PB83-113837
Hitachi Zosen NOx Flue Gas Treatment Process
Volume 2. independent Evaluation
Radian Corp.
Austin, TX
Prepared for
Industrial Environmental Research Lab,
Research Triangle Park, NC
Sep 82
U.S. DEPARTMENT OF COMMERCE
National Technical Information Service
'-
-------
PB83-113637
EPA-600/7-82-057b
September 1982
Hitachi Zosen NOX Flue Gas Treatment
Process; Vol. 2. Independent
Evaluation
by
RADIAN CORPORATION
8501 Mo-Pac Blvd.
Austin, TX 78759
EPA Contract No: 68-02-3171
Work Assignment 11
Project Officer: J. David Mobley
Industrial Environmental
Research Laboratory
Research Triangle Park, NC 27711
Prepared for:
U.S. ENVIRONMENTAL-PROTECTION AGENCY
Office of Research and Developmenitjg,-
Waahington, D.C. 20460
-------
TECHNICAL REPORT DATA
(Please read Inwuc lions on the reverie before completing)
1 REPORT NO
EPA-600/7-82-057b 12.
3 RECIPIENTS ACCESSIOI* O
P683 11 3,’3 7
DATE -
4 TITLE AND SUBTITLE Hitachi Zosen NOx Flue Gas Treat
ment Process; Vol. 2. Independent Evaluation
5 REPORT
September 1982
6 PERFORMING ORGANIZATION CODE
7 AUTHOR(S)
a PERFORMING ORGANIZATION REPORT NO
J.M. Burke
PRC GRAM ELEMENT NO
9 PERFORMING ORGANIZATION NAME AND ADDRESS
Radian Corporation
8501 Mo—Pac Boulevard
Austin, Texas 78759
il CONTRACT/GRANT NO
68—02—3171, Task 11
12 SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
13 TYPE OF REPORT AND PERIOD COVERED
Task Final; 2/80—2/82
CODE
14 SPONSORING AGENCY
EPA/600/13
15 SUPPLEMENTARY NOTES IERL—RTP project officer is J. David Mobley, Nail Drop
61, 919/541—2578.
15 AbSTRACT The report gives results of an independent evaluation of the
Hitachi Zosen (Hz) NOx flue gas treatment process, one of two selective
catalytic reduction (ScR) of NOx with ammonia processes (pilot-scale
tested under EPA sponsorship) treating flue gas slipstreams from coal—
fired boilers. Test results show that the process can reduce NOx emis-
sions by 90%. Initial tests resulted in plugging of the catalyst. But a
new catalyst with larger gas passages was tested: it operated for 5500
hours without any signs of plugging. Results of an energy analysis indi-
cate that the HZ process energy requirements are 0.3% of the boiler’s
capacity. Process costs were estimated based on the pilot plant test
results. Estimated capital investment and annual revenue requirements
for the HZ process are $44/kW and 2.91 mills/kWh, respectively. These
costs are slightly lower than previous estimates for the process.
17 KEY WORDS AND DOCUMENT ANALYSIS
a DESCRIPTORS
b IDENTIFIERS/OPEN ENDED TERMS
C COSATI F,eldGfoup
Pollution
Nitrogen Oxides
Catalysis
Ammonia
Flue Gases
Evaluation
Coal
13 DISTRIBUTION STATEMENT
Pollution Control
Stationary Sources
Hitachi Zosen Proces
Selective Catalytic
Reduction
Flue Gas Treatn ent
19 SECURITY CLASS (misReport)
133
07B
07D
21B
14G
?-1-D
21 NO OF PAGES
Release to Public
Unclassified
206
20 SECURITY CLASS (This page)
Unclassified
22 PRICE
EPA Form 2220-1 ($73)
: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 does not constitute endorse-
ment or recommendation for use.
ii
-------
CONTENTS
Section Page
SUMMARY 1
1.1 Introduction 1
1.2 Program Objectives and Approach 3
1.3 Results 4
1.3.1 Pilot Plant Test Results 5
1.3.2 Independent Evaluation Test Program Results 6
1.3.3 Results of the Conceptual Design of a
500 MW HZ Process 12
1.3.4 Results of the Material Balance Calculations
for a 500 MW HZ Process 14
1.3.5 Results of the Energy Balance Calculations for
for a 500 MW HZ Process 15
1.3.6 Results of the Cost Estimate for a 500 MW HZ
Process Application 16
1.4 Conclusions 21
2 INTRODUCTION 23
3 FIELD TEST PROGRAN 27
3.1 HZ Process 28
3.2 Hitachi—Zosen/Chemico Pilot Plant Description 29
3.3 Hitachi—Zosen/Cheird co Pilot Plant Test Program 33
3.3.1 Optimization Tests — Objectives 34
3.3.2 Optimization Tests — History 36
3.3.3 Demonstration Test 39
3.4 Independent Evaluation of the HZ Process 41
3.4.1 Quality Assurance 42
3.4.2 Continuous N0 Monitor Certification 42
3.4.3 Stack Sampling for Secondary Emissions 46
ill
-------
Contents (Continued)
Section Page
4 HITACHI ZOSEN PILOT PLANT TEST RESULTS 49
4.1 Optimization Tests 50
4.2 Demonstration Test Results 51
4.3 Quality Assurance Audit Test Results 58
4.4 Continuous NO Monitor Certification Test Results 62
4.5 Stack Sampling for Secondary Emissions Test Results 63
5 EVALUATION OF THE HITACHI ZOSEN PROCESS 69
5.1 Basis for the Evaluation of the Hitachi Zosen Process.... 70
5.2 Conceptual Design of the Hitachi Zosen Process for
a 500 MW Coal—Fired Boiler 72
5.2.1 Philosophy Used to Prepare the Conceptual Design.. 73
5.2.2 Pilot Plant Data Used to Prepare the Conceptual
Design
5.2.3 Conceptual Design Constraints and Procedures 75
5.2.4 Conceptual Design Results 77
5.3 Material Balance for a 500 MW Hitachi Zosen Process 80
5.4 Energy Requirements for a 500 MW Hitachi Zosen Process... 82
5.4.1 Hitachi Zosen Process Heat Credits 84
5.4.2 Energy Balance for the 500 MW Hitachi Zosen
Process 85
5.4.3 Electrical Energy Requirements of the 500 MW
HZ Process 86
5.4.4 Overall Energy Requirements for the 500 MW
Application of the Hitachi Zosen Process 87
5.5 Estimated Costs for a 500 MW Application of the
Hitachi Zosen Process 88
5.5.1 Estimated Total Capital Investment for a
500 MW Application of the HZ Process 89
5.5.2 Estimated Average Annual Revenue Requirements
for a 500 MW Application of the HZ Process 92
5.5.3 A Comparison of Capital Investment and Annual
Revenue Requirements of the HZ Process
with other SCR Processes 93
5.6 Overall Evaluation of the HZ Process 96
iv
-------
Contents (Continued)
Section Page
5.6.1 Pilot Plant Test Results 97
5.6.2 Results of the Independent Evaluation Test
Program 98
5.6.3 Results of the 500 MW Conceptual Design 101
5.6.4 Results of the Material Balance Calculations 102
5.6.5 Results of the Energy Balance Calculations 103
5.6.6 Results of the Cost Estimate for the 500 MW
Hitachi Zosen Process Application 103
5.6.7 Process Evaluation Summary 105
6 PROCESS DEVELOPMENT REQUIREMENTS 108
REFERENCES 113
APPENDIX A A-i
APPENDIX B B-i
APPENDIX C C4
APPENDIX D D-i
-------
ABSTRACT
Nitrogen oxide emissions from stationary sources may be reduced by 80 to
90 percent through the application of selective catalytic reduction (SCR) of
NO with ammonia. In the interest of furthering the development of this
technology, EPA sponsored pilot scale tests of two SCR processes treating
flue gas slip streams from coal—fired boilers. One of the processes tested
was the Hitachi Zosen (HZ) process. The results of an independent evaluation
of the pilot plant tests of the HZ process show that the process is capable
of reducing NO emissions from a coal—fired boiler by 90 percent. Initial
tests resulted in plugging of the catalyst. But a new catalyst with larger
gas passages was tested and it operated for 5500 hours without any signs of
plugging. Results of an energy analysis indicates the HZ process energy
requirements equal 0.3 percent of the boiler’s capacity. Process costs were
estimated based on the pilot plant test results. Estimated capital investment
and annual revenue requirements for the HZ process are $44/kW and 2.91 mills/kWh
respectively. These costs are slightly lower than previous estimates for
the process.
vi
-------
SECTION 1
SUMMARY
1.1 INTRODUCTION
Selective catalytic reduction (SCR) of NO with NH 3 is capable of reducing
NO emissions by 80 percent or more. As such, 5CR represents the most effective
process available for controlling stationary source NO emissions. For a
utility application of SCR, a catalytic reactor is located between the economizer
and air preheater sections of the boiler. At this location the flue gas tempera-
ture ranges from 300 to 400°C (570—750°F) which is optimum for the catalytic
activity. Ammonia is injected into the flue gas upstream of the catalyst and
reacts with NO on the catalyst surface to form elemental nitrogen and ‘water.
Most SCR processes were developed and are being operated commercially in
Japan, primarily on gas— and oil—fired sources. However, in the U.S., SCR
systems are now being installed on a limited basis. The most notable
application is a demonstration system that is being constructed to treat one-
half of the flue gas from Southern California Edison’s 215 MWe Huntington
Beach Unit No. 2 (an oil—fired boiler).’ The operation of this system is
expected to establish SCR as a commercially available technology for oil—
and gas—fired sources in the U.S.
In Japan, development efforts are currently aimed at applying SCR to coal—
fired sources. To date, most of the SCR process vendors in Japan have operated
pilot units on slip—streams from coal—fired boilers. In addition, there are
now four full—scale SCR systems treating flue gas from coal—fired boilers with
another eight units scheduled for start—up in 1982 and 1983. These development
1
-------
efforts are rapidly establishing SCR as commercially available technology for
controlling NO emissions from coal—fired sources in Japan.
The transfer of SCR technology from Japan to the U.S. for coal—fired
applications presents a potentially significant problem. Since most coal—
fired boilers in the U.S. operate ESP’s located downstream of the air preheater,
a typical SCR application would expose the catalyst in the reactor to the full
particulate concentration from the boiler. Although tests have been conducted
in Japan in which the catalyst was exposed to high particulate concentrations
with no adverse effects, the differences in the composition of particulates
from U.S. and Japanese coals could impact SCR operation.
In an effort to further the development of SCR technology and to determine
how differences between Japanese and U.S. coal/particulate properties impact
the performance of SCR processes, EPA has sponsored pilot scale (0.5 MW equiva-
lent) tests of two SCR systems. One of the SCR systems tested was the Hitachi
Zosen (Hz) process. The HZ pilot plant processed a flue gas slipstream from
a coal—fired boiler. The contractor responsible for the design and operation
of the pilot plant was Chemico Air Pollution Control Corporation (now
General Electric Environmental Services Corporation), which is the North
American licensee for the HZ process. Chemico was also responsible for
collection, evaluation, and reporting of the test data.
The primary objectives of the pilot plant test program sponsored by EPA
were: (1) to demonstrate the ability of the HZ process to achieve a 90 per-
cent reduction in NO emissions, and (2) to determine the long—term impacts
on catalyst performance which result from processing flue gas from a coal—
fired utility boiler.
In conjunction with the pilot plant test program, EPA contracted Radian
Corporation to prepare an independent evaluation of the processes tested based
on the pilot plant results. This section summarizes the results of the inde-
pendent evaluation of the HZ process. It includes a discussion of the results
of tests conducted by both Chemico and Radian and the results of Radian’s
independent evaluation of the HZ process. A separate report covering the
detailed results of the pilot plant test program has been prepared by HZ.
2
-------
1.2 PROGRAN OBJECTIVES AND APPROACH
The independent evaluation of the HZ pilot plant test program conducted
by Radian Corporation had three major objectives. The first was to provide
independent validation of the process measurements made by Chemico. The
second objective was to quantify any changes in the emission rates of secondary
pollutants (pollutants other than NOR) across the pilot plant reactor. And
the third objective was to complete a technical and economic evaluation of
the HZ process including identification of areas which require further develop-
ment or investigation.
In order to validate the measurements made by Chemico, a quality
assurance program was implemented. This program used EPA reference methods
and other standard measurement techniques to make independent audits of critical
process parameters such as flue gas flowrate, NH 3 injection rate, etc. In
conjunction with the quality assurance program, the continuous NO monitors
were subjected to certification tests designed to determine the monitors’
ability to make accurate, repeatable measurements. These certification
tests included measurement of the continuous monitors relative accuracy, drift,
calibration error, and response time.
Concurrent with the quality assurance program, a stack sampling program
was conducted to measure changes in secondary process emissions across the
SCR reactor. This approach required simultaneous sampling of the reactor inlet
and outlet for the species of interest. The samples were then analyzed and
differences between inlet and outlet concentrations determined.
Based on the results of the quality assurance program, the stack sampling
program, and the test data collected by Chemico, an evaluation of the HZ process
was completed. This evaluation consisted of several steps. First, the test
data were analyzed and reduced to a form which could be used to predict process
performance for a specified set of operating conditions. Then, using the
reduced test data and the results of the stack sampling program, material and
energy balance calculations were completed for a 500 MWe coal—fired application
3
-------
of the HZ process. The basis for these calculations was identical to that
used by TVA in developing cost estimates for the HZ process and presented
in “Preliminary Economic Analysis of N0 The Gas Treatment Processes”.’
The results of the material and energy balance calculations were then used
to develop a modified estimate of total capital investment and annual revenue
requirements for a 500 MW coal—fired application of the HZ process. Finally,
the test data were reviewed and areas requiring further investigation/quanti-
fication were identified.
1.3 RESULTS
Several areas which influence the technical and economic feasibility
of the HZ process were examined as part of this study including:
the pilot plant test results,
• the results of Radian’s independent tests,
• the results of a 500 MW conceptual design of the
HZ process,
• the material balance calculations for a 500 MW
process application,
• the energy balance calculations for a 500 MW
process application, and
• the estimated capital investment and annual revenue
requirements for a 500 MW HZ process application.
The following discussion summarizes the results of the evaluation of each of
these areas while Section 1.4 presents overall conclusions on the technical
and economic feasibility of applying the HZ process to a coal—fired boiler.
4
-------
1.3.1 Pilot Plant Test Results
The test program at the HZ pilot plant was initiated in June 1979, and
was completed in January 1981. During this period, the pilot plant processed
a flue gas slip stream from between the economizer and the air preheater of the
coal—fired unit No. 3 at Ceorgia Power Company’s Plant Mitchell Station.
Design flue gas flowrate to the pilot unit was 1700 Nm 3 /hr (1060 scfm) and
flue gas was processed for a total of about 10,000 hours during the program.
The pilot plant test program involved examination of three charges of
catalyst material under a variety of test conditions. In general, these tests
were divided into two categories: optimization tests and demonstration or
long—term tests. The objective of the optimization tests was to identify
operating conditions which would reduce N0 emissions by 90 percent at a minimum
total cost for operating the process. The major objective of the demonstration
tests was to document the ability of the process to achieve a 90 percent reduc-
tion in N0 emissions for 90 days.
The objectives of the pilot plant tests conducted by Chemico were met and
exceeded. The N0> reduction efficiency of the plant averaged over 90 percent
during the 90 day demonstration test and the average was 89.8 percent over a
5—month period of operation. This included several test periods during which
the NH 3 /N0 ratio was varied to determine its effect on N0 reduction
efficiency. If these test periods are excluded from the averages, the N0
reduction efficiency during the 5—months of operation would be greater than
90 percent.
Some other significant results of the test program showed that neither
temperature nor flowrate has as significant effect on N0 reduction efficiency
within a range about the design level. These results indicate that process
performance should not be impaired at boiler loads below the design level.
As a result, no temperature or flow control would be required for a full—scale
application of the HZ process.
5
-------
During the test program, 3 charges of catalyst were examined and two of
these (NOXNON 500) experienced severe plugging problems after about 2000 hours
of operation. When replaced with the N0 0N 600 catalyst, which has larger
gas passages, no further plugging problems were observed. The original plugging
was believed to be due to the adhesiveness of the fly ash. At high temperatures
fly ash samples collected from the power plant were found to agglomerate.
Tests with the NOXNON 500 catalyst did not last long enough to get a
good measure of catalyst activity, but results of the NOXNON 600 tests showed
a gradual decline in catalyst activity with time. After 5500 hours of operation,
activity of the NOXNON 600 had dropped slightly, but it was still possible to
achieve 90 percent N0 reduction. Since 5500 hours is close to one year of
operation ( 7000 hours) a catalyst life of 1 year seems reasonable based on
the test results. In fact, catalyst life may be extended well beyond 1 year
based on the results of the in situ regeneration test conducted at the con-
clusion of the test program. These tests showed the catalyst activity had
been restored to the activity of essentially new catalyst. Unfortunately,
since the regeneration test was conducted during the final week of the test
program, it is uncertain how long the effects of regeneration would last.
Overall, the results of the pilot plant tests indicate application of
the HZ process to a coal—fired boiler is technically feasible. The tests
demonstrated the ability of the process to achieve 90 percent N0 reduction
for over 90 days and also demonstrated a stable catalyst life of nearly one
year.
1.3.2 Independent Evaluation Test Program Results
The independent evaluation test program conducted by Radian had two
primary objectives: to insure the quality of the data collected at the HZ
pilot plant and to quantify changes in the concentrations of certain pollutants
across the HZ reactor. Data quality was determined by quality assurance (QA)
audits and continuous monitor certification tests, while changes in pollutant
6
-------
concentrations were determined by a secondary emissions sampling program. The
results of each of these elements of the independent evaluation program are
summarized below.
Quality Assurance Audits
The QA audits conducted by Radian were designed to insure that the process
data which are required to characterize the operation of the HZ pilot plant
were accurate. Radian employed reference methods for auditing process operating
parameters which were measured on a continuous or routine basis by Chemico. One
exception to this was the measurement of NH 3 emissions which were not routinely
monitored by Chemico, although the original design of the pilot unit included
an analyzer intended to determine NH 3 emissions.
The results of the NH 3 emissions sampling conducted by Radian indicated
an average N i - I 3 concentration at the reactor outlet of about 50 ppm under
operating conditions which result in 90 percent NO reduction. The NH 3 con-
centration at the reactor outlet was much higher than expected (previous work
in Japan indicated NH 3 concentrations of about 10 ppm). The relatively high
NH 3 concentrations are expected to have an impact on equipment located down-
stream of the catalytic reactor for a commercial application of the HZ process.
This is discussed in more detail later in this section.
The results of the QA audits conducted by Radian are summarized in
Table 1—1. As shown, all but the SO 2 concentration measurements were within
10 percent of the values recorded by Chemico. This indicates that with the
exception of the SO 2 monitor, the process data collected by Chemico accurately
characterize the operation of the HZ pilot plant.
7
-------
TABLE 1-1. RESULTS OF QA AUDITS AT THE
HITACHI ZOSEN PILOT PLANT
Measurement Audited
Relative Error 1 (%)
so 2 Concentration
—19.8
Flue Gas Flowrate
— 0.3
N B 3 Injection Rate
— 6.0
Reactor Pressure Drop
4.5
Reactor Temperature
4.8
1 Monitor Reading — Audit Measurement
Relative Error = x 1O0,
Audit Measurement
In the case of the flue gas SO 2 concentration at the HZ pilot plant,
the audit results were determined to be correct and the SO 2 monitor was in
error. This error is characteristic of the type of SO 2 monitor used (pulsed
fluorescence) when the instrument is calibrated with standard gases composed
of SO 2 in nitrogen.
Secondary Emissions Sampling
The secondary emissions sampling program was conducted by Radian during
July and August 1980, concurrent with the demonstration test conducted by Chemico.
The objective was to quantify changes in the emission rates of pollutants
other than NOR. For the most part, these tests were conducted during tests
in which steady state N0 reduction efficiency was maintained at 90 percent.
Table 1—2 summarizes the results of the secondary emissions sampling
program at the HZ pilot plant. As shown, concentrations of hydrocarbons,
carbon monoxide, hydrogen cyanide (HCN), and nitrosoamines at the reactor
outlet were below the detection limit of the analytical techniques employed.
For hydrocarbons and carbon monoxide no conclusions can be drawn concerning
the impacts of the HZ process on these pollutants. For HCN, the analytical
detection limit is equivalent to 10 ppbv and for N—Nitrosodimethylamine,
2 ppbv. In both cases, these concentrations are at levels which are considered
safe for emission sources.
8
-------
Table 1—2 does show an increase in SO 3 concentration across the HZ
reactor. This is due to oxidation of SO 2 in the reactor and was not unexpec-
ted since the catalyst contains vanadium pentoxide which is the catalyst used
in manufacture of sulfuric acid. The apparent change in particulate concen-
tration shown in Table 1—2 is believed to be due to unaccounted for strati-
fication in the ducts. It should be noted that no results for nitrous oxide
(N 2 0) are presented. This is due to the fact that the analytical technique
used to measure N 2 0 proved unsatisfactory for use in a flue gas stream.
TABLE 1-2. STACK SAMPLING RESULTS AT HZ PILOT UNIT
Flue Gas
Component
Reactor Inlet
Concentration’
Reactor Outlet
Concentration’
Nitrosoamines 2
(pg/dscm 2 )
5
5
Hydrogen Cyanide
(mg/dscm)
.01
.01
Aimnonia
(ppmv—dry basis)
Not measured
54.8
Sulfur Trioxide
(ppmv—dry basis)
8.4
20.7
Hydrocarbons 3 (C ,—C 6 )
(ppmv)
1.0
1.0
Carbon Monoxide 3
(%)
0.017
0.017
Particulate Loading
(gm/dscm)
7.1
7.7
Nitrous Oxide
—
—
‘Average of 3 or more tests
2 dscm — dry standard cubic meter
3 Below the detection limit
In addition to measuring the concentration of particulates in the flue
gas, an elemental analysis of the particulates was completed in an attempt to
determine if erosion of the catalyst has a measurable effect on the
9
-------
concentration of Vanadium (V) arid Titanium (Ti) in the particulates.
Table 1—3 presents the results of the elemental analysis of the particulates
collected at the HZ pilot plant. As shown, an apparent increase in all
elements occurs across the reactor, but the relative concentrations on V
and Ti remain constant. This indicates that there is no measurable change
in the concentration of V or Ti in the particulates exiting the reactor.
TABLE 1—3. RESULTS OF PARTICULATE ANALYSIS AT THE HZ PILOT PLANT 1
Component
In
Out
Out/In
Al
10.7%
13.0%
1.21
Ca
8200 ppm
9900 ppm
1.21
Fe
4.9%
6.0%
1.22
K
2.0%
2.5%
1.25
Mg
6300 ppm
7800 ppm
1.24
Mn
190 ppm
240 ppm
1.26
Sn
490 ppm
680 ppm
1.40
Na
4200 ppm
4700 ppm
1.12
Si
18%
23%
1.28
Zn
190 ppm
250 ppm
1.32
Cu
150 ppm
170 ppm
1.13
Ti
5800 ppm
6900 ppm
1.19
V
270 ppm
330 ppm
1.22
1 Concentrations are on a mass fraction basis
Continuous Monitor Certification Tests
Certification tests were conducted for the SO 2 and NO monitors used to
x
measure flue gas concentrations of pollutants at the inlet and outlet of the
reactor. These tests were included in the independent evaluation program to
insure the quality of the pilot plant performance data being collected by
Chetnico. Certification of continuous emission monitors involves a formal
procedure which has been developed by EPA to insure the accuracy of monitors
10
-------
measuring emissions from sources which must comply with new source performance
standard emission limitations. In order for a continuous emission monitor
(CEM) to be certified, it must be subjected to and pass a number of performance
tests, including:
• calibration error,
• response time,
• drift, and
• relative accuracy.
The performance specifications for each of the above certification tests are
presented in Table 1—4 along with the results of the certification tests.
The performance specifications are those contained in the Federal Register,
Vol. 44, No. 197, Wednesday, October 10, 1979 — “Proposed Rules: Standards
of Performance for New Stationary Sources; Continuous Monitoring Performance
Specifications”. 2 -
TABLE 1—4. CONTINUOUS MONITOR CERTIFICATION TEST
RESULTS AT THE HZ PILOT PLANT
Certification Performance
Test Specification
Inlet NO
Monitor
Outlet NO
Monitor
Calibration Error
—high level (%)
—mid level (%)
5
5
1.40
4.39
4.70
2.68
Response Time (mm)
15
1.4
1.6
Zero drift (2—hour) (%)
2
1.20
0.05
Calibration Drift (%)
(2—hour)
2
1.93
1.78
Relative Accuracy (%)
2O
14.1
10.5
1 Alternatively, <10 percent of the applicable emissions standard.
As shown in Table 1—5, the test results for both the NO continuous
monitors met the performance specifications. These data indicate that
continuous monitors were making accurate measurements of flue gas N0 con-
centrations.
11
-------
1.3.3 Results of the Conceptual Design of a 500 MW HZ Process
A conceptual design of a 500 MW HZ process was prepared based on the
pilot plant test results. This conceptual design served as a basis for
material and energy balance calculations and for a cost estimate for a 500 MW
application of the HZ process.
Table 1—5 summarizes the results of the conceptual design for a 500 MW
application of an HZ process. As shown, the key design variable levels are
presented for the SCR reactor and the downstream air preheater.
TABLE 1—5. RESULTS OF THE CONCEPTUAL DESIGN FOR A 500 MW
HITACHI ZOSEN PROCESS
Design Parameter Design Level
Reactor Design Parameter
• Number of Reactors 2
• Reactor Crossection (m 2 ) 96.5
• Catalyst Volume per reactor (m 3 ) 205
• Reactor System Pressure Drop (kPa) 1.28
• Soot Blowers per Reactor 4
• Soot Blowing Frequency 3/day
Air Preheater Design Parameters
Soot Blowers per Preheater 6
• Soot Blowing Frequency 6/day
• Element Configuration Combined Intertnediate and Low
Temperature Zone
• Element Construction Corrosion Resistant Material
in Intermediate—Low Temperature
Zone
12
-------
The conceptual design of the HZ process was prepared for a single
application of the process and it was based solely on the pilot plant test
results. The results of this design indicate that it is possible to reduce
N0 emissions by 90 percent using the HZ process. In fact, 90 percent NO
reduction was possible at space velocities greater than previous estimates
indicated (i.e., at a relatively lowercatalyst volume per unit volume of
flue gas treated). However, the greater space velocities were accompanied
by NH 3 emissions which were much higher than previous estimates.
One result of the high N i - i 3 emissions estimated for the conceptual
design was that special modifications to the air preheater are required to
mitigate problems associated with the formation of atnmonium sulfates downstream
of the reactor. These modifications were identified as part of a prior study 3
and they are based on Japanese experience with air preheater operation down-
stream of an SCR system. It should be noted that the modifications specified
for the air preheater were expected to minimize problems at relatively low
NH 3 and SO 3 concentrations at the reactor exit. The concentrations at the
reactor outlet for the conceptual design are much higher than anticipated in
previous studies of SCR technology and this could result in operational
problems which cannot be minimized by the preheater modifications included
in the conceptual design. This represents an area which requires durther
inves tigat ion.
Reactor pressure drop and other design parameters are fairly consistent
with previous estimates for the process. The design results also show that
the process can operate over a range of temperatures (340 to 410°C) and space
velocities (6,500 to 8,900 hr 1 ) without any significant effect on NO reduc-
tion efficiency. This indicates the process has good flexibility in processing
flue gas under conditions of changing boiler load.
In sun nary, the conceptual design indicates that the HZ process can
reduce NO < emissions by 90 percent. And, this NO reduction efficiency can
be achieved at a lower catalyst volume per unit of flue gas treated than
13
-------
previous estimates indicated. However, the lower catalyst volume of the
conceptual design is accompanied by a significantly higher NH 3 emission rate
which can result in severe operational problems in downstream equipment,
particularly the air preheater. Further work is required to determine if the
effects of these NH 3 emissions can be offset by the air preheater modifications
included in the conceptual design.
1.3.4 Results of the Material Balance Calculations for a 500 MW
HZ Process
Material balance calculations for a 500 MW application of the HZ process
were included as part of this study to identify raw material requirements for
the process and to serve as a basis for an estimate of capital investment and
annual revenue requirements. The material balance was based on the pilot plant
and secondary emissions sampling test results and thus reflects those results
in the estimated raw material requirements. The most significant results of
the material balance calculations include estimation of NH 3 requirements for
NO reduction, NH 3 and SO 3 emissions from the process, and steam requirements
for air preheater soot blowing.
The NH 3 requirements for the process were estimated to be 1.0 mole of NH 3
per mole of N0 in the flue gas entering the reactor. This requirement was
estimated based on the results of approximately 6 months of pilot plant
operation. During that 6 month period, the NH3INOX injection ratio averaged
0.98 while the N0 reduction efficiency averaged 89.8 percent. With the
NH 3 fN0 injection ratio of 1.0, estimated NH 3 requirements for the process
decreased about 10 percent for previous estimates.
Estimates of NH 3 and SO 3 emissions from the HZ process were significantly
higher than previous estimates indicated. As discussed earlier, this results
in the requirements for air preheater modifications and additional soot blowing.
The requirement for additional soot blowing results in a sevenfold increase in
HZ process steam requirements. However, this is not very significant from a
material balance standpoint, but it is important in terms of its effect on
14
-------
process energy requirements. It should be noted that HZ claims the NH3
emission and SO 2 oxidation rates can be reduced with no decrease in process
performance by adjusting the composition of the catalyst. However, since
this was not demonstrated during the pilot plant tests it was not considered
in preparing the material balances or the conceptual design.
In summary, the material balance calculations showed no significant
change in raw material requirements for the HZ process. The most important
result was the estimated NH 3 and SO 3 emission rates which were significantly
higher than previous estimates indicated.
1.3.5 Results of the Energy Balance Calculations for a 500 MW
HZ Process
An energy balance was completed as part of the evaluation of the HZ
process. This energy balance defined overall process energy requirements and
quantified the heat credits associated with the process. The results of the
analysis of energy requirements indicated that the HZ process has a net
energy consumption equivalent to about 0.3 percent of the energy input to
the boiler.
The individual components of the overall process energy requirements are
summarized in Table 1—6. Each of these components has been put on the basis
of heat input to the boiler. For steam, a thermal efficiency of 88 percent was
used to determine the energy input required to generate one Gcal of steam energy.
For electricity, a boiler heat rate of 2.27 Mcal/kWh was used. The heat credit
was assumed to replace heat input to the boiler on a l—to—l basis.
15
-------
TABLE 1—6. OVERALL ENERGY REQUIREMENT FOR A 500 MW
APPLICATION OF THE HZ PROCESS
Energy Area
Energy Requirement
(Gcal/hr)
Percent of Boiler
Capacity
Heat Credit
(3.15)
(0.28)
Steam
3.36
0.30
Electricity
3.50
0.31
Total
3.99
0.33
1.3.6 Results of the Cost Estimate for a 500 MW HZ Process
Application
An estimate of total capital investment and annual revenue requirements
for a 500 NW application of the HZ process was prepared as part of this
evaluation. The estimated costs reflect the results of the pilot plant tests.
When compared with the previous estimate prepared by TVA, the modified cost
estimates indicate the magnitude of the impact the pilot plant results had on
estimated process costs. In addition, comparison of the modified cost estimate
with cost estimates for other SCR processes is an indicator of the cost
effectiveness of the HZ process as tested in the pilot plant program.
Results of Capital Cost Estimate
Table 1—7 presents the individual components and the estimated total
capital investment for a 500 MW application of the HZ process. As shown, the
total capital investment was estimated to be approximately $22.1 x 106 which
is equivalent to approximately $44/kW of generating capacity. When compared
to TVA’s previous estimate, this represents a slight decrease in total capital
investment. The principal difference between the two estimates is the estima-
ted catalyst volume. The required catalyst volume based on the pilot plant
tests was estimated to be about 20 percent less, thereby decreasing the total
capital investment. However, the decrease in costs from reduced catalyst
volume requirements was somewhat offset by the costs of air preheater modi-
fications required to minimize ammonium sulfate deposition problems.
16
-------
TABLE 1—7. ESTIMATED CAPITAL INVESTMENT FOR A 500 MW APPLICATION
OF THE HITACHI ZOSEN PROCESSa
1 of
total dIrect
b Investment, S investment
Direct Investment
NB3 storage and injection 545,000 3.5
Reactor section 8,632,000 73.4
Gas handling 351,000 3.0
Air preheater modifications 1,461.000 12.4
Sub—total direct investment (DI) 11,089,000 94.3
Services, utilities (0.06 x Dt) 665,000 5.7
Total direct investment (TDI) 11,754,000 100.00
Indirect Investment
Engineering design and supervision 274,000 2.3
Architect and engineering contractor 69,000 0.6
Construction expense
= 0.25 (101 x i0_6)0 3 1,933,000 16.4
Contractor tees = 0.096 (Tnt x 10—6)0.76 625,000 5.3
Total indirect investment (IDI) 2,901,000 24.7
Contingency = 0.2 (TDI + IDI) 2,931,000 24.9
Total fixed investment (Tn) 17,586,000 149.6
Other Capital Char&es
Ailowance for startup and modifications
= (0.1) (TFI) 1,759,000 15.0
Interest during construction
= (0.12) (TPI) 2,110,000 17.9
Total depreciable investment 21,455,000 182.5
Land 5,000 -
Working capital 336,000 2 9
Royalty fee 300,000 2.6
TOTAL CAPITAL INVESTMENT 22,096,000 188.0
9asis: 500 MW new coal—fired power plant, 3.5% sulfur coal, 90% N0 removal.
Midwest plant location. Represents project beginning mid-1977, ending mid—1980.
Average basis for scaling, mid—1979. Investment requirements for Ely ash -
disposal excluded. Construction labor shortages with overtime pay incentive
not considered.
b ch item of direct investment includes total equipment costs plus installation
labor, and material costs for electrical, piping, ductwork, foundations,
structural, instrumentation, insulation, and site preparation.
-- 17.— -
-------
Results of the Annual Revenue Requirement Estimate
Table 1—8 presents the individual components and the total estimated
average annual revenue requirements for a 500 MW application of the HZ process.
As shown, the average annual revenue requirement was estimated to be approxi-
mately $10.2 x 126 which is equivalent to 2.91 mills/kWh. When compared to
TVA’s previous estimate this represents a 17 percent decrease in the annual
revenue requirements for the process.
As with the capital costs, the principal factor which decreased the annual
revenue requirements is the lower quantity of catalyst required in the reactor.
Again, this reduction in annual revenue requirements was somewhat offset by the
costs of increased air preheater soot blowing.
Cost Comparison and Summary
The capital investment and annual revenue requirements of the HZ process
have been estimated based on the results of the test conducted at the EPA
sponsored pilot plant in Albany, Georgia. The results of these cost estimates
indicate that the capital costs and annual revenue requirements are slightly
lower than the estimated costs prior to the test program. A more important
comparison, however, is the cost of the HZ process relative to the cost of other
SCR processes.
Since the same basis was used in preparing the modified HZ cost estimate
as TVA used in preparing preliminary economic estimates for other SCR processes,
it is possible to make a direct comparison with the costs of the Shell Flue
Gas treating (SFGT) Process which were developed under the EPA pilot plant
test program. Table 1—9 presents the estimated annual revenue requirements
for two pollution control systems which reduce emissions of particulates,
NO, and SO 3 by 99.5, 90, and 90 percent, respectively. As shown, the
pollution control systems include flue gas desulfurization capability and
have ESP’s located downstream in order to put the cost estimates on a
common basis.
18
-------
TABLE 1—8. ESTIMATED AVERAGE ANNUAL REVENUE REQUIREMENTS FOR
A 500 NW APPLICATION OF THE HITACHI ZOSEN PROCESSa
Annual
Unit
cost($)
Annual
cost($)
%
r
of annuaL
evenue ru.quuired
lies quantity
Direct Costa
Raw materials
NIl 3 5.25 x lO 6 kg 0.165/kg 866,300 8.47
Cata1y t 5,125,000 50.14
Total raw materIals 5,991,300 58.61
Conversion costs
Operating labor and supervision 8760 12.50/ 109,500 1 07
labor bra, labor hr.
Utilities
Steam 20,700 Gcal 7.94/Cco I 164,400 1.61
Eiectric .Ity 10,787,000 kWh 0.029/k Im 312,800 3.06
Heat credit 22,050 Ccal —7.94/Ccal (175,100) (1.71)
Maintenance .04 x TDI 470,200 4.60
Analyses 2,920 17.00/ 49,600 0 48
labor bra, labor hr. _______ —
Total Conversion coStS 931,400 9.11
Total direLt costs 6,922,700 67.72
Indirect Costs
Capital charges
Depreciation (0.06) (total
depreciable investment) 1,287,300 12.59
Average Cost of capital = (0.086) x
(total capitol investment) 1,900,300 18 59
Overheads
Plant = (0.5) (conversion COsts 314.700 3.08
minus utilities)
Administrative (0.1) 11 000 0.11
(operating labor costs)
Total Indirect cOsts 3,513,300 3437
Spent catalyst disposal (214,000) (2.09)
TOTAL ANNUAL REVtNUE RJiQU IREMCNTS 10,222,000 100 00
°BasIs:500K newcoal ...fjred power plant, 3.5% S coal. 90 percent NO >, reduction ’ 90 percent SO 2 removal. M1dwu st
power plant location, 1980 revenue requirements. Remaining life of power plant 30 years. Plant on line
7000 hr/yr. Plant heat rate equals 9.5 Hi/kWh. Investment and revenue requirement for disposal of fly
ash excluded. Total direct investment $11,754,000; total depreciable investment $21,455,000; and total
capital investment $22,096,000.
-------
TABLE 1-9. ESTIXATED A UAL REVENUE REQUIREMENTS FOR TWO
POLLUTION CONTROL SYSTEMS’
SCR Process
Annual Revenue
Requirements ($
x 1O )
SCR
FGD
ESP
Overall
SFGT
33.6
—
3.0
36.6
Hitachi Zosen
10.2
14.7
2.2
27.1
‘All costs except the HZ—SCR and the SFGT—SCR costs are from “Preliminary
Economic Analysis of NO Flue Gas Treatment Processes.” Tennessee Valley
Authority — Office of Power. EPA—600/7—80—02l, February, 1980.
As shown in Table 1—9, the estimated costs associated with the HZ -
processes are 30 percent lower than those of the SFGT process. These results
indicate that the HZ process, as tested in the pilot plant and presented in
the conceptual design, the most economical of the two SCR processes tested
in EPA’s pilot plant program within the constraints of the conceptual design
used in this study. It should be noted that the relative costs presented in
Table 1—9 are only valid for one specific application and they could change
for other applications.
Overall the results of the modified cost estimate indicate that, for the
particular application examined in this study, the HZ process is economically
competitive with other SCR processes. This is based on a conceptual design
which was representative of operating conditions demonstrated during the
pilot plant tests. It should be noted, however, that the costs can be
affected by the impacts of high NH3 and SO 3 emissions whose effects were not
examined during the pilot plant tests. Additionally, the estimates presented
in this evaluation were based on a 1—year catalyst life which was not demon-
strated. But, HZ will guarantee a 1—year catalyst life for coal—fired appli-
cations.
20
-------
1.4 CONCLUSIONS
The following discussion presents conclusions based on the work
performed during this study. For the most part, the information obtained
during the course of the study is summarized in this report and serves as
background for the conclusions presented here. The major conclusions of
this study are:
The HZ process can reduce NO emissions by 90 percent when
applied to a coal—fired boiler. This level of emissions
reduction was achieved over a 90—day period at an NH 3 /N0 ,<
injection ratio of 1.0 and space velocities greater than’
previous test work indicated. However, the excellent
performance of this pilot plant was accompanied by NH 3
emissions which were much higher than previous estimates
indicated.
The initial tests of the HZ process experienced problems
with catalyst plugging which resulted in failure of 2
charges of NOXNON 500 series catalyst. These problems
were eliminated by employing NOXNON 600 series catalyst
(a catalyst with larger gas passages) and the use of
compressed air (as opposed to superheated steam) for
reactor soot blowing. It appears likely that the good
performance of the NOXNON 600 catalyst was due to the larger
gas passages since the fly ash has a tendency to agglomerate
in dry environments.
A gradual decline in catalyst activity was recorded over
the duration of the test program which resulted in the
requirement for increased NH3/NO injection ratios to
attain 90 percent N0 reduction. Because the test program
was terminated after 5500 hours of operation, the catalyst
activity after 1 year of operation could not be determined.
A novel, in situ catalyst regeneration technique was
tested as part of the program. This test showed that the
regenerated catalyst had activity similar to fresh catalyst
and thus reversed some of the decline in activity observed
over the duration of the test program. Unfortunately, the
catalyst regeneration technique was tested toward the end
of the pilot plant test program and so it is uncertain
how long the effects of the catalyst regeneration will last.
21
-------
The independent evaluation test program indicated that
emission rates of most pollutants were not affected by
the HZ process. However, emission rates of both NH3
and SO 3 were relatively high and those emissions can
result in operational problems in downstream equipment.
The severity of any problems in this regard are very
site specific and could not be assessed as part of this
study. This should, however, be given careful con-
sideration in any planned applications of the HZ process.
The conceptual design and material balance calculations
indicated high NH 3 emission rates which will cause
severe operational problems in the air preheater located
downstream of the HZ process. The conceptual design
included air preheater modifications designed to minimize
those problems. But because the estimated N H 3 and SO 3
emission rates are much higher than previous estimates,
it is uncertain if the air preheater modifications
will prove adequate. Further investigation in this
area is required.
• The overall energy requirements for the HZ process
were estimated to be 0.3 percent of the boiler’s
capacity. This is a very small fraction of boiler
capacity and it does not significantly affect process
costs.
• The estimated capital investment and annual revenue
requirements for the HZ process were slightly lower than
TVA’s preliminary estimate. This indicates that the
HZ process is economically competitive with other SCR
processes when considered for application to a coal—fired
boiler. It should be noted that the cost estimates
assumed a 1—year catalyst life which was not quite
demonstrated during the pilot plant tests although it
would be guaranteed by HZ. The relative process costs
could change if a 1—year catalyst life is not possible.
tn conclusion, the pilot plant test results indicate the HZ process
is technically suited for application to coal—fired sources. However, the
tests did not demonstrate a 1—year catalyst life which is generally considered
a minimum for technical feasibility of an SCR process. In reality, a shorter
catalyst life would translate into increased annual revenue requirements.
In terms of costs, under the conditions of the cost estimate prepared as part
of this study, the HZ process is economically competitive with other SCR
processes.
22
-------
SECTION 2
INTRODUCTION
Selective catalytic reduction (SCR) processes have been shown to be the
most effective method for high levels of nitrogen oxide (NO>) control for
stationary sources. Numerous SCR units have been installed on both oil— and
gas—fired utility boilers, industrial boilers, and process heaters and furnaces
in Japan. Many of these units continuously achieve N0 reductions of >80%
giving SCR the highest NO reduction of any system available commercially.
SCR systems utilize ammonia (NH3) to reduce NO to molecular nitrogen
and water on the surface of a catalyst. The overall N0 reduction reactions
are summarized in Equations 2—1 and 2—2.
4N0 + 4NH 3 + 02 - 4N 2 + 6H 2 0 (2-1)
2N0 2 + 4NH 3 + 02 - 3N 2 + 6H 2 0 (2—2)
-_
The couunercially available SCR catalysts perform optimally at temperatures
of 300—400°C (570—750°F). For utility boiler applications, flue gas is at
these temperatures downstream of the economizer and upstream of the air pre-
heater. With industrial boilers these temperatures may exist upstream of the
economizer.
SCR processes were originally developed in Japan and have been applied
primarily to oil— and gas—fired units. SCR systems are being installed in
the U.S. on a limited basis. Several oil—fired process heaters and one
utility boiler in California have been equipped with SCR systems. The latter
23
-------
is under construction at Southern California Edison’s Huntington Beach
plant. This 5CR system will treat half of the flue gas from a 215 MWe oil—
fired utility boiler.
In Japan (and in the U.S.) current efforts are directed towards develop-
ing SCR systems for coal—fired applications. Presently, there are two
commercial SCR systems treating flue gas from coal—fired boilers with more
systems due to come on line later this year.
The transfer of SCR technology from Japan to the U.S. for coal—fired
applications presents a potentially significant problem. Since most coal—
fired boilers in the U.S. operate ESP’s located downstream of the air pre-
heater, the catalyst in an SCR reactor would be exposed to the full particulate
concentration from the boiler. And although tests have been conducted in Japan
in which the catalyst was exposed to high particulate concentrations with no
adverse effects, the differences in the composition of particulates from U.S.
and Japanese coals could impact SCR operation.
In an effort to further the development of SCR technology and in particu—
lar to determine how differences between Japanese and U.S. coal/particulate
properties impact the performance of SCR processes, EPA has sponsored pilot
scale tests of two SCR systems. The SCR systems tested were the Hitachi
Zosen (Hz) process and the Shell flue gas treating (SFGT) process, which is
also capable of removing SO 2 from the flue gas. In both cases, the pilot
plants processed a flue gas slipstream from a coal—fired boiler. The con-
tractors responsible for the design and operation of these pilot plants were
Chemico Air Pollution Control Corporation (now General Electric Environmental
Services Corporation), which is the North America licensee for the HZ process,
and the Process Division of UOP, who is the licensing agency for the SFGT
process. These contractors were also responsible for collecting, evaluation,
and reporting of the test data.
24
-------
The main objectives of the pilot plants were to demonstrate the capability
of continuously achieving 90% N0 removal (and 90% SO 2 removal simultaneously
for the SECT process) and to determine the effects of coal—fired flue gas on
long—term catalyst activity. Other objectives of the test program were to
determine the optimum operating conditions for the process and the raw
material and utility demands. This information is needed to make an accurate
evaluation of the design and costs for a commercial scale SCR system.
To perform the evaluation of the two SCR processes EPA contracted
Radian Corporation as an independent consultant. This report presents the
results of Radian’s evaluation of the HZ process. A separate report was
prepared for the evaluation of the SFCT process.
Radian Corporation’s independent evaluation of the HZ process had three
primary objectives. The first was to provide independent validation of the
data measurements made by Chetnico. Second, Radian was to quantify any change
(either production or reduction) in selected secondary pollutants (pollutants
other than NO) across the reactor. Finally, using both Radian and Chemico
data, a technical evaluation of the HZ process was to be prepared including
the identification of areas which require further development before the
process can be commercialized.
To validate the data collected by Chemico, Radian instituted stack
sampling and quality assurance programs. The relative accuracy of the
continuous NOx monitors were determined by stack sampling using EPA reference
Method 7. The NOx monitors were also tested for calibration error, response
time, and two—hour zero and calibration drift. The results for each of these
tests were checked against the minimum performance required by the New Source
Performance Standard (NSPS) presently in effect for continuous monitors on
utility boilers. If the monitors failed the test, the cause was determined
and the test was repeated. The quality assurance program also utilized
standard procedures for measuring important operating variables such as NH3
injection flow rate, flue gas flow rate, reactor pressure drop, reactor
temperature, etc.
25
-------
Along with the data validation efforts, Radian conducted a stack
sampling program to quantify the amounts of various secondary pollutants
entering and exiting the reactor. Samples from the reactor inlet and outlet
were collected simultaneously. These samples were then analyzed to determine
any change in concentration across the reactor.
Using the results of Radian’s stack sampling and quality assurance
programs and Chemico’s process data, a technical evaluation of the HZ process
was performed. The data were reduced and used to determine the operating
conditions necessary to obtain 90% N0 removal. Using these operating con-
ditions and the stack sampling results, material and energy balances for the
HZ system were prepared for a 500 MWe coal—fired utility boiler application.
The bases used in these calculations were the same as those used by TVA in
their “Preliminary Economic Analysis of NO Flue Gas Treatment Processes”. 1
The calculations were compared to TVA’s and incremental costs were estimated.
Areas requiring further development were then identified.
This report describes the evaluation of the HZ process pilot plant by
Radian Corporation. Section 3 characterizes the field test program including
the descriptions of the pilot plant and the test programs of Chemico and
Radian. Section 4 contains the results of field tests conducted by both
Chemico and Radian measurements. Section 5 presents the technical evaluation
of the HZ process including process performance variables, raw material and
utility requirements, and downstream impacts. Finally, Section 6 lists the
process development requirements.
26
-------
SECTION 3
FIELD TEST PROGRAM
To demonstrate the application of SCR processes on coal—fired boilers,
EPA contracted Hitachi—Zosen (Hz) to construct and operate an SCR pilot
plant in May 1978. Implementing the contract, HZ provided the basic design
of the pilot plant and the proprietary SCR catalyst. Chemico Air Pollution
Control Corporation, the major subcontractor, was responsible. for the detailed
design and engineering, construction, and operation of the pilot plant. This
pilot plant was designed to demonstrate the capability of the HZ process to
reduce NO emissions from a coal—fired boiler in the U.S. by 90 percent.
HZ completed the basic process design in Japan in June 1978. In September
1978 Chetnico completed Phase I activities including the detailed plant design
and equipment specifications. Phase II — the procurement of equipment, con-
struction subcontractor selection, and pilot plant construction and erection —
was finished by July 1979. Chemico then started up the pilot plant and began
debugging operations in preparation for the Phase III optimization tests.
This work covered the period from July 1979 to May 1980. (The length
Phase III was due to the use of three different catalyst charges. This is
explained in more detail in Section 4.) Finally, the long—term demonstration
test, Phase IV, was conducted from June 1980 to January 1981 at which time the
pilot plant was shut down.
Radian Corporation performed its field test programs during July and
August 1980. This report section describes in detail the test programs
conducted by Chemico and Radian. To provide background information, a dis-
cussion of the HZ process and a description of the pilot plant (including
details of flow rates, equipment sizes, process instrumentation, and plant
operations) are presented as a prelude to the test program discussions.
27
-------
3.1 HZ PROCESS
The HZ process is one of a number of dry SCR processes. With the appro-
priate catalyst, ammonia will selectively reduce NON, in the presence of oxy-
gen, to nitrogen and water on the catalyst surface. The homogeneous, gas—phase
reactions are summarized in Equations 3—1 and 3—2.
4N0 + 4N11 3 + 02 - 4N 2 + 61-120 (3—1)
2N0 2 + 4NH 3 + 02 - 3N 2 + 61120 (3—2)
For coal—fired flue gas applications and other gases containing large
amounts of sulfur oxides (SO ), the HZ process uses a catalyst composed of
titanium dioxide (Ti0 2 ) and vanadium pentoxide (V 2 0 5 ). This catalyst formula-
tion is resistant to “poisoning” by S0 and possesses an optimum reaction
temperature range of 350—400°C. At temperatures above this range, an addi-
tional reaction that becomes increasingly more significant is the oxidation
of NH 3 to N0 shown in Equation 3—3. High temperatures can also sinter the
catalyst thereby permanently reducing its activity.
4N}13 + 502 4N0 + 6HzO (3—3)
At temperatures below this range, the kinetics of the N0 reduction reactions
decrease, thus lowering the N0 reduction efficiency of the system. Another
negative impact of lower temperatures is the formation and deposition of
ammonium sulfates by the reaction of N}1 3 with sulfur trioxide (SO 3 ) present
in the flue gas (Equations 3—4 and 3—5).
NH 3 (g) + S0 3 (g) + H 2 0(g) - NH ,HS0 (1,s) (3—4)
2NH 3 (g) + S03(g) + H 2 0(g) —i - (NFI ,) 2 S0i+(s) (3—5)
These formation reactions become especially important iii downstream heat
recovery equipment where the flue gas is cooled.
28
-------
The formation of ammonium sulfates downstream will be enhanced by the
SO 2 oxidation capability of the SCR catalyst shown in Equation 3—6. One of
the primary components of the SCR catalyst, \T 2 0 5 , is the catalyst used
commercially to oxidize SO 2 in sulfuric acid plants.
S0 2 (g) + ½02(g) - S0 3 (g) (3—6)
Approximately 1 to 2 percent of the inlet (to the reactor) SO 2 is oxidized to
SO 3 by the SCR catalyst. This will be discussed further in Sections 4 and 5.
To design anSCRsystem, one must consider the specific characteristics
of the flue gas stream, particularly regarding fly ash loadings. For coal—
fired applications and other particulate—laden gases, a specfal physical -
design of the catalyst was developed by HZ to resist plugging by particulate
matter. The catalyst is made up of steel plates that are chemically and
thermally treated to allow coating of catalyst on the surface. Then the
plates are arranged in a “honeycomb” shape and assembled into units called
“cells”. The cells (the number of which depends on the size of the gas stream
being treated) are then stacked inside the reactor. The flue gas flows through
the open channels parallel to the surface of the catalyst—coated plates.
The fly ash can pass through the open channels without plugging the catalyst
bed. The flue gas components involved in NO reduction reactions diffuse
(by molecular diffusion for low flows and by eddy diffusion for more turbulent
flows) to the catalyst surface where the reactions occur.
3.2 HITACHI—ZOSEN PILOT PLANT DESCRIPTION -
The HZ pilot plant was located in Albany, Georgia at the Plant Mitchell
station of the Georgia Power Company. Flue gas for the pilot plant was
obtained from the Unit No. 3 duct after the boiler economizer but upstream of
the air preheater. The flue gas slipstream was first passed through an electric
heater to provide temperature control and to compensate for heat losses in the
duct from the boiler to the pilot unit. Then ammonia was injected into the gas
and the gas was passed down through the catalyst bed. The gas was then
29
-------
passed through a mechnical collector (cyclone) for particulate removal, an
induced draft fan, and then returned to the boiler duct. Figure 3—1 shows
a schematic of the HZ pilot plant.
The average flue gas flow rate of the pilot plant was 2210 Nm 3 Ihr
(1370 scfm) which is approximately equal to the flue gas flow from 0.7 MW
of electrical power generation. The concentrations of the primary flue gas
constituents entering the pilot plant are listed in Table 3—1. (As discussed
previously, the fly ash loading shown in Table 3—1 represents the full loading
from the boiler.)
TABLE 3-1. APPROXIMATE AVERAGE COMPOSITION OF FLUE GAS
AT HZ/CHEMICO PILOT PLANT INLET -- - -
Constituent Concentration
N 2
69%
by volume
CO 2
18%
by volume
H 2 0
8%
by volume
02
5%
by volume
SO 2
890
ppmv
NO<
450
ppmv
Particulates
7.1 gm/dscm
(3.1 gr/dscf)
The operating parameters of the pilot plant which are continuously
controlled are:
• flue gas flow rate,
flue gas temperatures, and
• NH3 flow rate (set by the NH3:NO mole ratio controller).
Table 3—2 lists the instruments used by Chemico to monitor the performance
of the pilot plant. This list is not a complete list of all the instruments
30
-------
Injection
Hitachi
Zosen Flue Gas
Electric Blower
Catalytic
Return to
Flue Gas
Reactor
Heater
Inlet
Sample
From Unit 3 Point _______ ____
clone
Econom zer
amp e
Figure 3—1. Schematic of the Hitachi Zosen pilot unit.
-------
on the pilot plant, but the parameters measured by these instruments are
sufficient to quantify the plant’s performance.
TABLE 3—2. PRIMARY INSTRUNENTS ON THE HZ PILOT PLANT
Parameter
Instrument
Location
Flue Gas
Flow Rate
Venturi
Reactor Outlet
Flue Gas
Temperature
Thermocouple
Reactor Inlet/Outlet
NH 3 Flow
Rate
Rotameter
Reactor Inlet
Flue Gas
N0> Concentration
TECO 10 AR Chemilumi—
nesc ent
Reactor Inlet/Outlet
Flue Gas
Flue Gas
SO 2 Concentration
02 Concentration
TECO 40 Pulsed
Fluorescent
Leads and Northrup 7863
Paramagnet Ic
Reactor Inlet or
Outlet -
Reactor Inlet
Reactor
Pressure Drop
Differential Pressure
Flow Transmitter
Chemico had originally planned to have a continuous NH 3 monitor in the system
to quantify NH 3 emissions from the reactor. However, this monitor never
performed accurately as NH3 reacted with S03 in the reactor outlet sample
line (see Equations 3—4 and 3—5) thus giving a lower than actual reading.
The controlling consideration of the pilot plant operation was protecting
the catalyst from being damaged or plugged. The basic precept of start—up
and shutdown procedures was to never expose the catalyst to flue gas or NH3
when the catalyst temperature was less than 230°C. To initiate a cold start-
up, the electric heater and the induced draft fan were started. The flue
gas valves on the boiler duct were in the closed position and the air intake
valve was in the open position). Heated air was drawn through the catalyst
until the catalyst temperature was 230°C. At this point, the valving was
switched so that the pilot plant was receiving flue gas from the boiler duct.
When the catalyst temperature reached 315°C, then NH 3 injection was begun
and testing began.
32
-------
During continuous pilot plant operations, a soot blower was used to keep
the catalyst surface clean of accumulated fly ash. This was initially performed
on an occasional basis using pressurized steam. Later, the frequency was
increased to once per shift. By the end of the program, compressed air was
being used at least once (often twice) per shift for soot blowing purposes.
This is discussed further in Section 4.
To shut down the pilot plant, first the NH 3 flow was stopped. Then
initial cooling was begun. At approximately 315°C, the pilot plant was
switched from flue gas to air and the reactor was purged with hot air before
cooling down completely. If the shutdown was due to a temporary outage such
as boiler maintenance, hot air (‘t 2OO—25O°C) would continue tobe recirculated
through the catalyst via the pilot plant’s recycle line.
3.3 HITACHI-ZOSEN PILOT PLANT TEST PROGR AN
Testing of the HZ pilot plant was begun in August 1979 and completed in
January 1981. The pilot plant was operated for a total of over 10,000 hours -
using three different catalyst charges (approximately 2500 hours on the first
charge, 2300 hours on the second, and 5500 hours on the third). The test
program can be divided into two parts, optimization tests and demonstration
or long—term tests. The purpose of the optimization tests was to determine
the best values of the controlling process parameters (e.g. , N1 -I3/NO injection
ratio) resulting in the desired NO reduction efficiency of 90%. The primary
purpose of the demonstration tests was to show the HZ proc.ess to be capable
of continuously reducing NO emissions from a coal—fired boiler for an extended
time period. Another important goal of the demonstration tests was to
quantify the material and utility demands of the process operating at 90% NO
reduction efficiency. The following discussion presents a brief operating
history of the HZ pilot plant including a discussion of each batch of catalyst
tested. Section 4 presents a detailed discussion and interpreation of test
results.
33
-------
3.3.1 Optimization Tests — Objectives
The objectives of the optimization tests conducted by Chemico were to
quantify the effect of key process operating parameters on the NO < reduction
efficiency of the HZ process and to determine the optimum value of process
control parameters required to achieve 90 percent NO reduction.
The principal variables examined in the optimization tests were:
NH3/NOx injection ratio,
• Flue gas flowrate,
• Reactor operating temperature, and
• Inlet NO concentration.
The effect of each of these variables on the performance of the HZ process
can vary in magnitude and direction. The following discussion briefly examines
each variable and identifies its directional impact on NO reduction efficiency.
NH 3 /NO Injection Ratio
The amount of NH 3 injected into the flue gas is the primary independent
variable which determines the amount of NO removed by the SCR system. The
Nl-13/NO mole ratio is the ratio of moles of NH 3 injected per mole of NO in
the flue gas at the reactor inlet. If no NH 3 is injected, the mole ratio
is obviously zero and no NO reduction occurs. As the NH 3 /NO mole ratio
is increased from zero to approximately 0.9 to 1.0 there is a parallel
increase in NO reduction by the HZ process. At NH 3 INOx mole ratios of 0.9—
1.0, the HZ process achieves approximately 90% NO reduction. As the mole
ratio is increased beyond this range, the NO reduction gradually approaches
its theoretically limiting value of 100%. However, for each additional amount
of NH 3 added beyond the 0.9—1.0 mole ratio range, it is reducing correspondingly
less NO thus resulting in more unreacted NH 3 exiting the reactor as a poten-
tial emission.
34
-------
Flue Gas Flow Rate
Flue gas flow rate can potentially impact the N0 reduction efficiency
of the HZ process in two ways. An increase in flow rate will reduce the
residence time of the flue gas in the reactor. So, if the NO < reduction
efficiency of the system is limited by chemical reaction kinetics, an increase
in flow rate will result in a decrease in NO reduction. However, if the
NO reduction efficiency of the system is limited by mass transfer (this
refers to the rate of transfer, or diffusion, of the reactants out of the
bulk gas to the catalyst surface), then an increase in flow rate may not
result in a decrease in NO reduction. For the range of flow rates tested
at the HZ/Chemico pilot plant, there was little change in the NO reduction
efficiency of the system for a significant change in flow rate. This suggests
that for the design of this pilot plant, the system is not limited by reaction
kinetics.
Reactor Operating Temperature
The reaction temperature can affect the NO reduction performance of the
HZ process. As the system temperature increases the NO reduction efficiency
increases due to an increase in the chemical reaction kinetics and the
diffusion rate of the reactant species. As the temperature increases beyond
the 750—800°F range, the rate of the NH 3 oxidation reaction increases thus
reducing the overall NO reduction. As discussed earlier, at temperatures
below about 600°F the rate of the NO reduction reactions -decrease and the
formation of anunonium sulfates becomes a potential problem. -
Inlet NO Concentration
Another process parameter that was investigated to determine its impact
on the system NO reduction performance was inlet N0 concentration. While
this parameter is not directly controllable in a commercial situation, pro-
visions were made as part of this pilot program to “spike” the flue gas with
NO to determine the effect of high N0 concentrations. While there was a
35
-------
slight increase in NO reduction, it was concluded that high inlet NO
concentrations will not adversely affect the processes removal efficiency.
3.3.2 Optimization Tests — History
- Optimization tests were conducted for each of the three charges of
catalyst installed in the HZ pilot plant, while the 90 day demonstration
test was conducted with the third catalyst. The following discussion
summarizes the optimization tests conducted on each charge of catalyst.
This includes a summary of the total time flue gas was processed by each
catalyst.
Catalyst Charge 1
The first catalyst charge was installed in the reactor in August 1979.
Shortly after catalyst installation the pilot plant was started up and then
exposed to flue gas. The optimization tests on this catalyst charge were
plagued by a series of minor problems, the type of which often occur during
the initial phases of pilot plant operation. In August, there were several
shutdowns due to leaks in the inlet sample line and leaks in the flanges
around the reactor. Very little testing was performed in September due to
analyzer problems and a two week boiler outage. Testing was delayed in
October due to malfunctions in the flowmeter (and its spare) that was used to
control the N i- I 3 injection rate. Also, the continuous analyzer for NH 3 emissions
from the reactor outlet did not work properly. It was then decided to use a
grab sampling/wet chemical analysis method for NH 3 emission -monitor±ng instead.
of the continuous analyzer. In November the plant was shut down to revise
the ductwork around the reactor to prevent fly ash buildup, install an NH 3
distributor (two crossed pipes with orifices in them), and clean the catalyst.
After the initial system debugging work, the optimization test for the
effect of NO concentration showed that the inlet NO concentration had little
or no effect on NO < reduction efficiency. During the NH 3 /NO mole ratio and
the flue gas flow rate tests it became apparent that the catalyst performance
36
-------
was decreasing since the percentage NO reduction decreased and the reactor
pressure drop increased. Because of this, the pilot plant was shut down -
again and the reactor was opened. The catalyst was taken out and cleaned.
When the catalyst was replaced the pressure drop- returned to normal but the
NO reduction was still lower. It became apparent that the activity of the
catalyst had decreased enough so that 90 percent NO reduction could not be
maintained; therefore, HZ decided to replace the catalyst. At this point
the reactor had processed flue gas for approximately 2500 hours.
Catalyst Charge 2
The second charge of catalyst was installed in the reactor in December
1979. It had the same composition, volume, and pitch (size of the gas
passages) as the first catalyst, but because of the plugging problems
experienced with the first catalyst, HZ modified pilot plant operation
to include steam soot blowing once per shift.
Optimization tests with the new catalyst charge were begun after instal-
lation in December 1979. These tests were intended to determine the effects
of NH3/NO mole ratio, flue gas flow rate, and reaction temperature on NO
reduction efficiency and the amount of NH 3 exiting the reactor.
The optimization tests on catalyst charge #2 were characterized by
unexpectedly high NO reduction efficiencies. It was determined later that
particles of catalyst had become imbedded in the outlet analyzer probe. As a
result, the probe was functioning as a catalytic reactor reducing NO and
then giving artifically low outlet NO concentrations resulting in high calcu-
lated N0 reduction efficiencies. The problem was corrected by replacing the
probe and periodically checking it for catalytic activity. The results of the
optimization tests identified the conditions at which the pilot plant would
operate during the demonstration test. These conditions are summarized in
Table 3—3.
37
-------
TABLE 3—3. RESULTS OF OPTIMIZATION TESTS ON CATALYST CHARGE 112
Parameter
Value
NH3INOx
Mole Ratio
1.0
Flue Gas
Flow Rate
1050 scfm
Flue Gas
Temperature
700°F
NO Redu
ction Efficiency
92%
As efforts were made to begin the final phase of the project, a contin-
uous long—term run under optimum conditions, the NO reduction efficiency
dropped unexpectedly. This phenomenon was similar to the experience with the
first catalyst charge as the decrease in efficiency occurred after a plant
shut down and was accompanied by a large increase in reactor pressure drop.
The shutdown (to replace flue gas heater elements) and start—up had followed
the prescribed procedures: the catalyst was never in contact with flue gas
or NH 3 at temperatures less than 230°C. The reactor was completely purged
with hot air before cooling. On start—up, hot air was used to heat the
catalyst to 230°C. Flue gas was then introduced and the temperature raised -
to 315°C at which point NH 3 was introduced. Numerous attempts were made to
clean the catalyst: increasing soot blowing pressure and frequency, passing
hot air through the catalyst, and removing the catalyst to clean it physically
and with compressed air. None of these efforts were successful. HZ then
decided to replace the catalyst with a new charge. At this point the second
catalyst had processed flue gas for approximately 2300 hours.
When the second catalyst was removed from the reactor, physical examina— -
tion revealed that many of the gas passages had become obstructed by
agglomerated deposits of fly ash. Initially, it was suspected that the steam
used for soot blowing was adding moisture to the reactor and resulted in
fly ash agglomeration. For this reason, HZ decided to use compressed air
for soot blowing in future tests. Subsequent examination of the fly ash,
however, indicated that the ash had a tendency to agglomerate without any
additional moisture being present.
38
-------
Catalyst Charge 3
The third catalyst charge was of a different type than the first two
charges. This catalyst had somewhat different characteristics, primarily a
wider pitch (spacing between plates) to help prevent plugging by fly ash.
The wider pitch of this catalyst necessitated a longer catalyst bed to provide
the same amount of surface area, so the reactor had to be lengthened. After
this work had been completed, the pilot plant was started up with the third
catalyst charge in May 1980.
The results of the optimization tests on the third catalyst charge are
summarized in Table 3—4.
TABLE 3—4. RESULTS OF OPTIMIZATION TESTS ON CATALYST CHARGE 3
Parameter
Value
NH3/NOx
Hole Ratio
1.0
Flue Gas
Flow Rate
1500 scfm
Flue Gas
Temperature
700°F
NO Redu
ction Efficiency
93%
Low flue gas flow rate tests were postponed because of a slight increase
in reactor pressure drop during the first test. Because of the plugging
problems experienced with the first two catalyst charges, it was decided to
run the long—term demonstration test at a higher flow rate.tp attempt to
keep the catalyst surface cleaner via the increased gas turbulence. NO
removals at the higher flow rates were above 90 percent so the demonstration
test was initiated in June 1980. -
3.3.3 Demonstration Test
One of the main purposes of this pilot plant program was to operate
the HZ process continuously for a period of 90 days. The primary objective
of this demonstration test was to achieve a daily average NO reduction of -
39
-------
90 percent for at least 75 out of 90 days. The initial 90—day demonstration
test at the HZ pilot plant began in May 1980 and continued through August
1980. During this 90—day period, only 75 days of data were obtained due to
malfunctions of the outlet N0 < analyzer which required repairs. Out of these
75 reporting days there were 57 daily average NO reduction efficiencies of
90 percent or greater. Average values of the descriptive process parameters
for the 90—day demonstration test are listed in Table 3—5.
TABLE 3—5. RESULTS OF 90—DAY DEMONSTRATION TEST: MAY—AUGUST 1980
Parameter
Average Value
NH3/NOx Mole Ratio
0.98
Flue Gas Flow Rate
.
1410 cfm
Flue Gas Temperature
713°F
NO Reduction Efficiency
90.7%
Inlet N0 Concentration
452 ppm
NH 3 Emissions
53 ppm
Because of the success of the initial 90—day test period, the demonstra-
tion test was extended. A major objective during the extended test period was
to determine the long—term effects of exposure to coal—fired flue gas on
catalyst activity. The N0 reduction performance of the system decreased
slightly, by about 2.5%, during the period from August to October 1980.
At approximately the same inlet conditions, the average NO removal for this
period was 88.3%. The results of the long—term demonstrat-ion test through
October are listed in Table 3—6.
40
-------
TABLE 3—6. RESULTS OF LONG-TERN DEMONSTRATION TEST: MAY—OCTOBER 1980
Parameter
Average Va1u -
NH3/NOx Mole Ratio
0.98
Flue Gas Flow Rate
1370 scfm
Flue Gas Temperature
7Ol F
NO Reduction Efficiency
89.8%
Inlet NO Concentration
451 ppm
N H 3 Emissions
57 ppm
For this approximately 5 month period, out of 144 calendar days there
were 118 days of reported data. Again, the 26 days where no data were-- -
reported were characterized by boiler outages and analyzer breakdowns.
Of these 118 reporting days, 63 days (over 50%) had average NO reductions
of 90 percent or greater.
3.4 INDEPENDENT EVALUATION OF THE HZ PROCESS
Radian Corporation conducted a sampling test program at the HZ
pilot plant to enable an independent evaluation of the HZ process. This
program featured the three following areas of testing:
• Quality Assurance
• Continuous N0 Monitor Certification
• Stack Sampling (secondary emissions)
This independent evaluation test program was performed during July and
August 1980, the later stages of the initial 90—day demonstration test on the
third catalyst charge.
The following discussion presents the test plans for each area of the
independent evaluation. The results of this evaluation are presented in
Section 4.
41
-------
3.4.1 Quality Assurance
An independent quality assurance program was conducted by Radian at the
HZ pilot plant to ensure the accuracy of the process measurements made by
Chemico. The measurements audited were those that are necessary to charac-
terize the performance of the HZ process, quantify raw material and utility
demands, and permit a technical evaluation of the HZ process for a commercial
installation. Those process measurements audited by Radian and their fre-
quencies are listed in Table 3—7.
TA.BLE 3-7. AUDITS OF THE PROCESS MEASUREMENTS AT
TilE HITACHI-ZOSEN PILOT PLANT
Process Measurement
Audits/Week
N i - I 3 Injection Rate
1
Reactor Pressure Drop
2
Flue Gas Flow Rate
2.5
Nil 3 Emissions
3
-
Reactor Inlet Temperature
1
- -
The results of the quality assurance audits and the procedures used are
presented in Appendix A. A summary and discussion of these results appears
in Section 4.
3.4.2 Continuous NO Monitor Certification
One of the most critical aspects of the independent evaluation test
program was to certify the continuous monitors that measured the reactor
inlet and outlet NO concentrations. The procedures used have been developed
by EPA to ensure the accuracy of continuous monitors at facilities which must
meet new source performance standards. A continuous monitor must pass a
number of tests which include:
42
-------
• relative accuracy,
• calibration error,
• response time,
• 2— and 24—hour calibration drift, and-
• 2— and 24—hour zero drift.
The specifications for these tests are in
197, October 10, 1979 — “Proposed Rules:
Stationary Sources; Continuous Monitoring
specifications are listed in Table 3—8.
the Federal Register, Vol. 44, No.
Standards of Performance for New
Performance Specifications.” These
These certification tests are discussed here and results are summarized
in Section 4.
TABLE 3—8. CONTINIJOUS MONITOR CERTIFICATION TESTS
Certification Test
Relative Accuracy
Calibration Error
Response Time
Calibration Drift (2—hour)
Calibration Drift (24—hour)
Zero Drift (2—hour)
Zero Drift (24—hour)
Performance Specifications
<20 percent of reference methods test
data in terms of emission standard
expressed in mass per unit heat input
<5 percent for mid—range and high
level calibration values
<15 minutes
<2 percent of span value
<2.5 percent of span value
<2 percent of span value
<2.5 percent of spanvalu
Relative Accuracy
Relative accuracy is the most important test of the continuous NO
monitors performance. This test is a comparison of the continuous monitors
readings with a stack gas analysis as determined by EPA reference methods:
Method 7 for NO and Method 3 for 02 and C02. Nine Method 7 tests are made,
43
-------
requiring three samples per test and the results are compared to the readings
of the continuous N0 monitor during the time the stack gas samples were
collected. Simultaneous Method 3 tests are performed so the N0 concentra-
tions can be converted to a mass emissions per unit heat input basis. Then,
the relative accuracy is calculated by using the following equations.
I I + Ic.I.gsI
Relative Accuracy = R.V. (3—7)
where = absolute value of the mean difference between the monitor
reading and the average value of the Method 7 tests,
R.V.= average N0 concentration of the Method 7 tests,
I c.I.g I = absolute value of the 95 percent confidence inter ial which
is defined by Equation 3—8.
t 9 7 5 L
C.I.gs = (n E — (EX.) 2 Y (3—8)
n(n-1) 1
where n = number of measurements (9),
t 9 7 5 = 2.306 for 9 tests,
X. = difference between the continuous monitor reading and
1 the Method 7 test result.
This procedure was followed exactly at the HZ pilot plant.
Calibration Error
Calibration error is a measure of the continuous monitor’s ability to
accurately measure the concentration of a calibration gas. A series of
15 measurements are made, 5 each of high— and mid—level calibration gases
and 5 of zero gas (air). The measurements are made such that no gas is
measured two or more times in a row. The calibration error is determined
using Equations 3—7 and 3—8 where the R.V., the reference value, is equal
to the calibration gas concentration and t = 2.776 (for 5 measurements).
.975
This procedure was followed exactly at the HZ pilot plant.
44
-------
Response Time
Response time measures the amount of time before a continuous monitor
responds to a change in the NO concentration in the duct. Zero gas (air)
is introduced into the continuous NO)< monitor’s sample line at the duct.
When the reading stabilizes at zero, the system is switched to monitor the
flue gas and the time required to reach a stable value is the upscale response
time. Downscale response time begins with high—level calibration gas and
the procedure is repeated.
Zero and Calibration Drift
Zero and calibration drift are measures of the change of the continuous
NO monitor’s response to zero and calibration gases with time. For the
two—hour drift test, zero and high level span gases are passed into the
monitor every two hours. The difference between the reading and the zero
or calibration gas concentrations is recorded. Fifteen sets of data are
taken. Drift is calculated using Equations 3—7 and 3—8 where the reference
value (R.V.) is equal to the monitor range and t 975 = 2.145 for 15 tests.
24—drift is measured similarly in 7 tests where t 75 = 2.447.
Since this pilot plant was an experimental unit, it was crucial for the
monitors to be accurate; therefore, they were calibrated every two hours,
sometimes even more frequently. The 24—hour drift test was irrelevant to
the operation of this pilot plant. Zero and calibration drift readings were
measured prior to Chemico’s routine calibration.
Detailed results of the continuous monitor certification tests are
presented in Appendix B while a summary of the results is presented in
Section 4.
45
-------
3.4.3 Stack Sampling for Secondary Emissions
In this study, NO emissions are of primary interest; however, emissions
of other compounds are also important. These were termed secondary emissions
and the sampling for these compounds was a part of the independent evaluation
test program. The objective was to quantify any change in concentration
due to passing across the HZ catalyst. Table 3—9 lists the pollutants and the
sampling plan of the secondary emission testing.
TABLE 3—9. SECONDARY EMISSIONS TESTING
Pollutant
Sampling
Location
Number of
samples at
each location
Particulate Matter Loading
Reactor
Inlet/Outlet
3
Particulate Matter Composition
“
3
Sulfur Trioxide (SO 3 )
“
6
Carbon Monoxide (CO)
“
3
Hydrocarbons (Ci—C 6 )
3
Hydrogen Cyanide (HCN)
“
3
Nitrosoamines
“
3
Nitrous Oxide (N 2 0)
“
3
Particulate matter loading was measured to quantify the loading of fly
ash to which the catalyst was exposed. This would demonstrate the catalyst’s
capability to operate under high fly ash loading conditions. Particulate
matter composition was analyzed to determine if catalyst erosion caused a
measurable change in particulate composition. The composition was analyzed
for any increase in titanium or vanadium (primary components of the-HZ
catalyst) concentration at the reactor outlet.
46
-------
SO 3 concentration in the flue gas was measured to quantify any produc-
tion of SO 3 due to SO 2 oxidation by the catalyst. Some degree of SO 2 oxida-
tion was expected as vanadium pentoxide is the catalyst used commercially to
oxidize SO 2 to 503 in sulfuric acid plants. -
Carbon monoxide and hydrocarbons were measured to determine any change
in concentration across the reactor. Since they both can be oxidized by
oxygen or NOR, it is possible that their concentrations might be reduced.
Hydrogen cyanide, nitrosoamines, and nitrous oxide were measured to
determine any change in concentration across the reactor. While no test
results indicate that these compounds are produced by SCR processes, there
are concerns that NB in the presence of the catalyst might result in the
production of one or more of the compounds.
The analyses of samples (except for nitrosoamines) was conducted onsite
to give quick feedback in case of unusual results. The sampling and analysis
procedures and their results are given in Appendix C. A summary of the
results appears in Section 4.
47
-------
SECTION 4
HITACHI ZOSEN PILOT PLANT TEST RESULTS
The HZ pilot plant tests. conducted by Chemico and the independent process
evaluation test program performed by Radian were designed to assess the techni-
cal feasibility of the HZ process and allow more accurate cost estimates to
be made. The results of these test programs were utilized to produce a
conceptual design for a commercial application of the HZ process... This e tion
summarizes and discusses the most significant results of the tesLprogram.
Test data are presented in this section as singular data points; however,
they do not represent a single test. Instead, they represent the average of
data collected at specific performance levels. For example, where NO removal
is presented for a particular temperature, mole ratio and flow rate, the NO
removal number represents the average of all appropriate tests at these con-
ditions. Likewise, the temperature, mole ratio and flow rate represent
averages of data lying within a specified range.
More detailed results of Radian’s test program are contained in the
Appendices. Detailed results from the Chemico test program are contained in
the final report prepared by Chemico and Hitachi Zosen. - - . .
In the sections which follow, the test results of each phase of testing
are discussed separately. The sections are organized in the following order:
• Optimization Tests,
• Demonstration Tests,
• Quality Assurance Audits,
Preceding page blank 49
-------
• Continuous NO Monitor Certification Tests, and
• Stack Sampling for Secondary Emissions.
4.1 OPTIMIZATION TESTS
Optimization tests conducted by Chemico were intended to determine the
optimum values of the controlling process parameters to obtain 90% NO reduc-
tion. The parameters of interest were as follows:
• Flue Gas Temperature,
• Flue Gas Flow Rate, and
• N1 3 /NO Mole Ratio.
x
Optimization tests had been initiated on the first catalyst charge and
completed on the second catalyst charge, both of which were the NOXNON 500
series. It was determined from these tests that the flue gas temperature
did not have a strong effect 0-i NO reduction, within the normal operating
temperature range of the catalyst (300 to 400°C). As the flue gas temperature
approached 430°C, there was a slight decrease in NO reduction probably due
to increased NH 3 oxidation. As the flue gas temperature drops below 340°C,
the NO reduction decreases due to slower reaction rates.
Some of the temperature opt:tinization test results are presented in
Table 4—1.
TABLE 4-1. SUMMARY OF TEMPERATURE OPTIMIZATION TEST RESULTS
Flue Gas Flow Rate NH
(scfm) Mole
3/NO Flue Gas Temperature
Ratio (°C)
N0
-
Reduction
(%)
1050
.91
410
92
1050
.89
378
91
1050
.92
339
50
-------
These results indicate that NO reduction is not a strong function of tempera-
ture. Therefore, for commercial operations temperature need not be controlled
as long as it remains in the 340 to 410°C temperature range.
The flue gas flow rate optimization tests showed that an increase in-
flow rate results in a slight decrease in NO reduction efficiency. These
results are shown in Figure 4—1. While there is a slight decrease iii percent
NO reduction with increased flow rate, the rate of the NO reduction
x
reactions increases significantly. This is probably due to enhanced mass
transfer which can occur as a result of increased gas turbulence in the reactor.
There is an economic tradeoff between increased flue gas flow rate (effectively
reducing the amount of catalyst required per unit mass of NO removed) and
increased reactor pressure drop. A good commercial design hould provide
the reactor with the capability of handling the maximum anticipated flow -
rate. Then, at lower flow rates, as could occur during low load operation,
the percent N0 reduction would not decrease.
The NH3/N0 mole ratio optimization tests were designed to quantify the
effect of NH3/NO injection ratio on NO reduction. With the flue gas tempera-
ture and flue gas flow rate held constant, the NH3/NO injection ratio was
varied and the resulting NO reduction was measured. Results of these Nl-I3/NO
mole ratio tests are shown in Figure 4—2. Also, Figure 4—2 shows the magni-
tude of NH 3 emissions exiting the HZ reactor. These results were determined
by stack sampling tests conducted individually by both Chemico and Radian.
At lower N}l3/N0 mole ratios, an increase in the mole ratio results in signif 1—
cant increase in NO reduction with only a moderate increase in-NH 3 emissions
However, at higher mole ratios less additional NO reduction is achieved and
NH 3 emissions significantly increase when the mole ratio is increased.
4. 2 DENONSTRATION TEST RESULTS
The primary objective of the demonstration test at the HZ pilot plant
was to reduce NO emissions by 90% continuously for 90 days. This demonstra-
tion test began in May 1980 and since it was successful, was extended 5 months
beyond the initial 90—day period.
51
-------
100
80 _______________________________________________________________________
I’ I I I I I I I I I I
1500 1600 1700 1800 1900 2000 2100 2200 2300 2400 2500 2600
Flowrate (Nui 3 /hr)
Figure 4—1. N0 reduction as a function of flue gas flourate.
-------
100_ _100
90_... _90
80_ _80
U
0
70 _70
rt
I-i
C i
60 _60
.,-I 0
C i
-1
4 - 4 CD
50 _5Q
‘1
CD
0 CD
U
40 _40g
ii
-0
a) 0
1 .-i
_30
o CD
Z rt
20 20°
— Ii
a
10
_0
I I I I I, I I
0.5 0.6. 0.7 0.8 0.9 1.0 1.1 1.2 1.3
NFI3/NOx injection ratio
Figure 4—2. NO reduction efficiency and NH 3 emissions as
a unction of NH3/NO injection ratio.
-------
Table 4—2 lists the average values of the key process variables during
the demonstration test. -
TABLE 4—2. KEY PROCESS VARIABLE VALUES DURING THE NO DEMONSTRATION TEST,
MAY-OCTOBER 1980
Operating Variable
Average Value
Standard Deviation
Flue Gas Flow Rate (scfm)
1370
97
Reactor Pressure Drop (in. H 2 0)
1.12
0.10
Inlet NO Concentration (ppm)
451
54
Flue Gas Temperature (°F)
701
24
NH3/NOx Mole Ratio
0.98
0.07
N0 Reduction (7.)
89.8
3.8
NH 3 Emissions (ppm)
54
15
Figures 4—3 and 4—4 graphically show the daily average performance of the
HZ pilot plant. The system successfully met the objective of continuous 90%
N0 removal of the 90—day demonstration test. However, Figure 4—3 shows that
the N0 reduction efficiency was slightly decreasing over the 5—month period.
This decrease in N0 reduction indicates a slight loss of catalyst activity
due to exposure to coal—fired flue gas.
Figure 4—5 shows the estimated relative reactivity of the catalyst as a
function of time. These curves were determined by calculating the N0
reduction reaction rate at an NH3/N0 mole ratio of 1.0 at various times
during the pilot plant’s operation. Figure 4—5 shows that after three months -
of exposure to coal—fired flue gas, the catalyst reactivity had decreased
by a few percent. Six months of exposure saw yet another decrease in reactivity.
Finally, after nine months of operation, just prior to the final shutdown and
disassembly of the pilot plant, the catalyst reactivity had decreased approxi-
mately 20% relative to its initial performance. At this point an experimental
regeneration technique was implemented in which the catalyst was water—washed
with warm water. This technique returned the catalyst from 77% of its initial
54
-------
‘I
I.
R
E
Fl
0
V
A
L
N
H
3
N
0
x
100
90
80
70
1.2
1.0
0.8
0.6
11
U,
:
-
+4 ++ -: -•
•+
+4+4+4+4+4+444
+++++•+ 4+ + 44+ 4+++ ++++++++ ++++4 ++ —+
+ 4+ +
.4
• + +
#4
+4
-
. 4
-
. 4
I
.
4+ ++ 444+4+
4++4+÷ ,+ +
+
44+.+444.4++++ 4+4+ #4 +44+44k
+ +# ++ + ++ 4 •+ +
+
+ +4
-
+
*
. 4
*Note Additional low mole ratio tests were
these two days.
run on
-
+ 4
+ 4+
4444444
+
.#4+ .44.4 4 . 1+H .+ 44+44444 4#44444#444+4+444*4444+ 144 1* 4
+
- +4+
1600
0 1500
w
1 400
G 1300
C
1 1200
m 1100
MAY ; JUNE
JULY
AUGUST
1 SEPTEMBER OCTOBER
Figure 4—3. Results of NO demonstration test of HZ pilot plant.
-------
1.1
1.5
E
L 1.3
T
A
P
0.9
730
710
M690
P
670
F 650
630
650
N
0550
x
I
N 450
350
+
-
+
—
-
+ +
++ 4+4+ +4+4* ++ + +
+
+ +
4
+ 44+
+ ,f +
4
+ +
. .++ +++
4+ +
.&
1 + +
+ + +
+ + 44+ 4 4+
+ 4+ +4+ 4 I. +
, +•
4++ +
+
+
4 4.
+ + +
4.
+ ‘ ++.+4.4411++ 44 4+ + 4e&ê+
•+
444+ + +4+44+ +
4 4.
+
++ 4 a l
4.
+444
4.
+
+
+ +
++
+
4
.4+
+
+ 44+
+
•+ + + ,
+ +44+ •+
run on these
two days.
4*
4
—
-
• ••+ + + + +
+4 4+ + +
+ 4
I:
+ +44+
1.
• + +44+ +44 + +44+
+ + 4• + ++ + +
+, 44+
+
+
++4++
+ +44,
4
MAY JUNE
JULY
AUGUST SEPTEMBER
OCTOBER
o I
Figure 4—4. Results of demonstration test of ciz pilot plant.
-------
1.0 - -‘
:=
c i
.6—
c j
.5—
Key
0 Before Regeneration
A After Regeneration
N•113/NOx 1.0
.2— NOXNON 600 Catalyst
.1—
0— I I I I I I I
0 1 2 3 4 5 6 7 8 9
Time from start—up (months)
Figure 4—5. Relative catalyst activity vs. time.
-------
activity to 85% of the initial activity. Although the demonstration test of
the HZ pilot plant did not show the process capable of maintaining 90% N0
reduction for a complete year, the success of the 90—day run and the potential
catalyst regeneration technique indicate the process is suited for application
to a coal—fired source.
4.3 QUALITY ASSURANCE AUDIT TEST RESULTS
The quality assurance audit tests were performed by Radian to ensure that.
the data collected by Chemico from the pilot demonstration of the HZ process
were accurate. To accomplish this evaluation, Radian used standard reference
methods to audit the process variables that characterize the performance of
the HZ process and that were monitored continuously by Chemico. The one
variable audited that was not monitored continuously by Chemico w s the
reactor outlet NH 3 concentration. The pilot plant was originally designed to
monitor the NH 3 emission rate continuously by using a total nitrogen analyzer.
However, the sample gas conditioning system to prevent aminonium sulfate/bi—
sulfate formation (by removing the SO 3 ) never functioned properly. As a
result, Chemico took batch NTh samples for spectrophotomic analysis on a
regular basis.
The results of Radian’s NH 3 emission tests are listed in Table 4—3. These
tests show an average NH 3 emission rate of 55 ppm for an NH 3 /N0x mole ratio
of 0.96 with 464 ppm N0 at the reactor inlet. This emission rate is signifi-
cantly higher than the 10—20 ppm expected by Chetnico. This may have been due
to an insufficient amount of catalyst or, conversely, an excess±ve flue gas
flow rate, either of which affect the space velocity and hence residence time.
Lower ammonia emissions would be anticipated if a lower space velocity was
implemented.
58
-------
TABLE 4-3. HITACHI ZOSEN PILOT PLANT NH 3 EMISSIONS
Test
NO In
(ppm wet)
NH 3 :NO
NH 3 Emission
(ppm dry)
1
459
0:91
45
2
451
0.97
46
3
466
0.92
49
4
5
453
469
0.92
0.99
60
44
6
479
0.99
63
7
8
440
450
1.05
0.89
54
58
9
10
460
493
1.01
0.89
74*
41*
11
482
0.98
68*
Average = 464
0.96
55
*Note: Sample analyzed by pH adjustment, direct nesslerization/spectrophotometry
instead of pH adjustment, distillation/titration as for other samples. Tests -
4 and 5 were run with both methods with good agreement. -
Normalized ammonia emission results from both Radian and Chemico test
data are shown in Figure 4—6. NH 3 emissions are normalized by dividing by
the inlet NO concentration. This method of presentation is more useful han
the standard NH 3 emissions vs. NH3/NO mole ratio plot since it includes the
effects of N0 concentration. N0 concentration will affect the concentration
of ammonia at the reactor outlet (at a constant N}I3/NO mol ratio) since an
increase of the inlet N0 concentration produces a corresponding increase in --
the quantity of NH 3 injected and a consequent increase in the quantity of NH 3
emitted.
The results of the other QA audits performed by Radian are suimnarized in
Table 4—4. (More detailed results, including the individual audit results,
are included in Appendix A). All the QA measurements except the reactor inlet
SO 2 concentrations were within 10 percent of the measurements made by Chemico.
59
-------
• 20—
1-i
x
0 .15—
I I I I I I I
0.5. 0.6 0.7 0.8 0.9 1.0 1.1 1.2
NH 3 /NO ratio
Figure 4—6. Normalized NH 3 emissions as a functiOn of NH3/NO injection ratio.
-------
TABLE 4-4. OTHER QA AUDIT TEST RESULTS
Process Variable
Audit Method
Relative Error (%)1
Flue Gas Flow Rate
EPA Reference Method Z
—0.3
NH 3 Injection Rate
Absorption in 50% H 2 SO& 4
followed by weight gain
measurement (similar to
EPA Reference Method 4)
—6.0
Reactor Pressure Drop
Magnehelic differential
pressure gauge
4.5
Reactor Inlet Temperature
Thermocouple with traverse
of reactor inlet
4.8
Reactor Inlet SO 2
EPA Reference Method 6
—19.8
Concentration
-
1 Relative Error = ( Process Measurement—Audit Result) 1007
Audit Result
Overall, the QA audit test results show that the process measurements
made by Chemico were accurate. However, an exception is the flue gas SO 2
concentration measurement by Chemico which was determined to be 20% low.
This was due to the quenching effect of 02 and CO 2 on the pulsed fluorescence
monitor used to measure SO 2 concentration. This error is typical for this
type of SO 2 instrument when calibrated with a reference gas of SO 2 in N2,
resulting in a lower SO 2 value when monitoring a gas stream containing oxygen
and carbon dioxide. Studies have shown that this error can be as high as 30%.
Chemico monitored the reactor inlet SO 2 concentration to determine the level
of SO 2 to which the catalyst was exposed. Fortunately, this was not a critical
measurement for characterizing the performance of the Hitachi—Zo en7Chemicc
pilot plant.
Based on the test results of Radian’s QA audits, Chemico’s process
measurements were within the accuracy of the methods utilized. Therefore,
there is no need to develop a correction factor for the process data. Chemico’s
process data can be used to characterize the performance of the pilot plant
and serve as a basis for a technical evaluation of the HZ process.
61
-------
4.4 CONTINUOUS NO MONITOR CERTIFICATION TEST RESULTS
Certification tests of the continuous NO monitors were conducted by
Radian to ensure the accuracy of the NO reduction data obtained by Chemico.
The certification tests entail rigorous procedures to test monitors used at a
commercial source for compliance with New Source Performance Standards.
Because the HZ pilot plant was an experimental unit, some of the procedures
were modified.
Table 4—5 lists the results of the continuous monitor certification
tests for the NO monitors at the HZ pilot plant. (More detailed results can
be found in Appendix B).
TABLE 4—5. CONTINUOUS NO MONITOR CERTIFICATION TEST RESULTS AT
HZ PILOT PLANT
Certification
Test
Performance
Specification
HZ
Pilot
Plant
Inlet NO
Monitor
.
Outlet NO
Monitor
Calibration Error
—high level
—mid level
5%
5%
1.40
4.39
4.70
2.68
Response Time
l5 mm
1.4
1.6
Zero drift (2—hour)
2%
1.20
0.05
Calibration Drift
(2—hour)
2%
1.93
1.78
Relative Accuracy
2O%
14.1
10.5
‘Alternatively, <10 percent of the applicable emissions standard.
The 24—hour zero and calibration tests were not performed as Chemico calibrated
the NO monitors every two hours; therefore, the 24—hour tests were irrelvant
for these monitors.
62
-------
The continuouS monitor certification test results show that the continuous
NO monitors at the HZ pilot plant were functioning properly, so the data
collected by Chemico accurately describes the pilot plant’s NO reduction
performance.
4.5 STACK SAMPLING FOR SECONDARY EMISSIONS TEST RESULTS
A stack sampling program for secondary emissions was conducted by Radian
to determine if any change ii i concentration occurred across the-reactor.
Table 4—6 summarizes the results of the secondary emissions sampling tests.
The following discussion of the results describes their impact on commercial
scale HZ process applications.
Sulfur Trioxide
Sulfur trioxide (SO 3 ) samples were collected at the reactor inlet and
outlet to determine if the HZ process produced SO 3 by the oxidation of SO 2 .
This was to be expected since the actual component of the HZ catalyst, vanadium
pentoxide, is used commercially to catalyze SO 2 oxidation.
The results of each of the SO 3 tests are listed in Table 4—7. The test
results indicate an average SO 2 oxidation of 1.5%. These results are in agree-
ment with previous experience at SCR installations in Japan showing 1—2% SO 2
oxidation.
This increase in SO 3 concentration exiting the HZ reactor haè potentially
significant impacts on a commercial application of the HZ process. With over
50 ppm NH exiting the reactor, the increased amount of SO 3 present will result
in greater production of ammonium sulfates downstream, especially iii the
boiler air preheater. Aimnonium sulfates can exist in the form of a liquid
and/or solid solution. They are corrosive and can cause plugging of the pre-
heater. The higher SO 3 concentration will also raise the initial formation
temperature of the ammonium sulfates. This could necessitate the use of
corrosion resistant material in the intermediate temperature zolie of the pre-
heater. Also, additional soot blowers may be needed along with increased soot
blowing frequency.
63
-------
TABLE 4-6. SUMMARY OF STACK SANPLING TEST RESULTS AT HZ PILOT PLANT
Flue Gas Reactor Inlet Reactor Outlet Measurement
Component Concentration 1 Concentration 1 Technique
Sulfur Trioxide 8.4 ppmv (dry basis) 20.7 ppmv (dry basis) Controlled condensation,
ion chromatograph
Particulate Loading 7.1 gm/dscm 7.7 gm/dscm In stack filter
Ilydrogen Cyanide 2 <.01 mg/dscm <.01 mg/dscm Absorption, distillation,
titration
Nitrosoamines 2 <5 jg/dscm <5 pg/dscm Absorption, extraction, gas
chromatograph w/nitrogen
specific detector
1Iydrocarbons 2 <1.0 ppmv <1.0 ppmv Gas chromatograph
(C 1 —C 6 ) flame ionization detector
Carbon MonoxIde 2 <0.017% <0.017% FIscher gas partitioner
1 Average of 3 or more tests
2 Below the detection limit
-------
TABLE 4-7. CHANGE IN SO 3 CONCENTRATION ACROSS THE REACTOR
AT ThE HZ PILOT PLANT
Test
SO 3 Inlet
(ppm dry)
SO 3 Outlet
(ppm dry)
SO 2 Oxidation
(%)
1
11.0
25.3
1.5
2
7.5
14.8
1.1
3
6.6
21.9
1.9
Particulate Matter (Fly Ash )
Particulate matter samples were collected from the HZ reactor inlet and
outlet to quantify the fly ash loading to which the catalyst was exposed and
to determine if catalyst erosion was occurring.
The results of the fly ash loading tests are shown in Table 4—8.
TABLE 4-8. PARTICULATE MATTER LOADINGS MEASURED AT ThE HZ PILOT PLANT
Test
Inlet
(gm/dscxn)
Outlet
(gm/dscm)
1
7.5
7.8
2
6.9
8.4
3
6.9
—
4
—
6.9
Particulate matter concentrations averaged 7.1 gm/dscm at the inlet and 7.7
gm/dscm at the outlet. These results are quite consistent, especially
considering that the sampling ports were located so that a traverse was
possible in only one plane.
65
-------
Table 4—9 lists the results of the elemental analysis of the particulate
matter samples taken at the HZ pilot plant. The elemental analysis test was
performed to determine if there was an increase in titanium (Ti) or vanadium (V)
concentration in the particulate matter exiting .-the reactor. The results show
the concentrations of titanium and vanadium to be constant relative to the
other elements, thus no measurable change in Ti or V concentration occurred.
There is an apparent consistent increase in all the elements across the
reactor. This was probably due to the presence of inert materials. There-
fore, while catalyst erosion may have occurred, the rate of erosion produced
no significant change in Ti or V concentration.
TABLE 4-9. RESULTS OF PARTICULATE MATTER ELEMENTAL
ANALYSIS AT THE HZ PILOT PLANT 1
Component
Reactor Inlet
Reactor Outlet
Out/In -
-
Al
10.7%
13.0%
1.21
Ca
8200 ppm
9900 ppm
1.21
Fe
4.9%
6.0%
1.22
K
2.0%
2.5%
1.25
-
Mg
6300 ppm
7800 ppm
1.24
Mn
190 ppm
240 ppm
1.26
Sn
490 ppm
680 ppm
1.40
Na
4200 ppm
4700 ppm
1.12
Si
18%
23%
1.28
Zn
190 ppm
250 ppm
1.32
Cu
150 ppm
170 ppm
1.13
Ti
5800 ppm
6900 ppm
1.1g
V
270 ppm
330 ppm
1.22
‘Concentrations are on a mass fraction basis
66
-------
jIydrogen Cyanide
Samples were collected and analyzed to determine if any HCN was being
produced in the reactor by the reaction of NTh and hydrocarbons. Both reactor
inlet and outlet samples were below the detection limit of the analytical
method, 10 ppb. This is orders of magnitude less than the 10 ppm threshold
limit value (TLV), maximum allowable for worker exposure. As a result, HCN
production Is not a problem with the HZ process.
NitrosoamineS
Samples from the reactor inlet and outlet were collected and analyzed
for nitrosoamine formation. Both samples were below the detectfon limit of
the analytical procedure, 5- pg/in 3 . This is well below the 65 pg/rn 3 source —
concentration (for n_Nitrosodimethylamine which has the lowest safe level of
aliphatic nitrosoamines) considered safe by EPA ’s multimedia environmental
goals. This means that potential nitrosoamine production in the HZ reactor,
if any, should not pose any environmental problems.
Hydrocarbons (MC) and Carbon Monoxide (CD )
The concentration of both HG and CO at the reactor inlet and outlet were
below the detection limit of the analytical method. No conclusion can be
made about the impact of the HZ catalyst on these compounds, but it is
unlikely that the HZ process produces them in significant amounts.
Nitrous Oxide (N )
N 2 0 was excluded from the sunuzary of results in Table 4—6 due to the
analytical method used to measure N 2 0 being unsuccessful. Several interfering
peaks on the infrared (IR) spectrum along with the low level of any N 2 0
present made detection impossible.
67
-------
SECTION 5
EVALUATION OF THE HITACHI ZOSEN PROCESS
The overall objective o this study was to determine the technical and
economic feasibility of applying the HZ process to a coal—fired utility
boiler based on the results of the pilot plant tests. The approach used to
meet this objective consisted of several elements. First, a basis was defined
to permit preparation of a conceptual design and completion of- materiararfd
energy balances for a specific application of the HZ process. The basis
is a 500 MW coal—fired boiler and is identical to that used by TVA in preparing
a preliminary economic analysis of several flue gas treatment processes in-
cluding the HZ process. 2 The basis identified material flows to the process,
the required reduction in NO emissions, and the characteristics of unit
operations not examined during the pilot plant tests. Using this information
in conjunction with the pilot plant test results, a conceptual design was
prepared.
Material and energy balances were completed for the process using the
conceptual design results and process data which are representative of the
pilot plant’s operation. The results of the stack sampling tests conducted
by Radian were also considered in preparing the material arid energy balances.
The results of this work identified the raw material and energy requirements
of the process for the 500 MW application. Finally, a cost estimate was
prepared for an HZ process applied to a 500 MW coal—fired boiler. This cost
estimate used the estimate prepared by TVA as a basis. Equipment costs were
adjusted for changes in process stream flowrates or other significant factors
(e.g., changes in required catalyst volume). Using adjusted equipment costs
and the calculated raw material and energy requirements, TVAt5 estimate was
modified and a new cost estimate developed.
Preceding page blank 69
-------
This section presents a discussion of the elements which comprised the
evaluation of the HZ process. The basis for the evaluation is defined and
the philosophy used in preparing the conceptual design is discussed. The
conceptual design results are presented along with details and results of
the material and energy balance calculations. The results of the cost
estimate are also presented with details of how the estimate was prepared.
Finally, a suary of results is presented with a discussion of how the results
impact the technical and economic feasibi] ity of the HZ process.
5.1 BASIS FOR THE EVAJ. UATION OF THE HITACHI ZOSEN PROCESS
This evaluation of the HZ process consisted of examining the raw
material and energy requirements and capital and annualized ost of th -
process for a specific application. This form was chosen for the evaluation- -
to provide a basis for comparing the costs of the HZ process to the costs of
other SCR processes which have been evaluated by TVA. In order to make these
comparisons meaningful, it was necessary to use the same basis in this
evaluation as was used by TVA in preparing their preliminary economic analysis.
TVA’s design basis is presented in detail in a report entitled “Preliminary
Economic Analysis of NO Flue Gas Treatment Processes” (EPA—600/8—80—021).’
The following discussion summarizes the aspects of the design basis which are
applicable to this evaluation of the HZ process.
Power Plant Characteristics
Only one power plant was used in preparing the evaluation df the HZ -
process. This is a new, 500 MW coal—fired boiler, with an assumed Midwestern
location for purposes of preparing cost estimates. The fuel for the plant
was a coal having a heating value of 5820 kcal/kg and containing 3.5percent
sulfur and 16 percent ash. The coal composition and the input coal require-
ments (based on a heat rate of 2.27 Mcal/kWh) for the 500—MW boiler are listed
in Table 5—1.
70
-------
TABLE 5—i. BASE CASE COAL COMPOSITION AND INPUT FLOW RATE
(500—MW new unit, 2.27 Mcal/kWh heat rate)
Component
Wt 7 ,
as fired
kg/hr
C
57.56
111,950
H 2
N 2
4.14
1.29
8,030
2,490
02
7.00
13,610
S
3.12
6,080
Cl
0.15
270
Ash
16.00
31,120
H 2 0
10.74
20,870
Total 1
00.00
194,420
The composition of the flue gas at the exit of the economizer is t iat
which will be processed by the HZ catalytic reactors since the temperature
required in the reactor is close to the economizer exit temperature. This
flue gas composition is presented in Table 5—2 and was estimated based on the
assumption that 95 percent of the sulfur in the coal is emitted in the flue
gas as SO < with 99 percent as SO 2 and 1 percent as SO 3 . It ‘isalso assumed
that 80 percent of the ash is emitted with the flue gas stream. The N0
concentration of the flue gas was estimates to be 640 ppm and consists of
approximately 95 percent nitric oxide (NO) and 5 percent nitrogen dioxide (NO 2 )
TABLE 5-2. FLUE GAS COMPOSITION AND FLOW RATE AT TEE ECONOMIZER OUTLET
(2.27 Mcal/kWh, 500—MW new unit, coal—fired, 3.5% 5,
5820 kcal/kg H 1-IV as fired, at 375CC)
Component
Vol %
kg1hr
-
N 2
73.39
1,412,000
02
3.23
71,000
CO 2
13.56
410,100
SO 2
0.26
11,400
-
SO 3
0.003
144
NO
0.064
1,365
HC I
0.012
300
H2O
9.48
117,300
100.00
2,023,600
71
-------
Emission Reduction Requirements
The NO reduction requirement for the HZ process in this evaluation is
90 percent. This level of NO reduction was selected to determine the cost
of achieving 90 percent NO reduction using SCR. It should be noted that
there are no federal regulations which require this level of NO reduction.
5.2 CONCEPTUAL DESIGN OF THE HITACHI ZOSEN PRGCES-S FOR A 500 MW COAL—FIRED
BOILER
The first step in the evaluation of the HZ process was to prepare a
conceptual design for a 500 MW application of the process using the results
of the pilot plant tests to define key process operating param ter and per
formance levels. In order to complete this step, it was necessar-y to evalua e
and reduce the pilot plant test results to a form whLch would serve as a basis
for the design. It was also necessary to make judgements concerning extra-
polation of results from the pilot plant to a 500 MW scale.
With the pilot plant data in the appropriate form, it was then possible
to define levels of key operating parameters such as the NH 3 /NO injection
ratio, flue gas space velocity, reactor operating temperature, etc. It was
also possible to determine the quantity of catalyst required tQ reduce NO
emissions by 90 percent for a 500 MW application. Once these items were
defined, the conceptual design was essentially complete and material and energy
balance calculations could be made based on the conceptual design results.
The following discussion presents the results of the conceptual design.
It includes details of how the pilot plant test data were reduced and how
those data were interpreted. It also includes a discussion of the general
philosophy followed in preparing the design along with specific details of
how the design was prepared.
72
-------
5.2.1 Philosophy Used to Prepare the Conceptual Design
Chetnico conducted a large number of tests with the HZ pilot plant. These
tests were conducted under a variety of conditions using three separate charges
of catalyst. For purposes of this evaluation it was necessary. to determine how
these test results would be interpreted to quantify process performance and,
in fact, which test results would be used to prepare the conceptual design.
The general philosophy followed in evaluating tesf data and-in preparing
the conceptual design for a 500 MW plant was to limit operating conditions- to -
those which had been well documented in the pilot plant tests. In some cases
it was necessary to extrapolate the test results, but this was minimized. The
following discussion presents specific details of which tests were selecteçi
as a basis for this evaluation.
Several factors were considered in selecting the tests whose results
would be used as a basis for this evaluation. First, it was considered
desirable to have a large number of tests with repeatable results as a basis.
The principal reason for this was that a greater number of tests allow a
more accurate characterization of the process performance and more accurate
estimates of process operating parameters such as reactor pressure drop,
NH 3 /NOx injection ratio, etc. Second, it was considered desirable to use
tests with operating conditions similar to those in which Radian conducted
the secondary emissions sampling program. For these reasons it was decided
to use the results of the demonstration test conducted during the summer
of 1980.
The demonstration test consisted of approximately 90 days of operation
under essentially identical conditions. This provided an accurate characteri-
zation of key process operating parameters associated with specific levels of
process performance. It was also during this test that Radian conducted the
majority of the quality assurance and secondary emissions sampling tests so
those results are directly applicable to a conceptual design based on the
demonstration test.
73
-------
As mentioned earlier, the general philosophy in preparing the conceptual
design was to use the pilot plant data to the maximum extent and limit any
extrapolation of results. More specifically, the design levels of key process
operating parameters such as temperature and flowrate were based strictly on
the levels measured during the demonstration test even though the optimization
tests indicated it might be possible to improve process performance by changing
some process operating parameters.
5.2.2 Pilot Plant Data Used to Prepare the Conceptual Design
The pilot plant data used to prepare the conceptual design were from the
demonstration test. Because of the quantity of data collected during this
test, it was possible to accurately characterize the average va1 tie of key
process operating parameters. The repeatability of the test results is
evidenced by the relatively small standard deviations of the key process
operating parameters during the demonstration test (see Table 4—2).
The pilot plant data used in preparing the conceptual design are sujiari—
zed in Table 5—3. The operating parameter levels shown represent averages of
7800 data points recorded during the demonstration test. One exception to
this is the N H 3 emissions data which are the results of a relatively small
number of manual sampling tests conducted by Chemico and Radian.
TABLE 5-3. AVERAGE OPERATING PARANETER LEVELS DURING THE 90-DAY
TEST OF THE 1-iITACHI ZOSEN PILOT PLANT
Operating Parameter
Average Valiie
Flue Gas Flowrate
2200 Nm 3 /hr
NH 3 /NO Injection Ratio
x
Flue Gas Temperature
0.98
0
370 C
-
Reactor Pressure Drop
0.28 kPa
NO< Removal Efficiency
89.8%
NH3 Emissions
54 ppmv
Inlet NO Concentration
451 ppmv
SO 2 Concentration
1120 ppmv 1
1 This concentration was adjusted based on the results of the QA audits.
74
-------
5.2.3 Conceptual Design Constraints and Procedures
Once the pilot plant data to be used in the conceptual design had been
selected, it was necessary to define design constraints and design procedures.
The constraints used in preparing the conceptual design are sutinnarized in
Table 5—4. The following discussion examines these constraints in detail
along with the procedures used in preparing the conceptual design.
TABLE 5—4. CONSTR.AIN S FOR THE HITACHI ZOSEN CONCEPTUAL DESIGN
Design Parameter
Level
Reactor Space Velocity
8860 hr’
NH 3 /NO Injection Ratio
1.0
Ni-I 3 Emissions
.133 ppm/ppm NO <
-in
SO 2 Oxidation
1.5% of inlet
SO 2
Reactor Operating Temperature
340° to 410°C
Reactor Space Velocity
The reactor space velocity used in preparing the conceptual design is identi-
cal to the space velocity during the demonstration test. Space velocity was
determined by dividing the flue gas volumetric flowrate (5200 m 3 /hr) by the catalyst
volume (0.587 m 3 ). Scale—up of the reactor is straightforward. One stack
of catalyst 0.276 in 2 x 2.13 in is required for each 5200 m 3 /hr of flue gas.
It is necessary for the space velocity and the catalyst configuration to
be constant for both the pilot unit and the conceptual design. This permits
direct use of other test data in preparing the conceptual design.
It should be noted that the space velocity specified for the cohceptual
design is a maximum and thus applies to the flue gas flowrate from the boiler
at full load. As discussed in Section 4, reduced space velocities (at reduced
loads) should not change the N0 reduction efficiency of the reactor.
75
-------
NH 3 fNO>çj jection Ratio
The NH 3 /N0 injection ratio during the HZ pilot demonstration test
averaged 0.98 while the N0 reduction efficiency-averaged 89.8 percent. Since
the conceptual design requires 90 percent NO reduction, a slightly higher-
NH 3 /NO injection ratio is needed. Examination of the data presented in
Figure 4—2 shows that the HZ process NO reduction efficiency should be
90 percent at an NH 3 /N0>< injection ratio of 1.0. This is the level selected
for the conceptual design. -
NH 3 Emissions
N Il 3 emissions from the HZ process are important from several standpoints.
First and most important, the level of NH 3 emissions determines Ehe severity
of impacts on downstream equipment. NH 3 emitted from an SCR reactor can result
in plugging and corrosion of the air preheater, which can increase the costs
associated with the application of SCR. The level of NH 3 emissions can also
determine the extent of adverse environmental impacts (if any) from the SCR -
process. Previous work has shown that excessive NH 3 emissions can result in
environmental problems.
The NH 3 emissions for the conceptual design were estimated to be- .133 ppm
NH 3 per ppm NO in. This estimate is based on an analysis of NH 3 emissions as
a function of N0 inlet and the NH3/NO injection ratio (Figure 4—6). For the
conceptual design with an inlet NO concentration of 640 ppm, this results in
an NH 3 concentration at the reactor oulet of 85 ppmv.
SO 2 Oxidation
SO 2 oxidation was measured during the demonstration test as part of the
secondary emissions sampling program conducted by Radian. Based on Radian’s
measurements SO 2 oxidation averaged 1.5 percent of the SO 2 in the flue gas.
Since the conceptual design has operating characteristics identical to the HZ
pilot unit, the fraction of SO 2 oxidized is also expected to be identical.
76- --
-------
For the conceptual design which has a flue gas SOz concentration of 2590 ppm,
this results in an additional 39 ppm of SO 3 in the flue gas at the reactor
exit.
SO 2 oxidation is important due to its potential impacts on downstream
equipment. By itself, SO 3 can cause severe corrosion in the extreme cold end
of the air preheater or in downstream duct work and equipment such as an ESP.
And when NH 3 is present in the flue gas, additional SO 3 canresult in the
formation of liquid aminonium bisulfate in the intermediate temperature zone
of the air preheater which, for typical air preheater designs, is more suscepti—
ble to corrosion and plugging. Thus the additional SO 3 in the presence of N i - I 3
can cause severe operating problems in the air preheater.
Reactor Operating Temperature
The reactor operating temperatures specified for the conceptual design
ranges from 340 to 410°C even though the operating temperature during the
demonstration test was maintained at 370°C. The range specified is based on
the results of the optimization tests which show little effect of operating
temperature on N0 reduction efficiency in this reactor. Since the reactor
can operate over a range of temperatures without impacting NO reduction,
the conceptual design does not include requirements for flue gas temperature
control at the reactor inlet.
5.2.4 Conceptual Design Results
The preceding discussion has summarized the data used and the limitations -
observed in preparing the conceptual design of a 500 MW HZ process. The
discussion which follows summarizes the results of the design calculations
and examines results in detail.
Table 5—5 summarizes the results of the conceptual design calculations.
As shown, key design variable levels are presented for the SCR reactor and
the downstream air preheater. The design levels presented for the reactor are
77
-------
based on the information outlined in the preceding discussion and the basis
defined in Section 5.1, while the air preheater design parameters are based
on the results of a study on animonium sulfate formation in air preheaters. 3
Design parameters for the air preheater are presented because the reactor design
results in high concentration of both NH 3 and SO 3 at the reactor outlet.
For a conventional air preheater design, the expected NH 3 and SO 3 concentrations
at the reactor outlet cause severe corrosion and plugging problems. As a
result, modification of the air preheater design is required to minimize those
problems. The following discussion examines the reactor and air preheater -
design parameter levels specified in the conceptual design.
Reactor Design Parameters
The basis for the reactor design has been outlined in this -Section. The.
number of reactors was selected to be identical to the design used in TVA’s
preliminary economic analysis of flue gas treatment processes. Reactor cros—
section was determined by dividing the flue gas flowrate for the 500 MW plant
by the flowrate to the pilot unit and multiplying by the c ossectional area
of the pilot plant reactor. Catalyst volume was obtained by multiplying the
crossectional area by the catalyst depth of the pilot unit (2.0 in).
Reactor pressure drop was 0.28 kPa during the pilot scale tests and was
assumed to be identical for the 500 MW design. An additional 1.0 kPa pressure
drop was assumed for ducts and flow distributors in the full scale design.
Only a single soot blower was used on the pilot plant while four soot blowers
were included in the conceptual design. The additional soa t blowets- were --
included in the conceptual design to reflect current practice in Japan where
a soot blower is placed between each layer of catalyst. (The pilot plant and
the conceptual design both have four layers of catalyst blocks with each block
0.5 in on a side). The soot blowing frequency specified for the conceptual
design is 3 times per day and this reflects actual operating experience at
the HZ pilot plant. -
78
-------
Air Preheater Design Parameters
The expected concentration of NH 3 and SO 3 at the reactor outlet are
85 ppm and 65 ppm respectively. An investigatfon of ammonium sulfates formation
in air preheaters has shown that these concentrations will result in severe
corrosion and plugging of the preheater. This results in the requirement for
corrective measures to minimize problems in the air preheater. The same
investigation of air preheater problems concluded that then optimum solution to
the problem is to employ available air preheater design options designed to -
minimize corrosion and plugging. These air preheater design options were
presented in Table 5—5.
TABLE 5—5. RESULTS OF THE CONCEPTUAL DESIGN FOR A 500 MW.
HITACHI ZOSEN PROCESS
Design Parameter Design Level
Reactor Design Parameters
• Number of Reactors 2
• Reactor Crossection (m 2 ) 96.5
Catalyst Volume per reactor (in 3 ) 205
• Reactor System Pressure Drop (kPa) 1.28
• Soot Blowers per Reactor 4
• Soot Blowing Frequency 3/day
Air Preheater Design Parameters
• Soot Blowers per Preheater 6 -
• Soot Blowing Frequency - 6/day -
• Element Configuration Combined Intermediate
and Low Temperature Zone
Element Construction Corrosion Resistant Material
in Intermediate—Low Temp Zone
79
-------
Use of additional soot blowers and increased soot blowing frequency is
intended to aid in cleaning deposits from the preheater. The combined heat -
transfer elements are also designed to aid in cleaning deposits from the
preheater by making the soot blower more effective in the intermediate zone
of the preheater where ammonium sulfates tend to form. Corrosion resistant
material is used to minimize corrosion due to the formation of liquid
ammoniuni bisulfate, the product expected downstream of the 500 I’IW application
of the HZ process.
It should be noted that these air preheater design options impact the
cost estimate for the HZ process and are included in the costs presented
later in this section. However, it should also be noted that these costs -
may not adequately reflect the costs which will result from the high N}1 3 and
SO 3 emissions estimated for -the HZ process. The emission rates estimated
for the conceptual design are much higher than any considered in the study
of air preheater problems, and therefore the use of available air preheater
design options may not solve the problems of ammonium sulfate plugging and
corrosion.
5.3 MATERIAL BALANCE FOR A 500 MW HITACHI ZOSEN PROCESS
The first step in completing the HZ process evaluation was preparation - -
of a conceptual design. Once this design was completed it was then possible
to complete material balance calculations which define raw material require-
ments for a 500 MW HZ system. This was done using TVA’s material balance as
a basis and the data generated during the pilot plant tests and secondary - -
emissions sampling to modify that material balance. Where appropriate, ratios
of modified stream flows to TVA’s original flows were made td determine new
flowrates (e.g., the quantity of air used to dilute/inject NH 3 ).
Figure 5—1 is a schematic of the HZ process which indicates process
streams and stream numbers considered in preparing the material balance. This
schematic and the stream numbers shown are consistent with those used by TVA.
80
-------
Coal
To Stack
Air
MU 3
Soot blowing
Figure 5—i. Hitachi Zosen proces fiowsheet.
-------
There are, however, two additional streams shown in Figure 5—1. These repre-
sent the compressed air used for soot blowing the reactor (stream 15) and the
additional steam required for soot blowing the modified air preheater (stream 16).
The results of the material balance calculations based on the pilot plant
data are summarized in Table 5—6. These represent the mass and volumetric
flows to all significant pieces of process equipment in the 500 MW application
of the HZ process. These flowrates were estimated based onthe data and
assumptions outlined in the preceding discussion. The results of the materiaL
balance calculations serve as a basis for the energy balance calculations which
are sunmiarized in the following section. These flows also serve as a basis
for estimating the size and consequent cost of process equipment for the
500 MW application of the HZ process.
In comparing the results of the material balance calculations presented
in Table 5—6 with those developed by TVA, there are no major differences in
material flowrates. The most significant difference is the estimated NH 3 and
SO 3 emissions from the HZ reactor and the consequent requirement for additional
soot blowing of the air preheater. The only other difference is the requirement
for compressed air for reactor soot blowing. TVA’s design did not include
reactor soot blowing, but it was included in this evaluation since soot blowing
was a routi-ne part of the pilot plant’s operation. These differences and the
-requirement for air preheater modifications result in a slight increase in
HZ process costs which are discussed in Section 5.5.
5.4 ENERGY REQUIREMENTS FOR A 500 MW HITACHI ZOSEN PROCESS
The material balance presented the mass flowrates of various process
streams in the HZ process. With this as a basis, it was possible tocomplete
an energy balance and to determine the energy requirements for a 500 MW
application of the process. Determination of the energy requirements involved
consideration of several factors including:
82
-------
TABLE 5—6. MATERIAL
BALANCE FOR 500 MW HI
TACHI
Stream Number
1
2
3
4
Description
Coal to Boiler
Combustion Air
to Air Heater
Combustion Air
to Boiler
Flue Gas to
Economizer
Mass Flow (lb/hr)
428,600
4,546,200
4,101,800
4,516,100
Volume Flow (scfm)
•
1,005,000
906,700
957,000
Stream Number
5
6
7
8
Description
Flue Gas to
Reactor
Flue Gas — N H 3
Mixture to Reactor
Flue Gas to
Air Heater
Flue Gas to
ESP
Mass Flow (lb/hr)
4,516,100
4,570,800
4,570,800
5,015,200
Volume Flow (scfm)
957,000
969,200
969,200
1,066,700
Stream Number
9
10
Il
12
Description
Gas to
FGD Unit
.
Treated Flue
Gas to Stack
NH 3 Air Mixture
Injected Into
Flue Gas
NH 3 From
Storage
Mass Flow (lb/hr)
4,960,700
5,160,000
54,700
1,650
Volume Flow (scfm)
1,066,700
1,139,900
12,200
—
Stream Number
13
14
15
16
Description
Steam to
Nil 3 Vaporizer
.
Fly Ash from
ESP for
..
Compressed Air
Reactor Soot Blowing
Steam for Air
Heater Soot
Blowing
Mass Flow (lb/hr)
Volume Flow (scfm)
1,011
355
54,600
—
2,530
555
7,700
2,700
Note: For ease of comparison with TVA’s results, the values in this tabl .e are presented in
English units.
co
-------
• Evaluation of any heat credits associated with the
HZ process,
• Determining the enthalpy of all streams entering and
exiting the process; and - -
• Determining the electrical energy required for fans and
compressors. -
Once each of these factors had been evaluated it was possible to determine
the overall energy demands of the HZ process.
The following discussion examines each of the above factors in detail
and summarizes the estimated energy requirements of the HZ process.
5.4.1 Hitachi Zosen Process Heat Credits
Because the HZ process is located between the economizer and air pre-
heater of a boiler, the potential exists to recover heat added to the flue
gas by the process. In the HZ process, the NO reduction reactions ancL the
NH 3 oxidation reaction are exothermic and so it is possible to recover the
heat of these reactions in the air preheater. Table 5—7 presents the reactions
which can occur in an HZ reactor. This table also identifies the heats of
reaction and, based on material balance calculations, the quantity of heat
which is added to the flue gas due to these reactions. As shown, no heat is
added to the flue gas as a result of NH 3 oxidation. This is due to the fact
that, within the accuracy of the process measurements, the pilot plant test
results indicated that no NH3 oxidation occurred in the reictor.
TABLE 5-7. REACTIONS WHICH CAN ADD HEAT TO FLUE GAS
IN THE HITACHI ZOSEN PROCESS
Reaction
400°C
Heat Added to Flue Gas
(Gcal/hr)
4N0 + 4NH 3
+ 02 ÷ 4N 2 + 6H 2 O
—96.98
kcal/gmol
NH 3
3.68
4NO 2 + 4NH 3
+ 02 - - 3N 2 + GH 2 O
—39.3LL
kcal/gniol
NH 3
0.16
4NH 3 + 302 - 2N 2 + 6H 2 0
—75.3
kcal/gmol
NH 3
0
84
-------
The heat-of the NO reduction reactions results in about a 7°C temperature
rise across the HZ reactor. This heat credit is offset to some extent by a
heat debit which results from the assumption that the NH 3 —air mixture is injec-
ted into the flue gas at 38°C. This heat debit is estimated to be approxi-
mately 0.7 Gcal/hr so the net heat credit for the HZ process is 3.15 Gcal/hr.
5.4.2 Energy Balance for the 500 NW Hitachi Zosen Process
Once the HZ heat credits were evaluated it was possible to use that
information in conjunction with the results of the material balance to complete
an energy balance for the HZ process. This energy balance considered all the
major streams entering and exiting the process. In addition, the heat credit
discussed in the preceding section was included. It should be noted that the
energy balance does not represent a comprehensive evaluation of the energy - -
requirements of the process since it does not include electrical energy require-
ments. This is considered later in this section. -
The results of the energy balance calculations for the 500 MW application
of the HZ process are suimnarized in Table 5—8. The energy associated with
major inputs and outputs are shown for major process streams. These energy
balance results include the heat credits and energy debit discussed in the
preceding section.
TABLE 5—8. SUMMARY OF ENERGY BALANCE CALCULATIONS FOR
500 MW APPLICATION OF THE HZ PROCESS
Enthalpy -
Stream Description (Gcal/hr)
Gas Streams to Process
Flue Gas to Blower 191.50
NH 3 —Air Mixture 0.23
Combustion Air to Air Heater 13.75
Gas Streams from Process
Flue Gas to ESP 84.68
Combustion Air to Boiler 133.13
Steam to Process
Additional Steam for Air Preheater Soot Blowing 2.90
Steam to NI-i 3 Vaporizer 0.30
85
-------
3 4.3 Electrical Energy Requirements of the 500 ‘1W HZ Process
The component of the overall energy requirements of the HZ process which
had not yet been defined is the electrical energy requirement. Electrical
energy is required by the process to operate fans and compressors. In con-
junction with process steam requirements and the net process heat credit,
the electrical energy requirements can be used to define the overall energy
requirement for the 500 MW application of the HZ -process. -
Electrical energy requirements were estimated for the fans and compressors
used in the HZ process. This estimate was prepared by determining the theore-
tical work required by each item of equipment and assuming an efficiency of
70 percent for conversion of electrical energy to mechanical wo k.
The estimated electrical energy requirements for a 500 MW application
of the HZ process are summarized in Table 5—9. As shown, the single largest
energy requirement is that associated with the main flue gas blower. This
accounts for nearly 90 percent of the total electrical energy. It sho xld be
noted that the flue gas fan energy requirement represents only the incremental
power required to overcome the pressure drop associated with the SCR system.
TABLE 5-9. ESTIMATED ELECTRICAL ENERGY REQUIREMENTS FOR A
500 MW APPLICATION OF ThE HITACHI ZOSEN PROCESS
Process Equipment
Item -
Theoretical
Requirement
Energy
(kW)
Est
Ener
.
imated Actual
gy Requirement
(kW)
N I ! 3 Compressor
84
120 -
Reactor Soot Blower
Compressor’
49
70
NH 3 Injection Blower
7
10
Flue Gas Blower 2
939
1341
Total
1079
1541
1 This represents average energy demand. The energy requirement is much
greater during soot blowing of the reactor.
2 Thjs represents incremental power required by the flue gas blower.
86
-------
5.4.4 Overall Energy Requirements for the 500 MW Application
of the Hitachi Zosen Process
The preceding discussion has defined the basis for and presented the
results of calculations which estimate the individual components of the energy
requirements for a 500 MW application of the HZ process. This section
summarizes these results and puts the individual components of the process
energy requirements on a common basis.
Table 5—10 presents the individual components and the estimated àvèrall
energy requirements for a 500 MW application of the HZ process. Each of
these components has been put on the basis of heat input to the boiler. For
steam, a thermal efficiency of 88 percent was used to deteri4r the energy
input required to generate one Ccal of steam energy. For electritity, a boiler
heat rate of 2.27 Mcal/kWb was used. The heat credit was assumed to replace
heat input to the boiler on a 1—to—i basis. -
As shown in Table 5—10, the overall energy requirement for the 500MW
application of the HZ process is about 4.0 Gcal/hr which represents only
0.33 percent of the boiler’s capacity. Both the steam and electrical energy
requirements are similar and the heat credit results in about a 50 percent
reduction in process energy requirements. -
TABLE 5-10. OVERALL ENERGY REQUIREMENT FOR A 500 MW
APPLICATION OF THE l IZ PROCESS
Energy Area
Energy Requirement
(Gcalfhr)
Percent of Boiler
Capacit y
Heat Credit
(3.15)
.
(0.28)
Steam
3.36
0.30
Electricity
3.50
0.31
Total
3.99
0.33
87
-------
5.5 ESTIMATED COSTS FOR A 500 MW APPLICATION OF THE HITACHI ZOSEN PROCESS
The final step-in the evaluation of the HZ process was to prepare a
modified cost estimate for a 500 MW application of the process. This cost
estimate includes an estimate of both capital costs and annual revenue require-
ments. It was prepared to show how the results of the pilot plant tests affec-
ted the economics of the process. Because the estimate used the same basis
as TVA’s preliminary economic- analysis, the results can be used to compare HZ
process costs with the costs of other SCR processes. It should be noted,
however, that the comparison is only for a specific application of the HZ
process and the results may not be valid for other applications.
The procedure used in preparing the estimate of capital-’costs and anñual
revenue requirements was to-first define equipment costs for the process. - -
This was done using TVA’s original equipment cost list and standard exponential
capacity factors for chemical process equipment. Installation costs for piping,
insulation, foundations, etc. were estimated as a percentage of the equipment -
costs. Once the installed equipment costs were defined, the total capital costs
and annual revenue requirements were estimated. The estimates of these costs
used factors developed by TVA along with the results of the conceptual design
and material and energy balances developed as part of this study.
An additional item included in the cost estimate was the cost of air pre-
heater modifications needed to minimize the impact of NH 3 and SO 3 emissions
from the process. These costs were taken from a study conducted by Radian
Corporation entitled “Ammonium Sulfate and- Bisulfate Formation in Air Pre— -
heaters”. These costs were based on a quote supplied by C—E Air Preheater.
The following discussion presents the results of the estimates of capital-
costs and annual revenue requirements for the HZ process. It includes a brief
discussion of how the estimates were derived and a comparison of the modified
88
-------
HZ cost estimate with the costs of other SCR processes as determined by
TVA. A detailed discussion of the economic premises used to prepare the cost
estimates along with an estimate of individual equipment item costs is pre-
sented in Appendix D of this report.
5.5.1 Estimated Total Capital Investment for a 500 MW Application
of the HZ Process
Procedure for Estimating Capital Investment
The capital cost estimate for the HZ process was based primarily on
the costs of process equipment. These costs were estimated far individual-
items of process equipment using the equipment costs developed by TVA and
Radian as a basis. Modified equipment costs were determined by comparing
process flowrates for individual equipment items and using cOst—size expc nents
which reflect equipment cost—size relationships typical of the chemical process
industry. In the case of the catalyst, it was assumed that the cost of the
NOXNON 600 catalyst is equal to that of the NOXNON 500 on a volume basis.
It was also assumed that no economy of scale was realized for the catalyst
(i.e., the cost per unit volume of catalyst is constant). Using this
procedure, the costs for individual equipment items were determined. These
costs should accurately reflect modified costs since, in nearly all cases,
the size did not vary significantly from the equipment sizes in the original-
cost estimate. These costs are representative of a Midwestern power plant
location and mid—1979 dollars. The costs were not escalated-to 1981 dollars --
in order to facilitate comparison with estimated costs for other SCR
processes.
Once equipment costs were determined, the various installation expenses -
such as piping, foundations, etc., were estimated as a percentage of equipment
costs. The sum of the process equipment costs and installation expenses con-
stituted the major components of the direct investment. The final component
89
-------
of the total direct investment includes a services and miscellaneous cost
which was calculated as a percentage of the installed equipment cost. The sum
of the installed equipment cost and the services and miscellaneous cost is the
total direct investment.
The total indirect investment (the sum of various indirect investments,
such as engineering design and supervision, A&E contractor, construction
expense, and contractor fees) was calculated based on the total direct invest-
ment using formulas described in Appendix D of this report. A contingency,
calculated as 20% of the sum of the total direct and total indirect investment,
was also included. The sum of the total direct investment, the total indirect
investment, and the contingency equals the total fixed investment.
Other capital charges included in the cost estimates consist of an allow-
ance for startup and modification, interest during construction, land costs,
working capital. The allowance for startup and modification and interest
during construction were calculated as a percentage of the total fixed —invest— -
menc, the previously calculated allowance for startup and modification, and -
interest during construction. The sum of total depreciable investment, land
costs, and working capital is the total capital investment.
Results of Capital Cost Estimate
Table 5—11 presents the individual components and the estimated total
capital investment for a 500 MW application of the HZ procéss As.shown, the--
total capital investment was estimated to be approximately $22.1 x 106 which is-
equivalent to $44/kW of generating capacity. This compares to TVA t s previous
estimate of $23.3 x lO and it represents a slight decrease in total-capital
investment. The principal difference between the two estimates is the lower
catalyst volume required for the design based on the pilot plant results. This
lower catalyst volume is a result of the greater space velocities in the pilot
90
-------
TABLE 5—11. ESTIMATED CAPITAL INVEST NT FOR A 500 MW APPLICATION
OF THE HITACHI ZOSEN pRocEssa
7. of
total direct
b - Investment, $ investment
Direct Investment
N}{3 storage and injection 645,000 5.5
Reactor section 8,632,000 73.4
Gas handling 351,000 3.0
Air preheater modifications 1,461,000 12.4
Sub—total diiect investment (DI) 11,089,000 94.3
Services, utilities (0.06 x DI) 665,000 - 5.7
Total direct investment (TM) u,754, 00 0 100.0
Indirect Investment -
Engineering design and supervision 274,000 2.3__
Architect and engineering contractor 69,000 0.6
Construction expense
= 0.25 (TM x 10—6)0.83 1,933,000
Contractor fees = 0.096 (TDI x 10 )0 .76 625,000 5.3
Total indirect investment (IDI) 2,901,000 24.7
Contingency 0.2 (TDI + IDI) 2,931,000 24.9
Total fixed investment (TFI) 17,586,000 149 ô.
Other Capital Charges
Allowance for startup and modifications
(0.1) (Tn) 1,759,000 15.0
Interest during construction
= (0.12) (TFI) 2,110,000 17.9
Total depreciable investment 21,455,000 162.5
Land 5,000 —
Working capital 336,000 2.9
Royalty fee 300,000 2.6
TOTAL CAPITAL INVESTMENT 22,096,000 168.0
5 Basis: 500 MW new coal—fired power plant, 3.5% sulfur coal, 90% NO 5 removal.
Midwest plant location. Represents project beginning mid—1977, ending mid—1980.
Average basis for scaling, mid—1979. Investment requirements for fly ash
disposal excluded, Construction labor shortages with overtime..pa 1 _iacentiVe
not considered.
b Ch item of direct investment includes total equipment costs plus installation
labor, and material Costs for electrical, piping, ductwork, foundations,
structural, instrumentation, insulation, and site preparation.
91
-------
plant tests compared to those used by TVA in preparing the preliminary cost
estimate. To some extent, the lower costs of catalyst are offset by the
capital costs for the air preheater modifications. The net effect is that
essentially no change in the process capital costs is expected based on the
pilot plant test results.
5.5.2 Estimated Average Annual Revenue Requirements for a 500 MW
Application of the HZ Process - -
Procedure for Estimating Annual Revenue Requirements
The annual revenue requirements were calculated based ona Midwestern.
power plant location and mid—1980 costs. These average annual revenue
requirements were divided into the direct costs, which consists of raw
materials costs and conversion costs, and indirect costs, which consist of
capital charges and overheads.
In calculating direct costs, the raw material, utility, and labor unit
costs used were those developed by WA while the required quantities of these
cost items were based on the conceptual design and the material and energy
balance results. The maintenance cost was calculated as a percentage of total
direct investment.
- The capital charges portion of the indirect costs consists of:
(1) depreciation, interim replacements, and insurance, whiGhwere estimated
based on the total depreciable investment; and (2) the average costs of capital
and taxes, which were determined based on the total capital investment. The
overhead charges consist of: (1) plant overhead, which was estimated as a
percentage of the cost of operating labor and supervision, maintenance, and
analyses; and (2) administrative overheads, which were estimated based on a
92
-------
percentage of operating labor and supervision. Specific details of how each
of the components of the annual revenue requirements was estimated are pre-
sented in Appendix 0.
The annual revenue-requirements were determined as the sum of the direct
and the indirect costs. Equivalent unit revenue requirements in mills/kWh
were then obtained by dividing by the on—stream time of 7000 hr/yr and the
plant MW rating.
Results of the Annual Revenue Requirement Estimate
Table 5—12 presents the individual components and the total estimated
average annual revenue requirements for a 500 MW application of t he HZ -
process. As shown, the average annual revenue requirement was estimated to -
be approximately $10.3 x 10 6 /yr which is equivalent to 2.91 mills/kWh. This
compares to TVA’s previous estimate of $12.2 x 10 6 /yr and represents nearly
a 1.6 percent decrease in the annual revenue requirements for the process.
As with the capital costs, the principal factor which changed the annual
revenue requirements is the decreased quantity of catalyst required in the
reactor. Again, the lower catalyst requirements are partially offset by the
increased costs associated with the air preheater modifications (i.e., in-
creased soot blowing).
5.5.3 A Comparison of Capital Investment and Annual Revenue
Requirements of the HZ Process with the SFGT Process -.
The capital investment and annual revenue requirements of the HZ process
have been estimated based on the results of the test conducted at the EPA
sponsored pilot unit in Albany, Georgia. The results of these cost estimates
indicate that the capital costs and annual revenue requirements are slightly
lower than the estimated costs prior to the test program. A more significant
indication of how the pilot plant results affect HZ process costs is a comparison
of HZ process costs to the costs of the other SCR process tested- during the EPA
NO pilot plant program, the shell flue gas treating (SFGT) process.
93
-------
TABLE 5-12.
ESTiMATED AVERAGE ANNUAL REVENUE REQUIREMENTS FOR
A 500 MW APPLICATION OF THE HITACHI ZOSEN PROCESSa
Annual
Unit
Annual
1
of cnu ’i.il
item quantity
cost($)
coat($)
revenue required
Direct Costs
Raw mat-ettais
Nil 3
Catalyst
Total raw materials
Conversion costs
Operating labor and supervision
8760
labor bra.
0.165/kg 866,300
5,125.000
5,991.300
12.50/
labor hr.
Iltilitles
Steam
Electricity
heat credit
Maintenance .04 x TD1
Analyses
Total Converbion coats
Total direct Costa
Indirect Costa
Capital charges
Depreciation (0.06) (total
depreciable Investment)
Average cost of capital = (0.086) x
(total capital investment)
Overheads
Plant (0.5) (conversion costs
minus utilities)
Administrative = (0.1)
(operating labor costs)
20,700 Ccal
10,787,000 kWh
22,050 Ccal
2,920
labor hrs.
7.94/Cca I
0.029/kwh
—7.94/Gcal
17.00/
labor 1w.
agaajs: 500 MW new coal—fired power plant, 3.5% S coal. 90 percent NO 5 reduction, 90 percent SO 2 removal. Midwest
power plant location, 1980 revenue requirements. Remaining life di power plant 30 years. P1 nt on line
7000 hr/yr. Plant heat rate equals 9.5 NJ/kWh. investment and revenue requirement for disposal of fly
ash excluded. Total direct investment $11,754,000; total depreciable investment $21,455,000, and total
capital investment $22,096,000.
5.25 x IO 6 kg
6.47
50 14
58 51
109,500 1 07
164,400
312,800
(175,100)
470,200
49,600
93].. 400
6, 922 700
l,287,30Q
1,900,30 11
314,700
11,000
3,513,300
(214,000)
Total indirect costs
Spent catalyst’ disposal
1.61
3.06
(1. 71)
4.60
o 48
9 11
67.72
12.59
18.59
3.08
0.11
34.37
(2 09)
100 00
‘L0T L ANNUAL REVENUE REQUIREMENTS 10,222,000
-------
Since the same basis was used in preparing the modified HZ cost
estimate as TVA used in preparing preliminary economic estimates for other.
SCR processes, it is possible to make a direct comparison using TVA’s
previously published results to account for process differences. Table
5—13 shows the estimated capital investment for two pollutioncontrol -
systems which reduce emissions of particulates, NOR, and SO < by 99.5, 90,
and 90 percent respectively. As shown, the three pollution control systems
employ two different SCR process, one-being the HZ process and the other,-
the SFGT process which also removes SO 2 from the gas. The HZ process cost
is added to the cost of a flue gas desulfurization system and both SCR
systems include ESP’s located downstream in order to put the cost estimates
on a coumion basis.
TABLE 5—13. ESTIMATED CAPITAL INVESTMENT FOR THREE
POLLUTION CONTROL SYSTEMS’
SCR
Process
Total
Capital
Investment
($
x
1O )
-
SCR
FGD
ESP
Overall
98.8
Shell
Flue
Gas
Treatment
84.2
—
14.6
Hitachi Zosen
22.1
50.4
10.8
83.3
‘All costs except the HZ—SCR and the Shell Flue Gas Treatment (SFGT)—SCR -
cost are from “Preliminary Economic Analysis of NO Flue Gas Treatment
Processes”. Tennessee Valley Authority — Office o Power, EPA — 600/7—80—021,
February, 1980. Costs for the SFGT—SCR estimated in “Independent Evaluation
of the Shell Flue Gas Process”, Radian Corporation — Draft Final Report.
EPA Contract No. 68—02—3171, Task 11.
As shown on Table 5—13, the HZ process requires lower capital
investment, having a capital cost nearly 20 percent lower than the SFGT
process. This indicates that the HZ process represents the more economic
SCR system for this particular 500 MW application within the accuracy of
the cost estimates.
95
-------
Table 5—14 presents the annual revenue requirements for the two
pollution control systems. As shown, the estimated costs associated with
the HZ processes are 30 percent lower than those of the SFGT process.
The results presented in Table 5—14 indicate that the HZ process, as
tested in the pilot plant and presented in the conceptual design, is
more economical than the SFCT process for the 500 MW application examined
in this study. It should be noted that the relative costs presented in
Table 5—14 are only valid for one specific application and they could
change for other applications.
TABLE 5-14. ESTIMATED ANNUAL REVENUE REQUIREMENTS FOR
THREE POLLUTION CONTROL SYSTEMS’ -
Annual Revenue Requirements ($ x 10 )
SCR Process
SCR FGD ESP Overall
SFGT
33.6
—
3.0
36.6
Hitachi Zosen
10.2
14.7
2.2
27.1
‘All costs except the HZ—SCR and the SFGT—SCR costs are from “Preliminary
Economic Analysis of NO Flue Gas Treatment Processes”. Tennessee Valley
Authority — Office of Power. EPA — 600/7—80—021, February 1980.
5.6 OVERALL EVALUATION OF THE HZ PROCESS
The preceding discussion has examined many factors which influence
the technical and economic feasibility of applying the HZprocess to a
coal—fired boiler. These factors include:
the pilot plant test results,
the results of Radian’s Independent Tests,
96
-------
• he results of the 500 MW conceptual design,
• the material balance calculations for the 500 1W
HZ process application,
• the energy balance calculations for the 500 MW HZ
process app1 cation, and
• the estimated capital investment and annual revenue
requirements for the 500 MW HZ process application.
In crder to put all these factors into perspective and thereby complete the -
evaluation of the HZ process, a summary of the most significant results was
prepared. This summary, which is presented in the following discussion,
identifies how the results of the tests and calculations influence the.feasi—
bility of applying the HZ process to a coal—fired power plant. In addition,
the summary identifies areas not addressed during the pilot plant tests.
5.6.1 Pilot Plant Test Results
The major objective of the pilot plant tests conducted by Chemico was -
to demonstrate 90 percent NO reduction during 90 days of continuous operation.
This objective was met and exceeded during the tests with the NOXNON 600 -
catalyst. The pilot plant operated for 90 days and averaged over 90 percent
NO reduction and over a 6 month period, NO reduction averaged 89.8 percent. -
- Some other significant results of the test program showed that neither
temperature nor.flowrate has as significant effect on NO reduction efficiency
within a range about the design level. - These results indicate that process
performance should not be impaired at boiler loads below the design level.
As a result, no temperature or flow control would be required for a full—scale
application of the HZ process.
During the test program, 3 charges of catalyst were examined and two of
these (NOXNON 500) experienced severe plugging problems after about 2000 hours
of operation. When replaced with the NOXNON 600 catalyst, which has larger gas
—97
-------
passages, no further plugging problems were observed. The original plugging
was believed to be due to the adhesiveness of the fly ash. At high ternperatures
fly ash sample collected from the power plant were found to agglomerate.
• Tests with the NOXNON 500 catalyst did not last long enough to get a- -
good measure of catalyst activity, but results of the NOXNON 600 tests showed
a gradual decline in catalyst activity with time. After 5500 hours of
operation, activity of the NOXNON had dropped &lightly, but-it was still
possible to achieve 90 percent N0 reduction. Since 5500 hours is close
to one year of operation ( 7000 hours) a catalyst life of 1 year seems reason-
able based on the test results. In fact, catalyst life may. be extended well
beyond 1 year based on the results of the in situ regeneration test con4ucted
at the conclusion of the test program. These tests showed the catalysr
activity had been restored to the activity of essentially new catalyst.
Unfortunately, since the regeneration test was cohducted during the final
week of the test program, it is uncertain how long the effects of regeneration
would last.
Overall, the results of the pilot plant tests indicate application of
the HZ process to a coal—fired boiler is technically feasible. The tests
demonstrated the ability of the process to achieve 90 percent N0 reduction
for over 90 days and also demonstrated a useful catalyst life of nearly one
year.
5.6.2 - Results of the Independent Evaluation Test Program
The independent evaluation test program conducted by Radian had two
primary objectives: to insure the quality of the data collected at the HZ
pilot plant, and to quantify changes in the concentration of certain pollutant-s
across the HZ reactor. As discussed earlier, the results of the quality
assurance audits and certification tests validated the process data collected
by Chemico. These results have no direct bearing on the overall feasibility
of the HZ process with one exception. As a result of the inability to monitor
NH 3 in the flue gas, the Mi 3 emission measurements, made as part of the QA
98
-------
audits, represent a significant portion of the data base on NH 3 emissions
from the process. These data, in conjunction with the results of the secondary
emissions sampling tests, quantify the secondary environmental impact associaied
with the HZ process as it was operated during the pilot unit tests.
The results of the secondary emissions sampling program indicated that
no measurable change occurred in the concentrations of trace elements in the
particulates passing through the reactor or in the concentrations of hydro-
carbons, GO, HCN, or nitroso mines in the flue gas. In the case of particu—
lates, the results show that no environmental problem will occur due to erosion
(if any) of the catalyst. In the case of HCN and nitrosoamines, the èoncen—
trations of these compounds were below the detection limit of the analytical
techniques employed and were also below the maximum safe con e’htrations for
emission sources. Again, this indicates that no environmental problem will -
occur due to production of these compounds.
Another result of the secondary emissions sampling program was that -
S03 is produced in the reactor as a result of SO 2 oxidation. This can cause
significant operational impacts in downstream equipment such as the air
preheater. The significance of the increase in flue gas SO 3 concentration -
across the reactor is magnified because of the high N i l 3 concentrations measured
at the outlet of the reactor. NH3 and S03 will react to form-atmnonium sulfates
in the air preheater which can result in plugging and corrosion. The severity
of the problem depends on the actual compound formed — ammonium bisulfate is
much more corrosive and more prone to cause plugging since it exists as a
liquid at air preheater temperatures while ammonium sulfate exists nly as
a solid.
The most significant factor in determining which ammonium compound will -
form is the NH 3 to SO 3 mole ratio. At NH 3 /S0 3 mole ratios below 2.0, some
liquid ammonium bisulfate will form since there is not enough NH 3 to completely
react with the SO 3 present. Because of this, the production of SO 3 in
the Hitachi Zosen reactor will promote ainmoniulu bisulfate formation and the
associated problems of plugging and corrosion. And the high levels of Nil 3 and
99
-------
SO 2 expected at the reactor outlet will result in the formation of significant
quantities of ammonium sulfates which are expected to cause severe problems
in the air preheater.
Any armnonium sulfates which do riot deposit in the preheater should solidify
and be collected in an ESP. Previous studies have indicated that environmental
problems can result from high N H 3 concentrations in fly ash depending on the
specifics of fly ash disposal techniques. However, for most applications,
no environmental problems would be expected.
One compound scheduled to be measured during the secondary emissions
sampling program was N 2 O. However, the analytical technique used to make
the N 2 0 measurements proved unsatisfactory. Therefore, no determination was
made on whether any significant quantity of N 2 0 was being produced in the - -
reactor. Any future development work on the HZ process should include pro-
visions for measurement of N 2 0. -
Overall, the independent evaluation tests indicated that:
no emissions problems resulted from application of
the HZ process with the exception of NH 3 and SOS,
• the levels of NH 3 and SO 3 emissions measured during
the pilot plant tests will cause severe operational
problems in the air preheater,
the estimated N i - I 3 emissions from the HZ process may
result in environmental problems depending on the
specific details of the fly ash disposal techniques.
However, for most applications, no problems are
expected, and
• it was not possible to determine the concentration of
N 2 O (if any) in the reactor outlet. This requires
further investigation.
-- 100_
-------
5.6.3 Res-ults of the 500 MW Conceptual Design
The conceptual design of the HZ process was prepared for a single
application of the process and it was based solely on the pilot plant test
results. The results of this design indicate that it is possible to reduce
NO emissions by 90 percent using the HZ process. In fact, 90 percent N0
reduction was possible at space velocities greater than previous estimates
indicated (i.e., at a relatively lower catalyst volume per unit volume of
flue gas treated). However, the greater space velocities were accompanied
by NH 3 emissions which were much higher than previous estimates.
One result of the high N}1 3 emissions estimated for the conceptual -
design was that special modifications to the air preheater a e r quired to
mitigate problems associated with the formation of an-unonium sulfates downstream
of the reactor. These modifications were identi-fied as partof a prior study
by Radian and they are based on Japanese experience with air preheater opera-
tion downstream of an SCR system. It should be noted that the modifications
specified for the air preheater were expected to minimize problems at
relatively low N i l 3 and SO 3 concentrations at the reactor exit. The concentra-
tions at the reactor outlet for the conceptual design are much higher than
anticipated in previous studies of SCR technology and this could result in
operational problems which cannot be minimized by the preheater modifications
included in the conceptual design. This represents an area which requires -
further investigation
Reactor pressure drop and other design parameters are fairly äonsistent -
with previous estimates for the process. The design results also show that
the process can operate over a range of temperatures (3400 to 410°C) and
space velocities (7,600 to 10,400 hr) without any significant effect on N0
reduction efficiency. This indicates the process has good flexibility in
processing flue gas under conditions of changing boiler load.
101
-------
In summary, the conceptual design indicates that the HZ process can
reduce N0 emissions by 90 percent. And, this N0 reduction efficiency can
be achieved at a lower catalyst volume per unit of flue gas treated than
previous estimates indicated. However, the lower catalyst volume of the
conceptual design is accompanied by a significantly higher NH 3 - emission -
rate which can result in severe operational problems in downstream equipment,
particularly the air preheater. Further work is required to determine if the
effects of these NH 3 emissions can be offset by the air preheater modifications
included in the conceptual design.
5.6.4 Results of the Material Balance Calculations
Material balance calculations for a 500 NW application of tiVe HZ process
were included as part of this study to identify raw material requ irements for
the process and to serve as a basis for the cost estimate which was presented
in Section 5.5. The material balance was based on the pilot plant and secondary
emissions sampling test results and thus reflects those results in the estimated
raw material requirements. The most significant results of the material
balance calculations include estimation of NH 3 requirements for N0 reduction,
NH 3 and SO 3 emissions from the process, and steam requirements for air
preheater soot blowing.
The NH 3 requirements for the process were estimated to be 1.0 mole of
NH 3 per mole of N0 in the flue gas entering the reactor. This requirement
was estimated based on the results of 5 months of pilot plant operation.
During that 5 month peTiod, the NU 3 /N0 in-jectiofl ratioaveraged 0 .98 while
the N0 reduction efficiency averaged 89.8 percent. With the NH3/NOx injection
ratio of 1.0, estimated NH 3 requirements for the pràcess decreased about
10 percent from previous estimates.
Estimates of NH 3 and SO 3 emissions from the HZ process were significantly
higher than previous estimates indicated. As discussed earlier, this results
in the requirements for air preheater modifications and additional soot blowing.
The requirement for additional soot blowing results in a sevenfold increase in
102
-------
HZ process steam requirements. However, this is not very significant from
a material balance standpoint, but it is important in terms of its effect -
on process energy requirements.
In summary, the material balance calculations showed no significant - -
change in raw material requirements for the HZ process. The most important
result was the estimated NH 3 and SO 3 emission rates which were significantly
higher than previous estimates indicated. -
5.6.5 Results of the Energy Balance Calculations
An energy balance was completed as part of the evaluation of the HZ
process. This energy balance defined overall process energyreqüirementS and
quantified the heat credits associated with the process. The results of the-
analysis of energy requirements indicated that the HZ process. hasa net
energy consumption equivalent to 0.3 percent of the energy input to the
boiler.
Energy requirements for the HZ process are about equally divided between
steam and electrical energy, each representing about 0.3 percent of the boiler’
capacity. The heat credit for the process is also about 0.3 percent of the
boiler’s capacity resulting in a 50 percent reduction in process energy requir
-rnents.
Overall, the energy requirements of the HZ process represent only a very
small fraction of the boiler’s capacity. Even without the heat crédit, these
requiremen -ts da not significantly influence process costs. -
5.6.6 Results of the Cost Estimate for the 500 MW Hitachi Zosen Pro ess
Application
An estimate óf total capital investment and annual revenue requirements
for a 500 MW application of the HZ process was prepared as part of this
evaluation. The estimated costs reflect the results of the pilot plant tests.
103
-------
When compared with the previous estimate prepared by TVA, the modified cost
estimates indicate the magnitude of the impact the pilot plant results had
on estimated process costs. In addition, comparison of the modified cost
estimate with a cost estimate for the SFGT process is a relative indicator
of the cost effectiveness of the HZ process as tested in the pilot plant. -
The capital costs for the HZ process were estimated to be $22.1 x 106
which represents a slight decrease compared to .TVA’s preliminary estimate.
This decrease is due to the reduced catalyst volwne required for the conceptual
design as compared to TVAtS design. It should be noted that the change in
capital costs is not significant when the accuracy of the estimates is
considered. Also, the costs which are associated with the high NH3 and SO 3
emissions from the HZ process may be considerably higher than the costs inclu-
ded in the estimate based on the pilot plant test results.
The estimated annual revenue requirements for the HZ p rocess were $10.2
x lO 6 fyr which again represent a slight decrease compared to TVA’s preliminary
estimate. This decrease is also due to the reduction in catalyst volume for -
the conceptual design based on the pilot plant test results. Another major
factor in determining the annual revenue requirements for the HZ process is
the useful catalyst life. The cost estimate prepared as part of this study
assumed a 1—year life, but this was not demonstrated during thTe test program.
Table 5—15 presents estimated annual revenue requirements as a function
of catalyst life. As shown, a 6—month catalyst life results in about a 50
percent increase in costs compared to the-base case (1—year li.fei ,hi1e a -
2—year life results in a 25 percent decrease in costs. Obviously, if a 2—year
catalyst life could be demonstrated, a significant eduction -in annual revenue
requirements would result. It should be noted that the 6—month catalyst life
is just an example and that Hitachi Zosen will guarantee a 1—year life for
coal—fired process applications.
104
-------
TABLE 5-15. ESTIMATED ANNUAL REVENUE REQUIREMENTS AS A
FUNCTION OF CATALYST LIFE
Catalyst
Life
Annual Rev
Requi ements
enue
$ x 10
Base Case
(1
year)
10.3
.
6 months
15.6
2 years
7.7
Overall the results of the modified cost estimate indicate that, for the
particular application examined in this study, the HZ process-is economically
competitive with other SCR processes. This is based on a conceptual design
which was representative of operating conditions demonstrated dui ing t r
pilot plant tests. It should be noted, however, that the costs can be
affected by the impacts of high NH 3 and SO 3 emissions whose effects were not
examined during the pilot plant tests. Additionally-, the est imates presented -
in this evaluation were based on a 1—year catalyst life which was not demonstra-
ted. Because of these two factors, HZ process costs could vary significantly -
from those estimated in this study. Further development work should focus
on defining the useful catalyst life and on demonstrating techniques for
minimizing the operational impacts of high NH3 and SO 3 emissions.
5.6.7 Process Evaluation Summary
- The preceding discussion has identified the factors which -influence the
technical and economic feasibility of the HZ process along-wi-th soxse of thìe
limitations of this evaluation. This section summarizes the most significant -
aspects of the evaluation and presents an overall conclusion concerning the
technical and economic feasibility of applying the HZ process to a coal—
fired power plant.
The most significant results of the pilot plant test program and this
evaluation are:
105
-------
• The HZ process can reduce NO emissions by 90 percent when
applied to a coal—fired boiler. - This level of emissions
reduction was achieved over a 90 day period at an NH 3 /N0
injection ratio of 1.0 and space velocities greater than’
previous test work indicated. However, the excellent
performance of this pilot plant was ac ’companied by NH 3
emissions which were much higher than previous estimates
indicated.
The initial tests of the HZ process experienced problems
with catalyst plugging which resulted in failure of 2
charges of NOXNON 500 serfes catalyst. - These prdblems
were eliminated by employing NOXNON 600 series catalyst
(a catalyst with larger gas passages) and the use of
compressed air (as opposed to superheated steam) for
reactor soot blowing. Using the NOXNON 600 catalyst,
about 5500 hours of operation were logged withou signs
of catalyst plugging.
A gradual decline in catalyst activity was recorded over
the duration of the test program which resulted in the
requirement f or increased NH3/N0 injection ratios, to
attain 90 percent NO reduction. Because- the test
program was terminated after 5500 hours of operation, the
catalyst activity after 1 year of operation could not be
determined. Determination of the useful catalyst 1 if a is
important in defining the capital costs into annual
revenue requirements for the HZ process. -
A novel, in situ catalyst regeneration technique was
tested as part of the program. This test showed that
the regenerated catalyst had activity similar to fresh
catalyst and thus reversed some of the decline in
activity observed over the duration of the test program.
Unfortunately, the catalyst regeneration technique was
tested toward the end of the pilot plant test program -
- and so it is uncertain how long the effects of the
catalSrst regeneration will last. Obviously the catalyst -
regeneratior technique has the otential to significantly
reduce process costs and this merits further investigatlot
The independent evaluation test program indicated
that emission rates of most pollutants were not affected
by the HZ process. However, emission rates of both NH 3
and SO 3 will be relatively high and those emissions can
result in severe operational problems in downstream
equipment. The severity of any problems is very
site specific and could not be assessed as part
of this study. This should, however, be given
careful consideration in any planned applicatipns
of the HZ process. -
106
-------
The conceptual design and material balance calculations
indicated high NH 3 emission rates which will cause
severe operational problems in the air preheater located
downstream of the HZ process. The conceptua-l design
included air preheater modifications designed to minimize
those problems. But because the estimated NH 3 and SO 3
emission rates are much higher than previous estitnates,
it is uncertain if the air preheater modifications
will prove adequate. Further investigation in this
area is required.
• The overall energy requirements for the HZ process -
were estimated to be 0.3 percent of the boiler’s
capacity. This is a very small fraction of boiler
capacity and it does not significantly affect process
costs.
• The estimated capital investment and annual revenue
requirements for the HZ process were slightly lower then
TVA’s preliminary estimate. This indicates that the
HZ process is economically competitive-with other SCR
processes when considered for operation to a coal—fired
boiler. It should be noted that the cost estimates
assumed a 1—year catalyst life which was not quite
demonstrated during the pilot plant tests. The relative
process costs could change if a 1—year catalyst life is
not possible. - -
In conclusion, the pilot plant test results indicate the HZ process
is technically suited for application to coal—fired sources. - However, the
tests did not demonstrate a 1—year catalyst life which is generally considered
aminitnum for technical feasibility of an SCR process. In reality, a shorter
catalysr life would translate into increased annual revenue requirements.
In terms of costs, under the conditions of the cost estimate preparea as partr
ofthis study, the HZ process is economically competitive with other SCR
processes.
107
-------
SECTION 6
PROCESS DEVELOPMENT REQUIREMENTS
The pilot plant test results and the evaluation completed as part of this
study have defined many of the factors which affect the technical and economic
feasibility of applying the HZ process to a coal—fired source. However, due
to limitations of the pilot plant test program, there are several key factors
which were not examined. These factors represent process development ieqi.iire—
merits of varying priorities; some must be examined in further test work to -
establish the technical or economic feasibility of the process and some
simply represent methods of improving process performance or lowering process
costs. This section presents a discussion of the process development require-
ments identified during the course of the independent evaluation program. It -
also defines priorities for further investigation of those requirements. -
The process development requirements identified during the course of this
study include:
• Determination of catalyst life for coal fired applications
- of the HZ process and investigation of catalyst regeneration.
• Investigation of measures to minimize the impa tsresu1t-iflg
from formation of ammonium sulfates downstream of the reactor --
• Quantification of N 2 0 emissions (if any) from the process.
• Demonstration of an automatic NH 3 injection control system
which includes a continuous NH 3 analyzer.
Each of these development requirements represents an area not addressed during
the pilot plant tests and therefore an unanswered question concerning the HZ
process. The following discussion examines each requirement in more detail
Preceding page blank 109
-------
and discusses its importance in terms of the technical and economic feasibility
of the HZ process.
Catalyst Life
The useful life of the NOXNON 600 Catalyst for a coal—fired application was
not determined during the pilot plant tests. The tests did show that the
catalyst activity declined gradually over the deviation of the test program.
But the catalyst was still able to achieve N0 reduction efficiencies of 90
percent after nearly 5500 hours of operation. In addition, the test of a
method of catalyst regeneration restored catalyst activity to close to the
level of new catalyst. These facts indicate that the useful -life of t I e
NOXNON 600 catalyst is close to year.
Determining the useful life of the NOXNON 600 series catalyst does not
appear to be critical to either the technical or economic feasibility of the-
HZ process. However further investigation of catalyst lif i icluding -the
longer term effects of catalyst regeneration can have a significant impact --
on the costs of applying the HZ process. As discussed earlier, demonstration
of a 2—year catalyst life would reduce annual revenue requirements of the
process by 25 percent. This could make the costs of applying the HZ process
significantly lower than the costs for other SCR processes. -
lmpacts of Axumonium Sulfates
The high NH 3 and SO 3 emissions from the HZ process are expected to.have
a severe impact on an air preheater located down stream of çbe process. For
this reason, the conceptual design of the HZ process specified air -preheater
modifications intended to minimize the impacts of ammonium sulfate formation.
In addition, the cost estimate prepared as part of this evaluation included
the captial and annualized costs associated with those air preheater modif i—
cations.
110
-------
Unfortunatly, there are no data to substantiate whether the proposed
air preheater modifications will be effective in mitigating the impacts
of amruonium sulfate formation especially at the concentrations of NH 3 and SO 3
expected downstream of the HZ process. The air preheater modifications which
were included in the conceptual design were identified as a possible solution
to the ammonium sulfate problem for concentrations of NH 3 and SO 3 on the order
of 30 ppm each. Therefore it is possible that, with the high concentrations
of NH 3 and SO 3 expected downstream of the reactor, the proposed air preheater
modifications will not be effective in controlling deposition and corrosion.
If this is the case, other more costly measures may be required to help
minimize the problem. This could significantly increase the estimated costs
for the HZ process.
The ability of air preheater modifications to minimize the effects of
ammonium sulfate deposition is an important question concerning the HZ process.
Further development work is needed to determine what measures are needed to
minimize the impacts of high NH 3 and SO 3 emissions and to 4uahtify ths costs -
associated with high emission levels.
N 2 0 Emissions
Originally, the secondary emissionS sampling program included tests to.
measure N 2 0 production in the HZ reactor. unfortunately, the techniques used
during the sampling program did not prove satisfactory and it was not possible
to measure N 2 0in the flue gas streams. As a result, thequèstion of N 2 0
production in the reactor remains unanswered.
Currently there are no regulations which limit N 2 0 emissions £rom
stationary sources. One reason for this is because significant quantities of
N 2 0 are not produced as a result of stationary source combustion. However,
prod ’ction of large quantities of N 2 0 could present problems not typically
encountered. For this reason, further development work should include tests
designed to characterize N 2 0 emissions (if any) from the HZ process. Based
on the poor performance of the infrared analyzer used for the secondary
emissions sampling program, the recommended technique is cryogenic trapping
ill
-------
followed by gas chromatography analysis. This involves concentradon of N 0
in the gas stream followed by quantitative analysis and is routinely used to
measure concentrations of N 2 0 as low as 0.3 ppm;
Since there are no regulations which limit N 2 0 emissions, the priority -
of development work designed to quantify these emissions is considered low.
NH 3 Injection Control/Continuous Nil 3 Analysis
Automatic control of the NH 3 /N0 injection ratio whic i includes the use
of continuous Nil 3 emission measurements as feedback to such a control system
represents an area not addressed during the pilot plant tests. The N injec-
tion control system employed during the test program was a feed foreword -
system. The NH 3 /NO injection ratio was setmahüally. The contro11er monitored
inlet N0 and flue gas flowrate and then determined the flowrate on Nil 3 required
to achieve the desired N1-I3/NO injection ratio. This control scheme operates -
without regard to NH 3 emissions.
The type of control system which has not been demonstrated is that which
uses N H 3 emission measurements as feedback. The reason for this is the lack
of a continuous NH 3 analyzer. This represents a research and development
requirement not only for the HZ process but for SCR processes in general.
- During the pilot plant tests, Chemico attempted to measure Nil 3 emissions
c ontinuously using a converter—analyzer system which oxidizes H toN0 and
then measures the N0 using a chemiluminesceflt analyzer. This technique did
not prove successful, although the exact problem(s) with thetechnique was
never determined. Currently, there is quite a bit of research being done in-
Japan and the U.S. to develop a continuous NH3 analyzer, but to date no
technique has been demonstrated. Further work in this area should focus on
evaluating the various techniques being developed and demonstrating the most
112
-------
promising ones. In the case of the HZ process a feedback control of the NH 3
injection rate could be used to minimize potential operating problems in tha
air preheater.
Summary
In summary, there are development requirements for the HZ process of
varying priorities. Of thes , the most signif-ic nt are: the determination
of the useful life of the NOXNON 600 catalyst and the demonstration of
techniques to minimize the impacts of high N I - i 3 and SO 3 emissions from the
process. Both these items can have a significant impact on the estimated
capital costs and annual revenue requirements of the HZ process.
113
-------
REFERENCES
1. Maxwell, J. D., et al. Preliminary Economic Analysis of N0 Flue Gas
Treatment Processes. EPA—600/7—80—021. February, 198(1.
2. Environmental Protection Agency. 40 GFR Part 60. Standards of- -
Performance for New Stationary Sources: Continuous Monitoring Performance
Specifications — Proposed Revisions. Federal Register/Vol. 44, No. 197/
Wednesday, October 10, 1979/Proposed Rules. -
3. Burke, J. M. and K. L. Johnson. Ammonium Sulfate and Bisulf .ate Fo na ion
in Air Preheaters. Draft Report. EPA Contract No. 68—02—3171. Task 27.
February 1981. -
4. Radian Corporation. Impact of N0 Selective Catalytic Reduction Prdcesses
on Flue Gas Cleaning Systems. Draft Report. EPA Contract No. 68—02—3171,
Task 27. February 1981. -
114
-------
Appendix A
TECHNICAL NOTE
RESULTS OF THE QUALITY ASSURANCE
AUDITS AT THE EPA SPONSORED
NO PILOT PLANTS
April 1981
Prepared By:
J.M. Burke
Radian Corporation
8501 Mo—Pac Boulevard
Austin, Texas 78759
Prepared For:
J. David Mobley
U.S. Environmental Protection Agency
Industrial Environmental Research Laboratory
Research Triangle Park, N.C. 27711 -
EPA Contract No: 68—02—3171, Task 11
A-i
-------
SECTION 1
INTRODUCTION
The Environmental Protection Agency sponsored pilot scale tests of
two selective catalytic reduction (SCR) processes for NO removal.- Thepro—
cesses tested were the Hitachi—Zosen (HZ) process and the Shell flue gas
treatment (SFGT) process which also removes SO 2 . The contractors responsible
for design and operation of these pilot units were Chemico Afr Po-llutiOfl T
Control Corporation (Hz process) and the Process Division of UOP (SFGT
process). In both cases, these contractors were responsible for collection,
evaluation, and reporting of test data.
As part of the pilot scale test program, EPA contracted Radian Corporation
to conduct an independent evaluation of the performance of both pilot units. -
This independent evaluation consisted of several steps including:
• A quality assurance program to check measurements being
made by the process operators,
• A sampling program to quantify changes in secondary process
- - emissions across the SCR reactors, and - -
• A program to certify the performance of the condn 5us NO
monitors (and SO 2 monitors at the SFGT pilot unit) using
EPA reference methods.
This technical note presents the results of the quality assurance (QA) audits -
conducted to insure the quality of the major process measurements made by
Chemico and UOP. Results of the secondary emissions sampling program and the
continuous monitor certification tests are contained in separate technical
notes.
A- 2
-------
SECTION 2
OBJECTIVES AND APPROACH
The objective of the Quality Assurance audits conducted by-Radian was
to provide an independent check of the major process measurements made on a
routine basis by both UOP and Chemico. Not all process measu rements were
subject to audit as part of the QA program. Only those measurements which
are most significant in terms of quantifying process raw matetial and ene gy
requirements were audited.
Table 1 summarizes the QA audit plan. It identifies the process measure-
ments to be audited, the planned audit frequencies, and the techniques used
to conduct the QA audits at both pilot plants. As shown, tltese process
measurements represent the most basic operating parameters required to
characterize the operation of the HZ and SFGT processes. The principal raw
material requirements are characterized by the quantity of NH 3 required to
reduce a specific quantity of NO and, in the case of the SFGT process,-the
quantity of H 2 needed to regenerate the acceptor. The major utility require-
ments are characterized by the power needed to overcome the reactor pressure
drop and the steam required. The flue gas flowrate is required to complete
material balances and to determine the capacity of the reactor and thus the -.
costs to treat a specific volume of gas. Ammonia emissions are important
since they can have potentially significant operational and .economic impacts
on equipment located downstream of an SCR system. And finally, flue gas SO 2
concentration was audited at the HZ pilot plant to document the ability of
the catalyst to operate treating flue gas with relatively high SO 2 concen-
trations.
A- 3
-------
The techniques used to conduct the QA audits represent independent
checks on the process measurements. In several cases the techniques u ed
were EPA reference methods while others represent standard instruments or
sampling and analytical methods. In the case of NH 3 emissions measurements,
care was taken to insui e that the formation of atnmonium sulfates did not -
interfere with the sampling and analysis procedure.
The audit frequencies presented in Table -3—8 were selected based on -
several criteria. One consideration in specifying the audit, frequencies was
the availability/reliability of continuous process measurements. The most
significant item which was not measured continuously at either pilot plant
was the NH3 concentration at the reactor outlet. As a conseq ience, the -.
planned audit frequency for this measurement was relatively’high.
Another consideration in determining the audit frequencies was the diffi
culty of making a particular measurement. If an audit is very time consuming
then obviously this reduces the number of audits which can be.conducted in
a given period of time. An example of this is the H 2 flowrate at the. SFGT
pilot unit. It was expected that auditing the H 2 flowrat measurement would
be difficult and so the frequency was lower.
It should be noted that the audit frequencies specified in-Tab1 1 were
initial frequencies and they changed as the QA program progressed. For the-
most part a change in audit frequency depended on the agreement -between the
audit result an 1 the value of the process measurement recorded byChemico or
- -
Figure 1 presents a decision chain which was used as one basis for
changing the initial audit frequencies, depending on the agreement between
process measurements and the audit results. As shown, if the agreement was
within ±5 percent, the audit frequency was decreased and the new frequency
determined by the on site QA program leader. If the audit agreement was
within ±10 percent of the measured vaJ2ue, no change in frequency was made.
A- 4
-------
TABLE 1. SUMMARY OF THE QA AUDIT PROGRAM
SO 2 Concentration!
Reactor Inlet
Flue Cas Flowrate
N Fl 3 Injection Rate
Reactor Pressure Drop
Reactor Operating
Temperature
Nil 3 Concentration!
Reactor Outlet
Steam Flowrate
lI Fl wrate
x 0.5
EPA Method 2
Absorption followed by
weight gain measurement —
analogous to EPA
Method 4
Magnehelic differential
pressure gauge
Thermocouple with
traverse of gas duct
Absorption, distillation
at d titration
Calibration of orifice
plate
Process Measurement!
Location
Pilot Plant Audited 1
Initial Audit
Frequency
(Audits/Week) Audit Procedure
HZ
SFGT
U’
K x
K x
x K
x K
x - x
x
3
3
3
3
10
0.5
1 X Indicates an aUdit was scheduled.
-------
Figure 1. Action plan based on QA audit results.
A- 6
-------
If the audit agreement was outside the ±10 percent range, corrective
action was initiated to determine the source of error in either the audit-
procedure or the process measurement. At the same time, the audit frequency
was increased. If the source of error was found and corrected, the audit
frequency was reduced. - However, if the error could not be resolved, the-
audit frequency was maintained at the maximum practical level within the
time and manpower limits of the program. In addition to the use of relative
accuracy of the QA audits asa basis for -niodifyitig the audit frequency,
several practical considerations resulted-in the reduction of the audit
frequencies. These are discussed in Section 3.
A- 7
-------
SECTION 3
RESULTS OF THE QA AUDITS
The planned QA audits and the audit procedures were outlined in Section 2
of this technical note. This section presents a summary of the Qk audits
and details of individual audit results. It also includes a lrief discussion
of the results.
Summary
Table 2 summarizes the QA audit results from both the HZ and the SFGT
pilot plants. As shown, most of the QA audit results were within 10 percent -
of the process measurements made by UOP and Chennco. This fndicates that these
process measurements were accurate and the test data collected can be used to
characterize process operation.
Two measurements which did not show good agreement with the QA audits
were the SO 2 concentration at the HZ pilot plant and the flue gas flowrate
at the SFGT pilot plant. In both cases, subsequent investigation resolved the
discrepancy between the process measurements and the QA audits.
In the case of the flue gas SO 2 concentration at the HZ pilot p1ant
the audit results were determined to be correct and- the SO 2 Wonitor was in
error. This error is characteristic of the type of SO 2 mon itor used- (pulsed
fluorescence) when the instrument is calibrated with standard gases composed
of SO 2 in nitrogen. The response of a pulsed fluorescence SO 2 monitor to
SO 2 in a gas stream is quenched by the presence of O and CO 2 . This means
that addition of 02 and CO 2 to a gas stream contining SO 2 and N 2 will result
in an apparent reduction in the concentration of SO 2 as measured by a pulsed
A— 8
-------
TABLE 2. RESULTS OF THE QA AUDITS AT BOTH THE
HZ AND SFGT PILOT PLANTS
Planned Audit
Process Measurement! Frequency Relative Actual Frequency Relative Actual Frequency
Location (Audits/Week) Error(%) (Audits/Week) Error(%) (Audits/Week)
SO 2 Concentration! 2 —19.8 2 NA’ NA
Reactor [ niet
Flue Gas Flowrate 3 —0.3 2 —14.4 2.5
NH, Injection Rate 3 —6.0 1 7.6 1
Reactor Pressure Drop 3 4.5 2 —1.4 2
Reactor Operating 3 4.8 2 —1.3 2
Temperature
Steam Flowrate 0.5 NA NA ND 2 0
112 Flowrate 0.5 NA NA ND U
‘NA — Not Applicable
2 ND — Not Determined
-------
fluorescence analyzer. The error which results from calibration with S0 2 —N 2
mixtures (this was ddne at the HZ pilot plant) can be as high as 30 percent.
In the case of the flue gas flowrate measuted at the SFGT pilot plant,
the audit was determined to be in error. Because of the discrepancy between
the process measurement and the audit result, UOP made an independent check
of the flue gas flow-rate measurement. The results of this independent check
agreed very well with the prbcess measurement —re ults. It-was concluded-that
the duct area used to calculate flow-rate for the EPA Method 2 tests (velocity
traverse) was in error. This may have been due to a measurement error or
blocking of a portion of the duct crossection by fly ash. -
Table 2 does not present the results of NH 3 emission tests conducted
at the pilot plants. These results are contained in Appendix B with the
secondary emissions sampling results. The reason for this is that continuous
NH 3 monitor at both pilot plants did not work and so, the NH 3 emission “audit
results” represented a large fraction of the NH 3 emission measurements at both
pilot plants and were not actually audits. (Both Chemico and UOP did conduç
batch sampling and analysis tests for NH 3 ).
Table 2 also indicates that two measurements scheduled to be audited
were not. Both the steam and H 2 flowrates were not audited because it was -
not possible to do so without shutting down the SFGT pilot plant operation.
However, following the completion of the demonstration tests,-the orifice
plates used tojneasure steam and H 2 flowrates were calibrated by UOP personnel.
Finally the planned and actual audit frequencies are presented in Table 2.
As shown, most of the actual audit frequencies are lower than the planned - -
frequencies. There are several reasons for this discrepancy. In the case of
reactor pressure drop, the flue gas inlet temperature, and the flue gas flow—
rate at the HZ plant, the agreement was excellent (±5 percent), and the
frequency was reduced according to the decision chain shown in Figure 1. In
the case of NH 3 injection rate, the agreement was good (±10 percent) and the
A-1O
-------
difference was constant so no benefit was derived from a large number of
audits. At the SFCT pilot plant, the audit frequency for the flue gas flow—
rate was ultimately limited by access to the flue gas stream. The sampling
locations at the reactor inlet and outlet had to be shared between personnel
conducting the QA audits and those conducting the stack samp]4ng and
continuous monitor certification tests.
Only the audit frequency for flue gas SO - concentratiQn was as originally
scheduled. However, the QA plan called for an increase in a dit frequency
if the relative error was greater than 10 percent. The audit frequency was
not increased in this case since the difference between the SO 2 monitor
results and the QA audits was relatively constant.
Overall, the QA audit results indicated that the process measurements
made at both pilot plants were accurate. One exception was-the SO 2 -measure—.
ments at the HZ pilot plant. The audit results and subsequent investigation
revealed the S02 monitor was in error by about 20 percent. This, however, does
not affect the overall process evaluation since SO 2 was measured to document
the concentration of SO 2 the catalyst was exposed to and not as a key process
parameter.
The following discussion presents the results of individ ual QA audits.
- SO 2 Concentration
Table 3 presents the results of the-QA audits of the flue gas SO 2
monitor at the HZ pilot plant. As shown, the relative error was quite high
averaging nearly 20 percent.
A-fl
-------
TABLE 3. HZ SO 2 MONITOR QA AUDIT RESULTS
Test
No.
Continuous Monitor
(ppm)
Method 6
(ppm)
Error
(%)
1
990
i248
—20.7
2
3
1000
960
1271
1256
—21.3
—23.6
4
5
99Q
1000
1207
1156
—18.0
—13.5
6
960
1216
—21.1
Average
983
1226
—19.8
This was due to the quenching effect of 02 and CO 2 on the pulsed f1uor scence
monitor used to measure flue gas SO 2 concentrations at the HZ pilot plant.
Investigators conducted an analysis of pulsed flQorescence analyzers and
determined the error could be as high as 30 percent. The following equation
was developed to correct the monitor’s results based on the flue gas composi-
t Ion. -
(K o Pso 2 + Kf) (I /l—i) = P 0 K 0 + co2 Kco + N2 7 ° Io/I)KN
Where I = Intensity without quenching
I = Intensity with quenching
- K. = Quenching rate for species i (i.e., S02, 0 ,C02, N2)
= Partial pressure of species i
Kf Fluorescence decay rate
Using the average flue gas composition and solving this equation for 10, the
corrected SO 2 concentration can be estimated. However, experimental values
for K.’s need to be determined. These data are not available at this time.
1
As a rough estimate, the results of the QA audits of the SO 2 monitor
indicate that the SO 2 concentrations measured by Chemic&were 20 percent
A-12
-------
lower than the actual flue gas concentration. Fortunately, these data do not
affect the overall process evaluation.
Flue Gas Flowrate
The results of the flue gas flowrate QA audits at the HZ and SFGT pilot -
plants are presented in Tables 4 and 5 respectively. As shown, the flowrate
audits at the HZ pilot plant-were in excellent.- agreement with the measurements-
made by Chemico, indicating that the process measurements were accurate.
However, the flowrate measurements at the SFGT pilot plant did not agree with
the audit results. As discussed earlier, it appears that the audit result
was inaccurate. The consistent error between the process measurements and
the audit results indicates that the actual audit measurement (flue gas
velocity rather than flowrate) was accurate and a consistent error was intro— -
duced in the flowrate calculation. The most-likely explanation for this i-s
that an inaccurate value for the duct crossection was used in the flowrate
calculations.
The QA flowrate audit results indicate that accurate-flowrate measure-
ments were made at both the HZ and SFGT pilot plants. -
TABLE 4. HZ FLUE GAS FLOWRATE QA AUDIT RESULTS
Flowmeter
Audit — Method 2
Error
Test No. (scftn)
(scfm) -
(%)
1 1324 1310 1.1 --
- 2 1341 1318 1.7
3 1345 1327 1.4
4 1542 1576 —2.2 -
5 1524 1568 —2.8
Average 1415 1420 —0.3
A-13
-------
TABLE 5. SFGT FLUE GAS FLOWRATE QA AUDIT RESULTS
Test No.
Flowmete
(scfm)
r
Audit — Method 2
(scfm)
Error
(%)
1
2
987
986
1102
1204
—1Q.4
-18.1
3
978
1165
-15.3
4
1026
•
.
1126
-13.5
5
6
1023
984
1236
1290
:17.2
-23.7
.
7
992
1123
—11.7
8
9
984
1000
1085
1116
—9.3
—10.4
10
1000-
1126
—11.2
•
Average
996
1163
—14 . 4
-
NH 3 Injection Rate
The results of the NH 3 injection rate QA audits at both the HZ and SFGT
pilot plants are presented in Table 6. As shown, the relative error at both
plants was consistent and had the same magnitude. This error. was within the
accuracy limits of the measurement technique (i.e., absorption followed by -
weight gain measurement). The results indicate that accurate (±10 percent)
NH 3 injection measurements were made at both pilot plants. -
A- 14
-------
TABLE 6. NH 3 INJECTION RATE AUDITS AT THE HZ AND SFGT PILOT PLANTS
Process
Test No. (gms
Measurement
NH 3 /min)
Audit
(gins
—Absorption
NH 3 /min)
Error
%
HZ Pilot Plant
1
11.8
12.7
—7.1
2
11.9
12.5
—4.8
Average
11.85
.12.-6
—6.0
SFGT Pilot Plant
10.4
9.7
6.9
1
2
9.7
9.0
- 7.3
3
9.8
9.1
•
7.8
Average
9.97
9.27
7.6
Reactor Pressure Drop
The results of the reactor pressure drop QA audits at the HZ and SFGT
pilot plants are presented in Tables 7 and 8, respectively. As shown, the
process measurements and audit results at both pilot plants were in excellent
agreement. This indicates that the pressure drop measurements made by Chemico —
and UOP during the pilot plant test program were accurate.
TABLE 7. HZ REACTOR PRESSURE DROP AUDIT RESULTS
Process Measurement. . Audit—Magnehelic Error
Test No. (in H 2 0) (in H 2 0) ( ‘I .)
1
1.05
l.Ô
5.0
2
1.05
1.0
5.0
.
3
1.04
1.0
4.0
4
1.04
1.0
4.0
5
1.04
1.0
4.0
6
.
1.05
1.0
5.0
Average
1.045
1.0
- 45
A- I 5 -
-------
TABLE 8. SFGT REACTOR PRESSURE DROP AUDIT RESULTS
Process
Measurement
Audit—Magnehelic
E or
Test No. (in
H 2 0)
(in
I1 O)
(%)
1 10.75 10.5 2.4
2 10.75 10.5 2.4
3 10.25 11.0 —6.8
4 1O.OQ 10.5 —4.8
5 9.75 10.0 -2.5
6 9.50 9.75 —2.6
7 10.50 10.25 2.4
Average 10.21 10.36 —1.4
Reactor Operating Temperature
The results of the reactor operating temperature QA audits at the HZ and
SFGT pilot plants are presented in Tables 9 and 10 respectively. As shown,
the process measurements and the audit results at both plants were in excellent
agreement. This indicates that the reactor inlet temperatures measured by
Chemico and UOP were accurate. Since Chemico demonstrated reactor performance
over a range of temperatures, these results are not critical to the HZ process
but they are to the SFGT process. UOP did not document SFGT process perfor— -
mance at reduced temperatures, so the process measurements define a lower limit
for process operation.
TABLE 9. HZ REACTOR OPERATING T ’IPERATURE AUDIT RESuLTS
Test
No.
Process
Measurement
(°F)
Audit Thermocou
(°F)
pie
Erro
(%)
r
1
717
681
5.3
‘
2
720
691
4.2
3
719
684
5.1
4
722
690
4.6
5
718
684
5.0
Average
719.2
686
4.8
A — I 6
-------
TABLE 10. SFGT REACTOR OPERATING TENPERATURE AUDIT RESULTS
Process Measurement Audit Thermocouple - - -. Error
Test No. (°F) (°F) - (%)
1 772 - 780 —1.0
2 762 769 —0.9
3 781 794 —1.6
4 781 790 —1.1
5 785 769 2.l
6 787 801 —1.7 -
7 780 802 —2.7
8 776 799 —2.9
9 778 788 —1.3
Average 778 - 788 —1.3 -
A-17
-------
DCN 80—203—001—11—14
Appendix B
TECHNICAL NOTE
RESULTS OF SECONDARY
EMISSION SAMPLING
AT THE
EPA SPONSORED NO PILOT PLANTS
December, 1980
Prepared By:
B.M. Ekiund
K.L. Johnson
J.M. Burke
Radian Corporation
8501 Mo—Pac Boulevard
Austin, Texas 78759
Prepared for;
J. David Mobley
U. S. Environmental - Prbtection Agency
Industrial Environmental Research Laboratory
Research Triangle Park, N.C. 27711
EPA Contract No: 68—02—3171, Task 11
B—i
-------
SECTION 1
INTRODUCTION
The Environmental Protection Agency sponsored pilot scale tests of
two selective catalytic reduction (SCR) processes for NO removal.. The pro-
cesses tested were the Hitachi—Zosen (Hz) process and the Shell flue-gas
treatment (SFGT) process which also removes SO 2 . The cont ractors responsible
for design and operation of these pilot units were Chemico A±r Pollut in
Control Corporation (HZ process) and the Process Division of UOP (SFGT
process). In both cases, these contractors were responsible for collection,
evaluation, and reporting of test data.
As part of the pilot scale test program, EPA contracted Radian
Corporation to conduct an independent evaluation of the performance of both
pilot units. This independent evaluation consisted of several steps including:
A quality assurance program to check measurements being made -
by the process operators, . -
A sampling program to quantify changes in secondary process
emissions across the SCR reactors, and
A program to certify the performance of the coñtiftugus NO
monitors (and SO 2 monitors at the SFGT pilot unit) using
EPA reference methods.
This technical note presents the results of the sampling program t6
quantify changes in secondary process emissions. Results of the quality
assurance program and the certification tests are contained in separate
technical notes.
B— 2
-------
SECTION 2
POLLUTANTS OF INTEREST
SCR processes have been operated commercially in Japan for several
years, but there is a surprising lack of data on secondary emissions-from
these processes. Therefore, one objective of EPA’s pilot plant test-
program was to quantify changes in secondary emissions acfoss the SCR
reactor. Based on available information about SCR processes ‘-and_ on
concerns expressed by the professional community, a group of- poLlutants was -
selected for detailed sampling and analysis at the EPA sponsored pilot plants.
These pollutants include: —
• ammonia (NH 3 ),
sulfur trioxide (SO 3 ),
particulate loading and composition,
• carbon monoxide (CO),
hydrocarbons (C 1 through C 6 ),
hydrogn cyanide (HCN),
nitrosoamines, and
- nitrous oxide (N 2 0).
The following discussion briefly examines the reasons for measuring-
each of these pollutants, while Section 3 contains-a description of the
pilot units and the process streams which were sampled during the test
program. Section 4 presents the results of the sampling effort, along
with a brief description of the sampling and analysis procedures used.
B— 3
-------
Ammonia (NH 3 )
Ammonia was measured at the reactor outlet at each pilot plant to quant-
ify the emission rate under normal operating conditions. Ammonia is injected
upstream of the cata1y t in an SCR process to react with NO reducing it-to
molecular nitrogen. Under normal operating conditions, it is expected that
some NH 3 will pass through the catalyst unreacted. This unreacted NH 3
represents a secondary process emission which caxi present significant oper-
ational and possibly environmental problems, depending on the emission rate.
Sulfur Trioxide (SO 3 )
Sulfur trioxide was measured at each pilot unit to determine if any -
change in concentration occurs in the NO reduction reactor. The reason for -
measuring S03 across the reactor differs for each plant. The Hitachi—ZOs n
system contains a catlyst which can promote oxidation and -it is possible
that SO 3 may be produced in the reactor due to SO 2 oxidation.— On the other
hand, UOP has indicated that the SFGT system may actually reduce flue gas
SO 3 concentrations due to reaction with the material used in the reactor.
Measurement of the SO 3 concentration at the reactor inlet-and outlet was -
made to quantify the magnitude of any changes which occur in the reactor at
bothpilot plants.
Particulate Loading and Composition
Particulate samples were collected áhd analyzed at the pilot units.
Particulate loading was measured to document the loading the reactor is
exposed to and thus demonstrate the ability of the reactor to operate
under high particulate loadings. Particulate composition was measured to
attempt to determine if any catalyst is being eroded from the reactor.
This was done by examining particulate composition at the reactor inlet and
outlet and determining if any increase in the concentration of the elements
present in the catalyst occurs.
B-4
-------
Carbon Monoxide (CO) and Hydrocarbons (HC )
Carbon monoxide and hydrocarbons were measured at the reactor inlet
and outlet to determine if any change occurs in the concentration of these
pollutants across the reactor. These two pollutants are of interest since
they can both be oxidized by oxygen or NON. As a consequence, it is possible
that the reactor may also reduce the emission levels of CO and/or HC.
Hydrogen Cyanide (HCN), Nitrosoainines, and Nitrous Oxide (N O )
Hydrogen cyanide, nitrosoamines and nitrous oxide were measured to
determine if any change in the concentration of these compoun 1soccurs-irL
the pilot system reactors. While no concrete evidence exists to indicate
that HCN, nitrosoamines, or N 2 0 are produced, concern has been expressed
that the presence of NH 3 and the catalyst may result in production of oñe
or more of these compounds. The sampling program was designed to measure
each of these compounds and document any change in their concentrations
across the reactor.
B- 5
-------
SECTION 3
PILOT PLANT DESCRIPTION
EPA has sponsored the demonstration of two SCR processes applied to
coal—filled boilers. The size of these units was nominally 0.5 MWe
equivalent and each was designed to remove 90 percent of the incoming
NO . The SFGT pilot unit was also designed to achieve simultaneous
removal of 90 percent of the SO 2 in the flue gas. The follp ing discussion
presents a brief description of each process in general and a specific
description of the EPA sponsored pilot plants.
3.1 CHEMICO/HITACHI—ZOSEN PLANT
Chetnico Air Pollution Control Corporation’s pilot demonstration of the
Hitachi—Zosen selective catlytic reduction (SCR) process is located at
Georgia Power Company’s Plant Mitchell Station in Albany, Georgia. The
process has been commercially applied to numerous oil and gas—fired boilers
in Japan. This pilot plant is a demonstration of the process for coal—fired
applications.
The SCR process utilizes antmonlg (NH 3 ) to selectively educe NO to
molecular nitrogen (N 2 ) in a catalytic reactor. The overall chemical
reactions are shown in Equations 1 and 2.
4 NO(g) + NH 3 (g) + 0 2 (g) - 4 N 2 (g) + 6 H 2 0(g) (1)
2 N0 2 (g) + 4 NH 3 (g) + 02(g) - 3 N 2 (g) + 6 H 2 0(g) (2)
B-G
-------
In the process, NH 3 is injected into flue gas which is at a temperature
between 300°C and 400°C. The gas then passes through a catalytic reactor
where the reduction reactions (Equations 1 and 2) occur rapidly.
The HZ pilot plant processed a 1000—1500 SCFM slipstream of flue gas
which was withdrawn between the economizer and the air preheater of the
coal—fired unit 3 at the Plant Mitchell Station. At this point, the flue
gas contained the full particulate loading from the boiler. The slipstream
from the boiler was first passed up through a flue gas heater for temperature
control. NH 3 was then injected into the flue gas which flowed down through
the reactor. After exiting the reactor, the gas passed through a cyclone
particulate collector, a fan, and it then returned to the boiler flue gas
duct. The particulate collector was included in the pilot unit to minimize
abrasion in the fan.
Figure 1 is a schematic of the HZ pilot plant which shows the
location of the sampling ports used during the test program. The flue gas
temperature was approximately 315°C at the inlet sample port and 370°C at the
outlet port. Full particulate loadings were present at both sampling
locations.
Return to
:
Inlet Heater Reactor Port
Sample
Port
Figure 1. Schematic of the HZ pilot plant.
B-
-------
3.2 IJOP/SFGT PILOT PLANT
The Process Division of UOP has operated a pilot demonstration of the
Shell Flue Gas Treatment (SFGT) process at Tampa Electric Company’s Big Bend
Station in North Ruskin, Florida. The process has been commercially applied
on an oil—fired boiler in Japan, while the pilot plant was a coal—fired
demons trat ion.
This SCR process is unique in that it removes both SO 2 and NO . Copper
supported on alumina is used as an SO 2 “acceptor”. SO 2 and 02 react with the
acceptor as shown in Equation 3.
S0 2 (g) + ½ 02(g) ÷ CuO(s) - -CuSOi+(s) (3)
Copper oxide and, to a greater extent, the copper sulfate (CuSOt,)
produced by the SO 2 acceptance reaction, act as a catalyst for the reduction
of NO by NH 3 according to Equations 1 and 2.
As SO 2 is removed in the reactor, the acceptor becomes “loaded” and
SO 2 emissions begin to increase. Once the time averaged SO 2 removal reaches
the design level (e.g., 90 percent), flue gas flow to the reactor is shut off
and the CuS0 /Cu0 are reconverted to Cu metal. In this regeneration step,
a steam/hydrogen mixture is passed through the reactor and a concentrated
SO 2 stream is produced. The regeneration reactions are illustrated by
Equations 4 and 5.
CuSOt 4 + H 2 CuO + SO 2 + H 2 O (4)
CuO + H 2 Cu + H 2 O (5)
The SO 2 produced by the regeneration of the reactor can be further
processed to recover the sulfur as sulfuric acid or elemental sulfur.
B-B
-------
The SFGT pilot plant processed a 1000 scfm slipstream of flue gas which
was withdrawn between the economizer and the air preheater of the coal—fired
unit 2 at Tampa Electric Company’s Big Bend Station. Flue gas from the boiler
contained full particulate loading and was heated by in—line burners before
entering the reactor. Flue gas flowed up through the reactor during accep-
tance of SO 2 and bypassed the reactor during the regeneration of the acceptor.
During the program, some tests were conducted in which the reactor was fully
loaded with SO 2 and N0 was the only pollutant removed. While these tests
were conducted, the reactor processed flue gas continuously. During cyclic
operation, the acceptance period ranged from 50 to 70 minutes, with an average
of 65 minutes, while the regeneration period ranged from 22 to 40 minutes,
with an average of 29 minutes.
Figure 2 is a schematic of the SFGT pilot plant which shows the
location of the sampling ports used during the test program. Flue gas
temperature was approximately 400°C at both the inlet and outlet sampling
ports. Full particulate loadings were present at both sampling locations.
Return
to Boiler
Duct
Outlet
Sample
Port
Reactor
Reactor
By—Pass
Flue Gas In—line
frnm Propane
Boiler Fired Flue t lnlet
Gas Heater jSample
Port
Figure 2. Schematic of the SFGT pilot plant.
B- 9
-------
SECTION 4
RESULTS
The stack sampling program was conducted by Radian during the period
June through August, 1980. A mobile laboratory was used during the program
and most of the analytical work was completed on—site. The results of the
stack sampling work are suuunarized in Tables 1 and 2 which show inlet and
outlet concentrations of the species measured at the HZ and SFGT pilot plants
respectively. It should be noted that rio results for N 2 0 are presented due
to the fact that the measuring technique proved unsatisfactory.
As shown in Table 1, the concentrations of CO, HC, HCN and nitrosoamines
at the HZ pilot plant were all below the detection limit of the techniques
used. This indicates that no significant change in the concentrations of
these species occurred in the reactor. Table 1 does show significant
increases in the concentrations of N H 3 and SO 3 emitted from the reactor,
while the apparent change in particulate loading is believed to be the result
of random error or unaccounted for stratification in the ducts.
Table 2 shows the results from the sampling program of the SFGT unit.
Concentrations of 1-ICN and nitrosoaniines were below the detection limit, again
showing that no significant change in the concentration of these species
occurred in the reactor. Decreases in both the HC and CO concentrations were
measured across the reactor. This change is probably due to oxidation of
these pollutants across the reactor. It is believed that the relatively high
levels of CO and HC’s measured were due to emissions from the in—line propane
burners located upstream of the reactor. As shown, a decrease in SO 3 concen-
tration was also measured across the reactor. This decrease is due to removal
of SO 3 from the flue gas via reaction with the acceptor. The apparent change
B- 10
-------
TABLE 1 - STACK SANPLING RESULTS AT HZ PILOT UNIT
Flue Gas Reactor Inlet Reactor Outlet
Component Concentration 1 Concentration 1 Measurement Technique
Nitrosoamines 2 <5 igIdscm <5 pg/dscm Absorption, Extraction, Gas
Chromatograph w/nitrogen
specific detector
Hydrogen Cyanide 2 <.01 mg/dscm <.01 mg/dscm Absorption, distillation,
titration
Ammonia Not measured 54.8 ppmv (dry basis) Absorption, distillation,
titration
Sulfur Trioxide 8.4 ppmv (dry basis) 20.7 ppmv (dry basis) Controlled Condensation,
Ion Chromatograph
Hydrocarbons 2 <1.0 ppmv <1.0 ppmv Gas Chromatograph
(C 1 —C 6 ) Flame Ionization Detector
Carbon Monoxide 2
-------
TABLE 2 - STACK SA}IPLING RESULTS AT UOP PILOT PLANT
Flue Gas
Component
Reactor Inlet
Concentration’
Reactor Outlet
Concentration 1
Heasurement Technique
Nitrosoam±nes
<5 pg/dscm
<5 pg/dscm
Absorption, Extraction, Gas
Chroniatograph w/nitrogen
specific detector
Cyanides
<.01 ingldscm
<.01 mg/dscm
Absorption, distillation,
titration
Ammonia
Not measured
15.3 ppinv (dry
49.0 ppniv (dry
basis) 2 Absorption, distillation,
basis) 3 titration
Sulfur Trioxide
11.4 ppinv (dry
basis)
0.1 ppmv (dry
basis)
Controlled condensation,
ion chromatograph
Hydrocarbons
28.5 ppmv
21.0 ppmv
Gas Chromatograph
Flame Ionization Detector
(C ,—C&)
Carbon Monoxide
0.13%
-------
in particulate concentration is believed to be due to unaccounted for
stratification in the ducts.
The following discussion examines the stack sampling results in more
detail. In addition, a brief discussion of the analytical techniques are
presented.
Ammonia
Ammonia samples were collected at the reactor outlet at both the HZ and
SFGT pilot plants. At the SFGT pilot plant, samples were collected during
both cyclic and N0 only operation. The samples collected during cyclic
operation were taken over the course of an entire cycle yielding an average
NH 3 concentration at the reactor outlet during the cycle.
Ammonia emission samples were collected by filtering the flue gas in
the duct and drawing the sample through a glass—lined probe into two
impingers containing 5 wt percent H 2 S0 4 . A composite sample was produced
by combining the two impinger solutions with the washings from the sample
probe. A 250 ml aliquot of the composite was taken for analysis of NH3
nitrogen by a distillation/titration procedure. The aliquot was buffered
with a sodium borate buffer solution, and the pH adjusted to 9.5 using sodium
hydroxide. This mixture was then distilled for approximately one hour, until
more than 200 ml of distillate was collected. The distillate was captured
subsurface in a boric acid indicating solution. This product solution was
then titrated with a standard .02 N sulfuric acid titrant.
The results of the NH 3 emissions sampling at the HZ pilot plant are
presented in Table 3. As shown, the N i- ! 3 emissions were quite high, averaging
about 55 ppm for all the tests. This emission rate is much higher than the
expected rate of 10 to 20 ppm, and it could have serious impacts on equipment
located downstream of the SCR process. In addition, NH 3 emissions of 50 ppm
could result in formation of an axnmonium sulfite/bisalfite plume downstream
B—13
-------
TABLE 3 — HITACHI-ZOSEN/CHENICO PILOT PLANT NH 3 EMISSIONS
Date
Time
NO In
(ppm wet)
N}13:NO
NH 3 Emission
(ppm dry)
7—18
13:45 — 14:05
459
0.91
45
7—18
15:11 — 15:31
451
0.97
46
7—29
14:54 — 15:20
466
0.92
49
7—30
10:37 — 10:57
453
0.92
60
8—1
12:23 — 12:48
469
0.99
44
8—2
11:17 — 11:42
479
0.99
63
8—2
18:50 — 19:10
440
1.05
54
8—5
15:29 — 15:57
450
0.89
58
8—6
16:53 — 17:13
460
1.01
74*
8—7
11:17 — 11:42
493
0.89
41*
8—7
13:15 — 13:40
482
0.98
68*
*Note: Sample analyzed by p11 adjustment, direct nesslerizatiori/spectrophotometrY Instead of
pH adjustment, dIstillation/titratiOn as for other samples. Tests from 7—29 and 8—1
were run with both methods with good agreement.
-------
of a wet FGD system. The reason for the high NH 3 emission rate is unclear
at this time, but is probably due to an inadequate amount of catalyst in the
reactor. Other possibilities, such as a high NH 3 injection rate have been
checked and do not explain the measured NH emission rate.
The results of the NH 3 measurements at UOP are presented in Table 4.
As shown, the samples taken during the NO only tests indicate an average
NH 3 emission rate of 49 ppm which is quite similar to the results from the
HZ NH 3 emission tests. The samples collected during the cyclic operation
of the SFGT process exhibit results which are quite different from the non—
cyclic operation. Samples collected during cyclic operation show an average
NH 3 concentration of about 15 ppm in the gas exiting the reactor. This
emission rate is more in line with expectations prior to the start of the
pilot scale tests, and should not present serious problems downstream of the
SCR reactor. It should be noted, however, that the measured NH 3 concentra-
tion is an average over the acceptance period and no information is available
on the instantaneous N I - I 3 emission rate over the course of a cycle.
Currently, no explanation has been found for the marked difference in
NH 3 emission rates during cyclic and non—cyclic operation at essentially the
same operating conditions (i.e., flue gas flow, temperature, NH 3 injection
rate). One possibility is that Cu and CuO promote NH 3 oxidation at a higher
rate than CuSOL does. Another possibility is that the metal surfaces in the
reactor strongly adsorb N I - I 3 . During non—cyclic operation this adsorption
would occur only once and the surfaces would quickly become saturated with
NH 3 . During cyclic operation N I -I 3 would be desorbed during regeneration.
Consequently, the reactor internals would always have some capacity for
adsorbing NH 3 , thus lowering the NH 3 emission rate during acceptance.
B-15 -
-------
TABLE 4 — SFGT PILOT PLANT NH 3 EEISSIONS
UOP
Date Time Cycle 1/
NO.
(ppi
In NH 3 Emissions
dry) NH3:NO (ppm dry)
6—21
16:12
—
17:12
875
380
1.17
15
6—23
12:49
—
13:52
903
420
1.21
16
6—24
13:26
—
14:36
918
390
1.31
22
6—26
13:23
—
14:20
948
375
1.25
17
6—26
17:52
—
18:48
951
No
UOP
Data
16
6—27
15:06
—
16:02
965
No
UOP
Data
11
6—28
15:49
—
16:52
982
392
1.17
12
6—28
17:27
—
18:32
983
402
1.19
13
6—30
15:27
—
17:06
1004
No
UOP
Data
16
7—11
7—15
7—15
14:01
14:50
18:28
—
—
—
14:23
15:15
18:53
NO 0nly
NO 0nly
NO 0nly
420
327
224
1.18
1.21
1.38
47
52
48
-------
TABLE 5 — CHANGE IN SO 3 CONCENTRATION ACROSS THE REACTOR IN THE HZ PROCESS
—I
Date
Time
SO 3 Inlet
(ppm dry)
SO 3 Outlet
(ppm dry)
SO 2 Oxidation
(%)
7—23
16:30
— 17:20
11.0
25.3
1.5
7—24
16:50
— 17:40
7.5
14.8
1.1
7—25
16:25
— 17:10
6.6
21.9
1.9
TABLE
6 — CHANGE
IN
SO 3
CONCENTRATION
ACROSS
THE REACTOR
IN
ThE
SFGT
PROCESS
Date
Time
SO 3 Inlet
(ppm dry)
SO 3 Outlet
(ppm dry)
Removal
(%)
10-8
14:40
- 15:40
10.4
0.095
99.1
10—8
17:15
— 18:45
12.4
0.108
99.1
-------
Sulfur Trioxide
Sulfur trioxide was collected using a controlled condensation system.
The SO 3 was selectively condensed at 80°C in a modified Graham condenser.
The sample was recovered by rinsing the condenser coil with deionjzed water
and the sulfate concentration of the coil rinse was measured by ion
chrotnatagraphy. The results of the SO 3 sampling are presented in Tables 5
and 6.
For the three test results from the HZ pilot plant, shown in Table 5,
an average of 1.5 percent of the inlet S02 is oxidized to SO 3 across the
reactor. This effect is to be expected since a primary component of the
catalyst, vanadium pentoxide, is used commercially as a SO 2 oxidation
catalyst. The increase in SO 3 concentration, by an average of 12.3 ppm, has
potentially important downstream impacts. First, the resultant higher
sulfuric acid (H2SOk) dewpoint decreases the amount of heat that a boiler
air preheater can safely extract from the flue gas without suffering increased
corrosion. This can be a significant energy penalty. Increased maintenance
costs and decreased service life for the preheater or downstream equipment,
due to H 2 SO attack could result. Another potential impact is an increase
in atmuonium bisulfate (NH 1 +HSOz 4 ) formation in the preheater from the reaction
of SO 3 and water with residual ammonia exiting the SCR reactor. This can
result in an increased pressure drop for the system. Maintenance costs
could increase due to more frequent cleaning of the preheater. Finally,
any additional SO 3 that passed through the preheater would most likely
result in increased H 2 S0t emissions. Conventional flue gas desulfurization
(FGD) systems are not very effective for capturing the H 2 SO , aerosols or
mist that would form.
For the two tests run at UOP’s plant, an average of 99% of the SO 3 was
removed from the flue gas. This SO 3 removal capability of the system is
also important. By lowering the H 2 S0 dewpoint, it may be possible to
B—i 8
-------
operate the air preheater at a lower flue gas exit temperature and thus
recover more energy in the preheater. This can be a significant energy
savings. Lower SO 3 concentrations should minimize downstream impacts,
decreasing emissions and maintenance and increasing equipment life.
Particulates
Particulates were collected by filtering the flue gas in stack (EPA
Reference Method 17) with a 30 by 100 mm glass fiber sampling thimble.
Traverses of the duct were conducted to ensure a representative sample.
The samples were collected isokinetically in accordance with established
EPA methods. The samples were weighed to determine grain loadings and a
composite of three samples was analyzed to determine the elemen€al
composition of the particulates.
The results of the particulate loading measurements are presented in
Tables 7 and 8. At the HZ pilot plant, the grain loading averaged 7.1 gm/dscm
at the inlet and 7.7 gm/dscm at the outlet. No particulate matter is produced
within the system, so the 8% difference in grain loading values is probably
attributable to small random errors. It is also possible that stratification
of the particulates existed in the ducts, since the sample ports were located
close to downstream disturbances.
At the SFGT plant, the grain loading averaged 8.9 gm/dscm at the inlet
and 6.3 gin/dscm at the outlet. No large air leaks are present in the
system to dilute the flue gas, so the 29% difference in grain loading values
is due to an error in sampling. The magnitude of the difference suggests
a systematic error in the sampling procedure. Because of the location
of sample ports on the SFGT pilot unit, it was only possible to traverse
the duct in a single plane. Since the sample ports were also
located close to disturbances at the SFGT plant, it is conceivable that
B — i 9
-------
TABLE 7 - GRAIN LOADINGS MEASURED AT THE HZ PILOT PLANT
Lt, 3
0
Date
Time
Inlet
(gm/dscm)
Outlet
(gm/dscin)
7—29
17:50
— 18:50
7.5
7.8
7—30
17:02
— 18:02
6.9
8.4
7—31
14:09
— 15:09
6.9
—
7—31
18:08
— 19:08
—
6.9
TABLE 8
- GRAIN
LOADINGS
MEASURED
AT
SFGT PILOT
PLANT
Date
Time
Inlet
(gm/dscf)
Outlet
(gni/dscf)
6—19
10:47
— 11:27
8.3
6.7
6—19
13:55
— 14:35
11.2
6.8
6—19
17:05
— 17:45
7.4
5.5
-------
representative sampling of the duct was not possible using this single plane.
Table 9 presents the results of the elemental analysis of the
particulates collected at the HZ pilot plant. This analysis was conducted
to determine if catalyst erosion was causing an increase in the concentration
of Vanditim (V) or Titanium (Ti) in the particulates leaving the reactor.
As shown in Table 9, an apparent increase in all elements occurs across the
reactor, but the relative concentrations on V and Ti remain constant. This
indicates that there is no measurable change in the concentration of V or
Ti in the particulates exiting the reactor. The apparent change could be
due to several factors. One possibility is the presence of an inert material
(e.g., unburned carbon) in the inlet sample. This material may have entered
the sample due to contamination, or if it is indeed unburned carbon, it may
have been burned in the flue gas heater located between the reactor inlet
and outlet sample ports. In any event, it can be concluded that no signif-
icant increase in the concentrations of V or Ti in the particulates occurred
due to catalyst erosion. -
Table 10 presents the results of the elemental analysis of the
particulates collected at the SFGT pilot plant. At this plant, the elements
of interest are Alumina (Al) and Copper (Cu). As shown, no significant
change in the composition of the particulates was measured with respect to
these elements. The changes which are shown in Table 10 are probably due
to random errors in the sampling on analysis technique and are not
representative of real changes in particulate composition. Therefore, it
can be concluded that no significant increase in the concentrations of
Al or Cu in the particulates occurred due to catalyst erosion.
Carbon Monoxide and Hydrocarbons
Samples were collected by cooling a filtered flue gas stream to about
200°F and then passing it through a Perma—Pure dryer to remove moisture.
Samples were collected in Mylar gas bags that had been purged several times
with the conditioned gas.
B— 21
-------
TABLE 9 - RESULTS OF PARTICULATE ANALYSIS AT ThE HZ PILOT PLANT
Component 1
In
Out
Out/In
Al
10.7%
13.0%
1.21
Ca
8200 ppm
9900 ppm
1.21
Fe
4.9%
6.0%
1.22
K
2.0%
2.5%
1.25
Mg
6300 ppm
7800 ppm
1.24
Mn
190 ppm
240 ppm
1.26
Sn
490 ppm
680 ppm
1.40
Na
4200 ppm
4700 ppm
1.12
Si
18%
23%
,
1.28
Zn
190 ppm
250 ppm
1.32
Cu
150 ppm
170 ppm
1.13
Ti
5800 ppm
6900 ppm
1.19
V
270 ppm
330 ppm
1.22
‘Concentrations are on a mass fraction basis
B-2 2
-------
TABLE 10 — RESULTS OF PARTICULATE ANALYSIS AT THE SFGT PILOT PLANT
Component’
In
Out
Out/In
Al
86%
807
93
Ca
-
1.8/
1.8/a
1.00
Fe
12%
11.1/0
0.93
K
1.5%
1.4%
0.93
Mg
5100 ppm
5000 ppm
0.98
Mn
300 ppm
320 ppm
1.07
Sn
270 ppm
270 ppm
1.00
Na
4300 ppm
5100 ppm
1.19
Si
20%
16%
0.80
Zn
410 ppm
720 ppm
1.76
Cu
96 ppm
100 ppm
1.04
Ti
5400 ppm
5100 ppm
0.94
V
255 ppm
340 ppm
1.33
‘Concentrations are on a mass fraction basis
B-23
-------
The samples were analyzed on—site using a Fisher Gas Partioner Model
1200 TCD—GC and an AID portable Gas Chromatograph. The gas samples were
introduced into the instruments, and the retention times and peak heights
were compared to those obtained from mixes of standard gases. At the HZ
pilot plant the GC became inoperative and Ci—C 6 hydrocarbon samples were
collected in snout* bags and transported to Radian’s Austin Laboratories
where they were analyzed within a week of collection on a Tracor Model
560 FID—CG. The results are presented in Tables 11 and 12.
As shown in Table 11, the concentration of both CO and HC’s at the
HZ pilot plant were below the detection limit of the analysis techniques
so no conclusion could be drawn concerning the impact of the catalyst on
these pollutants. However, at the SFGT pilot plant, decreases in the
concentrations of both CO and HC were measured across the reactor. This
decrease is attributed to oxidation of the CO and HC’s by the catalyst
forming CO 2 and H 2 0. The oxidation of these compounds was not unexpected
since the NO reduction reaction is actually an oxidation of NH 3 by NO
and the catalyst is basically an oxidation catalyst. It should be noted
that the high levels of CO and HC’s at the SFGT pilot plant were probably
a result of incomplete combustion in the propane burners used for flue gas
temperature control upstream of the reactor.
Hydrogen Cyanide (HCN )
Cyanide samples were collected by bubbling filtered flue gas through
a 10% sodium hydroxide solution. Samples were shipped to Radian’s Austin
Laboratories for analysis. It was originally proposed to measure the
concentration of cyanide spectrophotometrically with a pyridine—barbituric
*Corisists of layers of polyester, polyvinyl chloride, aluminum, polyamide
and polyethylene.
B- 24
-------
TABLE 11 - FISHER GAS PARTIONER AND G.C. DATA FROM THE HZ PILOT PLANT
Sampling Species (as % of total flue gas)
Location
CO 2 O N 2 CO C 1 —C 6
Inlet 17.8 4.4 69.7 <0.017 <1 ppm
Outlet 17.2 4.9 68.5 <0.017 <1 ppm
TABLE 12 - FISHER GAS PARTIONER AND G.C. DATA FROM THE SFGT PILOT PLANT
Sampling Species (as % of total flue gas)
Location
CO 2 02 N 2 CO CH
Inlet 17.5 4.8 70.5 0.13 28.4 ppm
Outlet 17.6 4.8 70.3 <0.017 21.0 ppm
The total C 1 —C 6 hydrocarbon concentration was the same as the CH&,
concentration.
B— 25
-------
acid reagent, but this method proved to be unfeasible due to inferference
from SO 2 absorbed in the sodium hydroxide solution. A silver nitrate
(AgNO 3 ) titration method was used instead. The silver ions complex with the
cyanide to form Ag(CN)z. The excess Ag is detected by paradimethylamino—
benzalrhodanine, which undergoes a yellow to salmon color change in the
presence of Ag+ ions.
The AgNO 3 titrametic method employed by Radian could detect cyanide at
a level equal to or greater than 0.01 rng/dscm. The concentration of
cyanide at Chemico and UOP pilot plants was below this limit of detection
at both the inlet and outlet sampling ports. This indicates that no
significant increase in the flue gas HCN concentration occurs across the
reactor at either pilot plant.
Nitrosoamines
Nitrosoaxnine samples were collected by bubbling filtered flue gas through
a iN potassium hydroxide solution. The samples were packed in ice and
shipped to Radian’s Austin laboratories. There each sample was serially
extracted with distilled—in—glass methylene chloride to remove the organic
species from the sample matrix. The methylene chloride extracts were then
combined and concentrated by evaporation to less than 10 ml. Hexane (10 ml)
was added and each extract was reconcentrated to remove the methylene
chloride. The resulting hexane extracts were then analyzed by GC.
The samples were initially analyzed by a nitrogen specific detector. The
chromatogrphic data was generated using a 6’ x 4mm 10 percent carbowax 20M
+ two percent KOH column. Two separate injections were conducted for each
sample. The first injection at 110°C isothermal was employed for low
molecular weight nitrosoamiries such as N—nitrosodimethylamine. A second
injection at 220°C isothermal was employed for higher boiling species such
as N—nitrosodiphenylamine. Several small peaks were observed; none of these
matched the retention time of known nitrosoamines in the standard used. The
B— 26
-------
samples were then analyzed by a Hall electroconductivity detector operated
in the nitrosoamine specific mode using a 8’ x 2 nun 15 percent LAC—2R—446
Column at 145° isothermal. No peaks were observed in these analyses.
The CC/MS analysis could detect nitrosoamines at a level equal to or
greater than 5 ug/dscin. The concentration of nitrosoamiries at the
Chetnico and UOP pilot plants was below this limit of detection at both
the inlet and outlet sampling ports. This shows that no significant
change in nitrosoainfne concentrations occurred across the reactor at
either pilot plant.
Nitrous Oxide (N 2 0 )
Grab samples of flue gas were collected in Mylar bags. The gas was
cooled and filtered to remove condensed ammonium salts and water vapor.
The samples were then injected into a Miran 1A Gas Analyzer (Wilks
Scientific Corporation) and the concentration of N 2 0 was determined by
infra red (IR) spectral analysis.
Nitrous oxide is not normally measured in flue gas studies and there is
no generally accepted procedure. The method used in the sampling program
was not sensitive enough to detect the presence of N 2 0. Several inter-
fering peaks combined with the low level of any N 2 0 present made deter—
mination of the N 2 O concentration impossible.
In order to measure N 2 0, it would have been necessary to concentrate
it prior to analysis. One technique which has been used for measuring low
N 2 0 levels is cryogenic trapping and gas chromatography. A silica gel
column is used to concentrate the N 2 0 in the gas sample. After withdrawal
from the duct, particulates, water, and CO 2 are successively removed from the
sample. The gas is then passed through a gel—concentration column maintained
at —70°C which further concentrates the N 2 0. The N 2 0 is desorbed by flushing
with helium. The gas sample passes through a separation column and the N 2 0
B-27
-------
concentration is measured with a thermal conductivity detector.
This method is routinely used to measure atmospheric level of N 2 0
(250—300 ppb), but can detect levels down to at least 50 ppb.
SUNMARY
The stack sampling program conducted by Radian was designed to quantify
changes in the emission rate of secondary pollutants from the SCR process.
The results of this program showed that:
A significant quantity of NH 3 is emitted from the HZ
process and from the SFGT process during non—cyclic (N0 only)
operation,
The HZ process produces SO 3 by oxidation of SO 2 at a
rate equivalent to 1 to 2 percent of the SO 2 in the flue
gas.
The SFGT process removes nearly all the SO 3 in the flue
gas entering the reactor. This could have a beneficial
impact on downstream equipment which may be subject to
acid corrosion.
The particulate loadings in the processed flue gas at
both pilot plants was about 7 gm/dscm which is typical of
many coal—fired applications.
• No significant change in particulate composition was
measured across the reactor at either pilot unit.
• The concentrations of both CO and HC’s were reduced across
the SFGT reactor. It is believed this reduction was
due to oxidation. No change was measured in the
concentration of CO or HC’s at the HZ pilot unit due to
the fact the concentrations were below the detection
limit of the analytical techniques.
• No significant change in the concentrations of HCN or
nitrosoamines occurred at either pilot unit.
• The technique used for measuring N 2 O proved unsatisfactory
for the quantity of N 2 0 (if any) present in the flue gas
at either pilot unit.
- 28
-------
Appendix C
TECHNICAL NOTE
RESULTS OF THE CONTINUOUS MONITOR
CERTIFICATION TESTS AT THE
EPA SPONSORED NO PILOT PLANTS
April, 1981
Prepared By:
J.M. Burke
B.M. Eklund
Radian Corporation
8501 Mo—Pac Boulevard
Austin, Texas 78759
Prepared For:
J. David Mobley
U.S. Environmental Protection Agency
Industrial Environmental Research Laboratory
Research Triangle Park, N.C. 27711
EPA Contract No: 68—02—3171, Task 11
C — I
-------
SECTION 1
INTRODUCT ION
The Environmental Protection Agency sponsored pilot scale tests of
two selective catalytic reduction (SCR) processes for NO removal. The pro-
cesses tested were the 1-litachi—Zosen (HZ) process and the Shell flue gas
treatment (SFGT) process which also removes SO 2 . The contractors responsible
for design and operation of these pilot units were Cheinico Air PQllutiorl
Control Corporation (HZ process) and the Process Division of UOP (SFGT
process). In both cases, these contractors were responsible for collection,
evaluation, and reporting of test data.
As part of the pilot scale test program, EPA contracted Radian Corporation
to conduct an independent evaluation of the performance of both pilot units.
This independent evaluation consisted of several steps including:
A quality assurance program to check measurements being made
by the process operators,
A sampling program to quantify changes in secondary process
emissions across the SCR reactors, and
A program to certify the performance of the continuous NO
monitors (and SO 2 monitors at the SFGT pilot unit) using
EPA reference methods.
This technical note presents the results of certification tests to assure
the accuracy of the data collected by the continuous NO and S02 monitors.
Results of the secondary emissions sampling program and the quality assurance
program are contained in separate technical notes.
c—2
-------
SECTION 2
OBJECTIVE AND APPROACH
The objective of the continuous monitor certification tests was to insure
the accuracy of the data generated by the continuous monitors at the two pilot
plants. In conjunction with the quality assurance program, the certification
tests were expected to provide accurate characterization of the pilot plant’s
performance permitting the data collected by Chemico and UOP to be used as a
basis for a more detailed process evaluation.
The approach used in conducting the continuous monitor certification
tests was based on a formal procedure which has been developed by EPA to insure
the accuracy of monitors measuring emissions from sources which must comply
with new source performance standard emission limitations. According to the
procedure developed by EPA, a continuous emission monitor must pass a number
of performance tests in order to be certified. These tests include;
• calibration error,
• response time,
• 2— and 24—hour zero drift,
• 2— and 24—hour calibration drift, and
• relative accuracy.
The performance specifications for each of the above certification tests are
shown in Table 1. These specifications are those contained in the Federal
Register, Vol. 44, No. 197, Wednesday, October 10, 1979 — “Proposed Rules:
Standards of Performance for New Stationary Sources; Continuous Monitoring
Performance Specifications”.
c—3
-------
TABLE 1. CONTINUOUS MONITORING SYSTEM PERFORNANCE
AND EQUIPMENT SPECIFICATIONS
Certification Test Performance Specification
5 percent for mid—range and
high level calibration values
< 15 minutes
2 percent of Span’ value
2.5 percent of Span’ value
2 percent of Span’ value
2.5 percent of Span’ value
20 percent of reference
test data in terms of emission
standard 2 or 10 percent of the
applicable standard, whichever
is greater
Because the pilot plants were operated as experimental systems, the
operating characteristics of the continuous monitors were different than those
of a continuous monitor system for a typical commercial scale application. In
particular, the pilot plant continuous monitors received considerably more
operator attention in an effort to insure the quality of the data being collected.
As a consequence, not all the continuous monitor performance specifications were
strictly adhered to. The major emphasis of the certification tests conducted at
the pilot plants was to insure the relative accuracy of the continuous monitors.
The other certification tests were considered of secondary importance.
The following discussion presents the procedure for conducting each certi-
fication test, identifies areas where the certification tests at the pilot plants
differed from those outlined in the performance specifications, and presents
reasons for the differences. The certification results are presented in detail
in Section 3. For more details of the certification test procedures, the
reader is referred to the Federal Register mentioned above.
Calibration Error
Response Time
Zero Drift (2—hour)
Zero Drift (24—hour)
Calibration Drift (2—hour)
Calibration Drift (24—hour)
Relative Accuracy
‘Span refers to the range of the emission monitor (e.g., 0—500 ppm).
2 Emission standards are expressed in terms of pollutant mass emissions
per unit of heat input.
C- 4
-------
Calibration Error
Calibration error is a measure of the ability of a continuous monitor
to accurately analyze the concentration of a calibration gas. To determine
calibration error, a series of 15 measurements are made; 5 each of high and
mid—level calibration gas concentrations and 5 of zero gas (air). The
measurements are made such that the same gas concentration is not measured
two or more times in a row (i.e., non—consecutively). The calibration error
is then determined by the difference between the concentration of the cali-
bration gas (which has been verified by one of several techniques defined in
the CEM performance specifications) and the continuous monitor readings.
Specifically the error is defined by Equation 1.
Calibration Error = Xl + i d 95 I (1)
RV
where lxi = absolute value of the mean difference
between the monitor reading and the
calibration gas concentration.
RV = concentration of the calibration gas.
C l 95 = absolute value of the 95 percent
confidence interval which is defined
by Equation 2.
Cl 95 = (ni)½ (n Ex _(Ex ) 2 ) ½ (2)
where n = the number of measurements (in this case 5)
t = 2.776 (for 5 measurements)
X. difference between the continuous monitor
reading and the calibration gas concen-
tration for individual readings.
For the continuous monitor to pass the calibration error test, the error
as defined by Equation 1 must be less than or equal to 5 percent for both the
mid— and high—range calibration gases. There was no deviation from this
procedure in conducting the calibration error test at either pilot plant.
C- 5
-------
Response Time
Response time is a measure of the time which elapses before a continuous
monitor responds to a change in the pollutant concentration in a duct. To
determine response time, zero gas is introduced into the continuous monitor’s
sampling system at the duct/sample line interface. Once the monitor reading
stabilizes at zero, the sampling system is switched from zero gas to monitoring
the stack effluent and the time required for the monitor to reach a stable
reading is determined. This is the upscale response time. Subsequently, the
downscale response time is determined by introducing high—level calibration gas
into the continuous monitor sampling system at the duct/sample line interface.
Once the monitor’s reading stabilizes at the high level gas concentration,
stack gas is introduced into the sample line and the time required for the
monitor’s reading to reach a stable value downscale is determined. This pro-
cedure is repeated three times and the mean upscale and downscale response times
are determined. The maximum of these values must be less than 15 minutes for
the continuous monitor to pass this test.
This method of measuring response time represents a slight change from
previous EPA performance specifications which were the ones followed during
the certification tests at the pilot plants. In the previous certification
procedures, response time was determined by alternately introducing zero and
high—level calibration gas into the sample line at the duct/sample line inter-
face. Upscale response time was that required to stabilize the CEM system
reading after the switch from air to high—level calibration gas while down—
scale response time was measured in the opposite manner. This measure of
response time should provide an accurate indication of system response pro-
vided care is taken to prevent excessive span gas pressure in the sample line.
Zero and Calibration Drift
Zero and calibration drift are measures of how the continuous monitor’s
response to zero and calibration gases change with time. To measure two—
hour drift, zero and high level calibration gases are introduced into the instru-
ment at two—hour intervals. The differences between the monitor readings and
C- 6
-------
the zero or calibration gas concentrations are recorded and a total of 15 sets
of data are collected. Drift is then calculated using Equations 1 and 2 where
the reference value (RV) equals the range of the continuous monitor and
t. 9 75=2.145 (for 15 tests). 24—hour drift is measured in a similar mariner.
At the start of a 24—hour period the monitor is calibrated and zeroed and then
no further adjustments are made. After 24 hours, zero and calibration drift
are recorded and the instrument is recalibrated. A total of 7 measurements
are required to determine 24—hour drift and Equations 1 and 2 are used.
For 7 measurements, t.975=2.447.
Because the pilot plants were operated as experimental units, the accuracy
of continuous monitorswas of particular concern. As a consequence, the cali-
bration and zero settings on the continuous monitors were adjusted about
once per hour for the SFGT pilot plant (generally at the start and completion
of an acceptance period), and every two hours at the liZ pilot plant. Obviously,
the 24—hour drift test is not meaningful under these circumstances arid, in the
case of the SFGT pilot plant, the two—hour drift test could not be completed
strictly according to performance specifications. Instead, the zero and cali-
bration drift was determined for the period between initial and final cali-
bration and zeroing of a monitor.
Relative Accuracy
Relative accuracy is the most important test in determining the perfor-
mance of the continuous monitors. This test is comparison of the continuous
monitor’s analysis of stack gas with a stack gas analysis as determined by
EPA reference methods (Method 3 for 02 and C0 2 , Method 6 for SO 2 , and Method 7
for NO ). Relative accuracy is determined by making a minimum of 9 reference
method tests (for Method 7, one test requires three samples) and comparing
the results to the continuous monitor readings during the time the reference
tests were collected. In addition, for determining the relative accuracy of
NO and SO 2 continuous monitors, simultaneous reference method tests for
02 are required so the NO and SO 2 concentrations can be converted to a mass
emissions per unit of heat input (e.g., lb/b 6 Btu) basis.
c- I
-------
After the reference method tests are completed and the continuous
monitor and reference method test results are converted from concentrations
to a mass emission per unit heat input basis, the relative accuracy can be
determined. Relative accuracy is calculated using Equations 1 and 2 where
the reference value (RV) is equal to the mean of the reference method tests
and t. 975 =2.306 (for 9 tests).
This procedure was followed exactly at the HZ pilot plant. And 9 sets
of reference method tests for both SO 2 and NO were completed at the inlet
and outlet of the SFGT pilot plant reactor. However, reference method tests
for O or CO 2 were not included because UOP did not operate a diluent monitor.
As a result, the CEM data could not be converted to a mass emission per unit
of heat input basis. In the case of the SFGT pilot plant this does not
impact the validity of the relative accuracy tests because: (1) the reactor
operated under 7.5 kPa (30 in H 2 0) pressure so no dilution occurred between
the reactor inlet and outlet, and (2) the measurements made by the continuous
monitors were intended to demonstrate the ability of the process to achieve a
specific percentage reduction in NO and SO 2 emissions based on pollutant
concentrations rather than a reduction in emissions below a specific mass
emission rate per unit heat input.
c . 8
-------
SECTION 3
CONTINUOUS MONITOR CERTIFICATION TEST RESULTS
The certification tests conducted at the HZ and SFGT pilot plants were
discussed in Section 2 of this technical note. This section presents a summary
and Individual results from the certification tests. It also includes the
results of reference method tests on the calibration gases used during the
certification test program.
Summary
Table 2 summarizes the results of the continuous monitor certification
tests for both the HZ and SFGT pilot plants. As shown, the test results for
all the continuous monitors met the performance specifications with one
exception; the relative accuracy of the outlet NO monitor at the SFGT pilot
plant was over 50 percent while the performance specifications require a
relative accuracy of 20 percent or less. These data indicate that, with the
exception of the SFGT outlet NO analyzer, the continuous monitors were making
accurate measurements of flue gas NO and SO 2 concentrations. It should be
noted that 24—hour zero and calibration tests were not included in the
certification tests since the instruments were calibrated at least every
two hours, thus making the 24—hour drift tests meaningless. In addition,
at the SFGT plant, only one SO 2 monitor was used, so the values for all tests
except response time and relative accuracy are identical.
The poor relative accuracy of the outlet NO analyzer shown in Table 2
would tend to indicate that the monitor was making inaccurate measurements of
flue gas NO concentrations. However, there are several factors which must be
considered when evaluating these test results. First, the absolute error in
C- 9
-------
TABLE 2. CONTINUOUS MONLTOR CERTIFICATION TEST RESULTS
Certification
Test
Performance
Specif ic.atiofl
SF61’ Pilot Plant
SO 2 ’ Inlet NO
Outlet NO
HZ
Inlet NO
Pilot
Plant
Outlet NO
Inlet SOP ’
Monitor
Monitor
Monitor
Monitor
Monitor
Monitor
Calibration error
—high level
—raid level
<5%
<5%
1.4
0.7
1.4
0.7
3.85
4.62
2.52
2.62
1.40
4.39
4.70
2.68
Response time
<15 mi i i
0.8
1.3
1.7
0.8
1.4
0.05
Zero drift (2—hour)
<2%
0.25
0.25
0.64
1.04
1.20
1.78
Calibration drift
(2—hour)
<2%
0.49
0.49
1.35
1.18
1.93
Relative accuracy
<20 12
14.0
8.6
12.6
52.3
14.1
10.5
‘One Instrument wds ueed to measure both inlet and outlet of the reactor.
2 Altenndtively, <10 percent of tire applicable emission8 standard.
-------
the Method 7 versus the monitor measurements was only 16 ppm on the average.
This is a relatively small difference considering the accuracy of the Method 7.
procedures. At low NO levels, similar to those encountered at the reactor
outlet, EPA found the Method 7 analysis to be accurate to only ±35 percent.
A second factor, which indicates that the outlet NO monitors performance was
within acceptable limits, is that the performance specifications require the
relative accuracy be less than or equal to 20 percent or 10 percent of the
applicable standard, whichever is greater. Use of the standard as a basis for
relative accuracy calculations is in recognition of the inaccuracy of the
Method 7 analysis at low levels. For example, using the NSPS for bituminous
coal fired sources as a basis, the relative accuracy of the outlet NO monitor
Is approximately 5 percent, which is within acceptable limits. Because the
relative accuracy of the outlet N0 monitor was within 10 percent of the
applicable standard and because of the documented poor performance of Method 7
at low N0 concentrations, it was determined that the outlet N0 monitor was
performing acceptably.
Overall, the results of the certification tests indicate that the
continuous N0 and SO 2 monitors at both pilot units were performing acceptably.
Therefore, the data collected during the pilot plant tests by Chemico and
UOP are representative of the pilot plant’s performance. This is especially
true in light of the extensive monitor maintenance program carried out by
both Chemico and UOP which was designed to insure the accuracy and quality of
the performance data collected.
The preceding discussion presented a summary of the monitor certification
test results. Details of individual measurement results follow. This includes
the results of reference method analysis of calibration gases.
c—il
-------
Calibration Gases
The calibration gases used during the certification tests were analyzed
prior to the tests to insure the manufacturer’s reported concentrations were
accurate. The results of these analyses are presented in Table 3. As shown,
the reference method test results were within 5 percent of the manufacturer’s
reported concentrations except for the mid—range SO 2 calibration gas. In
this case, the SO 2 concentration was assumed to be that determined by the
reference method results while the other gas concentrations were assumed equal
to the manufacturer’s reported concentration.
TABLE 3. COMPARISON OF MANUFACTURER’ S REPORTED CALIBRATION GAS
CONCENTRATIONS WITH REFERENCE MET}IOD TEST RESULTS
Calibration Gas
Concentration 1 Reference Method Test Result (ppm ) Error 2
Description ppm 1 2 3 Average (%)
High—Range SO 2 2690 2631 2844 2792 2756 —2.4
Mid—Range SO 2 1470 1612 1618 1613 1614 —8.9
Inlet
High—Range NO 930 914 914 911 913 +1.9
Mid—Range NO 515 464 557 525 515 0
Outlet
High—Range NO 91.9 87.2 88.2 95.0 90.1 +2.0
Mid—Range NO 50.4 54.2 51.9 43.5 49.9 +1.1
1 This is the manufacturer’s reported concentration.
2 Error = (manufacturer’s reported concentration — reference method result)
— (reference method result)
Calibration Error
The calibration error measurements for the HZ pilot plant are presented
in Table 4 while the SFGT pilot plant measurements are presented in Table 5.
As shown, only one set of data was taken for the SO 2 monitor at the SFGT pilot
plant since this instrument was used to measure both inlet and outlet concen—
C—12
-------
trations. The work—up of the HZ pilot data is presented in Tables 6—9 while
the work—up of the SFCT pilot plant data is presented in Tables 10—15. As
shown, all the monitors passed the calibration error tests.
TABLE 4. RAW DATA: CALIBRATION ERROR TESTS AT
TBE l iZ PILOT PLANT
Inlet NO Monitor 1
Calibration Measurement
Outlet NO Monitor
Calibration
Gas
Concentration
(ppm)
Measurement
System
Reading
(ppm)
Run
No.
Gas
Concentration
(ppm)
System
Reading
(ppm)
1
955
948
0
0.4
2
0
2
91.9
87.2
3
955
947
50.4
49.3
4
515
533
91.9
89.5
5
0
4
0
0.3
6
515
532
50.4
49.2
7
955
948
0
0.3
8
0
6
91.9
89.8
9
955
941
50.4
49.3
10
515
538
0
0.4
11
0
5
50.4
49.3
12
515
536
91.9
89.0
13
0
6
50.4
49.9
14
955
955
91.9
91.0
15
515
534
0
0.4
1 This monitor failed the initial test and was later retested with a
different high—range calibration gas.
C-13
-------
TABLE 5.
RAW DATA:
Run
No.
Inlet NO
Calibration
Gas
Concentration
(ppm)
Monitor
Measurement
System
Reading
(ppm)
Outlet NO
Calibration
Gas
Concentration
(ppm)
Monitor
Measurement
System
Reading
(ppm)
SO 2 Monitor
Measurement
Gas System
Concentration Reading
(ppm) (ppm)
1
930
905
0
1
0 0
2
0
14
515
505
2690 2740
3
930
902
930
916
1614 1620
4
515
495
515
504
2690 2688
5
0
9
930
908
0 —4
6
515
494
0
1
1614 1604
7
930
922
930
907
0 —8
8
0
11
515
501
1614 1616
9
930
922
0
1
2690 2664
10
515
520
515
504
0 0
11
0
10
0
1
2690 2676
12
515
529
930
916
1614 1616
13
0
5
515
507
2690 2680
14
515
532
930
919
0 —4
15
930
958
0
1
1614 1604
-------
TABLE 6. HZ NO INLET MONITOR: HIGH—RANGE
CALIBRATION ERROR RESULTS
Run
No.
Calibration
Concentration
Gas
(ppm)
Measurement System
Reading (ppm)
Difference
(ppm)
1
955
948
—7
3
955
947
—8
7
955
948
—7
9
955
941
—14
14
955
955
0
2X. =
3 -
—36
jxj =
7.2
C 1 95 1 =
6.17
Relative
Accuracy =
1.4%
TABLE 7.
HZ NO INLET
CALIBRATION
MONITOR: MID—RANCE
ERROR RESULTS
Run
No.
Calibration
Concentration
Gas
(ppm)
Measurement System
Reading (ppm)
Difference
(ppm)
4
515
533
18
6
515
532
17
10
515
538
23
12
515
536
21
15
515
534
19
x. =
1
98
lxi =
19.6
C l 95
=
-,
1 .
Relative
Accuracy =
4.39%
c-15
-------
TABLE 8. HZ NO OUTLET MONITOR: HIGH-RANGE
CALIBRATION ERROR RESULTS
Run Calibration Gas Measurement System Difference
No. Concentration (ppm) Reading (ppm) (ppm)
2 91.9 87.2 —4.7
4 91.9 89.5 —2.4
8 91.9 89.8 —2.1
12 91.9 89.0 —2.9
14 91.9 91.0 —0.9
EX = —13
I I = 2.6
ICI 95 I = 1.72
Relative Accuracy = 4.70%
TABLE 9. HZ NO OUTLET MONITOR: MID-RANGE
CALIBRATION ERROR RESULTS
Run Calibration Gas Measurement System Difference
No. Concentration (ppm) Reading (ppm) (ppm)
3 50.4 49.3 —1.1
6 50.4 49.2 —1.2
9 50.4 49.3 —1.1
11 50.4 49.3 —1.1
13 50.4 49.9 —0.5
-5
3 -
I I= 1
1c 1 95 1 = 0.35
Relative Accuracy = 2.68%
C— 6
-------
TABLE 10. SFGT NO INLET MONITOR: HIGH-RANGE
CALIBRATION ERROR RESULTS
Run
No.
Calibration
Concentration
Gas Measurement System
(ppm) Reading (ppm)
Difference
(ppm)
1
930
905
—25
3
930
902
—28
7
930
922
—8
9
930
922
—8
15
930
958
+28
EXi =
—41
l l =
8.2
ci 95 1
=
27.65
Relative
Accuracy =
TABLE 11.
SFGT NO INLET MONITOR: MID—RANGE
CALIBRATION ERROR RESULTS
3.85%
Run
Calibration
Gas Measurement System
Difference
No.
Concentration
(ppm) Reading (ppm)
(ppm)
4
515
495
—20
6
515
494
—21
10
515
520
5
12
515
529
14
14
515
532
17
EX =
—5
lxI=
1
1CI 95 1 =
Relative Accuracy =
22.77
4.62%
C-I?
-------
TABLE 12. SFGT N0 OUTLET MONITOR: HIGH—RANGE
CALIBRATION ERROR RESULTS
Run
No.
Calibration
Concentration
Gas Ieasurement System
(ppm) Reading (ppm)
Difference
(ppm)
3
930
916
—14
5
930
- 908
—22
7
930
907
—23
12
930
916
—14
14
930
919
—11
EX. =
1
—84
ixi =
16.8
Ici 95
=
6.65
Relative
Accuracy =
2.52%
TABLE 13.
SFGT NO OUTLET MONITOR: MID-RANGE
CALIBRATION ERROR RESULTS
Run
No.
Calibration
Concentration
Gas Measurement System
(ppm) Reading (ppm)
Difference
(ppm)
2
515
505
—10
4
515
504
—11
8
515
501
—14
10
515
504
—11
13
515
507
—8
EX. =
1
—54
=
10.8
1c1 95 1
=
2.69
Relative Accuracy =
2.62%
C-18
-------
TABLE 14. SFGT SO 2 MONITOR HIGH-RANGE
CALIBRATION ERROR RESULTS
Run
No.
Calibration
Concentration
Gas Measurement System
(ppm) Reading (ppm)
Difference
(ppm)
2
2690
2740
50
4
2690
2688
—2
9
2690
2664
—26
11
2690
2676
—14
13
2690
2680
—10
=
1
-2
lxi =
0.4
Ici 9 d
=
36.59
Relative
Accuracy =
1.38%
TABLE
15. SFGT SO 2 MONITOR MID—RANGE
CALIBRATION ERROR RESULTS
Run
No.
Calibration
Concentration
Gas Measurement System
(ppm) Reading (ppm)
Difference
(ppm)
3
1614
1620
6
6
1614
1604
—10
8
1614
1616
2
12
1614
1616
2
15
1614
1604
—10
x.=
1
—10
l I=
2
1C 1 95 1
=
9.29
Relative Accuracy =
0.70%
C-19
-------
Response Time
Table 16 presents the results of the response time measurements at the
HZ and SFGT pilot plants. As shown all the continuous monitors had response
times much lower than the 15 minute limit required by the performance speci-
fications. The relatively rapid response times recorded are due, in part,
to the short sample lines (less than 25 meters) at the pilot plants. These
results indicate that even though the response time measurements were made in
accordance with previous certification procedures (and not the current
proposed procedures) the response times of the pilot plant continuous monitors
are acceptable.
TABLE 16. RESPONSE TThIE DATA FROM THE HZ AND SFGT PILOT PLANTS
Test
No.
Response Time
(mm) 1
HZ Monitors
SFGT
Monitors
NO>< In
NO Out
SO 2 In
SO 2 Out
NO In
NO
Out
1
1.3
1.8
1.5
0.7
0.7
1.3
2
1.4
1.5
3.2
0.8
0.8
1.6
3
1.5
1.6
0.4
0.8
0.8
1.0
Av
e.
1.4
1.6
1.7
0.8
0.8
1.3
1 Response time was defined as the time it took the instrument to register 95
percent of a step change in concentration of gas to the monitor.
Zero and Calibration Drift
As discussed earlier, the pilot plants were operated as experimental units.
As a result, the accuracy of the continuous monitors was of critical importance
and they were calibrated quite frequently (as often as once an hour). By the
strict definition, both the 2— and 24—hour drift tests were not applicable to
the pilot plants’ continuous monitors. Instead, zero and calibration drift
data were collected for the monitors prior to the routine calibration carried
out by Chemico and UOP. This gives an indication of the drift experienced
by the monitors between routine calibrations and is approximately equivalent
to the 2—hour drift tests.
c- 20
-------
TABLE 17. HZ PILOT PLANT ZERO AND CALIBRATION
DRIFT TEST RESULTS
Run No.
Inlet NO
i Zero (ppm)
Monitor 1
L Cal (ppm)
Outlet
t Zero (ppm)
NO
Monitor 2
Cal (ppm)
1
—12
—12
0
2.7
2
40
7
0
—5.0
3
8
—15
0
—3.0
4
—20
23
—0.1
2.5
5
0
40
0
3.5
6
7
—10
0.1
3.6
7
—4
13
0
—3.0
8
—10
10
0
0
9
—35
15
0
4.0
10
—1
9
0
—4.0
11
—3
40
—0.1
2.0
12
—15
—30
—0.1
—2.0
13
—17
—42
0
0
14
18
39
0
2.0
15
45
—25
—0.1
—2.0
X.
1
1
72
—0.3
1.3
XI
0.07
4.8
0.02
0.09
1ci 95 1
11.91
14.45
0.03
1.69
Error (%)
1.20
1.93
0.05
1.78
1 The range of this monitor was 0—1000 ppm.
2 The range of this monitor was 0—100 ppm.
c— 21
-------
TABLE 18. SFGT
Inlet NO Monitor Outlet NO Monitor 1 SO 2 Monitor
i Zero ACa1 AZero ACa1 AZero ACa1
Run No. (% of Span) (% of Span) (ppm) (ppm) (7. of Span) (% of Span)
1 1.0 —0.5 —0.7 —4.3 —0.3 —0.9
2 —0.5 0.3 —4.7 —2.3 0.3 —0.5
3 0.4 1.0 —4.7 —3.0 0.2 —0.5
4 —0.3 1.5 1.0 —2.7 0.4 —0.5
5 0.1 1.0 0 —3.0 0.3 —0.3
6 —0.5 —1.3 0.7 —2.3 0.2 —0.3
7 —0.2 —1.8 —2.0 2.0 0.3 —0.4
8 —0.1 —0.1 —2.0 1.3 0.4 —0.4
9 —0.5 —2.0 —4.3 10.0 0.2 —0.7
10 —0.5 —3.5 1.0 3.7 —0.2 1.0
11 —0.5 1.1 1.3 1.7 —0.4 0.3
12 —1.0 0.4 2.0 —10.0 0 0.1
13 —1.1 —2.5 4.0 1.0 0.3 1.0
14 —0.4 2.0 —3.3 —1.7 0 —0.3
15 —1.0 —2.0 —4.3 3.0 —0.4 —0.3
—5.1 —6.4 —16 —6.6 —1.3 —2.7
lxi 0.34 0.43 1.07 0.44 0.09 0.18
1CT 95 I 0.30 0.92 1.54 2.51 0.16 0.31
Error (%) 0.64 1.35 1.04 1.18 0.25 0.49
‘The range of this monitor was 0—250 ppm.
-------
Relative Accuracy
Relative accuracy is the most important of the certification tests con-
ducted since this compares the continuous monitor results to reference method
analyses of the flue gas. In addition, the relative accuracy tests provide
an independent check of the pilot plants’ performance while the other certi-
fication tests only yield performance Information about the instruments.
Tables 19, 20, and 21 present the raw data and results for the HZ
relative accuracy tests while Tables 22 and 23 present this information for the
SFGT pilot plant. As shown, all monitors except the NO outlet at the SFGT
pilot plant meet the performance specifications. The fact that the SFGT
outlet NO analyzer does not meet the relative accuracy specification is not
significant in terms of the performance data collected. The average NO
reduction recorded by the instruments during the relative accuracy tests
was 84.9 percent, while the Method 7 results indicate an average N0 reduction
of 87.4 percent.
TABLE 19. HZ NO INLET MONITOR: RELATIVE ACCURACY TEST DATA
Test
No.
Monitor
Readings
Reference Method
Results
NO)<
(ppm wet)
(%
02
wet)
NO
(ppm dry)
(%
O
wet)
Moisture
(%)
1
485
3.4
469
3.0
7.8
2
443
3.6
442
3.0
7.0
3
479
3.4
430
2.5
8.0
4
469
3.3
443
2.5
7.9
5
489
3.8
439
3.0
7.5
6
460
3.5
453
3.0
8.4
7
435
2.8
442
2.0
8.1
8
447
2.9
469
2.5
7.9
9
484
3.6
524
3.0
8.5
c-2 3
-------
TABLE 20. HZ NO OUTLET MONITOR: RELATIVE ACCURACY TEST DATA
Test
No.
Monitor
Readings
Reference Method
Results
NO
(ppm wet)
(%
02
wet)
NO
(ppm dry)
(%
02
wet)
Moisture
(%)
1
73.7
3.4
73.2
3.5
8.1
2
66.6
3.6
62.8
3.3
7.4
3
57.3
3.4
52.4
2.7
8.2
4
57.8
3.4
56.4
2.5
8.1
5
63.9
3.2
55.0
2.5
7.8
6
64.0
3.9
70.1
4.0
8.7
7
79.4
2.9
81.4
2.0
8.4
8
82.5
2.5
87.9
2.0
8.1
9
78.8
3.3
87.5
3.5
8.8
C-24
-------
TABLE
21. HZ N0
Inlet Monitor
Outlet Monitor
Difference
Monitor
Method 7
Difference
Monitor
Method
(lb/lO 6 Btu)
Test No. (lb/lO 6 Btu)
(lb/lO 6 Btu)
(lb/lO 6 Btu)
(lb/lO 6 Btu) (lb/lO 6 Btu)
1 0.711 0.657 0.054 0.105 0.106 —0.001
2 0.657 0.605 0.052 0.097 0.088 0.009
3 0.687 0.576 0.111 0.082 0.071 0.011
4 0.668 0.590 0.078 0.082 0.075 0.007
5 0.717 0.601 0.116 0.091 0.073 0.018
6 0.663 0.621 0.042 0.094 0.102 —0.008
7 0.603 0.572 0.031 0.111 0.105 0.006
8 0.623 0.624 —0.001 0.112 0.114 —0.002
9 0.702 0.718 —0.016 0.112 0.123 —0.011
EXj 0.467 — — 0.029
0.618 0.052 0.095 0.003
1C1 95 1 0.035 0.007
Relative
Accuracy (%) 14.1 10.5
-------
DATA AND RESULTS 1
TABLE 22.
SFGT NO MONITORS:
Inlet Monitor
Outlet Monitor
Difference
Monitor
Method
(ppm)
Monitor
Method 7
Difference
Test No. (ppm)
(ppm)
(ppm)
(ppm)
(ppm)
1 433 395 38 68.7 79.4 —10.7
2 408 394 14 94.7 79.2 15.5
3 438 391 47 71.9 50.6 21.3
4 395 334 61 70.0 38.6 31.4
5 398 345 53 48.0 35.1 12.9
6 427 425 2 42.3 40.7 1.6
7 387 357 30 48.3 34.8 13.5
8 313 299 14 65.0 34.5 30.5
9 357 351 6 28.3 20.6 7.7
— — 265 — — 123.7
I I 366 29.44 45.9 13.7
IC 9 st 16.56 10.29
Relative
Accuracy (%) 12.6 - _________________ ____
‘All concentrations are expressed on a wet basis.
-------
SFGT SO 2 MONITOR: RELATIVE ACCURACY TEST DATA AND RESULTS’
TABLE 23.
Monitor
Inlet Monitor
Method 7
Difference
Monitor
Outlet Monitor
Method 7
Difference
Test No. (ppm)
(ppm)
(ppm)
(ppm)
(ppm)
(ppm)
1 2412 2207 205 202 192 10
2 2336 2162 174 198 189 9
3 2408 2145 263 210 205 5
4 2440 2765 175 206 203 3
5 2436 2103 333 144 148 —4
6 2360 2120 240 120 119 1
7 2460 2223 237 186 224 —38
8 2472 2278 194 182 164 18
9 2444 2048 396 158 141 17
_______________________________________________________________
EX 1 — — 2217 — — 21
II 2172 246.3 176.1 2.3
101951 57.7 12.9
Relative
Accuracy (%) 14.0 8.6
‘All concentrations expressed on a wet basis.
-------
It should be noted that the results of the SFGT relative accuracy tests
are expressed as ppm instead of mass emissions of unit heat input to the boiler.
This was due to the fact that there was no diluent monitor used at the SFGT
pilot plant. While these results do not strictly conform to the performance
specifications, they are still valid in terms of establishing process per—
formance. This is due to the fact that the SFGT pilot plant reactor operated
under positive pressure and thus no dilution occurred across the reactor.
Thus the measured removals represent actual removals and not apparent changes
in concentration due to dilution.
c—2 8
-------
APPENDIX D
ECONOMIC PREMISES AND EQUIPMENT
COSTS USED FOR COST ESTIMATE
SECTION 1
INTRODUCTION
This appendix presents the economic premises developed by TVA and used
as a basis for the cost estimate prepared as part of this evaluation. It
also includes the estimated equipment costs which were used to determine the
total capital investment.
ID—i
-------
SECTION 2
ESTIMATED EQUIPMENT COSTS
Table 1 presents the modified estimated equipment costs for the HZ
process and compares those costs with the original TVA estimates. For the
HZ process, the modified equipment cost estimates were developed based on
the ratio of material flowrates or energy requirements and standard exponen-
tial factors which relate equipment cost ratios to equipment size ratios.
For all equipment except the HZ catalyst, the exponents were taken from the
article: “Capital Cost Estimating” by K.M. Guthrie, Chemical Engineering,
March 24, 1969. For the HZ catalyst, the costs were determined by assuming
that no economy of scale is realized for the catalyst. Costs for the air
preheater modifications were based on a previous study completed by Radian
entitled “Amnionium Sulfate and Bisulfate Formation in Air Preheaters”.
As shown in Table 1, most of the equipment items are projected to de-
crease slightly in cost. But, the equipment costs based on the pilot plant
design includes the cost for air preheater modifications. As a result, there
is essentially no change in total equipment costs between the Radian and TVA
estimates.
D-2
-------
TABLE 1. MODIFIED EQUIPMENT COST ESTIMATES
Total Equipment Cost 1979$
Item Quantity Description TVA Radian
Area 1 — NH 3 StoraRe and Inlection
1 2 NH 3 Unloading Compressor 50,000 46,000
2 8 NH3 Storage Tanks 230,000 216,000
3 2 NH3 Vaporizer 36,800 34,400
4 2 NH3 Air Mixture Blower 13,500 12,600
Sub—Total Area 1 330,400 309,500
Area 2 — Reactor Section
1 2 Reactor 1,029,000 1,005,200
2 Catalyst 6,370,000 5,125,000
Sub—Total Area 2 7,399,000 6,130,200
Area 3 — Flue Gas Fans
1 4 Flue Gas Blower 405,000 300,900
Sub—Total Area 3 405,000 300,900
Area 4 — Air Preheater Modifications
1 2 Modified Air Preheater 1,243,400
Sub—Total Area 3 0 1,243,400
TOTAL, Areas 1—4 8,134,400 8,006,700
D-3
-------
SECTION 3
ECONOMIC PREMISES
This section presents the economic premises developed by TVA and used
as a basis for the cost estimates prepared as part of this study.
Capital Investment
Capital investment estimates were based on a midwestern plant location
and represent projects beginning in mid—1977 and ending in mid—1980, with an
average cost basis for scaling of mid—1979. Direct investments were calculated
using the average annual Cheniicai Engineering cost indices for the period up
to 1975 and the recent TVA projections for the period 1976—1981 shown in
Table 2. Actual equipment cost estimates are based on 1978 cost information
obtained from engineering—contracting, processing, and equipment companies.
Costs related to equipment, material, and construction—labor shortages
with accompanying overtime pay incentive and costs for the generation
facilities for electricity used by the NO removal processes and FGD unit
were not included.
TABLE 2. COST INDICES AND PROJECTIONS
Year
1970
1971
1972
1973
1974
1975
1976 a
1977 a
1978 a
1979 a
19808
1981 a
Plant b
Material
LaborC
125.7
123.8
137.4
132.3
130.4
146.2
137.2
135.4
152.2
144.1
141.9
157.9
165.4
171.2
163.3
182.4
194.7
168.6
197.9
210.3
183.8
214.7
227.1
200.3
232.9
245.3
218.3
251.5
264.9
237.9
271.6
286.1
259.3
293.3
309.0
282.6
8 Pro)ections.
bsame as index in Chenrtca Engineeri.ng for “equipment, machinery, supports.”
CS as index in Chem a1 Engineering for ‘construction labor.”
D-4
-------
Direct Investment— —
The purchase cost of the process equipment and the cost of materials and
labor for the installation of this equipment were included. Also, the cost
of piping and insulation, ductwork, excavation, site preparation, foundation,
structural, roads and railroads, electrical, instrumentation, buildings,
painting, and services required for each unit area were estimated. The costs
for the HZ process was estimated based on the flow diagram and equipment lists.
Services, utilities, and miscellaneous costs were estimated at 6% of the
total direct investment. This expense covers allocated costs for the use of
such power plant facilities as maintenance shops, stores, communications,
security, and offices. Parking lots, walkways, landscaping, fencing, and
vehicles are also included in the service facility estimate.
Indirect Investment— —
The indirect investment Includes costs for engineering design and
supervision, architect and engineering contractor expenses, construction
expenses, contractor fees, and contingency. The engineering design and
supervision and contingency factors are based on proven design, not first—of—
a—kind installation.
Engineering design and supervision——This indirect investment factor was
estimated using a technique that correlates the number of major pieces of
process equipment with drafting room man—hour and engineering design costs.
Architect and engineering (A&E) contractor expenses——This cost was based
directly upon the engineering design and supervision costs. A&E expense was
assumed to be 25% of the portion of engineering design and supervision costs
associated with major equipment.
D-5
-------
Construction expense——Construction expense was estimated as a function of
the direct investment using the following equation.
0.83
Construction expense = 0.25 (x)
where x = direct investment in $ x 106
Construction facilities (which include costs for mobile equipment, temporary
lighting, construction roads, raw water supply, safety and sanitary facilities,
and other similar expenses incurred during construction) are considered a
part of construction expenses.
Contractor fees——The relationship between contractor fees and total
direct investment used to estimate the cost of contractor fees was:
0.76
Contractor fees = 0.096 (a)
where a = direct investment in $ x i 6
Contingency——The contingency was assumed to be 20% of the sum of total
direct investment, engineering design and supervision costs, A&E expenses,
construction expenses, and contractor fees.
Other Capital Charges— —
The sum of the total direct and total indirect investment is total
fixed investment. Other capital charges which must be added to this total
fixed investment to obtain the total capital investment are: (1) allowance
for startup and modification, (2) interest during construction, (3) land,
(4) working capital, and (5) royalty fees.
Allowance for startup and modifications——This expense was estimated to
be 10% of the total fixed investment.
D- 6
-------
Interest during construction——This item was estimated to be 12% of the
total fixed investment. This percentage is calculated as the simple interest
which would be accumulated at a 10%/yr rate assuming a debt—equity ratio
of 60:40 on the incremental capital investment and a 3—year project expendi-
ture schedule as shown in Table 3.
Land——The cost of land was estimated at $3500/acre.
Working capital——Working capital consists of: (1) money invested in
raw materials, supplies, and finished products carried in stock and semi-
finished products in the process of being manufactured; (2) accounts receivable;
(3) cash retained for payment of operating expenses, such as salaries, wages,
and raw material purchases; (4) accounts payable; and (5) taxes payable.
For these premises, working capital was defined as the equivalent cost of
3 weeks of raw material costs, 7 weeks of direct costs, and 7 weeks of overhead
costs. The raw material and direct costs do not include the costs of catalyst
replacement.
TABLE 3. PROJECT EXPENDITURE SCHEDULE
Year 1 2 3 Total
Fraction of total expenditure as
borrowed funds 0.15 0.30 0.15 0.60
Simple interest at 10%/yr as % of
total expenditure
Year 1 debt 1.5 1.5 1.5 4.5
Year 2 debt — 3.0 3.0 6.0
Year 3 debt 1.5 1.5
Accumulated interest as % of
total expenditure 1.5 4.5 6.0 12.0
Royalty fees——A royalty fee of $300,000 was assumed for the HZ process.
This fee is on the same order as that associated with similar processes.
D— 7
-------
Revenue Requirements
Annual revenue requirement calculations are based on 7000 hours of
operation per year using mid—1980 operating costs and an average capital charge.
Process operation schedules are assumed to be the same as the power plant
operating profiles.
Direct Costs— —
Direct operating costs include raw material, labor, utility and mainten-
ance costs. The 1980 projected unit costs for the raw materials were for
delivered costs to a midwestern location.
Projected 1980 utility costs are shown in Table 4. Unit costs for
electricity and steam generated by the power plant are based on actual pro-
duction cost including labor, fuel, depreciation, taxes, and rate base return
on investment. The electricity rates are based on purchase from an independent
source with full capital recovery provided and are adjusted for the quantity
used.
Maintenance costs were estimated as a percentage of the direct invest-
ment. The maintenance factor depends on operating characteristics of the
process and was estimated based on either actual operating experience or
maintenance needs in similar process areas. The estimated maintenance factor
for the HZ process is 4 percent of indirect investment.
TABLE 4. PROJECTED 1980 UNIT COST FOR UTILITIES
Utility
Cost, $
Steam
2.00/MBtu
Electricity
0.O29/k Th
D- 8
-------
Indirect Costs— —
Indirect costs include capital charges, projected to 1980, and overheads.
Following power industry practice, regulated company economics were used for
calculating the capital charges. A breakdown of the capital charges is given
in Table 5. The depreciation rate is straight line based on the life of
the power plant.
TABLE 5. ANNUAL CAPITAL CHARGES FOR POWER INDUSTRY FINANCING
Percentage of
total depreciable
capital investment
Years Remaining Life 30
Depreciation—straight line (based on
years remaining life of power unit) 3.3
Interim replacements (equipment having
less than 30—year life) 0.7
Insurance and property taxes 2.0
Total rate applied to original investment 6.0
Percentage of
unr ecover ed
capital investmenta
Cost of capital (capital structure
assumed to be 60% debt and 40% equity)
Bonds at 10% interest 6.0
Equityb at 14% return to stockholder 5.6
Income taxes (Federal and State)C 5.6
Total rate applied to depreciation base
aoriginal investment yet to be recovered or “written off”.
bContains retained earnings and dividends.
CSince income taxes are approximately 50% of gross return, the amount of
taxes is the same as the return on equity.
dApplied on an average basis, the total annual percentage of original fixed
investment for new (30 year) plants would be 6.0% + 1/2 (17.2%) = 14.6%.
D—9
-------
Capital Charges——In estimating the regulated capital charges associated
with FGT, the conventional method of considering the overall life of the power
plant was used. The Federal Power Commission (FPC) recognizes the conclusion of
the National Power Survey that a 30—year service life is reasonable for steam—
electric plants. Because some life spans are less than 30 years, however, FPC
has designated interim replacements as an allowance factor to be used in
estimating annual revenue requirements. Use of this allowance following FPC—
recommended practice provides for financing the cost of replacing such short—
lived units. An average allowance of about 0.35% of the total investment is
normally provided. However, to provide for the unknown life span of S02—N0
control facilities, a larger allowance factor of 0.70% is used for new units.
An insurance allowance of 0.5% is also included in the capital charges based
on FPC practice. Property taxes are estimated at 1.5% of the total depreciable
capital investment.
Debt—equity ratio is another component of capital charges for which
variations of ratios may be expected. FPC data indicate that the long—term
debt for privately owned electric utilities varied only slightly from 51.5
to 54.8% of total capitalization during the period 1965—1973. However, recent
economic trends have changed the incremental debt—equity ratio because
utilities are more dependent on bonds and bank loans for project funding. For
these economic premises the capital structure was assumed to be 60% debt and
40% equity. The interest rate for bonds was assumed to be 10% and the return
to stockholders on equity 14%. Costs of capital and income tax charges were
applied to the uncovered portion of capital investment. Income taxes and
return on equity are each about 50% of gross return. Since return on equity
is 5.6% of total capital investment, income taxes are also 5.6% of total
capital investment. Since most regulatory commissions base the annual permIs-
sible return on investment on the remaining depreciation base (that portion
of the original investment yet to be recovered or “written off”) a portion
of annual capital charge included in the lifetime operating costs declines
uniformly over the life of the power plant.
D— 10
-------
Overheads——Plant and administrative overheads vary from company to
company. Based on the various methods used in industry and illustrated in
a variety of cost estimating sources, the following method of estimating
overheads was used. Plant overhead was estimated as 50% of the subtotal
conversion costs less utilities, which include the projected costs for labor,
maintenance, and analyses. Administrative overhead was estimated as 10% of
operating labor and supervision.
Spent Catalyst Disposal— —
For the HZ process, the catalyst base support has scrap metal value.
Thus, this scrap value yields a credit toward the annual revenue requirement.
D-l1
------- |