United States Office of Air Quality EPA-340/1 -83-012
Environmental Protection Planning and Standards January 1983
Agency Research Triangle Park NC 27711
Stationary Source Compliance Series
A Compilation
of SO2 and IMOx
Continuous
Emission
Monitor
Reliability
Information
-------
EPA-340/1-83-012
A Compilation of SO2 and NOx Continuous
Emission Monitor Reliability Information
Prepared by:
James W. Peeler
Entropy Environmentalists, Inc.
Research Triangle Park,
North Carolina
Prepared for:
Louis R. Paley
Stationary Source Compliance Division
and
Anthony Wayne
Region VII
United States Environmental Protection Agency
SSCD Contract No. 68-01-6317
and
Darryl von Lehmden
and
Thomas Logan
Quality Assurance Division
United States Environmental Protection Agency
QAD Contract No. 68-02-3431
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Air Quality Planning and Standards
Stationary Source Compliance Division
Washington, D.C. 20460
January 1983
U.Z. Environmental Protection Agency
;ic-i ;;, Library (5PL-16)
'•:•'. j S, Dearborn Street, Room 1670
^icago, IL 60604
-------
The Stationary Source Compliance series of reports is issued by the
Office of Air Quality Planning and Standards, U.S. Environmental
Protection Agency, to assist Regional Offices in activities related to
compliance with implementation plans, new source emission standards,
and hazardous emission standards to be developed under the Clean Air
Act. Copies of Stationary Source Compliance Reports are available -
as supplies permit - from Library Services, U.S. Environmental
Protection Agency, MD-35, Research Triangle Park, North Carolina
27711, or may be obtained, for a nominal cost, from the National
Technical Information Service, 5285 Port Royal Road, Springfield,
Virginia 22151-
This report has been reviewed by the Office of Air Quality Planning
and Standards, U.S. Environmental Protection Agency, and approved for
publication as received from Entropy Environmentalists, Inc. Approval
does not signify that the'contents necessarily reflect the views and
policies of the U.S. Environmental Protection Agency, nor does mention
of trade names or commercial products constitute endorsement or
recommendation for use.
ii
-------
EXECUTIVE SUMMARY
As air pollution control agencies emphasize the implementation of
continuous emission monitoring (CEM)* programs and effective utilization of
CEM results, the reliability of the monitoring data (i.e., the availability,
accuracy, precision, and representativeness of monitoring results) increases
in importance for both the affected source owner/operator and the control
agency. This report presents the latest information on the reliability of
S02 and NOX CEM results, i.e.: (1) achievable CEM availability, (2)
point-in-time CEM accuracy, and (3) the long-term capability of CEMs to
provide accurate emission data.
The information in current literature shows that the CEM availability
(data capture rate) levels are between 6? and 95 percent for SC>2 and NOX
CEMs. CEM availability is expected to vary on source-specific and
source-category bases and to increase with additional operational experience.
Analysis of more than 119 relative accuracy tests (RATs) of S02 and
NOX CEMs shows that the four most common commercially available SC>2 and
NOX CEMs are capable of meeting the 20 percent "point-in-time" relative
accuracy specification (both on a concentration and lb/10° Btu basis) on
coal- and oil-fired steam generators with, and without wet FGD systems.
The results of numerous RATs conducted after the initial demonstration of
compliance with the Performance Specifications show that both S02 and NOX
CEMs continue to be capable of obtaining accurate data on a long-term basis
under a wide variety of source conditions. No general deterioration in the
S02 and NOX CEMs1 accuracy was apparent during the 24-month period
following the initial demonstration of compliance.
* Throughout this document, the acronym CEM is used to mean both "continuous
emission monitor" and "continuous emission monitoring." The specific
connotation should be clear from the context in which it is used.
iii
-------
TABLE OF CONTENTS
I. Introduction 1
II. S02 and NOX CEM Availability 3
Background 3
CEM Availability Estimates 5
References g
III. S02 and NOX CEM Accuracy 9
Background 9
Compilation of S02 and NOX RAT Results 12
Long-Term CEM Performance 16
APPENDIX 23
-------
I. INTRODUCTION
The Environmental Protection Agency and many State air pollution control
agencies are currently expanding the scope and implementation of S02 and
NOX continuous emission monitoring programs for stationary sources. Greater
importance is being placed on CEM results for evaluating the adequacy of the
operation and maintenance practices of source processes and control systems.
In some cases, such as those sources subject to the recently promulgated NSPS
(40 CFR 60, Subpart Da) requirements, S02 and/or NOX CEM results are used
to determine compliance with emission limitations. Clearly, the increased
utilization of CEM reflects control agency efforts to achieve sustained
emission reductions from stationary sources.
As control agencies emphasize the implementation of CEM programs and
effective utilization of CEM results, the importance of obtaining reliable
monitoring data increases for both the affected source owner/operator and the
control agency. More specifically, the level of long-term monitor performance
in terms of CEM availability, accuracy, precision, and representativeness is
of fundamental importance.
Until recently, there has been a lac^c of long-term monitor performance
evaluations. Although many CEMs have met the applicable Performance
Specification requirements during an initial performance test, very few
monitoring systems have been retested to demonstrate long-term accuracy. In
addition, various CEM users have reported widely differing values of CEM
availability (data capture rates). However, only very limited quantitative
information has been available to determine the causes of CEM unavailability
(i.e., unreliable CEM equipment, inappropriate applications of CEMs, and/or
the inexperience of CEM users).
-------
This report presents additional, recently available information about the
long-term reliability of S02 and NOX CEM results. The term "CEM
reliability" is considered to be the degree to which source operators and
control agencies may depend on CEM data to yield a consistent and valid
measure of S02 and NOX emission levels. This report addresses (1)
achievable CEM availability, (2) point-in-time CEM accuracy, and (3) long-term
capability of CEM to provide accurate data. The results and conclusions in
this report are based on CEM availability values reported in current
literature, numerous performance specification test results, and a number of
performance audit results. This document does not attempt to specify
achievable CEM availability and accuracy.
-------
II. S02 and NOX GEM AVAILABILITY
This section presents background information on CEM availability,
including definitions of terminology and a brief discussion of the major CEM
availability factors. Several estimates of availability are discussed for
SC>2 and NOX monitoring systems installed at electric utility steam
generators. These estimates bracket the expected range of CEM availability
for contemporary gas monitoring programs.
Background
The term "CEM availability" is the data capture rate exhibited by a
continuous emission monitoring system. It is defined as the percentage of
time a monitoring system is operating (i.e., sampling, analyzing, and
recording emission data) relative to the total time the system is required to
operate. This percentage does not incorporate the accuracy or precision of
the CEM data.
The availability of S02 and NOX CEM data affects the characterization
of emission levels and/or process and control system performance at a given
source. No CEM system can be expected to, operate continuously over an
extended period of time. Unforeseen malfunctions of various components and
necessary routine maintenance and repairs will reduce the availability.
Quantification of the achievable level of CEM availability will aid the
control agency and the source operator in determining accurately both
acceptable monitor performance and when to initiate corrective action.
-------
CEM availability is dependent on a variety of factors, including the
following.
(1) Reliability of specific instrumentation,
(2) Instrument maintainability and availability of spare
parts.
(3) User's operational, repair, and trouble-shooting expertise
(and/or the availability of service from the monitor
vendor).
(4) Environmental conditions (exposure to weather, corrosive
gases, and/or dusty environments, etc.).
(5) Site conditions (accessibility, adequacy of power supply
and other utilities, presence of mechanical vibration,
existence of microwave and/or radio frequency
interference, etc.).
(6) Effluent conditions (temperature, pressure, flow rate,
particulate loading, etc.).
7. User's dedication of manpower and prioritization of repair
efforts during periods of monitor malfunction.
In view of the number of factors that affect CEM availability, the
infinite combinations of these factors, and the varying degree of severity of
specific problems encountered in actual practice, CEM availability must be
expected to vary over time and from source to source. CEM availability is
expected to increase over time on an industry-wide basis as unreliable
instrumentation is excluded from the competitive CEM market, as unsuccessful
approaches to the installation and operation of CEMs are identified, and as
monitor operators gain necessary experience and develop appropriate quality
control programs and cost effective preventive maintenance programs. Because
of these time-dependent factors, the averaging period of all CEM availability
estimates must be clearly identified to characterize adequately the many
problems that detract from CEM availability.
-------
GEM Availability Estimates
All of the factors discussed above increase the difficulty and decrease
the usefulness of developing "average" CEM availability percentages to
represent the level of monitor performance that can be reasonably achieved. A
more realistic approach is to establish an expected or acceptable range of CEM
availability values.
One notable attempt to characterize CEM availability was conducted by the
Utility Air Regulatory Group (UARG) in response to the EPA September 18, 1979
proposal of Subpart Da requirements.^ During the fall of 1978, UARG surveyed
71 electric utility companies. Survey responses from 63 companies provided
information on a total of 539 emission monitoring channels (S02, NOX, 02,
C02, and opacity) installed at 237 generating units throughout the U.S. In
this survey, in which monitor availability is expressed on a monthly basis, 55
responses indicated an average availability of 67 percent for S02 monitoring
channels, 50 responses indicated an average availability of 67 percent for
NOX monitoring channels, and 85 responses indicated an average availability
of 76 percent for 02 and/or C02 monitoring channels. It is emphasized that
(1) the UARG survey included some monitors installed before the EPA
promulgation of CEM requirements, (2) considerable advancements in CEM
application technology and additional operating experience have been made since
the survey was conducted, and (3) the survey results reflect to an
indeterminate degree the relatively low level of control agency emphasis on the
implementation of CEM programs that existed at the time the survey was
conducted. Thus, although currently achievable CEM availability is expected to
be somewhat greater than that indicated by the "average" results of the 1978
UARG survey, these results are nevertheless indicative of a minimum level of
achievable CEM a/ailability.
-------
CEM availabilities achieved by two utility companies actively involved in
CEM programs for some time are summarized in Table 1. The two companies are
the Texas Utilities Generating Company2 and The Montana Power Company^.
The results over the five year period from 1977 to 1981 show significant
improvement in SC>2 and NOX availability. The improvement is attributed to
increased overall experience of the monitor operators. Table 1 also shows that
five of the six monitoring systems exhibited approximately 95 percent
availability during 1980 and 1981.
A study of CEM data reported to the California South Coast and the Bay
Area Air Quality Management Districts^ provides further estimates of SOp
and NOX CEM availability. This study included data reported from November
1978 to April 1980 for 62 CEMs in the South Coast District, and data reported
from July 1979 to December 1980 for 38 CEMs in the Bay Area District. A total
of 33 SC>2 CEMs had achieved an average availability of 97.2 percent, and a
total of 67 NOX CEMs had achieved an average availability of 96.7 percent.
These high availability figures may be the result of all the CEMs being
installed on gas or oil-fired steam generators with no FGD systems. Therefore,
these results may not accurately represent the achievable CEM availability for
coal-fired boilers and/or units equipped with wet FGD systems.
The results of these three investigations probably represent the ultimate
level of CEM availability; whereas, the results of the UARG 1978 survey
probably represent the minimum acceptable level of CEM availability. The
minimum data collection requirements, which are included in EPA's June 11, 1980
promulgation of Subpart Da (i.e., 18 hours per boiler operating day, and 22
days per 30 operating days, or approximately 75 percent) fall within the CEM
availability range bounded by the above references.
-------
TABLE 1.
REPORTED GEM AVAILABILITY
Station Unit Monitor
Percent Availability
1977 1978 1979 1980 1981
Martin Lake
Martin Lake
Martin Lake
Colstrip 1 and 2
**
Colstrip 1 and 2
**
scu
so.
so.
NO,
86.1 88.6 100 97.7 97.4
90.6 100 97.5 98.6
49 94.1 36.9
84.1 92.5 95.7 96.6 96.2
88.9 90.1 94.8 96.5 95.7
**
Texas Utilities Generating Company,
Montana Power Company
-------
References
1 "Section IV - October/November 1978 Continuous Monitor Survey," An
Evaluation of the Continuous Monitoring Requirements of the September 19, 1978
Subpart Da NSPS Proposal, pp. 72-96.
2 Bill Bonner, "Continuous Emission Monitoring - Texas Utilities Generating
Company's Experience," Proceedings. Continuous Emission Monitoring: Design,
Operation and Experience Specialty Conference, pp. 62-69.
3 David K. Nation, "Continuous Emission Monitoring Experience at Colstrip
Units 1 & 2," Proceedings, Continuous Emission Monitoring: Design, Operation
and Experience Specialty Conference, pp. 25-38.
4 R.P. Hebert, W.C. Kelly, and'w.J Mitchell, "Assessment of the Reliability
of Continuous Emission Monitors Installed in Two California Air Quality
Regions," Proceedings, Continuous Emission Monitoring: Design. Operation and
Experience Specialty Conference, pp. 18-24.
-------
III. S02 and NOX GEM ACCURACY
Background information relevant to the measurement of gas CEM accuracy
and the interpretation of CEM relative accuracy test results is presented in
this section. Also, a compilation of relative accuracy test results is
provided, giving both "point-in-time" CEM accuracies and capabilities of SC>2
and NOV CEMs to achieve the level of performance required by both the
A
existing and proposed relative accuracy specifications of Appendix B (40 CFR
60). Finally, the results of performance audits of S02 and NOX CEMs are
presented and discussed to provide a preliminary evaluation of long-term CEM
accuracy for a number of sources.
Background
The accuracy of CEM data is of critical importance to both the control
agency and the source owner/operator. To protect the interests of both
parties, the CEM data must be sufficiently accurate to ensure valid
conclusions and decisions. Clearly, the uncertainty-level of all conclusions
based on the data decreases as the accuracy of the data improves.
Measurement accuracy is generally defined as the degree of agreement of a
measured value with the "true" value (or the degree of agreement of the mean
of a set of measurements with the corresponding "true" values).
Traditionally, the accuracy of S02 and NOX CEM data has been determined by
conducting relative accuracy tests (RATs) in accordance with the requirements
of Performance Specification 2, Appendix B, 40 CFR 60. The term "relative
accuracy" reflects the fact that the accuracy of the CEM data is determined
relative to the results obtained by performance of Reference Method Test
procedures; the Reference Method results are presumed to be the "true" values
-------
for this test. Historically, RATs have only been conducted during CEM
Performance Specification Tests (PSTs); however, requirements to perform
periodic RATs will probably be mandated as basic elements of most quality
assurance programs. (See "A Compilation of Quality Assurance Procedures for
SC>2 and NOX Continuous Emission Monitoring Systems.") Also RATs may be
conducted by the control agency as a central part of SC>2 and NOX CEM
performance audit programs.
To place the interpretation of CEM RAT results into the proper
perspective, one must clearly understand how several procedural test
l
requirements and a number of constraints affect the representativeness of the
test results. According to the existing Performance Specification 2 (Appendix
B, 40 CFR 60), the relative accuracy of S02 and NOX CEM data is determined
in units of concentration (ppm) from a series of nine measurements using
Method 6 for S02 and Method 7 for NOX. (Concurrent moisture sampling is
also conducted, if needed, to adjust either the wet basis CEM data or the dry
basis Reference Method data to the same moisture basis.)
The relative accuracy of S02 and NOX CEMs is computed as the sum of
(1) the absolute value of the mean from the differences between the 9 pairs of
concurrent CEM and Reference Method results and (2) the 95 percent confidence
interval associated with the observed differences. This sum is divided by the
mean Reference Method value in order to express the relative accuracy as a
percentage. The relative accuracy calculated using this procedure is actually
expressed in terms of error; smaller relative accuracy values indicate better
monitor performance. Performance Specification 2 requires the result of the
RAT to be less than, or equal to, 20 percent for acceptable CEM performance.
Relative accuracy is affected by errors in the CEM data and the Reference
Method data. The Reference Methods are not totally precise because of the
10
-------
inherent variability and the expertise and ability of the tester. The
confidence interval of the RAT determination reflects the precision among the
paired GEM - Reference Method results; i.e., it is the sum of (1) the
imprecision of the CEM data, (2) the inherent variability of the Reference
Method results, and (3) the ability of the tester to conduct the Reference
Method tests. For a particular RAT, the individual errors cannot be
quantified. Therefore, the cause of an excessive confidence interval value
(e.g., poor testing practices or poor CEM performance) cannot be objectively
determined. Thus, while meeting the relative accuracy specification
demonstrates that a CEM has provided accurate data during the test, failure tc
do so does not necessarily demonstrate that a CEM has provided inaccurate
data.
The RAT covers an approximate nine-hour period and represents only the
accuracy of the CEM data during that test period and at the effluent
conditions encountered during the RAT. Thus, the RAT provides a
"point-in-time" measure of CEM. accuracy. A successful RAT only demonstrates
that the CEM is capable of obtaining sufficiently accurate data for its
intended use. The CEM accuracy actually achieved on a day-to-day basis is
primarily dependent upon (1) the validity,of the calibration procedure
employed for the CEM, and (2) the manner in which the CEM operator performs
the daily calibration procedure.
As a final note regarding the interpretation of RAT results, it is
important to recognize that revisions to Performance Specifications 2 and 3
proposed first in the October 10, 1979, Federal Register, and subsequently
reproposed in the January 26, 1981, Federal Register, significantly change th
scope of the RAT. The presently promulgated specifications require that the
relative accuracy of pollutant SC>2 and NOX monitors be determined in units
of concentration (ppm). There are no RAT requirements on the diluent (02 or
11
-------
COp) monitor which is used to convert pollutant concentration measurements
to units of the standard (lb/106 Btu) at fossil fuel-fired steam
generators. The proposed revisions require that the RAT determination be
conducted in units of the standard (Ib pollutant/106 Btu) and thereby,
provide a measure of the combined pollutant-diluent monitoring system
accuracy. The proposed revisions provide for evaluation of the accuracy of
the total CEM system (both pollutant and diluent monitors) in the same units
of measurement that are used for reporting excess emissions to the control
agency.
Compilation of S02 and NOY RAT Results
The results of M1 RATs (in units of concentration) of 22 installed S02
CEMs and the results of 25 RATs (in units of concentration) of 15 installed
NOV CEMs are tabulated in the Appendix of this report. In addition, the
A
results (expressed in units of lb/106 Btu) of 34 S02 RATs and the results
of 19 NOX RATs are provided for a subset of the same population of
monitors. Table 2 summarizes the results of these tests in terms of meeting
the existing and proposed relative accuracy specifications.
In evaluating the RAT results provided in the Appendix and summarized in
Table 2, the following qualifications of the data base should be kept in mind:
(1) All tests were performed at coal- and oil-fired steam
generators; some of the coal-fired installations were
equipped with wet FGD systems.
(2) All RATs were performed either during the initial
demonstration of compliance with Performance Specifications
2 and 3 or during announced performance audit programs
conducted by various control agencies.
12
-------
TABLE 2.
SUMMARY OF SO2 AND NOX RELATIVE ACCURACY TEST RESULTS
Monitor Pollutant
LSI SM810
DuPont 460
Contraves Goerz
GEM-100
EDC DIGI 1400
CSI Monitor
All Monitors
so2
S02
NOX
NOX
SO2
S02
NOX
NOX
so2
S02
N°x
N0x
so2
S02
NOX
NOX
so2
SO2
SO2
NOX
NOX
FGD
Controls
NO
YES
NO
YES
NO
YES
NO
YES
NO
YES
NO
YES
NO -
YES
NO
YES
NO
NO
YES
NO
YES
Existing RA Requirements
No. of No. of % of Tests
Monitors Tests <_20% RA
7
2
4
2
6
1
5
0
2
1
2
0
1
1
2
0
1
17
5
13
2
11
9
6
9
6
3
5
0
3
4
3
0
1
2
2
0
2
23
18
16
9
91%
89%
50%
100%
100%
67%
100%
— — —
67%
50%
67%
— — ~
100%
100%
50%
— ™ —
100%
91%
78%
69%
100%
2
Proposed RA Requirements
No. of No. of % of Tests
Monitors Tests ^20% of RA
6
2
3
2
6
1
5
0
1
1
1
0
1
1
2
~"
1
15
5
11
2
9
7
4
7
6
3
5
0
1
4
1
0
1
1
2
~~
2
19
15
12
7
100%
86%
25%
100%
100%
100%
100%
— — —
100%
50%
100%
_. _ —
0%
100%
__ _ —
100%
95%
80%
75%
100%
Performance Specification 2, Appendix B, 40 CFR 60 (RA test performed in units of pollutant
concentration - ppm)
2 Proposed revisions to Performance Specifications 2 and 3, Federal Register, Vol. 46, No. 16
26. 1981 (RA test performed in units of the standard - lbs/10b Btu)
-------
(3) Source maintenance personnel and/or monitor manufacturer
service representatives are known to have made adjustments
to the subject CEMs prior to the RATs in some cases.
In view of the above, the RAT results may not necessarily represent the
performance of similar CEMs under dissimilar conditions. In addition, valid
statistical inferences regarding the ability of randomly selected CEMs to meet
relative accuracy specifications cannot be derived from the data base.
Within the limitations of the available data base, analysis of the
/
included RAT results indicates:
(1) There is no apparent significant difference between the
capabilities of S02 and NOX CEMs to meet RAT
specifications. (NOX CEMs may have a slightly higher
failure rate; however, this finding may be due to the
limited data.)
(2) There is no apparent significant difference between the
capabilities of CEMs installed at oil- and coal-fired
facilities to meet RAT specifications.
(3) There is no apparent significant difference between the
capabilities of CEMs installed at FGD equipped and non-FGD
equipped facilities to meet RAT specifications. (S02
monitors did perform better at non-FGD sources; however,
NOX monitors generally performed better at FGD equipped
sources. Again, these results may be attributable to the
limited data.)
14
-------
(4) CEMs manufactured by Lear Siegler, DuPont, Contraves Goerz,
Environmental Data Corporation, and Columbia Scientific
Industries have been shown to be capable of meeting RAT
requirements of £20 percent, both in units of concentration
(ppm) and in units of the standard (lb/106 Btu).
(5) Of the 53 S02 and NOX RATs that had results in both .
units of concentration (ppm) and units of the standard
(lb/106 Btu), (1)6 CEMs failed to meet both the
concentration and system relative accuracy specifications,
(2) 3 CEMs failed only the concentration relative accuracy
specification, and (3) 2 CEMs failed only the system
relative accuracy specification. Therefore, changing the
RAT specifications from the existing concentration basis to
the proposed system basis would affect the status of only 5
CEMs (approximately 9 percent of the population) with
respect to compliance with the relative accuracy
specification.
(6) The minimum 95 percent confidence interval value observed
in the results of 41 SC>2 RATs in units of concentration
was 1.4 percent of the mean SC>2 concentration value; the
minimum 95 percent confidence interval value observed in
the results of 34 S02 RATs in units of lb/106 Btu was
2.2 percent of the mean SC>2 emission value.
(7) The minimum 95 percent confidence interval value observed
in the results of 25 NOX RATs in units of concentration
was 2.0 percent of the mean NOX concentration value; the
miniraun 95 percent confidence interval value observed in
the results of 19 NOX RATs in units of lb/106 Btu was
1.3 percent of the mean NOX emission value.
15
-------
Long-Tertn GEM Performance
The results of RATs conducted for 20 S02 CEMs at varying time intervals
after the initial successful demonstration of compliance with Performance
Specification 2 are shown in Figure 1. The same type of results for 7 NOY
X
CEMs are shown in Figure 2. For some of the S02 and NOX CEMs, more than
one RAT was conducted subsequent to the initial PST; at one source, a total of
7 RATs have been conducted over a period of 42 months.
In evaluating the long-term RAT results, the following qualifications
should be kept in mind:
(1) All of the S02 and NOX CEMs are installed at coal-fired
steam generators except for the 2 NOX CEMs installed at
oil-fired facilities. Five of the S02 CEMs and two of
the NOX CEMs are installed on sources equipped with wet
FGD systems.
(2) All of the RATs performed after the initial demonstration
of compliance with Performance Specifications 2 and 3 were
announced CEM performance audits.
(3) Source maintenance personnel and/or monitor manufacturer
service representatives are known to have inspected and/or
made adjustments to the subject CEMs prior to the RATs in
eight cases.
(4) In many cases, the initial RAT results were determined in
units of concentration; these results are used even though
all subsequent RAT values were determined in units of
lb/106 Btu.
16
-------
FOR SO2 CEM PERFORMANCE AUDITS
CONDUCTED AFTER SUCCESSFUL PERFORMANCE SPECIFICATION TESTS (PSTs)
Monitor
Type
1
2
3
4
-v
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
LSI
LSI
LSI
DuPont
Contraves
Goerz
EDC
DuPont
CSI
LSI
LSI
EDC
LSI
LSI
LSI
LSI
DuPont
LSI
LSI
Contraves
Goerz
Contraves
Goerz
FGD
Yes
Yes
No
Yes
Yes
Yes
No
No
No
No
No
No
No
No
No
No
No
No
No
No
•
4
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42
Pass;Pass
PassjPass
LSI = Lear Siegler SM810; DuPont = DuPont 460; Contraves Goerz = Contraves Goerz GEM-100;
CSI = Columbia Scientific Industries; EDC = Environmental Data Corporation DIGI 1400.
-------
FOR NOX CEM PERFORMANCE AUDITS CONDUCTED
AFTER SUCCESSFUL PERFORMANCE SPECIFICATION TESTS (PSTs)
3
r
J
Monitor
Typo
J.JO 1
LSI
LSI
LSI
hiUC
EDC
LSI
FGD
Vt\ c
Yes
No
NO
Mc-v
NO
NO
0 2 3 4 5 f, 7 H 9 10 11 12 13 14 ] 5 16 17 18 19 20 21 22 23 24 25
Pa*"1" i** pQ'jj^^PLiG" ^^ Friil >»- Pass
LSI - Loar Sicqlor SMfUO; EDC = Environmental Data Corporation DIGI 1400.
-------
(5) All RATs performed after the initial PST were conducted by
personnel from Entropy Environmentalists, Inc. in
accordance with all applicable regulatory testing
requirements and in accordance with Entropy's internal
quality assurance program. Thus, a significant level of
experimental control is applicable to all test results
obtained after the initial PST.
The non-routine adjustments of the subject CEMs made just prior to the
performance of the announced RATs may have influenced the level of monitor
performance achieved. These non-routine adjustments and services probably
increased the apparent capability of the CEMs to obtain valid data. The fact
that all RATs performed after the initial PST were conducted by essentially
the same testing personnel ensured the consistency of the test results and
eliminated interlaboratory- variance effects.
Overall, the gas CEM performance audits provided the following results:
(1) The 20 installed S02 CEMs met the applicable relative
accuracy specification for 25 of the 33 audits conducted (H
of the 8 relative accuracy test failures were consecutive
tests of the same monitoring system).
(2) 15 of the 20 S02 CEMs audited passed all of the relative
accuracy tests.
(3) The 7 installed NOX CEMs met the applicable relative
accuracy specification for 9 of the 13 audits conducted (2
of the 4 relative accuracy test failures were consecutive
tests of the same monitoring system) .
(4) 4 of the 7 NOX CEMs audited passed all of the relative
accuracy tests that were conducted.
20
-------
In evaluating the above results, it is important to note that: (1) there
are no promulgated quality assurance requirements applicable to the audited
CEMs, and (2) there are no promulgated requirements for the GEM operators to
periodically test installed CEMs or to take necessary corrective actions where
unacceptable performance is observed.
Although other results can be derived from the long-term CEM accuracy
data obtained to date, there is insufficient information upon which to base
any significant statistical analysis. Additional RATs of these and other
installed SC>2 and NOX CEMs are currently being planned and conducted. The
results of these future tests will enhance the existing data base, both in
terms of the number of monitors represented and the length of time for which
results are available for the same monitors. However, the inclusion of these
additional results is not expected to dramatically affect the overall
characterization of CEM reliability.
21
-------
APPENDIX
RELATIVE ACCURACY
TEST RESULTS
-------
AVERAGE MONITOR PERFORMANCE DATA
Monitor
LSI
LSI
LSI
LSI
DuPont
DuPont
DuPont
DuPont
DuPont
Contraves
Contraves
Contraves
EDC
EDC
EDC
CSI
Fuel
Coal
Coal
Coal
Coal
Coal
Coal
Oil
Coal
Oil
Coal
Coal
Coal
Coal
Coal
Oil
Coal
Pollutant
Gas
so2
SO2
N0x
NOX
so2
SO2
S02
NOX
NOX
so2
SO2
NOX
SO2
so2
NOX
so2
FGD
System
YES
NO
YES
NO
YES
NO
NO
NO
NO
YES
NO
NO
YES
NO
NO
NO
Pollutant
Concentration
Relative
Accuracy
(%)
13.9
11.5
11.1
17.2
16.7
10.1
10.1
16.6
13.3
30.6
16.4
17.5
14.1
18.0
13.2
8.9
95~%
Confidence
Level
(%)
4.6
4.2
5.4
6.1
11.9
5.5
3.1
4.0
5.2
9.0
3.5
3.7
4.6
3.5
4.9
4.0
System
[lbs/106 Btu]
Relative
Accuracy
(*)
13.4
12.3
13.9
24.4
11.2
10.8
13.4
16.4
14.4
47.3
7.5
13.1
11.9
32.0
12.5
9.1
95%
Conf idenc
Level
(%)
5.2
4.1
5.3
7.2
7.2
6.0
3.6
3.8
5.1
12.0
4.6
7.8
5.5
6.7
6.5
4.1
A-2
-------
LSI MONITOR PERFORMANCE
SOURCES WITH PGD EMISSIONS CONTROL SYSTEMS
ource
^ode
^-l
3-1
?-l
?-l
?-l
?-l
?-2
?-2
?-2
?-2
?-2
?-2
3-2
3-2
?-l
?-l
?-2
3-2
Test P
Date
06/80
07/81
12/81
06/80
07/81
12/81
06/80
05/81
07/81
12/81
06/80
05/81
07/81
12/81
05/82
05/82
04/82
04/82
ollutant
Gas
S02
SO2
so2
NOX
NOX
NOX
so2
SO 2
SO2
S02
NOX
NOX
NOX
NOX
so2
NOX
S02
NOX
Fuel F
Burned f
Coal
Coal
Coal
Coal
Coal
Coal
Coal
Coal
Coal
Coal
Coal
Coal
Coal
Coal
Coal
Coal
Coal
Coal
Pollutant
Concentration
Relative
Accuracy
(*)
17.5
10.7
7.3
5.8
7.6
15.8
18.8
14.6 *
8.6
9.9
17.0
12.7
2.5
13.1
28.3 *
13.5
*
9.6
11.6*
95%
Confidence
Level
(%)
3.5
4.3
4.8
4.6
6.1
7.7
7.5
5.8
3.9
"2.2
6.1
4.6
2.2
6.8
3.5
4.9
6.0
5.9
System
[lbs/106 Btu]
Relative
Accuracy
(X)
13.2
9.8
6.8
18.6
16.0 *
12.2
4.3
14.4
5.0
24.6
26.7 *
16.4
11.3*
*
11.6
95%
Confidence
Level
(X)
5.0
5.4
3.8
8.8
5.9
4.2
3.5
5.2
2.4
7.0
4.9
5.0
7.3
5.0
6 Reference Method runs instead of 9 runs
A-3
-------
LSI MONITOR PERFORMANCE
SOURCES WITHOUT FGD EMISSIONS CONTROL SYSTEMS
Source
Code
C-34
C-34
C-34
C-34
C-34
C-34
C-5
C-5
G-9
H-l
M-5
M-5
R-l
R-l
S-l
S-2
S-3
Test F
Date
10/77
10/78
12/79
10/77
10/78
10/79
08/80
08/80
09/81
12/81
10/80
10/80
12/81
12/81
03/82
02/82
03/82
ollutant
Gas
SO2
so2
SO2
NOX
NOX
NO
X
SO2
NOX
SO2
so2
SO2
NOX
SO 2
NOX
SO2
so2
so2
Fuel
Burned
Coal
Coal
Coal
Coal
Coal
Coal
Coal
Coal
Goal
Coal
Coal
Coal
Coal
Coal
Coal
Coal
Coal
Pollutant
Concentration
Relative
Accuracy
(*)
19.7
2.2
10.1
15.6
22.7
25.6
9.5
8.3
9.0
8.9
18.4
7.3
18.5
23.5
3.8
4.1
22.5
95%
Confidence
Level
6.0
1.4
4.3
3.8
6.7
7.5
3.8
5.3
2.7
4.7
9.0
3.2
3.5
10.2
2.0
2.2
6.3
System
fibs/106 Btul
Relative
Accuracy
(*)
12.8
9.8
32.4
26.7
5.6
7.7
15.5
10.2
14.2
28.1
16.0
11.1
17.7
95%
bonf ide
Level
2.3
2*. 6
8.0
6.7
4.8
4.8
8.3
3.2
3.6
10.9
3.3
2.2
4.6
A-4
-------
DUPONT MONITOR PERFORMANCE
SOURCES WITHOUT FGD EMISSIONS CONTROL SYSTEM
Source
Code
A- 3
N-7
N-8
N-8
N-9
N-9
N-10
N-10
T-2
Q-4
Q-4
Test
Date
08/80
02/81
12/80
12/80
11/80
11/80
12/80
12/80
02/82
01/82
03/82
Pollutant
Gas
NOX
NOX
SO 2
NOX
SO 2
NOX
SO 2
NOX
so2
SO2
so2
Fuel
Burned
Coal
Oil
Oil
Oil
Oil
Oil
Oil
Oil
Coal
Coal
Coal
Pollutant
Concentration
Relative
Accuracy C
(X)
16.6
9.0
11.7
17.5 1
14.1
10.0
4.5
16.6
9.3
3.5
17.5
95X
;onfidence
Level
(X)
4.0
6.0
3.4
6.5
2.7
3.6
3.3
4.7
8.4
2.8
5.4
System
[lbs/106 Btu]
Relative
Accuracy C
(X)
16.4
11.6
15.5
13.4
17.8
13.6
7.0
19.0
10.1
10.2
12.1 '
95X
Confidence
Level
(X)
3.8
6.1
3.5
8.6
4.0
1.3
3.4
4.5
8.3
3.8
5.9
A-5
-------
DUPONT MONITOR PERFORMANCE
SOURCES WITH FGD EMISSIONS CONTROL SYSTEM
Source
Code
B-2/3
B-2/3
B-2/3
Test
Date
01/81
11/81
04/82
Pollutar
Gas
so2
SO2
so2
t Fuel
Burned
Coal
Coal
Coal
Pollutant
Concentration
Relative
Accuracy
(X)
passed
10.1
23.2
95%
Confidence
Level
(X)
3.5
16.8
System
[lbs/106 Btu]
Relative
Accuracy
(%)
6.7
7.2
19.7
95%
Conf iden
Level
(%)
2.3
3.0
16.4
A-6
-------
CONTRAVES MONITOR PERFORMANCE
SOURCES WITH FGD EMISSIONS CONTROL SYSTEM
Durc
3ode
3-1
3-1
3-1
D-l
3 Test
Date
09/80
07/81
12/81
03/82
>ollutant
Gas
so2
SO2
S02
so2
Fuel
Burned
Coal
Coal
Coal
Coal
Pollutant
Concentration
Relative
Accuracy
(X)
5.8
9.2
55.2
52.0 *
> 95*
r Confidence
Level
(%)
5.3
5.0
19.2
6.3
System
[lbs/106 BtuJ
Relative
Accuracy
(X)
18.8
9.5
99.0
61.8 *
95%
Confidence
Level
(X)
10.0
3.5
24.5
10.0
6 Reference Method runs instead of 9 runs
A-7
-------
CONTRAVES MONITOR PERFORMANCE
SOURCES WITHOUT FGD EMISSIONS CONTROL SYSTEM
Source
Code
F-4S
F-4S
F-4N
F-4N
F-4N
F-4N
J Test I
Date
04/81
04/81
04/81
04/81
09/81
09/81
=ollutani
Gas
so2
NOX
so2
NOX
so2
NOX
: Fuel
Burned
Coal
Coal
Coal
Coal
Coal
Coal
Pollutant
Concentration
Relative
Accuracy
(%)
3.7
15.6
29.1
32.3
16.4
8.9
95%
Confidence
Level
(X)
1.8
2.5
3.7
2.0
4.9
6.7
System
[lbs/106 Btu]
Relative
Accuracy
(X)
7.5
13.1
95%
Conf idef
Level
(X)
4.6
7.8
A-8
-------
EDC MONITOR PERFORMANCE
SOURCES WITHOUT FGD EMISSIONS CONTROL SYSTEM
Source
Code
E-123
E-45
J-l
j Test 1
Date
08/80
08/80
12/81
'ollutant
Gas
N0x
N°x
so2
Fuel
Burned
Oil
Oil
Coal
Pollutant
Concentration
Relative
Accuracy
21.2
5.1
18.0
95%
Confidence
Level
6.9
2.8
3.5
System
[lbs/106 Btu]
Relative
Accuracy
17.1
7.8
32.0
95*
Confidence
Level
7.8
5.2
6.7
A-9
-------
EDC MONITOR PERFORMANCE
SOURCES WITH FGD EMISSION CONTROL SYSTEM
Sourc
Code
K-3
K-3
e Test
Date
06/80
07/81
Pollutanl
Gas
so2
SO2
: Fuel
Burned
Coal
Coal
Pollutant
Concentration
Relative
Accuracy
(X)
19.6
8.6
95%
Conf idenc
Level
(X)
4.6
System
[lbs/106 Btu]
Relative
e Accuracy
(X)
11.9
95%
Confidence
Level
(X)
5.5
A-10
-------
CSI MONITOR PERFORMANCE
SOURCE WITHOUT FGD EMISSION CONTROL SYSTEM
'XX IT* C^
'ode
3-4
M
! Test P
Date
01/82
03/82
Dllutant
Gas
so2
so2
Fuel
Burnec
Coal
Coal
Pollutant
Concentration
Relative
Accuracy
4.7
13.1
95%
Confidence
Level
3.4
4.5
System
[lbs/106 Btu]
Relative
; Accuracy
10.5
7.7
95%
Confidence
Level
3.3
4.9
A-11
-------
r
TECHNiCAL REPORT DATA
EPA- 340/1-83/012
4. ! 1 fi_t AND SUBTITLE
A Compilation of 502 ancl N0x Continuous Emission
Monitor Reliability Information
James W. Peeler
9. PERFGHMING ORGANIZATION NAME AND ADDRESS
Entropy Environmentalists, Inc.
P.O. Box 12291
Research Triangle Park, NC 27709
12. SPONSORING AGENCY NAME AND ADDRESS
OAQPS
Stationary Source Compliance Division
Waterside Mall, 401 M Street, SW
Washington, DC 20460
3. RECIPIENT'S ACCESSION NO.
5. REPORT DATE
January 1983
6. PERFORMING ORGANIZATION CODE
8. PERFORMING ORGANIZATION REPORT 1
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
68-01-6317
13. TYPE OF REPORT AND PERIOD COVER
FINAL - IN-HOUSE
14. SPONSORING AGENCY CODE
EPA/200/04
15. SUPPLEMENTARY NOTES —
This report presents the latest information on the reliability of S02 and NO
CEM results, i.e.: (1) achievable CEM availability, (2) point-in-time CEM
accuracy, and (3) the long-term capability of CEMs to provide accurate emission
data.
"• KEY WORDS AND DOCUMENT ANALYSIS
a. DESCRIPTORS
Air Pollution
Continuous Emission Monitoring
18. DISTRIBUTION STATEMENT
Release to Public
b.lDENTIFIERS/OPEN ENDED TERMS
Reliability Information
Data Analysis
19. SECURITY CLASS (This Report)
unclassified
20. SECURITY CLASS I This page)
unclassified
c. COSATI Field/Group
21. NO. OF PAGES
42
22. PRICE
EPA Form 2220-1 (R«v. 4-77) PREVIOUS EDITION is OBSOLETE
A-12
-------
Protection Agency
U.S. Bnylrome Prot
i ,'- n 0 , 1J _L 'J ' ^•-i t/ V " -, r r;r\
, - - Dcp-r.-t,om Street, Room i670
clo, IL 60604
-------
United States Office of Air Quality Planning and Standards
Environmental Protection Stationary Source Compliance Division
Agency Washington, D.C 20460
Official Business Publication No EPA-340/1-83-012
Penalty for Private Use Postage and
$300 Fees Paid
Environmental
Protection
Agency
EPA 335
If your address is incorrect, please change on the above label,
tear off, and return to the above address
If you do not desire to continue receiving this technical report
series, CHECK HERE Q , tear off label, and return it to the
above address
------- |