TYPE OF SIZE DATA (1=MEASURED, 2=CALC) NDIST 2
SIZE DISTRIBUTION MMD D50 16.30 DM
STANDARD DEVIATION SIGMAP 3.40
SIZE BAND END POINTS ENDPT .010 DM
.070
.100
.140
.220
.340
.500
.700
1.000
1.400
2.200
3.400
5.000
7.000
10.000
14.000
22.000
34.000
100.000
SECTIONAL DATA
NUMBER OF ESP SECTIONS NUMSEC 4
NO. INCREMENTS PER SECTION LSECT 10 INCREMENTAL LENGTH LINCS .90 FT
10 .90
10 .90
10 .90
SECTION NO. 1
VOLTAGE VOS .431E+05 V
CURRENT TCS .490E+00 A
PLATE AREA AS .270E+05 FT2
B-2
-------
TOTAL WIRE LENGTH
CORONA WIRE RADIUS
NUMBER OF WIRES
WIRE-TO-PLATE SPACING
WIRE-TO-WIRE SPACING
WLS 27000.00 FT
ACS .0750 IN.
NWS 18
BS 4.50 IN.
SYS*2 6.00 IN.
GAS VOLUME FLOW RATE
GAS VELOCITY
GAS TEMPERATURE
GAS PRESSURE
GAS VISCOSITY
SECTION NO. 2
VGS .338E+06 ACFM
VGASS 5.00 FT/S
TEMPS 300.00 DEG F
PS I.000 ATM
VISS .230E-04 KG/M/S
VOLTAGE
CURRENT
PLATE AREA
VOS
TCS
AS
.420E+05 V
.737E-5-00 A
.270E+05 FT2
TOTAL WIRE LENGTH
CORONA WIRE RADIUS
NUMBER OF WIRES
WIRE-TO-PLATE SPACING
WIRE-TO-WIRE SPACING
WLS
ACS
NWS
BS
SYS*2
27000.00 FT
.0750 IN.
18
4.50 IN.
6.00 IN.
GAS VOLUME FLOW RATE
GAS VELOCITY
GAS TEMPERATURE
GAS PRESSURE
GAS VISCOSITY
VGS .338E+06 ACFM
VGASS 5.00 FT/S
TEMPS 300.00 DEG F
PS 1.000 ATM
VISS .230E-04 KG/M/S
SECTION NO. 3
VOLTAGE
CURRENT
PLATE AREA
VOS
TCS
AS
. 398E+05 V
.748E+00 A
. 270E+05 FT2
TOTAL WIRE LENGTH
CORONA WIRE RADIUS
NUMBER OF WIRES
WIRE-TO-PLATE SPACING
WIRE-TO-WIRE SPACING
WLS
ACS
NWS
BS
SYS*2
27000.00 FT
.0750 IN.
18
4.50 IN.
6.00 IN.
GAS VOLUME FLOW RATE
GAS VELOCITY
GAS TEMPERATURE
GAS PRESSURE
GAS VISCOSITY
VGS .338E+06 ACFM
VGASS 5.00 FT/S
TEMPS 300.00 DEG F
PS 1.000 ATM
VISS .230E-04 KG/M/S
SECTION NO. 4
VOLTAGE VOS .382E+05 V
CURRENT TCS .981E+00 A
PLATE AREA AS .270E+05 FT2
B-3
-------
TOTAL WIRE LENGTH
CORONA WIRE RADIUS
NUMBER OF WIRES
WIRE-TO-PLATE SPACING
WIRE-TO-WIRE SPACING
GAS VOLUME FLOW RATE
GAS VELOCITY
GAS TEMPERATURE
GAS PRESSURE
GAS VISCOSITY
WLS 27000.00 FT
ACS .0750 IN.
NWS 18
BS 4.50 IN.
SYS*2 6.00 IN.
VGS .338E+06 ACFM
VGASS 5.00 FT/S
TEMPS 300.00 DEG F
PS 1.000 ATM
VISS .230E-04 KG/M/S
NON-IDEAL PARAMETERS
NUMBER OF NON-IDEAL DATA SETS NONID 3
SNEAKAGE FRACTION .05 GAS VEL. SIGMA .15 NO. BAFFLED STAGES 4.00
.10 .25 4.00
.00 .00 4.00
RAPPING PARAMETERS
*******************************************************************************
TYPE OF RAPPING CALCULATION (0,5=OLD ESTIMATION, 1-3=DYNAMIC) 0
NUMBER OF RAPPING DATA SETS
SOURCE OF RAPPING DUST
FIRST RAPPING DUST MMD
FIRST RAPPING DUST SIGMA
NRAPD 1
NEFF 1
ARD50 6.0 UM
ARSIGM 2.5
OPACITY PARAMETERS
*******************************************************************************
OPACITY PATH LENGTH PATHL 15.00 FT
NO. INDICES OF REFRACTION (0-10) NCOMPS 0
NO. WAVELENGTHS FOR OPACITY CALC (0=10,1=1) NWAVES 0
NO. INDICES OF REFRACTION FOR VARIABLE INDEX (0-10) NLAMDA 0
INPUT MASS LOADING - 4.580E-03 KG/ACM
*** ESP INCREMENT NO.
MASS REMOVED (KG/M3)
UNCORRECTED % PEN
1 -kick
= 2.247E-03
48.3463
CUM. MASS REMOVAL = 2.247E-03
*** ESP INCREMENT NO. 2 ***
MASS REMOVED (KG/M3) = 7.354E-04
UNCORRECTED % PEN - 31.4444
CUM. MASS REMOVAL = 2.983E-03
*** ESP INCREMENT NO.
MASS REMOVED (KG/M3)
UNCORRECTED % PEN
3 k-k-k
- 3.814E-04
22.6795
CUM. MASS REMOVAL - 3.364E-03
B-4
-------
*** ESP INCREMENT NO. 4 ***
MASS REMOVED (KG/M3) = 2.357E-04
UNCORRECTED % PEN - 17.2632
*** ESP INCREMENT NO. 5 ***
MASS REMOVED (KC-/M3) - 1.589E-04
UNCORRECTED % PEN - 13.6118
*** ESP INCREMENT NO. 6 ***
MASS REMOVED (KG/M3) - 1.135E-04
UNCORRECTED % PEN - 11.0036
ESP INCREMENT NO. 7 ***
MASS REMOVED (KC-/M3) - 8.446E-05
UNCORRECTED % PEN = 9.0625
*** ESP INCREMENT NO. 8 ***
MASS REMOVED (KG/M3) = 6.478E-05
UNCORRECTED % PEN - 7.5736
*** ESP INCREMENT NO. 9 ***
MASS REMOVED (KC-/M3) - 5.095E-05
UNCORRECTED % PEN = 6.4025
*** ESP INCREMENT NO. 10 ***
MASS REMOVED (KG/M3) = 4.081E-05
UNCORRECTED % PEN - 5.4645
CUM. MASS REMOVAL = 3.600E-03
CUM. MASS REMOVAL = 3.759E-03
CUM. MASS REMOVAL = 3.872E-03
CUM. MASS REMOVAL - 3.957E-03
CUM. MASS REMOVAL - 4.021E-03
CUM. MASS REMOVAL = 4.072E-03
CUM. MASS REMOVAL = 4.113E-03
*** CORRECTION FOR .05 SNEAKAGE ***
SNEAKAGE MASS (KG/M3) = 2.289E-04
MASS PENETRATION = 4.668E-04
CORRECTED % PENETRATION - 10.1913
CUM. MASS REMOVAL
INLET MASS LOADING
4.113E-03
4.580E-03
*** ESP INCREMENT NO. 11 ***
MASS REMOVED (KG/M3) = 1.419E-04
UNCORRECTED % PEN
6.6182
CUM. MASS REMOVAL = 4.255E-03
*** ESP INCREMENT NO.
MASS REMOVED (KG/M3)
UNCORRECTED % PEN
12 ***
= 6.138E-05
5.2710
CUM. MASS REMOVAL - 4.316E-03
*** ESP INCREMENT NO. 13 ***
MASS REMOVED (KG/M3) - 4.068E-05
UNCORRECTED % PEN = 4.3782
CUM. MASS REMOVAL - 4.357E-03
*** ESP INCREMENT NO. 14 ***
MASS REMOVED (KG/M3) = 3.046E-05
UNCORRECTED % PEN - 3.7098
CUM. MASS REMOVAL - 4.388E-03
*** ESP INCREMENT NO.
MASS REMOVED (KG/M3)
UNCORRECTED % PEN
15
- 2.395E-05
3.1842
CUM. MASS REMOVAL - 4.412E-03
B-5
-------
*** ESP INCREMENT NO. 16 ***
MASS REMOVED (KG/M3) = 1.940E-05
UNCORRECTED % PEN - 2.7585
*** ESP INCREMENT NO. 17 ***
MASS REMOVED (KG/M3) = 1.601E-05
UNCORRECTED % PEN - 2.4071
*** ESP INCREMENT NO. 18 ***
MASS REMOVED (KG/M3) - 1.342E-05
UNCORRECTED % PEN -= 2.1127
*** ESP INCREMENT NO. 19 ***
MASS REMOVED (KG/M3) - 1.138E-05
UNCORRECTED % PEN = 1.8629
*** ESP INCREMENT NO. 20 ***
MASS REMOVED (KG/M3) = 9.730E-06
UNCORRECTED % PEN - 1.6494
*** CORRECTION FOR .05 SNEAKAGE ***
SNEAKAGE MASS (KG/M3) - 2.333E-05
MASS PENETRATION = 9.849E-05
CORRECTED % PENETRATION - 2.1505
*** ESP INCREMENT NO. 21 ***
MASS REMOVED (KG/M3) - 1.444E-05
UNCORRECTED % PEN = 1.7296
*** ESP INCREMENT NO. 22 ***
MASS REMOVED (KG/M3) - 9.737E-06
UNCORRECTED % PEN = 1.5168
*** ESP INCREMENT NO. 23 ***
MASS REMOVED (KG/M3) - 7.852E-06
UNCORRECTED % PEN = 1.3451
*** ESP INCREMENT NO. 24 ***
MASS REMOVED (KG/M3) - 6.650E-06
UNCORRECTED % PEN = 1.1998
*** ESP INCREMENT NO. 25 ***
MASS REMOVED (KG/M3) = 5.722E-06
UNCORRECTED % PEN = 1.0747
*** ESP INCREMENT NO. 26 ***
MASS REMOVED (KG/M3) = 4.985E-06
UNCORRECTED % PEN = .9657
*** ESP INCREMENT NO. 27 ***
MASS REMOVED (KG/M3) - 4.367E-06
UNCORRECTED % PEN = .8703
CUM. MASS REMOVAL = 4.431E-03
CUM. MASS REMOVAL = 4.447E-03
CUM. MASS REMOVAL - 4.460E-03
CUM. MASS REMOVAL = 4.472E-03
CUM. MASS REMOVAL = 4.482E-03
CUM. MASS REMOVAL - 4.482E-03
INLET MASS LOADING - 4.580E-03
CUM. MASS REMOVAL - 4.496E-03
CUM. MASS REMOVAL - 4.506E-03
CUM. MASS REMOVAL - 4.514E-03
CUM. MASS REMOVAL = 4.520E-03
CUM. MASS REMOVAL - 4.526E-03
CUM. MASS REMOVAL - 4.531E-03
CUM. MASS REMOVAL = 4.535E-03
-------
*** ESP INCREMENT NO. 28 ***
MASS REMOVED (KG/M3) = 3.855E-06
UNCORRECTED % PEN = .7860
CUM. MASS REMOVAL = 4.539E-03
*** ESP INCREMENT NO. 29 ***
MASS REMOVED (KG/M3) = 3.414E-06
UNCORRECTED % PEN = .7114
CUM. MASS REMOVAL = 4.543E-03
*** ESP INCREMENT NO. 30 ***
MASS REMOVED (KG/M3) = 3.036E-06
UNCORRECTED % PEN - .6450
CUM. MASS REMOVAL = 4.546E-03
*** CORRECTION FOR .05 SNEAKAGE ***
SNEAKAC-E MASS (KG/M3) = 4.923E-06
MASS PENETRATION - 3.444E-05
CORRECTED % PENETRATION = .7519
CUM. MASS REMOVAL - 4.546E-03
INLET MASS LOADING - 4.580E-03
ESP INCREMENT NO. 31 ***
MASS REMOVED (KG/M3) - 3.364E-06
UNCORRECTED % PEN = .6411
CUM. MASS REMOVAL - 4.549E-03
*** ESP INCREMENT NO. 32 ***
MASS REMOVED (KG/M3) = 2.845E-06
UNCORRECTED % PEN - .5789
CUM. MASS REMOVAL = 4.552E-03
¦kick ESP INCREMENT NO. 33 ***
MASS REMOVED (KG/M3) = 2.505E-06
UNCORRECTED % PEN = .5242
CUM. MASS REMOVAL = 4.554E-03
ESP INCREMENT NO. 34 ***
MASS REMOVED (KG/M3) = 2.227E-06
UNCORRECTED % PEN - .4756
CUM. MASS REMOVAL = 4.557E-03
*** ESP INCREMENT NO. 35 ***
MASS REMOVED (KC-/M3) = 1.989E-06
UNCORRECTED % PEN - .4321
CUM. MASS REMOVAL = 4.558E-03
*** ESP INCREMENT NO. 36 ***
MASS REMOVED (KG/M3) = 1.779E-06
UNCORRECTED % PEN - .3933
CUM. MASS REMOVAL - 4.560E-03
*** ESP INCREMENT NO. 37 ***
MASS REMOVED (KG/M3) - 1.602E-06
UNCORRECTED % PEN = .3583
CUM. MASS REMOVAL = 4.562E-03
*** ESP INCREMENT NO. 38 ***
MASS REMOVED (KG/M3) » 1.443E-06
UNCORRECTED % PEN = .3268
CUM. MASS REMOVAL - 4.563E-03
*** ESP INCREMENT NO. 39 ***
MASS REMOVED (KG/M3) = 1.302E-06
UNCORRECTED % PEN = .2983
CUM. MASS REMOVAL - 4.565E-03
B-7
-------
ESP INCREMENT NO. 40 ***
MASS REMOVED (KG/M3) = 1.175E-06 CUM. MASS REMOVAL = 4.566E-03
UNCORRECTED % PEN - .2727
*** CORRECTION FOR .05 SNEAKAGE ***
SNEAKAGE MASS (KG/M3) = 1.721E-06
MASS PENETRATION = 1.420E-05
CORRECTED % PENETRATION - .3101
CUM. MASS REMOVAL = 4.566E-03
INLET MASS LOADING = 4.580E-03
ESP OUTLET STATISTICS
*** OUTLET DATA CORRECTED FOR SNEAKAGE, SIGMA G,
AND SMALL PARTICLE MIGRATION VELOCITY ERROR ***
DIAMETER UM
EFFICIENCY
PENETRATION
EFFECTIVE MIG
.0400
.999731
.000269
.130644
.0850
.994865
.005135
.083786
.1200
.989648
.010352
.072645
.1800
.982905
.017095
.064671
.2800
.978976
.021024
.061383
.4200
.980865
.019135
.062879
.6000
.985836
.014164
.067661
.8500
.991178
.008822
.075186
1.2000
.995030
.004970
.084307
1.8000
.997619
.002381
.096000
2.8000
.998845
.001155
.107499
4.2000
.999076
.000924
.111038
6.0000
.999116
.000884
.111749
8.5000
.999785
.000215
.134203
12.0000
.999931
.000069
.152239
18.0000
.999963
.000037
.162186
28.0000
.999967
.000033
.163892
67.0000
.999967
.000033
.163965
CORRECTED TOTAL EFFICIENCY = 99.950
PENETRATION - .050
*** CORRECTION FOR RAPPING REENTRAINMENT ***
DIAMETER UM
EFFICIENCY
PENETRATION
EFFECTIVE MIG
.0400
.999684
.000316
.128116
.0850
.994769
.005231
.083492
.1200
.989492
.010508
.072406
.1800
.982640
.017360
.064427
.2800
.978561
.021439
.061073
.4200
.980290
.019710
.062409
.6000
.985121
.014879
.066878
.8500
.990345
.009655
.073752
1.2000
.994119
.005881
.081630
B-8
-------
1.8000
.996684
.003316
.090736
2.8000
.997969
.002031
.098531
4.2000
.998320
.001680
.101547
6.0000
.998499
.001501
.103331
8.5000
.999308
.000692
.115643
12.0000
.999584
.000415
.123750
18.0000
.999739
.000261
.131129
28.0000
.999842
.000158
.139113
67.0000
.999932
.000068
.152462
TOTAL EFFICIENCY WITH RAPPING - 99.917
PENETRATION - .083
TEST PLANT 2: 4 SECTION ESP WITH 320 SCA, BITUMINOUS COAL SIZE DISTRIBUTION
ESP SCA - 319. FT2/KACFM
SNEAKAGE - .05
SIGMA G - .15
RAPPING MMD = 6 UM WITH SIGMA =2.5
TOTAL EFFICIENCY WITHOUT RAPPING = 99.950
PENETRATION - .050
TOTAL EFFICIENCY WITH RAPPING = 99.917
PENETRATION = .083
DUST LOADINGS: INLET = 3.12856 GR/DSCF
OUTLET = .00258 GR/DSCF
.00435 LB/MBTU
VISIBLE EMISSIONS: OPACITY
- 1.60
%
(NO-RAP)
- 1.33
%
(RAPPING PUFF)
.27
%
STACK DIAMETER
= 4.58
M
EXTINCTION COEFF. = .0035 1/K
B-9
------- |