United States
Environmental Protection
Agency
Industrial Environmental
Research Laboratory
Research Triangle Park NC 27711
Research and Development
EPA-600/S7-84-054 May 1984
Project Summary
Computer Programs for
Estimating the Cost of Paniculate
Control Equipment
Andrew S. Viner and David S. Ensor
The report describes an interactive
computer program, written to estimate
the capital and operating expenses of
electrostatic precipitators (ESPs) fabric
filters, and venturi scrubbers used on
coal-fired boilers. The program accepts
as input the current interest rate, coal
analysis, emission limit, and design and
operating parameters of the control
device. The installed cost of the collector
and the annual fixed and variable
operating and maintenance costs are
estimated. Based on the interest rate
specified, an annual payment of interest
and principal is calculated for the
amount of capital required. This annual
capital cost is added to the annual
operating and maintenance costs to
yield a total annual cost of the collector.
A comparison between reported and
predicted costs indicates that the model
is capable of ±25 percent accuracy.
This Project Summary was developed
by EPA's Industrial Environmental
Research Laboratory, Research Triangle
Park, NC, to announce key findings of
the research project that is fully docu-
mented in a separate report of the
same title (see Project Report ordering
information at back).
Introduction
Predicting the lowest cost of meeting a
proposed particulate emissions standard
requires a rigorous engineering analysis
of the available control options. A
complete engineering analysis of a
project includes process design, estima-
tion of equipment costs, and investment
analysis. Although the methods used in
process design and investment analysis
are well developed, there are no conven-
ient means for estimating the cost of
pollution control equipment.
This report describes a set of empirical
models developed to aid in estimating
capital and operating costs of particulate
control systems for coal-fired boilers. The
models predict process costs based on
the design and operating conditions of
the control device and are capable of ±25
percent accuracy. The output from the
models includes an itemization of the
capital costs, the annual operating and
maintenance costs, annual capital cost,
and total annual cost. These models have
been implemented into a computer
program that is available on the Radio
Shack® TRS-80 Models I and III compu-
ters. Complete details of the models and
the computer program are included in the
full report.
Cost Modeling
The simplest way to estimate the
capital cost of equipment of known size is
to scale it according to the size and cost of
an existing unit of the same design. The
equations used/ in the cost estimation
models are similar to Williams' "six-
tenths factor"1 (where the size increase
of the equipment is raised to a fractional
power) and (are based on estimated
equipment costs. There is a separate
equation for each major equipment item or
group of items; e.g., one equation is used
to estimate the cost of a coldside ESP,
including the cost of the ESP and hoppers,
hopper heaters, electrical hardware,
control room, and ESP support steel.
Other equations are used to estimate the
cost of insulation, ducting and supports,
ash handling system, ash pond, and
induced draft fans. Sometimes, more
-------
than one size factor is used in the
calculation with a separate exponent for
each. For example, the ESP cost model
predicts capital cost for an ESP based on
its specific collection area and flow rate
through the ESP. These equations were
developed by Chapman et al.2 The data
from which the ESP and baghouse
models are derived are contained in a
report by Campbell et al.3 The data for the
scrubber cost model are from reports by
Ponder et al.4 and Kinkley and Neveril.5
The calculation of operating and
maintenance costs is much more straight-
forward because these costs are directly
proportional to system size and the
amount of throughput of the control
device. For example, the cost of mainte-
nance for an ESP is estimated to be
$0.769/m2 of collector per year at full
capacity. For an ESP operating at only 70
percent of capacity, the cost would be
proportionately less. The cost of operating
labor and materials and electricity
required to run the control device is
calculated similarly.
Accuracy of the Model
Predictions
Since the models for capital cost are
based on estimates of equipment costs, it
is prudent to compare the predictions of
the models with accurate and well-
documented data on the cost of existing
units. These stringent requirements for
data quality eliminate guesses and
estimates from models similar to our
own. Table 1 lists available data.
Predictions of costs for the units listed
in Table 1 are from several sources.
Pertinent design and operating informa-
tion was obtained from the reports cited
in Table 1. Where information was not
available, reasonable values were assumed.
For parameters where it was necessary to
assume a value, the same assumption
was made for all similar control devices
(e.g., all fabric filter bags have a life of 2
years). Engineering and contingency costs
were each assumed to be equal to 20
percent of the total field cost of the project.
All of this information was input to the
computer program (described below) to
obtain the results. For a comparison
between these predicted costs and the
actual costs to be made, it was necessary
to subtract the retrofit costs from the
reported costs. In the absence of an
itemized retrofit cost, a value equal to 10
percent of the reported total field cost was
subtracted, based on recommendations
by Ensor et al.8
Results of an overall comparison
between the actual and predicted costs
Table 1. Sources of Cost Data on Actual Installations
Plant
Cherokee #3 Ir)
Nucla (rj
Kramer (r)
Sunbury (r)
Shawnee {r)
Harrington #2
George Neal #3
Plant A
Navajo
Type of collector
TCA wet scrubber
Baghouse
Baghouse
Baghouse
Baghouse
Baghouse
ESP coldside
ESP hotside
ESP hotside
Size
MW
160
38
US
175
1.750
350
500
576
750
Year of
installation
1971
1974
1977
1973
1978
1977
1975
1975
1977
Citation
Ensor et a/.6
Ensor et al.7
Ensor et a/.8
Cass and Brad-
way9
Hudson et al.™
Turner"
Ensor et a/.12
Sparks'3
Merchant and
Gooch"
(r) = retrofit.
are shown in Figure 1. The dashed lines in
the figure represent 25 percent bounds
on the x=y line. That is, the upper dashed
line represents a predicted cost that is 25
percent greater than the corresponding
reported cost, and the lower dashed line
indicates a predicted cost that is 25
percent less than the reported cost. Some
of the data points have error bars around
/O"
them to denote a range of values for
predicted costs because the reported
costs for these data points include the
costs of two or more control devices at
one site. The upper value of the range
represents the cost of a single large
control device equivalent in size to the
actual units, and the lower value repre-
sents the predicted cost of two or more
I 7°7
ts
1
4
Z-TITTZI) ±25% Limits
O Scrubber Cherokee
A ESPs
1. George Neal C75)
2. Navajo (77)
3. Plant A (7S)
Q Baghouses
1. Nucla C74)
2. Kramer (77)
3. Sunbury (73)
4. Harrington (77)
5. Shawnee (78)
figure 1.
70'
Reported Cost, $
Comparison between actual and predicted costs.
10*
-------
independent projects. The best prediction
should fall somewhere within this range,
accounting for both the added cost of
separate installations and the savings
associated with a large-scale project.
No attempt was made to verify the
accuracy of the predictions of operating
and maintenance costs, but they are
believed to be adequate. Since the annual
capital cost represents the major portion
of the total annual cost, even a large error
in prediction of operating expenses would
not affect the ±25 percent accuracy of the
total annual cost.
Discussion
Most of the points in Figure 1 lie within
the ±25 percent error bounds with two
notable exceptions: the Sunbury baghouse
and the George Neal precipitator. For the
Sunbury baghouse there is reason to
question the accuracy of the reported
costs. The Sunbury unit was the first
fabric filter installed on a utility boiler.
Conventional wisdom says that first-time
installations have high costs relative to
later projects of the same scope. Costs
tend to decline as one progresses along
the "learning curve" and acquires
experience with the construction, install-
ation, and start-up of control devices. In
spite of this reasoning, the reported cost
is lower than expected. One slight
advantage that the Sunbury plant can
claim is the use of a stripped ESP shell for
the fabric filter housing. However, this
salvage is not enough to account for the
very low cost. It can only be speculated
that this project went very smoothly,
without costly errors, or that the reported
costs are incomplete.
The George Neal ESP is the other point
that lies outside of the ±25 percent error
bounds. To understand this discrepancy,
it is helpful to look at the more detailed
breakdown of actual and reported costs in
Table 2. The predicted and reported costs
for each item agree fairly well except for
the miscellaneous charges and the
engineering costs. There is no ready
explanation for the difference between the
reported and predicted miscellaneous
costs. This category is difficult to predict
because it includes a variety of items
such as earthwork, concrete foundations,
painting, and all of the indirect charges.
The model developed for the miscellane-
ous cost may need some fine tuning;
however, insufficient information is
available to make any changes.
The other source of discrepancy
between the reported and predicted costs
is the engineering cost. As mentioned
earlier, the engineering cost is estimated
Table 2. George Neal Costs Summary (1975 Dollars x 1C?)
Reported
Predicted"
Predicted*
Collector and supports
Ducting and supports
Ash removal system
Insulation
Miscellaneous
Total field cost
Engineering
Contingency
Turn-key cost
11.200
2,100
1.700
1.200
3,200
19.400
600
C
20,000
12.1OO
3,360
1.450
1,390
7.580
25.900
4,520
C
30,400
72,300
3,760
1.700
1.420
7,490
26.700
4.720
C
37,400
"Cost computed for one 530-MW ESP.
toCost computed as twice the cost of one 265-MW ESP.
^Contingency has been factored into equipment costs.
as a fraction of the total field cost. In this
case a value of 20 percent was assumed
to be reasonable. Obviously this charge is
too high, but note that it is difficult to
predict the engineering charges without
a priori knowledge of what obstacles may
be faced during the course of the project.
When in doubt it is wise to err to the
conservative side in cost estimation.
Overall, the models do a fairly good job
of predicting the cost of particulate
control devices, usually within ±25
percent. It must be emphasized, however,
that the accuracy of the model predictions
is highly dependent on the quality of the
design and operating information supplied
by the user.
Computer Program
The COST program was written to
facilitate use of these cost estimation
techniques. It is written in the BASIC
computer language and is designed with
the user in mind. Program features
include: interactive data entry with
default values available, English and
metric units, fast execution, hard copy
printout, and storage of input and output
parameters in files for easy retrieval. To
ensure the ease of use and clarity of
documentation, the program was dis-
tributed to a small group of users whose
experience ranged from beginner to
advanced programmer. Their helpful
suggestions and comments have been
incorporated into the final version of the
program.
The current program has one important
limitation: the COST program will not
determine the design and operating
criteria required to meet a given emissions
limit. The user of the program is required
to do the necessary modeling to determine
the specific collection area and electrical
characteristics for ESPs, air-to-cloth
ratio and pressure drop for fabric filters.
and liquid-to-gas ratio and pressure drop
for venturi scrubbers. It is hoped that the
ability to model these devices will be
incorporated into a later version of the
COST program.
Figure 2 is a simplified flowsheet of the
COST program. The first step in the
execution of the program is the input of
the plant, coal, economic, and emissions
data. Figure 3 shows a sample input
worksheet that lists the parameters
required by the program at this point.
Similar worksheets aid the user in
collecting input parameters for each
control device. After the plant information
has been entered, the program will do
some preliminary calculations and then
save all of the input data on a disk. The
next step is selection of the desired
control device. The user selects either
the fabric filter, ESP, or venturi scrubber
model. Finally the design and operating
information must be entered. This
information can be based on an operating
plant, or it can be obtained from a
computer model of the device. For
example, the GCA fabric filter model,15
the venturi scrubber performance model,16
and the SoRI or RTI EPA models16'17 can
be used to obtain the necessary design
and operating parameters for a given
emission limit. It is up to the user to
supply meaningful input to the program.
Once the necessary information has been
entered, the program calculates the
capital costs, fixed annual operating cost,
variable annual operating cost, cost of
electricity, annual capital cost, and the
total annual cost. The results are displayed
on the computer's video screen and
stored on a disk to be recalled when
necessary. The program can also produce
a hard copy of the results (see Figure 4) on
a line printer.
Most of the time required to get results
from the COST program is spent filling in
-------
Fabric
Filter
Costs
^
Coal Data
Plant Data-*.
Emission -
Limit ,-
Economic
Data
A — Select a control device.
B = Data saved on disk for future use.
Figure 2. COST program flow sheet.
PLANT INPUT WORKSHEET
Plant. _!!£?
Date: 6/12/81
R marks- Demonstratlon of tne COST program
PLANT DA TA
Boiler size
Capacity factor
Chemical Engineering Plant Cost Index
Emission limit
Plant altitude
Interest rate
Cost of electricity
Contingency as percent of field cost
Engineering as percent of field cost
Stem cycle efficiency
Coal heating value
Fraction excess air
Percent carbon in coal
Percent hydrogen in coal
Percent oxygen in coal
Percent sulfur in coal
Percent nitrogen in coal
Percent ash in coal
Percent water in coal
Figure 3. Plant input worksheet.
500
70
-MWnet
-%
270
43
300
-ng/J
or
or
Jb.MBtu
Jt
15
30
-%/yr
-mill/kWh
20
20
38
23.240
0.25
-%
JtJ/kg
-Btu/lb
60
5.4
11.2
0.6
1.6
7.6
13.6
the worksheets. The program itself can be
run in less than 5 minutes.
System Requirements
The COST program is written for a
Radio Shack® Level II, Model I or Model III
TRS-80 microcomputer with 48 kilobytes
of RAM and at least one SVi-inch floppy
disk drive. The program requires Radio
Shack's TRSDOS Disk Operating System
(DOS) or any other DOS that is capable of
running Microsoft Disk Basic and is file-
compatible with TRSDOS. The COST
program is also designed to provide a
listing of results on a line printer.
Complete documentation for the program
is contained in the full report.
Summary
The empirical models that have been
developed provide a convenient and
reliable means for estimating capital and
operating costs for paniculate control
systems. The results from these models
can be used in an investment analysis to
compute levellized cost or some other
measure of merit. The computer program
that was written to implement these
models provides a quick and easy way to
compare alternative designs and check
estimates.
At present, the models cannot consider
any site-specific factors that would affect
costs, nor can they consider the effect of
new technologies on capital and operating
costs. It is possible to expand the models
to provide this flexibility. The utility of the
computer program can be enhanced by
incorporating control device performance
models to allow a user to predict the
collector performance more easily before
going on to the cost estimation procedure.
References
1. R. Williams. Chemical Engineering,
54(12):124, 1947.
2. R.A. Chapman, D.P. Clements, LE.
Sparks, and J.H. Abbott. Cost and
Performance of Paniculate Control
Devices for Low Sulfur Western
Coal. In: Second Symposium on
the Transfer and Utilization of
Paniculate Control Technology,
Vol. I. EPA-600/9-80-039a (NTIS
PB 81-122202), September 1980.
3. K.S. Campbell, et al. Economic
Evaluation of Fabric Filtration
Versus Electrostatic Precipitation
for Ultrahigh Paniculate Collection
Efficiency. Electric Power Research
Institute Report No. EPRI FP-775,
June 1978.
4. T.C. Ponder, Jr., et al. Simplified
Procedures for Estimating Flue
-------
OUTPUT FROM fARTICULATE CONTROL COST MODEL VEF.SION 1.1
PLANT : KRAMER
DATE : 09/23/81
REMARKS : COMPARISON UITH ACTUAL COSTS
PACE 1 OF 3
PLANT DATA
BOILER SIZE
CAPACITY FACTOR
CHEM. ENC. PLANT COST INDEX
EMISSION LIMIT
PLANT ALTITUDE
INTEREST RATE
COST OF ELECTRICITY
CONTINGENCY AS PERCENT OF FIELD COST
ENGINEERING AS PERCENT OF FIELD COST
STEAM CYCLE EFFICIENCY
CAS FLOW RATE
BOILER EMISSIONS
COAL HEATING VALUE
FRACTION EXCESS AIR
PERCENT CARBON IN COAL
PERCENT HYDROGEN IN COAL
PERCENT OXYGEN IN COAL
PERCENT SULFUR IN COAL
PERCENT NITROGEN IN COAL
PERCENT ASH IN COAL
PERCENT HATER IN COAL
115 MH NET
69.0 X
202.0
0.0 LB/H6TU
973 FT
20.00 X/YR
30.00 HILU/KWH
20.0 X
20.0 X
38.0 X
558 KACFH
0.0 LB/M6TU
10,911 BTU/LE
0.25
67.50 X
5.30 X
10.60 Z
0.30 X
1.51 X
4.28 X
11.00 X
OUTPUT FROM PARTICULATE CONTROL COST MODEL VERSION 1.1
PLANT : KRAMER
DATE : 09/23/81
PACE 2 OF 3
SACHOUSE NAME : e*ci
REMARKS : ONE LARGE CACHOUSE
6ACHOUSE DATA
CAS TEMPERATURE
EXFECTED BAGHOUSE LIFE
AIR TO CLOTH RATIO
MAXIMUM PRESSURE DROP
AVERAGE PRESSURE DROP
NUMBER OF MODULES
REVERSE AIR FAN SIZE (KH/10C5 M2 BAG AREA)
BAG LIFE
fAG LIFE EXPONENT
PAG REPLACEMENT COST
LAE'OR RATE
MATERIAL OVERHEAD FRACTION
FAN LOAD FACTOR
FAN EFFICIENCY
HOPFER HEATER DUTY FACTOR
ACCESORIES DUTY FACTOR
ASH REMOVAL SYSTEM DUTY FACTOR
SCHEDULED fc»G REPLACEMENT TIME
UNSCHEDULED EAC REPLACEMENT TIME
FRACTION OF UNSCHEDULED EAC REPLACEMENTS
NON TAG MAINTENANCE TIME (HR/YR/1QOO CMS)
365
20
1.690
6.00
4.50
46
165
2.0
0.600
* 0.65
» 14.00
0.100
0.820
80.0
0.300
o.aoo
0.600
2
7
o.oso
560
F
YRS
FT/MIN
IN H20
IN H20
YEARS
/FT2
/HR
X
MIN/M2
MIN/M2
OUTPUT FROM PARTICULATE CONTROL COST MODEL VERSION 1.1
•PLANT : KRAMER
DATE t 09/13/81
PACE 3 OF 3
FACHOuse NAME : IACI
REMARKS : ONE LARGE BACHOUSE
RESULTS
COLLECTOR ( SUPPORTS
DUCT INC i SUPPORTS
ASH REMOVAL SYSTEM
INSULATION
ASH POND
ID FAN
MISCELLANEOUS
ENGINEERING
CONTINGENCY
FIXED OPERATING COSTS
VARIABLE OPERATING COSTS
COST OF ELECTRICITY
jure 4. An example of the hard copy printout.
2.92E-06
3.13E»05
S.39E+OS
9.69£*03
Z.43E»OS
S.74E*04
1.77E»06
1.36E»06
1.36E-06
7.S2E»04
1.46E»05
1.27E-05
• • 6.81E+06
Gas Desulfurization System Costs.
EPA-600/2-76-150 (NTIS PB
255978), June 1976.
5. M.L. Kinkley and R.B. Neveril.
Capital and Operating Costs of
Selected Air Pollution Control
Systems. EPA-450/3-76-014(NTIS
PB 258484), May 1976.
6. D.S. Ensor, et al. Evaluation of a
Paniculate Scrubber on a Coal-
Fired Utility Boiler. EPA-600/2-
75-074 (NTIS PB 249562), Novem-
ber 1975.
7. D.S. Ensor, R.G. Hooper, and R.W.
Sheck. Determination of the Frac-
tional Efficiency, Opacity Charac-
teristics, Engineering and Economic
Aspects of a Fabric Filter Operating
on a Utility Boiler. Electric Power
Research Institute Report No. EPRI
FP-297, November 1976.
8. D.S. Ensor, et al. Kramer Station
Fabric Filter Evaluation. Electric
Power Research Institute Report
No. EPRI CS-1669, January 1981.
9. R.W. Cass and R.M. Bradway.
Fractional Efficiency of a Utility
Boiler Baghouse: Sunbury Steam-
Electric Station. EPA-600/2-76-
077a (NTIS PB 253943), March
1976.
10. J.A. Hudson, et al. Design and
Construction of Baghouses for
Shawnee Steam Plant. In: Second
Symposium on the Transfer and
Utilization of Particulate Control
Technology, Vol. I. EPA-600/9-
80-039a (NTIS PB 81-122202),
September 1980.
11. J.H. Turner. Application of Fabric
Filtration to Combustion Sources.
EPA-600/J-80-112 (NTIS PB 81-
126484), AlChE Symposium Series
76, No. 196, pp. 369-379, 1980.
12. D.S. Ensor, et al. Evaluation of the
George Neal No. 3 Electrostatic
Precipitator. Electric Power Re-
search Institute Report No. EPRI
FP-1145, August 1979.
13. Personal communication with L.E.
Sparks (EPA/IERL-RTP), 1982.
14. G.H. Merchant, Jr., and J.P.Gooch.
Performance and Economic Eval-
uation of a Hot-side Electrostatic
Precipitator. EPA-600/7-78-214
(NTIS PB 292648), November
1978.
15. R. Dennis and H.A. Klemm. Fabric
Filter Model Format Change; Vol-
ume I. Detailed Technical Report.
EPA-600/7-79-043a (NTIS PB
293551), February 1979.
16. S.J. Cowen, D.S. Ensor, and L.E.
Sparks. TI-59 Programmable Cal-
culator Programs for In-Stack
-------
Opacity, Venturi Scrubbers, and
Electrostatic Precipitators. EPA-
600/8-80-024 (NTIS PB 80-193147),
May 1980.
17. P.A. Lawless, J.W. Dunn, and LE.
Sparks. A Computer Model for ESP
Performance. In: Third Symposium
on the Transfer and Utilization of
Paniculate Control Technology,
Vol. II. EPA-600/9-82-005b (NTIS
PB 83-149591), July 1982.
Andrew S. Viner and David S. Ensor are with Research Triangle Institute,
Research Triangle Park, NC 27709.
Leslie E. Sparks is the EPA Project Officer (see below).
The complete report, entitled "Computer Programs for Estimating the Cost of
Paniculate Control Equipment," (Order No. PB 84-183 573; Cost: $16.00,
subject to change) will be available only from:
National Technical Information Service
5285 Port Royal Road
Springfield. VA 22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
Industrial Environmental Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
-------
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
Official Business
Penalty for Private Use $300
"V Vs
' i'j.^.ft.. .VJ'.1-'
?_••••! •.-- •> nj i»
U.S. GOVERNMENT PRINTING OFFICE: 1984• 759-102/9691
------- |