NOTE --- When unzipping the ESPMODEL.ZIP file be sure to use the "-d"
option (ie., PKUNZIP -d ESPMODEL.ZIP). If this option is not used the
setup program included will not be able to find the files it needs.

(FYI - The "-d" option in PKUNZIP restores the directory structure stored
into the ZIP file.)

---------------------------------------------------------------------------

ÉÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍ»
º º
º ESPVI VERSION 4.0a º
º º
ÈÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍ¼

This file contains information about relatively minor changes that have been
made to ESPVI 4.0 for the purposes of improving its operation and stability.
The primary description of the use and operation of ESPVI 4.0 remains the
documentation provided with the original version, but this file contains
additional information about the program.

In general, the changes have been made to improve the operation of the model
with regard to data entry and automation. Several months' use of the model
showed that some operations were performed repeatedly with manual adjustments
of parameter values. Some of these have been automated to obtain the same
results faster and more accurately.

INTERNAL CHANGES

In a recent paper, Hutchins, et al., (Aerosol Sci. Tech. 22:202-218), present
measurements of the Cunningham correction factor and give viscosity and mean
free path formulas they use in their data reduction. All these factors have
been incorporated into ESPVI 4.0a, because they now seem to be the best values
to use. The mean free path and Cunningham Factor do not show up directly in
the results, but the gas viscosity calculated with the new expression is
roughly 10 percent larger than previously. The new viscosity value should be
transferred to your data files and saved (see below about data files). The
overall change in collection efficiency is small, but the changes do affect the
submicron particles noticeably.

In the file DATA.CNF, a number of values related to coal have been removed
because they were not used. In addition, the number of available wave forms
has been increased. Therefore, if you want to maintain the previous version of
ESPVI 4.0, you must also maintain a copy of the old version of DATA.CNF.

There are more data stored in some of the data files now. Every effort has
been made to allow data files from the previous version (ESPVI 4.0) to be used
in this version, but some values may require manual editing. When files are
saved under this version, the newly added data are appended to the end of the
file. In some cases, it has been noted that data files from the previous ver-
sion give poor and inaccurate results. Therefore, it is recommended that the
results from files created with ESPVI 4.0 be checked. If there are any doubts,
the data files should be recreated.

The Help system has been updated to be more sensitive to the context of the
current screen. There are more help topics and improved descriptions of the
options available. Where help is not appropriate, it does not display when
<F1> is pressed.

There have been a few other internal changes for improved predictions. First,
the PM10 values have been corrected to be based on aerodynamic particle
diameter (as described in the Federal Register) rather than physical diameter
(as done in ESPVI 4.0). The resulting changes are relatively small, but if you
see a discrepancy from previously calculated PM10 values, this is one reason.

Second, particle charging has been modified to use an improved algorithm. The
algorithm is faster, more accurate, and allows for effects of bipolar
charging. The predicted charges are slightly smaller than in the previous
version.

Third, since the new charging algorithm easily accounts for bipolar charging
effects, a back corona model has been incorporated. The back corona model
detects the onset of back corona as in the previous version, but now adds a
back corona charging current to the charging model and reduces the corona
onset voltage slightly. Together, these effects mimic most of the known
effects of back corona.

In cases with severe back corona (high resistivities and high current
densities), the model will probably predict very much higher current densities
than ESPVI 4.0 would with the same peak-to-average ratios. The peak-to-average
ratios should be decreased to bring the calculated current densities into
better agreement with the measured values.

Despite the improvements in back corona prediction with ESPVI 4.0a, there are
still cases of back corona that remain difficult to explain. Such cases may
have back corona in only a small fraction of the electrodes, but to such a
degree that it determines the voltage and current density for the whole
section.

In the file DATA.CNF, there is a parameter under the title "Back Corona
Multiplier" that can be used to change the program's response to back corona.
If the parameter is set to 1000 or larger, all the back corona modeling is
turned off (detection of back corona is still reported.) If the parameter is
set to any other number, the modeling is on but the number describes the
maximum charge degradation that can occur. The value 10 has been chosen to fit
a number of cases where back corona is a problem. Although the user can
modify the back corona multiplier with a text editor, we can give no guidance
about a better value to use.

Fourth, even though the dust layer thickness was a parameter is the previous
version, its only effect was to modify the effective wire-plate spacing. In
this version, the voltage drop in the dust layer is also computed. The dust
layer voltage drop affects the electric field in the wire-plate gap.
Generally, you will see little effect until the layer is 2 or 3 mm thick and
the resistivity is above 10E11 ohm-cm. When back corona is detected, the
voltage drop across the layer is held fixed at the value of the breakdown
voltage.

CHANGES IN FEATURES

The following features have been added, in the order of their appearance in
ESPVI.

COLOR - The visual appearance of the program has been modified by changing the
screen colors. Whether or not the colors are an improvement, a new program,
COLORSET, has been added to the package for the purpose of allowing you to
change the colors. COLORSET reads the color values from the file MENU.CNF and
displays them on the screen. By choosing and saving new colors to MENU.CNF,
you may customize the program's appearance to a large degree. The graphs have
been modified to use a white background and cannot be modified.

MAIN MENU - The previous main menu contained the entry "Calculate V-I &
Performance." These two functions have been split into separate items for
"Calculate V-I Curves" and "Performance Calculation." The two functions have
been separated because they are used differently. In addition, arrow symbols
() have been added beside menu items that lead to another menu, while the
absence of an arrow means that a direct action will be taken or a data entry

form will be obtained. The bottom line of the display is used to indicate
active function keys or to display messages that may be of help during
calculations.

FILE OPERATIONS MENU - two options have been added to the file choices:
"Create New Master File" and "Access Data Directory." The first of these
choices will make a whole new set of data files from the data currently in
use. The user is prompted for a file name that has not been used before and
the files are created. This should make the creation of a new ESP data set
from an existing one much easier. Of course, the newly created files will need
to be edited to put in correct values. "Access Data Directory" calls an
auxiliary program to allow direct access to data files. This option has been
added to allow for deletion of obsolete data files and to provide a quick way
to see all the data files together. The access makes use of the Ziff-Davis
program DR.COM (used with permission.)

FILE/BROWSE MENU - another option has been added, "Incremental Results." This
option allows viewing of a file that stores electric field, current density,
and space charge values at each increment of the calculation. It may be
helpful in understanding the detailed operation of the ESP. The increments in
the file are the upstream and downstream values for each element. Sparking
conditions are indicated with an asterisk (*) next to the gas field values;
back corona conditions are indicated with and asterisk next to the dust field
values. The local current density and particulate space charge are also shown
in the file. Values of local electric field in the gas and in the dust layer
are displayed in files named "master.INC", with "master" being the name of the
Master file.

Some increment files are too long for the simple file browse viewer to display
entirely. These files may be viewed whole under "Access Data Directory" as
described above.

DATA ENTRY - There have been no changes in the submenu, but there have been
several changes in the actual data entry menus and in some of the default
actions.

DATA/DESIGN/OVERALL PARAMETERS - Previously, when the number of sections in an
ESP was changed, the user was required to save each of the affected files
separately: design, operating voltages, and sneakage/rapping/turbulence.
Failure to do so would cause problems in subsequent attempts to use the file.
Now, when the number of sections is changed, all three files will be saved
together, using whatever values are current when the save operation is
performed. We recommend performing the save while still in the Design menu;
then modify whatever values are needed in the voltage and SRT menus and save
again in each of those. Remember that whenever the number of sections is
increased, the new sections receive all the values from the previous final
section.

DATA/OPERATING VOLTAGES - Two items on the menu have been changed or added.
There is a bias reset option and the wave form input is now a section-wise
menu. The bias reset removes the effect of one of the automatic performance
calculations. When the corona onset is adjusted automatically, all the corona
onset factors for the section are shifted by the same amount (biased);
usually, the biases are different for each section. The Bias Reset simply
returns all the bias values to zero. The WAVE FORM menu now requires entering
a distinct "wave form number" for each section. This is aimed toward the use
of intermittent energization, where common practice is to adjust the duty

cycle for optimum section performance. With the copying ability of the data
entry menu, the setting of values is not much harder than previously.

DATA/OPERATING VOLTAGES/JINPUT&CINPUT - Three new inputs are /requested in
both these menus related to the power supplies: maximum voltage (average),
maximum peak voltage, and maximum current (average). These values can usually
be obtained from the manufacturer's specifications. The maximum peak voltage
can also be estimated as 1.414 times the maximum average voltage. If the
values are not known, then set them to 0 as a sign that there are no valid
numbers for the parameters. These power supply limits are used to terminate
V-I calculations and adjustment calculations when they are reached. However,
values of zero will not terminate any calculations.

The maximum current rating for a section should be the sum of the ratings for
all TR sets that supply the section plate area. The total current may exceed
the rating of one power supply, but should not exceed the rating of all
supplies together. Some common TR set maximum ratings are 45 or 55 kV average
voltage, 63 or 70 kV peak voltage, and 500, 1000, or 1500 mA average current.
ESPs that require more current will have multiple units operating in parallel.

DATA/ELECTRODE - In this version, when the displayed section number is
changed, the electrode data for that section is automatically loaded. This
means that some operations will need to be performed in a different order, but
review and modification of existing data will be easier. Specifically, to edit
the electrode file for a section, display its section number and proceed to
ELECTRODE description. However, to copy the electrode data from one section to
another, move to the destination section, load the data from the other
section's file, and then, save the destination's data under a new file name.

The graph option now displays whatever data is current, not the data that
exists in the current file. This allows the user to see the effects of changes
without saving the changed data to the current file. Once the changes are
satisfactory, however, the user must save them explicitly before moving to
another section, or the changes will be lost.

Another menu, OPTIONS for input, has been added to facilitate electrode
spacing. The OPTIONS allow for setting the element locations with a constant
spacing (Average) or setting an inlet and outlet offset distance and spacing
the remaining elements equally among those locations. This menu may help more
with wire electrodes than with mast electrodes.

DATA/SNEAKAGE/BY SECTION - There are three changes in this menu. The
turbulence parameter has been designated "Turbulent core number", and the two
misalignment inputs have been expanded from single values for the whole ESP to
sectional values in the SNEAKAGE menu to allow for section-to-section
variations.

The Turbulent Core Number represents a different way to calculate the effects
of turbulence. Essentially, the change is from a Highly turbulent Deutsch-type
collection to a series of laminar collection zones separated by regions of
complete mixing. The core number describes the number of laminar zones per
element and generally should have the value 1. That is, as particles pass
under and element (wire), they experience an electrical force much stronger
than the turbulent forces that cause mixing and are collected as if the mixing
did not exist; once past the element, the mixing dominates again.

To allow the mixing to continue under each element, set the core number to 5

or 10. Then, there would be five or ten mixing zones as the particles pass the
element. To allow for very low turbulence, set the core number to 0.2 to 0.5;
then the particles would pass two to five elements before mixing would occur.

In ESPVI 4.0, the Turbulent Core Fraction served the same purpose but
represented a different concept and could not exceed 1. The new Turbulent
Core Number allows a meaningful parameter to cover the range from full laminar
flow to full turbulent flow.

The files for the Sneakage/Rapping/Turbulence now contain velocity
maldistribution factors, saved into the file whenever maldistribution is
calculated (see below). These values are then used to modify the ideal
performance without requiring the maldistribution calculation each time. A
consequence of this storage is that all the SRT values are saved when the
maldistribution is calculated; you therefore need to be certain the right
parameter values are current when the operation is performed.

DATA/FLUE GAS/COMPOSITION - The EXCESS AIR entry has been removed since it was
not used in the calculations. The gas composition should be the composition at
the inlet of the ESP, where any excess air or leakage would be included.

DATA/PARTICLE - A new input menu has been added, FREQUENCY. This is the
differential form of the HISTOGRAM input menu that deals with the cumulative
distribution. The cumulative values are used to derive relative frequency
values that can be edited in FREQUENCY. If FREQUENCY is used for editing, the
resulting cumulative values may change slightly, because the frequency values
are normalized to 100 percent (sum of frequency percentages = 100), while the
cumulative input does not require that the final cumulative total be 100
percent.

FREQUENCY may be useful in emphasizing or reducing particular particle size
contributions, such as in submicron sizes, or in transferring differential
size distributions from published graphs. HISTOGRAM is still the preferred
method for entering data from impactors. With FREQUENCY, it is easy to zero
out a particular size contribution; with HISTOGRAM, you must consider the
effect on the whole distribution.

The cumulative distribution graph has been changed to a logarithmic scale to
show more detail at the small diameters and the relative frequency of each
particle size has been added to the display.

V-I CALCULATION - There is one new choice on the menu, "Automatic V-I
Calculation." This choice will calculate a V-I operating point by stepping the
peak voltage upward in the increments determined by Set V-I Options. However,
"Automatic" will terminate the ramping of voltage according to three criteria:
sparking onset, heavy back corona onset, or power supply limits. This option
differs from "V-I Curve Calculation" in having the termination procedure. The
computation may be viewed as printed values or graphical values, according to
the Graphics On/Off switch. A message describing the reason for the
termination is written at the bottom of the screen each time the V-I curve for
one section is finished.

The power supply limits (voltage and current) are new inputs requested under
"Operating Voltages," as described later. If zero values are used for the
power supply limits (to indicate unknown values), the termination will be only
for sparking or back corona.

PERFORMANCE CALCULATION - Most of the changes and additions to ESPVI 4 appear
in this menu. What was formerly a single choice, "Performance Calculation,"
has been expanded into several. There is "Standard Method," three types of
adjustments, and "Compute Maldistribution." "Standard Method" comes closest to
the original "Performance."

The reasons for the additions are as follows. Repeated testing with ESPVI has
convinced us that the model will accurately predict total section currents if
all the input parameters are correct (voltage, peak-to-average ratio, corona
onset factor, particle size distribution, etc.) Unfortunately, in most cases,
one or more of the parameters are not known, but the total current or current
density is known. The adjustment calculations are designed to seek out values
for parameters that match the predicted current to the known current.

In addition, the velocity maldistribution correction that was used in the
original ESPVI was an approximation. As we gained better understanding of the
effects of maldistribution, it became apparent that another type of
calculation was needed. The present approach computes the ESP performance at
seven gas velocities distributed around the mean value according to the
velocity standard deviation. The individual computed performances are then
combined in a weighted manner to obtain the total maldistribution effect.
Factors that indicate the ratio of performance with maldistribution to the
performance without maldistribution are computed and stored for future
computations.

"Compute Maldistribution" performs the seven-fold computation explicitly and
should be performed whenever the total efficiency of the ESP changes by more
than a few percent. "Standard" will also perform the maldistribution
computation if 1) it has never been performed before, or 2) certain parameters
have changed enough to require a new computation. The parameters that are
checked for changes are SCA, plate area, gas volume flow, gas velocity, and
rapping or steady reentrainment factors. All of these require a recomputation
if they are changed by more than a few percent. The maldistribution factors
are stored in the Sneakage/Rapping/Turbulence files automatically after each
seven-fold computation. If the Standard computation is used without the
maldistribution computation being required, then the prestored maldistribution
factors are used to modify the ideal performance calculation.

The "Adjust xxx" choices use the given current density and the computed
current density to adjust a parameter. "Adjust Peak Factor" varies the
peak-to-average ratio for a match. "Adjust Onset Factor" changes the corona
onset factors for the whole section by the same amount to reach a matching
condition. "Adjust Particle Dist" changes the number concentration at 0.2 æm
to match only the first section's current density. These are explained in more
detail below.

The peak-to-average ratio is a critical parameter whose value is rarely known.
Very few measurements have been reported in the literature, and it is
sensitive to the electrode geometry and power supply configuration. Therefore,
we think it reasonable to adjust the parameter for a current density match,
within bounds. If power supply limits have been provided, the peak-to-average
ratio will not be allowed to produce voltages above the supply peak voltage
limit.

The corona onset for each element is set in the electrode input menu. The
adjustment is applied to all elements of the section at one time in the form
of an additive constant, called a corona onset bias. The values of the bias

may be displayed and stored in the Operating Voltage menu; they may also be
rest to zero there. By using a section-wise bias, the same electrode file may
be used for many sections without requiring a separate electrode file to
accommodate each corona onset value. It is also difficult to justify
adjusting the corona onsets of single elements to match the current for the
whole section, at the present level of knowledge of ESP operation.

In fitting ESP measurements with the model, cases of high voltages and low
currents in the first section of an ESP have been encountered. These cases
suggest the presence of a strong particulate space charge, yet opacity
measurements do not show abnormally high particle concentrations. It seems
likely that submicron particles are present, contributing to the space charge,
but invisible as far as opacity is concerned. Such submicron peaks have been
observed experimentally. "Adjust Particle" varies the concentration of
particles of about 0.2 æm diameter to match the current in the first section
only. Later sections might have to be adjusted by other means, but often fall
in line when the first section is made to match. This choice is really only
effective when the computed current density is well above the measured current
density.

Each of the Adjust xxx operations is independent of the others. They can be
applied sequentially in any order, but the results may well be different. For
instance, if adjusting the corona onset improves the match between calculated
and measured currents but does not match them precisely, the peak-to-average
ratios can then be adjusted to finish the match. This may result in a
different set of operating conditions and collection performance than if the
peak-to-average ratios were adjusted first. Remember, the adjustments are
included to compensate for imperfect knowledge of the operating conditions.

If the corona onset voltages are known, even roughly, the corona onset factors
should be set to give the proper onset voltages and not be adjusted to match
the measured currents. If the peak-to-average ratios have been measured, they
should be set at the measured values and not be adjusted to match the measured
currents. The user will have to develop some awareness of appropriate and
inappropriate values for the adjusted parameters.

VIEW RESULTS - In the particle description page, the differential inlet
concentrations have been added. These can be used with the penetrations to
compute the outlet differential distributions.

GRAPH RESULTS - A new graph has been added, Wave Form. This graph shows the
computed wave form one section at a time. When half-wave or intermittent wave
forms are displayed, the inactive half-cycles are displayed at low amplitude
along the lower axis. The internal TR set voltage is shown as a dotted
sinusoidal curve. The actual electrode voltage is the solid curve. A
double-dotted horizontal line represents the average voltage of the wave form.
A black, dashed line below the wave form represents the lowest voltage reached
by one wire in the section; other wires may not be as low. The red, dot-dashed
line represents the corona onset voltage for the section under clean gas
conditions.

In the size-dependent penetration and cumulative mass graphs, the pure log-
arithmic scale has been changed to a bounded logarithmic scale. A bounded
logarithmic scale has the values in the lowest decade compressed so that all
values down to and including zero can be plotted. This scale has been incor-
porated because the graphs used fixed lower bounds anyway, so many low values

were simply not plotted. Now they will be plotted, but at very low resolution.
Except for the baseline (0 value) on the plot, all the other grid lines shown
are precisely the same as their counterparts on a pure logarithmic scale.

UTILITIES & CONFIGURATION - The only change here is that the configuration is
automatically saved when the menu is exited, whether or not any changes have
been made. The same configuration file is read whenever entrance is made into
the menu.

ADDITIONAL OUTPUTS

The internal values of several calculations are now stored in output files for
further use. In general, if you don't want these files, you can turn them off
by setting Print Details to OFF in the configuration menu.

The charge on particles at the outlet of each section is stored in the file
named "master.CHG", where "master" is the Master File name. The values are
stored as the number of electrons on each particle and as the ratio of actual
charge to the calculated saturation charge. These files may be viewed from
within the program under FILE/BROWSE/CHARGE.

Values of local electric field in the gas and in the dust layer are displayed
in files named "master.INC", with the same naming convention as for the charge
files. The increments in the file are the upstream and downstream values for
each element. Sparking conditions are indicated with an asterisk (*) next to
the gas field values; back corona conditions are indicated with and asterisk
next to the dust field values. The local current density and particulate space
charge are also shown in the file. These files may be viewed from
within the program under FILE/BROWSE/INCREMENTAL (or FILE/ACCESS if the
results are too long.)

Finally, outlet particle concentrations for each particle size are displayed
in files named "master.PTn", where n is a digit from 0 to 9. These files
represent the last ten outputs that have been calculated. The files also
contain the gas velocity used during the calculation.

These files are intended to be used in cases where the flow or electrical
conditions are so different that one pass of the model is not adequate.
Consider a 4 chamber ESP with unequal division of flows. The collection
performance of each chamber can be predicted with the velocity and voltages
appropriate for each, one after another. The output concentrations would be
stored in "master.PT0", "master.PT1", "master.PT2", and "master.PT3".

If you then load each file into a spreadsheet program (or calculate by hand),
you should multiply each particle concentration by the gas velocity of that
data set, then add the velocity-concentration products for each size in all
the data sets, and then divide by the sum of the gas velocities for the four
sets. This produces a velocity-weighted average outlet concentration that can
be used to obtain total emissions, opacity, etc.

Three files have been supplied for use in printing the graphical screens:
VGAEGRAF.COM, VGAHGRAF.COM, and VGAHRSYS.SYS. The file VGAEGRAF.COM can be
installed with a line in the AUTOEXEC.BAT file; it translates the graphics
screens into a form that EPSON-compatible printers can print. The file
VGAHGRAF.COM can be installed the same way and translates the screens for
Hewlett-Packard-compatible printers. Either of the two files should be
installed in a location such as the DOS directory from where it could be
loaded by the appropriate line in the AUTOEXEC.BAT file. Some printers such
as the Hewlett-Packard Deskjet will reverse the black/white values on the
printed page with the above files. In that case the file VGAHRSYS.SYS should
be installed in the ESPVI 4.0a directory, and the following line inserted in
the CONFIG.SYS file:

DEVICE=C:\ESPVI\VGAHRSYS.SYS

This file translates for Hewlett-Packard Deskjet type printers, reversing the
black/white values on the page. If VGAHRSYS.SYS is used neither of the other
two files should be loaded. If either a drive other than C or a directory
other than ESPVI is used for ESPVI 4.0a, appropriate changes should be made to
the CONFIG.SYS line.

Another option is to install the DOS program, GRAPHICS.COM, contained in
MSDOS 6.x, with an appropriate command in the AUTOEXEC.BAT file. With this
option use the appropriate profile (see DOS help) for the printer you have.
An advantage with this option is that it provides full size printouts with a
higher level of detail than does the other options, which only provides
half-page printouts.