United States Office of Air Quality
Environmental Protection Planning and Standards
Agency Washington, D.C. 20460
Stationary Source Compliance Series
c/EPA SUMMARY REPORT
A Pilot Project to
Demonstrate the
Feasibility of a
State Continuous
Emission Monitoring
System (CEMS)
Regulatory Program
-------
-------
EPA-340/1 -86-007
SUMMARY REPORT:
A Pilot Project to Demonstrate the
Feasibility of a State Continuous
Emission Monitoring System (CEMS)
Regulatory Program
Prepared By:
Perrin Quarles
Perrin Quarles Associates, Inc.
Charlottesville, Virginia 22901
and
James W. Peeler
Entropy Environmentalists, Inc.
Research Triangle Park, North Carolina 27709
Under Contract No. 68-02-3960,
Work Assignment 3-121
With Engineering-Science
Fairfax, Virginia 22030
Prepared for
EPA Project Officer: John Busik
EPA Work Assignment Managers: Louis R. Paley
Anthony Wayne
U.S. ENVIRONMENTAL PROTECTION AGENCY
Stationary Source Compliance Division
Office of Air Quality Planning and Standards
Washington, DC 20460
June 1986
-------
Disclaimer
This report has been prepared for the Stationary Source Com-
pliance Division, Office of Air Quality Planning and Standards, U.S.
Environmental Protection Agency, and the Air Branch, Air and Toxics
Division, U.S. Environmental Protection Agency, Region VII. The
opinions, suggestions, and conclusions expressed in this report are
those of the authors, and do not necessarily represent those of the
U.S. Environmental Protection Agency.
-------
- 11 -
SUMMARY REPORT
A Pilot Project to Demonstrate the Feasibility
of a State Continuous Emission Monitoring
System (GEMS) Regulatory Program
Table of Contents
Executive Summary ........................ v
I. Introduction ........................ 1
II. Major Program Related Findings and Recommendations ..... 5
A. The Importance of EER Review and Follow-up ...... 5
B. Audit Program ..................... 6
C. Quality Assurance Procedures ............. 8
D. Power Company Practices ................ 9
(1) Capability ................. • * • 9
(2) Use of CEMS ................... 10
E. State Program Constraints ............... 10
(1) State Agency Distrust for CEMS Technology .... 11
(2) The Need for Technical Assistance ........ 11
(3) The Data Reliability Issue ............ 12
(4) The CEMS Usefulness Issue ............ 12
F. EPA State Assistance
(1) State Assistance During the Pilot Project .... 13
(2) Structural Constraints .............. 15
(3) EPA Managerial Techniques ............ 15
III. CEMS Reliability ...................... 16
A. Opacity CEMS ..................... 16
(1) Evaluation Plan ................. 16
(2) Major Findings and Conclusions .......... 17
-------
- Ill -
B. S02 GEMS's 19
(1) Evaluation Plan 19
(2) Major Findings and Conclusions 20
C. Unavoidable or Excusable CEMS Downtime and Inaccuracy . 24
IV. Excess Emission Report Review 25
A. Introduction 25
B. Reporting Practices 25
(1) Findings 25
(2) Recommended EER Changes 26
(3) Unresolved Issues for Agency Consideration .... 27
C. Recommended Agency EER Review Procedures 28
D. Opacity Emission Trends in Missouri 30
V. The Coal Sampling and Analysis (CSA) Project 33
Appendix A; Bibliography of Technical CEMS Pilot
Project Reports 36
Appendix B; Bibliography of Written Comments on CEMS
Pilot Project Reports 39
Appendix C; Project Reports to be Finalized for General
Distribution 41
-------
- IV -
Illustrations
Illustration 1 Average Opacity GEMS Downtime for Six
Quarters by Unit 17
Illustration 2 CEMS Downtime Trend Analysis for Six
Consecutive Quarters 18
Illustration 3 Summary of EER Reporting Practices 25
Illustration 4 Changes Affecting All or Most Sources .... 26
Illustration 5 Changes Affecting One or a Limited
Number of Sources 27
Illustration 6 Trend of Exceedances Reported in Opacity
EERs from 1980-1984 31
Illustration 7 Quarterly Trend of Opacity Exceedances
Prior to and During the Project Timeframe . . 31
Illustration 8 A Comparison of Individual Source Opacity
Exceedance During the One Year Project
Timeframe (Annual Average) 32
Illustration 9 Opacity Exceedance Trends for the Three
Worst Performers During the One Year
Project Timeframe 32
Illustration 10 The Major Reasons for Opacity Exceedances
Among the Three Worst Performers During
the One Year Project Timeframe 33
-------
- v -
SUMMARY REPORT
A Pilot Project to Demonstrate the Feasibility
of a State Continuous Emission Monitoring
System (CEMS) Regulatory Program
Executive Summary
I. Background
In October 1975 EPA promulgated comprehensive requirements for
State continuous emission monitoring system (CEMS) regulations. Six
years later almost half of the States had not fully complied with the
EPA requirements, and most who had complied were not actively imple-
menting or enforcing their CEMS regulations. According to a series of
EPA surveys, much of this reluctance could be attributed to assump-
tions regarding the unreliability of CEMS, the burden of CEMS opera-
tion and maintenance, and the difficulty of reviewing and interpreting
reported data.
The CEMS pilot study charter was to identify and evaluate the
basis for and accuracy of these concerns in one or more States, and to
fashion, if possible, an approach to resolve them. Planning for a
major long-term multi-phased study began in the spring of 1982; field
work was initiated in March 1983 and completed in September 1984; and
the final data review and analysis was completed in August 1985. This
report includes a summary of major findings and recommendations drawn
from 27 technical project reports and reflects the comments provided
by State and industry participants. A bibliography of project reports
is included as Appendix A, and a bibliography of written comments is
included as Appendix B.
Principal participants in the study included: EPA's Stationary
Source Compliance Division; EPA Region VII's Air Branch; Missouri's
Air Pollution Control Program; Iowa's Department of Water, Air and
Waste Management; Entropy Environmentalists, Inc.; Perrin Quarles
Associates, Inc.; Union Electric Company (opacity and SO2 CEMS's);
City Utilities of Springfield (opacity and SO2 CEMS's); Kansas City
Power & Light Co. (opacity CEMS's); St. Joseph Light & Power Co.
(opacity CEMS's); Board of Municipal Utilities, Sikeston Power Station
(SO2 CEMS's); Iowa Public Service Co. (S02 CEMS's and coal sampling);
and Muscatine Power arid Water (SO2 CEMS's and coal sampling). The
Utility Air Regulatory Group (UARG) and its consultant Kilkelly
Environmental Associates also participated in the study by assisting
Union Electric Company and Iowa Public Service Company (both members
of UARG).
-------
- VI -
II. Project Method
The overall CEMS pilot project strategy focused on State program
assistance and demonstration. It was initiated by assembling a team
of CEMS technical and programmatic specialists, which included both
EPA experts and consultants. EPA Region VII was selected based on its
willingness to support and actively participate in the study and its
interest in building State CEMS program capability within the Region.
Iowa and Missouri were selected based on the number and variety of
CEMS's installed in each State, and each agency's limited implementa-
tion experience.
Both States made a formal commitment to participate, and each
supplied staff support and a portion of its § 105 grant funds to the
project. Missouri also committed to begin implementing a CEMS program
if convinced of its feasibility and usefulness. Iowa had already
begun implementation of its CEMS program and was included to expand
the study to cover additional SO2 CEMS's and to evaluate potential
coal sampling and analysis (CSA) acceptability criteria as an
alternative to SO2 CEMS's.
Utility companies from each State who volunteered to participate
were selected for special study based on factors designed to include
the greatest variety of CEMS types and site specific circumstances.
The study focused on opacity and SO2 CEMS's because they are the most
prevalent and represent the principal monitoring technologies.
Ultimately, independent workplans were jointly developed for
parallel opacity and S02 CEMS projects in Missouri, and concurrent S02
CEMS and CSA evaluation projects in Iowa. Each workplan had three
central elements: (1) an evaluation of CEMS reliability through
audits, excess emission report (EER) review, and the development and
application of quality assurance procedures; (2) State agency train-
ing in conducting CEMS audits and EER review; and (3) the preparation
of informal guidelines and recommendations for addressing problems
identified during the project. Both technical and programmatic
considerations were emphasized.
III. Major Accomplishments
All major workplan objectives relating to CEMS's were achieved
during the pilot project. Technical reports document CEMS reliability
and problems experienced (most of them resolved) at all of the study
facilities; and the study has resulted in new recommendations relating
to audit techniques, industry EER reporting practices, State EER
review procedures, quality assurance procedures, and numerous other
aspects of CEMS regulatory program activities.
-------
- vii -
No unresolvable technical or programmatic constraints to CEMS
program implementation were identified during the project; and im-
portant benefits were documented. Moreover, at the end of the pro-
ject, Missouri began independent implementation of its CEMS program,
adopting many of the project recommendations.
Although acceptability criteria for CSA programs in Iowa have not
been finalized, recommendations for technical criteria have been
proposed, and any further development of final criteria will require
the resolution of policy questions involving possible changes to
Iowa's S02 emission regulations, the selection of an S02 emission
averaging time, and, if short term averages are retained, the degree
of uncertainty regarding potential SO2 violations that will be per-
mitted when reviewing CSA results.
Many of the project findings and recommendations are transfer-
rable to other State agencies (although some are unique to Missouri
and Iowa). Project findings and recommendations will be transferred
primarily by finalizing and distributing selected technical reports,
including this suiranary report. A list of these reports is included as
Appendix C.
!V. Significant Project Findings and Recommendations
A. Monitor Reliability
Technical evaluation of CEMS performance during the project
indicated very high levels of reliability. EERs indicated an overall
opacity CEMS quarterly average availability of 92.9% (ranging from
75.0% to 99.9%) for 21 CEMS's, and 95.7% of the quarter (ranging from
83.0% to 99.8%) for the special study CEMS's. All of the special
study CEMS's were operating within ±5% opacity during the performance
audits, and the majority were within ±3%. Quality assurance records
at the same CEMS's showed very few incidents requiring corrective
action; and, with the exception of one CEMS which required significant
repairs, corrective action was taken quickly and the systems were
returned to service with very little delay.
The SO2 CEMS's also performed well. In Missouri they achieved a
quarterly average availability of 94.3% (ranging from 33.2% to 99.6%)
for six CEMS's, and in Iowa they achieved 95.1% (ranging from 87.3% to
99.3%) for three CEMS's. Audits at four of the five study sources
showed a consistent accuracy within ±10% of emission levels (CEMS's at
the fifth source were inaccurate throughout the project). Stratifica-
tion tests at each source also demonstrated that CEMS measurements
were representative of the total emissions. Although there was a
significant level of quality assurance activity during the project,
this was not unexpected, because of the comparative complexity and
sensitivity of SO2 CEMS's.
-------
- vin -
For both opacity and SO2 GEMS's, actual availability on a time
percentage basis may be somewhat lower than the reported percentages
indicate because periods of source inoperation have not been taken
into account. Source operating time was generally not available
during the study.
Also, despite high levels of CEMS availability, the study in-
dicated that operation and maintenance problems sometimes occur that
cannot be reasonably avoided and should be excused. These include:
normal quality assurance activities (which will sometimes include
CEMS problems which are difficult to diagnose, requiring unusual
out-of-service periods in a given quarter); emergency unit outages;
unsafe monitor access conditions (lightning storms, dangerously cold
weather, etc.); catastrophic failure; and vendor repair schedules.
B. Excess Emission Report Review
EERs provided useful documentation of excess emissions and CEMS
performance during the project and were used to target sources warran-
ting follow-up. Simple review procedures and realistic screening
criteria were developed, which should require a staff time commitment
to EER review in Missouri of less than one week per quarter. Based on
project experience, review should take less than one hour per EER on
the average. This would be significantly reduced if recommendations
for a uniform EER format and summary are adopted by Missouri.
Based on interviews with company personnel, EER review and
follow-up are also important for providing feedback to companies who
are reporting emission and monitoring problems and who may be adjust-
ing the level of response based on their perception of the State's
concern. Some company personnel believe that systematic EER review
and follow-up procedures will also ensure a more even-handed applica-
tion of State regulations. Follow-up when problems are reported
should also promote the credibility of the State agency with respect
to its emission control and monitoring requirements.
C. Quality Assurance Procedures
Quality assurance procedures were considered appropriate and
desirable by sources included in the study. Most had already devel-
oped and applied such procedures before the study began, and the new
study procedures supplemented procedures already in use. At the end
of the study some, but not all, of the new procedures were adopted for
continued use. Key findings include:
o There are no technical barriers to the design and implemen-
tation of quality assurance procedures for CEMS's used for
surveillance, with the single exception of cross-stack in
situ gas CEMS's.
-------
- ix -
o Effective quality assurance procedures may be designed so
that after an initial trial or learning period, daily and
periodic checks should require an average time commitment of
less than 16 hours per quarter for opacity GEMS's and from
30 to 60 hours per quarter for SO2 CEMS's.
o Quality assurance procedures will enhance the reliability of
data from most GEMS's and serve as an important key to
identifying GEMS problems when they occur, thus facilitating
effective corrective action. The development and
application of QA procedures are particularly appropriate
for sources which have demonstrated chronic GEMS problems
that interfere with meaningful interpretation of reported
data.
o Quality assurance procedures, including their scope and
frequency of application, should vary depending on the
GEMS type,, plant conditions, the skills and experience of
plant personnel, the organizational structure of the utili-
ty, the experience in application, and other factors unique
to a specific company or plant. They should also be subject
to change based on experience in application. Standardized
quality assurance procedures applicable to all CEMS's in all
plant locations are not appropriate.
D. Self-Audits
During the project, most sources successfully completed opacity
self-audits and decisively demonstrated the capability to perform this
function independently. Self-audits should be a valuable aid to
agencies who are evaluating reported emissions or GEMS problems but do
not have the resources to follow up with GEMS performance audits.
E. Power Company Capabilities
Power companies typically have the staff and organizational
capability to ensure GEMS reliability. Companies with a management
commitment to achieve such results had voluntarily developed extensive
quality assurance programs prior to the project. For the most part,
these programs should ensure an acceptable level of reliability for
use of GEMS data in agency surveillance activities.
F. Use of GEMS Data by Power Companies
Both opacity and S02 CEMS's are periodically used by power com-
panies to monitor control system performance and to diagnose control
system problems. CEMS's are also frequently used to monitor process
efficiency (in the case of opacity) and sometimes to monitor fuel
sulfur content or fuel blending (in the case of S02). These uses of
GEMS data indicate the extent to which power companies trust GEMS
-------
- X -
technology. However, such applications rarely supply sufficient
incentive for utility companies to install GEMS's absent the re-
quirement to do so; and they do not serve as an incentive to attain
high levels of data capture.
G. State Agency Training Assistance
State agency training assistance should be a critical EPA concern
in any new and technically complex program area such as continuous
emission monitoring. Training assistance is necessary to carry out
such a program and was found to be most effective when conducted in
field locations under real-life circumstances. Better training
results were also observed when a State agency makes a management
commitment to implement its GEMS program.
H. State Agency Concerns
Much of the historical reluctance in Missouri to committing
agency resources to GEMS related to a fundamental lack of trust in
GEMS technology and a perception that EER review provided few, if any,
surveillance benefits. This bias already existed prior to EPA's first
major promulgation of GEMS requirements, and was strengthened by EPA
indecision on the content of technical performance specifications and
quality assurance procedures supporting GEMS technology, as well as
EPA's failure to proceed diligently to implement and enforce its own
requirements. A demonstration of both GEMS reliability and usefulness
was necessary to reverse this bias.
The reliability of CEMS's was demonstrated through the review of
EERs, audits and the review of quality assurance records, all of which
pointed to a high degree of GEMS availability, accuracy and precision
for most of the GEMS included in the study. The usefulness of CEMS's
was demonstrated by developing screening criteria and actually target-
ing sources for follow-up based on simple EER review procedures.
During the project, Missouri concluded that EER review could
provide a valuable surveillance aid in addition to on-site inspec-
tions, and ultimately developed an alternative plan (approved by EPA)
for meeting EPA's annual inspection frequency requirement by combining
EER review with periodic GEMS audits. The State has also indicated
that other program changes will be considered as a result of the pilot
project and that the cooperative dialogue with utility representatives
maintained during the pilot project will continue as these changes are
considered.
-------
-------
SUMMARY REPORT
A Pilot Project to Demonstrate the Feasibility
of a State Continuous Emission Monitoring
System (CEMS) Regulatory Program
I. Introduction
Background
The CEMS pilot project was proposed in January 1982 as an experi-
mental effort to accomplish EPA's CEMS program objectives in a State
by assistance and demonstration. The essential requirements were
(1) willing participation of a State which had not yet given full
support to its CEMS program (CEMS requirements must be in existence,
along with reasonable progress in implementation by industry);
(2) willing support by an EPA Regional Office; and (3) assembly and
adequate funding of a team of programmatic and technical specialists
that because of their expertise could address and resolve most if not
all issues that might constrain progress in implementing a CEMS
regulatory program.
Most States at the time were not actively implementing a CEMS
program — even where CEMS regulations were in place. In a 1981 EPA
survey many expressed concerns that the instruments were not intrin-
sically reliable; that the data could not be trusted; and that EER
review and follow-up were inappropriate given the limited usefulness
of the data and the difficulty of interpretation. At the same time,
EPA's technical experts had concluded that most CEMS's work very well,
if certain minimum, quality assurance procedures are observed; and it
was known that successful EER review programs were being implemented
in several State and EPA Regional Offices.
Two questions were raised: Was there some unresolved technical
or programmatic barrier in certain States that had not yet been
identified at the federal level? If not, could a State previously
unwilling to implement a CEMS program be persuaded to do so by demon-
strating the benefits of the program or by disproving or resolving
barriers presented by the State?
-------
Page 2
Project Approach
The introduction to the initial project workplan describes the
basic philosophical approach taken:
... It is premised on the concept that in the current regulatory environ-
ment the most positive environmental results will be obtained by effec-
tively encouraging State agencies to take regulatory initiatives that are
most suitable to circumstances in the individual State. These circum-
stances may be varied in character, including regulatory, political,
institutional, technical, and environmental considerations, among others.
It is assumed that effective federal assistance begins with a full
understanding of such factors, then proceeds with the application of this
knowledge in a manner that promotes inherent incentives to achieve an
effective compliance program (on the part of both State and source), while
avoiding inherent constraints. Accordingly, an important objective of the
initial phase of this project is to define the operative goals, concerns,
and constraints of State agency programs selected for assistance. It will
be an overall objective of the project to provide assistance within this
framework . . .
It is also assumed that State programs and regulatory policies may be
inconsistent with, and possibly less stringent than, federal programs and
policies. It is a goal of this project to avoid any confrontation on such
issues and, instead, to forge an alliance of resources and expertise that
is aimed toward overall improvement of air pollution control by electric
utilities. A major focus will be to develop innovative ways to overcome
problem areas posed by the States.
Overall Plan and Objectives
The overall project plan included the following steps:
o Select a Regional Office that would support the project;
o With the Regional Office select one or more States that had not
implemented its CEMS program and might be approachable;
o Approach the State; market the project idea; and make an initial
assessment of what type of assistance would be necessary to prove
the feasibility and value of the program;
o If the State expressed initial interest, develop a more detailed
workplan that addressed intelligently and constructively point by
point each concern or reservation expressed by the State; con-
tinue to refine the workplan until the State would agree to its
content and schedule (avoiding any appearance of arms length
transaction or negotiation).
o In general, the project would emphasize training and technical
assistance to the State, focusing on opacity and SO2 CEMS's in-
stalled at coal fired power facilities. This assistance would
include CEMS audits, quality assurance procedures, excess emis-
sions report review, CEMS performance trouble shooting, and an
evaluation of State program problems and needs.
-------
Page 3
The most important measure of success would be either the State's
implementation of a CEMS program or the identification and documen-
tation of unresolvable barriers. However, there were also other
project objectives not directly tied to State adoption of a CEMS
program. These included:
o Documentation of CEMS program benefits.
o Development and evaluation of effective State assistance
techniques.
o Identification and resolution to the extent possible of specific
technical, programmatic and regulatory constraints to the
effective implementation of a CEMS program.
o Dissemination of important information developed in the project
to other agencies.
EPA Regional Office and State Selection
EPA Region VII was selected for the study based in large part on
its willingness to support and participate actively in the project.
Also, development of State CEMS programs in the Region was already an
element of the Region's program plan; and the Region expressed
interest in technical and program-related assistance in accomplishing
its objectives.
Both Iowa and Missouri provided an opportunity in terms of the
number of sources currently required to monitor — mostly coal-fired
utilities — and the variety of CEMS's and CEMS applications. This
would allow for ready transferability of the technical study results
to other Regions and States. Missouri, which had been reluctant to
proceed with full implementation and enforcement of its CEMS program,
also provided the opportunity to study and possibly resolve agency
program-related constraints. Both States were ultimately selected for
the study.
The study was restricted at the outset to CEMS applications in
the coal-fired electric utility industry, because it was felt that
electric utilities provided a sufficient quantity and variety of CEMS
applications to make the study goals achievable. They also account
for the greatest number of CEMS applications in Missouri and Iowa, as
well as nationwide, and are an industry of long-standing significance
to most State agencies because of the large volume of air contaminants
they generate.
In December, 1982, the Missouri Air Pollution Control Program
agreed to participate. In doing so, Missouri added an additional
objective to the study — to evaluate the potential use of CEMS as a
substitute for on-site inspections. Missouri contributed staff time
equivalent to one-half man year per year and a portion of its § 105
grant contract funds to the study.
-------
Page 4
In May, 1983, Iowa's Department of Water Air and Waste Management
also agreed to participate, and like Missouri, contributed one-half
man year per year of staff time and § 105 grant contract funds to the
study. Iowa was principally interested in establishing criteria under
which coal sampling and analysis (CSA) procedures could be substituted
for SO2 GEMS's. Because an active opacity CEMS program already
existed, the study in Iowa was limited to S02 CEMS's and CSA
procedures.
Industry Participation
At the request of Missouri, industry volunteers for the project
were solicited, and a special study plan was developed which provided
for concentrated assistance and CEMS evaluation at a limited group of
utility plants in both Missouri and Iowa. This resulted in signifi-
cant benefits throughout the study. The cooperation and assistance
obtained from industry was genuine and thorough and enabled the
project team to identify important State activities needed to ensure
proper CEMS program implementation.
Electric utility participants in the project included: Union
Electric Company (opacity and S02 CEMS's); City Utilities of
Springfield (opacity and SO2 CEMS's); Kansas City Power & Light Co.
(opacity CEMS's); St. Joseph Light & Power Co. (opacity CEMS's); Board
of Municipal Utilities, Sikeston Power Station (S02 CEMS's); Iowa
Public Service Co. (S02 CEMS's and coal sampling); and Muscatine Power
and Water (SO2 CEMS's and coal sampling). The Utility Air Regulatory
Group (UARG) and -its consultant Kilkelly Environmental Associates also
participated in the study by assisting and representing Union Electric
Co. and Iowa Public Service Co. (both members of UARG).
Project Workplans
Project workplans were developed jointly with State and industry
participants to address specific needs and concerns identified in the
initial project meetings in Missouri and Iowa. There were three key
elements to these plans: (1) an evaluation of CEMS reliability by
audits, EER review, and the design and implementation of quality
assurance procedures; (2) Agency training in conducting audits and in
EER review; and (3) the preparation of informal guidelines and
recommendations for addressing problems identified during the project.
Separate workplans were prepared for the Missouri opacity CEMS
project; the Missouri S02 CEMS project; the Iowa SO2 CEMS project; and
the Iowa CSA project. Based on these plans, specific schedules and
activities were developed for each special study source. A key
objective during the project was to maintain maximum flexibility so
that individual workplans could be frequently adjusted to address
unanticipated problems and events.
-------
Page 5
Project Accomplishments
In large part the pilot project objectives have been accom-
plished. At the completion of field work Missouri began active
implementation of EER review and follow-up procedures, as well as a
CEMS audit program. Numerous technical and programmatic constraints
were identified, both at the State and at the federal level, and
either resolved or targeted for follow-up. Successful State assis-
tance techniques and strategies were also identified.
In the Iowa project, the evaluation of S02 CEMS's was used to
expand the study of S02 CEMS's in Missouri, and resulted in valuable
findings supporting the SO2 monitoring programs of both agencies. A
significant advance was also made in the evaluation of CSA programs
and the analysis of criteria that might be used to determine whether
certain CSA programs might be substituted for SO2 CEMS. However, no
final criteria have been established, because most of the technical
criteria are dependent on the resolution of policy issues in Iowa and
at EPA.
Numerous reports describing lessons learned from the project have
now been completed and made available. This summary report and other
reports referenced in Appendix C provide a principal means of dissemi-
nating project results.
II. Major Program Related Findings and Recommendations
A. The Importance of EER Review and Follow-up
Perhaps the most significant findings of the study is the
importance of EER review and follow-up by the State agency. For years
following the adoption of CEMS requirements in Missouri, plants with
CEMS's have been reporting exceedances on a quarterly basis. Some-
times the exceedances have been due to significant control system
problems. Facility personnel have been aware of these problems and
usually have taken some type of corrective action. When the State has
not followed up on reported problems, power company staff reported the
following reactions:
o Some questioned the relevance or importance of CEMS requirements
that are being ignored.
o One viewed the EER as a mechanism for wiping the slate clean each
quarter. If the State did not follow-up, this was viewed as a
discretionary acceptance of the reported exceedances as being
within acceptable limits.
-------
Page 6
o Another was concerned that the reported exceedances were being
accumulated and might be used by the State or EPA in some future
enforcement action.
o In one plant experiencing control problems, internal debate has
occurred regarding the type of corrective action needed, and the
company has waited for a reaction to its reported exceedances to
help resolve the debate. With no reaction from the State, the
company has selected the most environmentally protective but most
expensive course of action. This has been a cause for concern.
Company personnel have felt that other companies willing to take
a higher risk may be selecting less effective and less expensive
measures. The company has felt penalized by this higher standard
of responsibility. Company personnel have felt that more State
interaction would result in a more even-handed application of the
State regulations.
In all, the lack of agency EER follow-up clearly affects the
agency's credibility in the eyes of company personnel. As a result of
the pilot study, some level of State EER review and follow-up is
considered important not only to monitor exceedances and CEMS perfor-
mance, but also —
o To provide direct feedback to companies who are reporting
emission and monitoring problems and who may be adjusting the
level of response based on their perception of the State's
concern.
o To maintain the credibility of the State's environmental program
among sources targeted for self-monitoring.
B. Audit Program
Prior to the project, audit procedures for both opacity and SO2
CEMS's had been developed by EPA and applied on a limited scale
throughout the country. These procedures were applied in the study
and modified based on the experience gained in their application.
The primary goal of the audit program was to verify the operating
status and accuracy of CEMS's that were subject to quality assurance
procedures during the project. The audits clearly achieved this
objective. In the case of opacity CEMS's, the audits confirmed that
each CEMS in the study could be operated and maintained within ±5%
opacity. In the case of S02 CEMS's, inherent design and software
problems were identified in a small number of the systems as a direct
result of the audits; and, except in the case of the cross-stack in
situ CEMS's, the audits confirmed that S02 CEMS's could also be
operated and maintained within ±10% of the emission level for the
types of power plants included in the study.
-------
Page 7
Other important benefits were also derived from the audit phase
of the study:
o Combiner system audits. One opacity CEMS configuration was
encountered for which no audit procedures had been developed. At
each unit owned by one participant, separate monitors are in-
stalled on dual ducts. Measurements from both monitors are
combined into a single measurement by a computer. It was neces-
sary to develop a new procedure to audit this type of system, and
the procedure is now available for similar applications.
o Improved opacity CEMS audit procedures. During the project
opacity audit procedures were streamlined and improved based on
the experience of the project team, with input from the indus-
trial participants. For example, panel meter checks were elimin-
ated where not m;eded, the post test zero compensation was
relaxed, and the overall time length of the audit was substan-
tially reduced.
o Improved S02 CEMS audit procedures. A significant finding was
that the injection of calibrated gasses as an audit technique can
provide significant time and cost savings with results comparable
to a relative accuracy audit. In this context the best approach
is for the State to provide calibrated gasses of a known but
undisclosed value for the source to inject. The procedure,
itself, requires expertise to perform successfully. Other
significant S02 CEMS audit cost savings were demonstrated by the
use of transportcible CEMS's — as much as one-half the normal
cost per relative accuracy audit can be achieved by transporting
an audit monitor from site to site.
o Self-audits. A highly significant finding involving audits in an
overall CEMS regulatory program is that they can be conducted
effectively by industry staff. These self-audits can be very
useful to an agency that has a limited number of qualified staff
or limited time to conduct the audits. The audit can be used to
confirm the reliability of exceedance data, or as a precaution
where continuing CEMS problems are being reported. The project
team recommends the use of self-audits as an EER follow-up
procedure where significant CEMS or exceedance problems are being
reported.
o Training. One goal of the project was to train State personnel
in the use of these audit procedures so that they could verify
the accuracy of CEMS's at their discretion. With this capability
State personnel could determine on their own which CEMS's were
providing accurate data and which ones were not — this would
give the State a basis for identifying where more rigid quality
assurance activities were needed, and it would remove doubt about
the quality of exceedance data when raised. To accomplish this
goal, training audits were conducted.
-------
Page 8
This phase of the project was very successful. At the end of the
project, State personnel began an independent opacity CEMS audit
program, choosing facilities with emissions and monitoring
problems identified in recent EERs. An SO2 CEMS audit program is
currently under development.
A significant finding by the project team relates to the impor-
tance of field training. Prior to the project, classroom train-
ing had been attempted without noticeable success. In the field,
however, the element of real life conditions and the practical
importance of the results seems to be more highly motivating —
especially if the training is conducted at sources within the
jurisdiction of the trainees and the trainees actually partici-
pate in the training audit.
C. Quality Assurance Procedures
During the study, quality assurance procedures were developed for
six opacity CEMS's and five S02 CEMS's. They were not constrained by
quality assurance regulations (except to include pertinent maintenance
requirements and performance standards). They were designed to
address monitor and location specific considerations and contained
built-in flexibility to allow for change based on experience in
application.
For the most part, these procedures were not difficult to imple-
ment. From records provided by plant personnel, the average time
required to implement opacity CEMS QA procedures was less than 5
minutes per day for the daily checks, and from 2-8 hours per quarter
for the periodic checks; for S02 CEMS's it took from 10 to 20 minutes
each day for the daily checks (at one plant with monitor location
constraints, it took 73 minutes a day) and from 4-10 hours per month
for the periodic checks.
In interviews following the project, plant personnel confirmed
that the procedures were simple and useful, and elements have been
adopted voluntarily by several of the plants. The QA procedures also
resulted in quick and accurate identification of CEMS problems and in
most cases also led to appropriate resolution of the problems.
Three major findings during this phase of the study include:
First, the application of effective QA procedures enhances the
reliability of data from most CEMS's and serves as a key to the
identification of CEMS's that do not provide acceptable performance.
Our study did not test the hypothesis that QA procedures are necessary
(or that QA plan requirements might be needed). However, if contin-
uing CEMS performance problems are experienced, simple, non-burdensome
QA procedures may be designed and implemented to address these prob-
lems on a monitor-specific basis.
-------
Page 9
Second, QA procedures should vary from plant to plant and in some
cases from CEMS to GEMS. Certain minimum procedures should be in-
cluded in any QA plan? however, these procedures should reflect the
actual conditions at the plant, giving consideration to the type of
plant, the monitor, its location, the skills and experience of plant
personnel, the organizational structure of the utility, etc. In
addition, the initial procedures should be subject to change as
specific implementation experience with the CEMS is gained.
Standardized quality assurance procedures applicable to all CEMS's in
all plant locations are not appropriate.
A major outcome of the project has been the development of
example QA plans for principal types of opacity and S02 CEMS's. It
was also determined that cross-stack in situ CEMS's for S02 are not
readily subject to QA procedures because of inherent design problems.
Third, there are no technical barriers to successful quality
assurance (with the single exception of cross-stack in situ gas
CEMS's). Instead, most barriers are readily resolved by a management
commitment to quality assurance. It was discovered that the existing
CEMS O&M procedures at many of the plants were advanced and working
well to ensure a high level of CEMS availability and accuracy. Where
such procedures were not in place, as familiarity with the project QA
procedures increased, progressively better performance was achieved,
especially with S02 CEMS's.
Essentially, successful QA is directly related to the technical
capability of plant personnel; this capability increases with ex-
perience; and QA experience is the product of a management commitment
to proceed with QA. Industry commentors also point to the importance
of training and support provided by vendors, and one commentor indi-
cated that a natural growth of capability should occur as skilled
people are distributed throughout the electric utility and other
industries.
D. Power Company Practices
(1) Capability
It was evident that personnel at each of the study plants in-
cluded in the opacity and SO2 CEMS studies are highly capable of
operating and maintaining CEMS's in a manner that will minimize
downtime and inaccuracy within limits that are acceptable to State
agency and EPA compliance program staff.
Skilled engineers and instrument technicians were usually avail-
able at each company to maintain the numerous monitoring and engineer-
ing systems necessary to assure efficient electric power production.
Although there is necessarily a learning period during which CEMS
-------
Page 10
performance may suffer — ultimately, the operation and maintenance of
these systems should not present an unusual or unmanageable burden to
these staff.
The project team observed that companies with a management
commitment to do so had established an effective organizational
structure, procedures and staff to identify and respond to monitoring
problems in a quick and effective manner. This commitment usually
included the in-house formulation of quality assurance procedures for
GEMS's. Moreover, during the quality assurance assessment phase of
the project, many of these procedures proved to be sufficient to
ensure acceptable GEMS performance during the project.
(2) Use of GEMS
Every company included in the study indicated that GEMS's are
used to monitor control system performance. In the case of one
company, standard operating procedure required taking progressive
steps to reduce emissions if the opacity began to increase — these
steps, including load change and a shift in coal blending procedures,
were keyed to 2% opacity increments, starting well below the opacity
limit.
GEMS's are also routinely used to diagnose control system prob-
lems. As corrective action is taken, the GEMS will be used to assess
the effectiveness of the corrective action. Because of the often
complex nature of particulate and SO2 control systems, the GEMS
facilitates a quick trial and error approach to diagnosis, often with
only low levels of emissions involved.
Power company staff also indicated that opacity CEMS's are often
used to monitor production efficiency (since an increase in opacity
often signals poor combustion efficiency and a corresponding higher
cost in electricity production), and S02 CEMS's are sometimes used to
check the sulfur content of the coal, especially where a minor per-
centage difference will have a significant impact on the cost of the
fuel. No personnel indicated, however, that cost savings would
justify installation of CEMS's in the absence of any requirement to do
so (with the possible exception of the Subpart Da facility).
E- State Program Constraints
No unresolvable constraints to implementing a CEMS regulatory
program in Missouri were identified during the project. However,
numerous inhibiting factors were identified. The following are
noteworthy:
-------
Page 11
(1) State Agency Distrust for GEMS Technology
Missouri did little to supervise the implementation of its CEMS
regulations after they were adopted in 1976. Managers explained that
their personal experience in earlier years with CEMS's had been
unsatisfactory, and that they had serious questions regarding CEMS
reliability. When EPA promulgated comprehensive CEMS requirements in
1975, they felt their concerns relating to unreliability had not been
adequately addressed; and they questioned the soundness of EPA's
actions. They also felt that the States were not adequately con-
sulted, and that EPA had failed to provide satisfactory technical
support for these regulations. The delay in promulgating revised
performance specifications and in the development of quality assurance
guidelines, along with EPA's lack of diligence in enforcing its own
CEMS requirements, reinforced these doubts and concerns.
In sum, the perception of State managers was that EPA had requir-
ed the States to proceed with a questionable monitoring program,
without adequate State agency consultation, and then had not provided
sufficient technical and policy support to make the program work.
This perception was clearly presen,t as the pilot project began, and
was only resolved as EPA stood behind its offer to provide the tech-
nical consultation and assistance promised as a part of the pilot
project, and ultimately demonstrated the reliability and usefulness of
CEMS's during the project.
(2) The Need for Technical Assistance
The State agency's need for CEMS technical assistance involving
its CEMS program was stated at the outset of the study. Therefore,
much of the study workplan in the early stages was devoted to iden-
tifying and evaluating specific needs; then, as the study proceeded,
an effort was made to provide needed assistance in the most effective
manner. Specific areas of assistance included training in conducting
audits and evaluating audit results, reviewing EERs, developing QA
procedures and reviewing QA results, and responding to numerous
specific technical problems that arose.
In general, technical assistance was provided with training as
the principal objective, so that at the completion of the project
agency personnel would be confident of their own capability to imple-
ment aspects of their CEMS regulatory program requiring technical
expertise. The fundamental significance of training as an integral
aspect of technical assistance was demonstrated by the end of the
project as agency personnel began successfully to assume the respon-
sibilities of implementing a CEMS program.
Two important characteristics of a successful CEMS training
program were identified: (1) training should take place in a field
location regulated by the agency, with the cooperation of industry
personnel, and with agency staff actively participating in the session
-------
Page 12
(training should not rely solely on the distribution of manuals and
guidelines accompanied by classroom instruction); and (2) training is
more effective where there is a clear agency management commitment to
the program.
Other areas of useful technical assistance in the project included;
o Low cost assistance was provided by supplying opacity CEMS audit
devices and calibrated gasses for use by the State.
o A constructive critique of the State's EER review program re-
sulted in a guides that the State has used in initiating changes
to its program.
o Quality assurance procedures were designed and field tested for
major opacity and SO2 GEMS's.
(3) The Data Reliability Issue
Questions concerning the reliability of CEMS's and the accuracy
of reported data were clearly a constraining factor in Missouri prior
to this study. A perception that the CEMS's might be inaccurate, and
if inaccurate, biased high, made the State agency unwilling to rely on
reported exceedance data as a true indication of actual exceedances.
This issue was effectively addressed by the following project activ-
ities:
o Audits demonstrated the inherent reliability of most CEMS's at
plants included in the study.
o Quality assurance plan design and implementation demonstrated
that industry personnel at most plants included in the study
could maintain acceptable accuracy.
o Many (but not all) EERs, when analyzed, tended to confirm the
validity of exceedance data by (1) identifying exceedances
associated with events for which exceedances would be expected,
and (2) identifying CEMS problems that might interfere with the
accuracy of exceedance data.
(4) The CEMS Usefulness Issue
Questions concerning the usefulness of CEMS data were mostly
related to questions involving reliability. Unreliable data could not
be used by the agency; therefore, why invest significant time in EER
review and follow-up? The State also expressed interest in the
potential usefulness of CEMS as a less resource intensive substitute
for inspections. If convinced of its reliability, the State wanted to
substitute EER review for routine inspections.
-------
Page 13
When, during the field work phase of the project, the State was
convinced that the data could be a reliable indication of the continu-
ous compliance status of a facility, it requested and received EPA
approval to substitute a combined .program of EER review and audits for
EPA's annual major source inspection requirement. Later the State
realized that the GEMS program would require more time to implement
than the inspection time saved. Despite this realization, the State
is proceeding with a full scale implementation program — not because
it offers time savings, but because in the State's view GEMS's offers
a superior basis for determining the continuous compliance status of
utilities.
F. EPA State Assistance
(1) State Assistance During the Pilot Project
It was correctly perceived at the outset that the focus should be
on State assistance and not culminate in results which could not then
be duplicated by State agency staff. Thus, the project design includ-
ed training sessions and technical tasks which combined contractors
and State agency staff. The significance of this concept became fully
appreciated only after the project was well underway. Key observa-
tions were:
o Effective assistance cannot be confined to a series of pre-planned
activities.This is the case even though the State may participate
in the planning of these activities. Effective assistance
involves a commitment to evaluate on a continuous basis whether
the pre-planned activities actually provide the assistance
needed, and the willingness and flexibility to change approach if
they do not.
o Effective assistance requires a commitment to continue until the
objectives are achieved. The appropriate elements of assistance
should be worked out jointly between the State and EPA, beginning
with a careful and realistic assessment by the State of its
needs, and proceeding with an effort by EPA to provide assistance
according to those needs. Then, if the State agency is still not
proceeding satisfactorily, further assistance may be necessary in
identifying inhibiting factors, and assistance in addressing them
as well. It should be realized that significant program con-
straints may not be easily recognized by the State or EPA, and an
ongoing evaluative effort will usually be justified.
o Assistance with strings attached. "We will show you how to do
this, and help you get started, but then you have to do it our
way and achieve a level of performance that we decide is appro-
priate." This approach risks causing resentment. The resentment
is all that much more acute if EPA demands fail to consider State
problems and constraints that cannot be easily overcome. There
-------
Page 14
are two levels to the problem. The State may feel that EPA has
imposed the condition without looking into the matter carefully
enough to recognize that the constraint exists. This decreases
the credibility of EPA staff .persons and is annoying to the State
staff who must take time away from mission activities to build a
case that will convince EPA to see things a different way.
The second level problem exists when EPA understands the situa-
tion well enough and has decided to force the State into a
particular course of action anyway. From EPA's point of view
this may be necessary as a matter of national or regional policy.
However, it is destined to create anger in the State agency and
may overflow into resistance or lack of cooperation in other
areas. Before proceeding along this course, the groundwork
should be laid to establish and preserve a friendly and objective
management relationship between EPA and the State. Good inter-
personal management practices should be observed. These should
include a significant level of encouragement and support and
constructive criticism only after a careful evaluation of all the
facts.
Assistance Versus Demand and Negotiation. Since State agency
programs are largely funded by EPA, a significant portion of EPA
effort is devoted toward overseeing State agency activities to
ensure that federal objectives are being achieved. This over-
sight includes an annual negotiation of performance commitments;
a program audit; the submission and review of quarterly status
reports; EPA participation in selected State inspections; EPA
evaluation of specific, ongoing State enforcement actions; EPA
review of State regulatory actions, permits, and penalties; and
other management-related activities.
The current project offers insight into this State EPA management
relationship by virtue of the results achieved: there were no
federal CEMS program implementation performance standards when
the project was initiated; there was no formal federal audit
which found the State CEMS program to be deficient; there was no
arms length negotiation to impose accountable performance stan-
dards on the State; yet, the State, at first reluctant, has now
begun implementation of its CEMS program.
In essence, an effort was devoted to convincing the State of the
merit of the program. No strings were attached. The only
commitment was that the State and federal project team members
complete the project, and in the process be open minded about
actual CEMS benefits and problems identified. If problems were
encountered, they should be addressed constructively and a
reasonable attempt should be made to overcome them. If at the
end of the project the State remained convinced that the program
had no value, there would be no obligation to implement it. This
approach resulted in a forthright admission and analysis of real
EPA State relationship constraints during the project.
-------
Page 15
(2) Structural Constraints
The Missouri Air Pollution Control Program (APCP) is subject to a
complex management structure that affects its relationship with EPA.
Like most State agencies it has evolved over the years subject to the
unique management and organizational concerns of the State government.
APCP staff are hired and paid by the Department of Natural Resources,
and the APCP's political direction is primarily responsive to the
State legislature and the Air Conservation Commission, which has a
concurring role in the hiring of the APCP Staff Director and imposes
performance requirements of its own on the Agency. Including EPA,
there are, in effect,, three separate managers of APCP activities.
To add complexity, the APCP has no direct line relationship with
its field offices — the regional offices (multi-media) are managed
through a contractual relationship; and the four independent local
offices are managed more distantly through the pass-through of federal
§ 105 funds.
One impact of this complex arrangement is the increased trans-
action cost of ordinary program implementation activities. For
example, each managing agency has its own overlapping and inconsistent
reporting requirements (usually justified by its special needs). A
change in the required EPA format or content may ripple through the
entire system, forcing further negotiations with Regional and local
agencies and attempts to derive the required new information from
existing reports instead.
A potentially useful EPA management approach in this situation is
to anticipate transaction problems when new performance commitments
and reporting requirements are to be imposed, and to look for methods
to reduce the burden and to accommodate the State's desire for restrict-
ing changes only to those that are essential.
(3) EPA Managerial Techniques
Missouri agency staff perceive that they are on the front line
addressing air pollution problems from the industry in their State,
and that their efforts, and not necessarily the efforts of EPA staff,
result in direct air pollution control benefits. EPA support is
appreciated; however, EPA efforts often seem primarily aimed at
documentation of State activities or at performance evaluation. State
agency personnel frequently perceive that these levels of documen-
tation and interaction detract from important mission-related activ-
ities.
Whether EPA agrees with the State's assessment or not, EPA staff
have the ability to facilitate effective State agency efforts by good
managerial practices, including significant levels of encouragement,
assistance and conscientious efforts to minimize activities or require-
ments that detract from principal surveillance and enforcement goals.
-------
Page 16
III. CEMS Reliability
A. Opacity CEMS
(1) Evaluation Plan
CEMS performance was evaluated in three project activities: EER
review; performance audits; and through the design and application of
quality assurance procedures. The principal performance evaluation
involved six CEMS's installed on six generating units at four facili-
ties in Missouri. EER data from other power facilities was also
evaluated.
For each of the units selected for special study, preliminary
draft opacity CEMS QA procedures were developed based on the specific
features of individual monitoring systems, information from the
operator's manuals provided by the monitor manufacturers, and previous
experience in testing and auditing similar CEMS's. An initial audit
of each of the six opacity CEMS's was performed. Each initial audit
included a systems audit (a qualitative evaluation of operation,
maintenance, and record keeping practices) and a performance audit (a
quantitative evaluation of CEMS accuracy and precision).
Based on information gained from the initial audits and comments
provided by source personnel, the preliminary QA procedures were
revised to address source-specific factors that were unknown prior to
the initial audit, and to incorporate pre-existing routine maintenance
procedures at each source.
For each source, the QA procedures included: (1) simple daily
checks to be performed from the monitor control unit location, (2)
periodic QA checks to be performed both at the monitoring location and
at the control unit, and (3) general corrective action procedures to
be used when daily or periodic QA check control limits were exceeded
or when repairs to the opacity CEMS were necessary.
Agreement with source representatives was obtained on step-by-
step QA procedures; then, station personnel implemented the procedures
for approximately 6 to 8 months. The QA documentation was then
reviewed by the project team, and recommendations for revisions to the
QA procedures were made. In most cases, suggestions were made to
simplify the QA procedures.
Each station also conducted an audit of its CEMS's during the
project and provided copies of the audit results to the project team.
In addition to these results and the QA documentation obtained during
the study, data and information on opacity CEMS performance were also
obtained from the review of EERs submitted during the project.
The four special study facilities continued the implementation of
opacity CEMS QA procedures (with modifications in some cases) for
-------
Page 17
approximately one year; and the field study was concluded when the
project team completed a final performance audit of the GEMS.
(2) Major Findings and Conclusions
This study documented excellent opacity CEMS performance.
EER review for six quarters of data including the study period
indicated an overall performance record of 92.9% uptime during the six
quarters (ranging from 75.0% to 99.0%) for 21 CEMS's. Most of the
downtime, in fact, was due to factors other than monitor malfunction
— e.g., computer malfunction, strip chart malfunction, operator
error. Monitor equipment malfunctions were clearly identified for
only 1.2% of the six quarters. At the six units subject to the more
comprehensive study the average uptime for the same six quarters was
95.7% (ranging from 83.0% to 99.8%) and the average downtime specif-
ically attributable to monitor equipment failure was only 0.7%. The
actual percentage of uptime would be somewhat lower if unit operating
time were taken into account; however, operating time was not avail-
able during the study for most sources.
ILLUSTRATION 1
Average Opacity CEMS Downtime for Six Quarters by Unit
DURATION
(% OF THE
QUARTER)
oa _
33 _
20 -
ta -
10 -
11 -
10 -
o -
G -
4 -
2 -
A _
— .
._,
— n
•
r~"
VS/MMSSSSSSS//J
"t
g
f:
d
t
C.
-.
r -
. ,
. "
1 '
r
'n
--
"~^
s
v
v •
BCDIFGH I J KLMNOPQRSTU
SPECIAL STUDY CEMS >s
ALL OTHER CEMS 's
-------
Page 18
ILLUSTRATION 2
CEMS Downtime Trend Analysis
for Six Consecutive Quarters
(% OF THE
QUARTER)
2nd O T363 3rd O
19
1st Q 1984- 2nd O 1984-
Audits of the eight transmissometers included in the special
study also confirmed their accuracy. In all, 16 CEMS's performance
audits were conducted by the project team. The CEMS's were within the
calibration error specification (i.e., ±3% opacity) in 12 of the 16
audits at the low,, mid and high range levels. The greatest cali-
bration error observed during any of the audits was 4.9% opacity,
which exceeded the calibration error specification by only +1.9%
opacity. Excessive dust accumulation on the optics was a factor in
only two instances? in both cases the units were not operating when
the audit was performed, and the actions taken by source personnel
prevented these problems from affecting the determination of excess
emissions. Internal s;pan or zero error exceeded the ±2.5% opacity in
only two instances, also by a very minor amount (-0.4% opacity-span,
in one instance; +1.5% opacity-zero, in the other). No optical
alignment problems were observed during the audits.
Perhaps the best indicator of monitor performance was recorded as
a part of the daily and periodic QA checks undertaken at the eight
CEMS's in the special study. Although the data from QA logs were not
complete, they showed a need for only 12 zero or span adjustments
based on the primary data recorder (i.e., computer printout or strip
chart readings) for all eight CEMS's checked during the one year
period. Only 25 incidents requiring corrective action were noted. In
almost all cases the problems were discovered during the daily or
periodic checks, and except in a very few instances, the corrective
-------
Page 19
action was taken very quickly and the GEMS was calibrated and returned
to service with minimum delay.
Based on these findings, it is the project team's conclusion that
electric utilities in circumstances similar to those included in the
study have the capability of operating and maintaining opacity CEMS's
at a high level of accuracy and availability or they should be able to
develop that capability.
B. SO, CEMS's
(1) Evaluation Plan
(a) Missouri
A field study was conducted at three sources in Missouri to
evaluate the reliability of S02 GEMS data and to facilitate the
development and evaluation of quality assurance procedures for S02
GEMS's.
The major goals of this phase of the study were: (1) to develop
effective quality assurance procedures for the major types of S02
CEMS's used at electric utility steam generators in Missouri; (2) to
determine the accuracy and reliability of SO2 GEMS data when effective
QA procedures are implemented; and (3) to solve specific technical
problems with S02 CEMS's that were identified prior to and during the
project.
This study did not attempt to identify or define the ideal or
minimum QA plan. Instead, the study attempted to develop, for a
number of CEMS's and source-specific situations, relatively simple,
cost effective QA procedures. It was intended that the process of
developing these procedures and the flexibility of the approach would
provide examples that could be easily adapted to other sources and
situations.
(b) Iowa
In Iowa, the S02 CEMS study was conducted in conjunction with a
coal sampling and analysis (CSA) study. Iowa's primary interest was
to develop criteria that could be used to evaluate a CSA program
proposed as a substitute for SO2 CEMS's. As in the case of S02
CEMS's, the CSA program would be used as an indication (but not as a
demonstration) of compliance with SO2 emission limits.
A demonstration program involving a comparison of CSA and CEMS
data was designed (1) to establish the relationship between the two
measurement methods, and (2) to demonstrate that CSA results repre-
senting coal use during a twenty-four hour period could prove to be an
adequate surveillance method relative to exceedances of emission
limits with two hour (Iowa's) and three hour (NSPS) averaging periods.
-------
Page 20
Two CSA demonstration projects were conducted: one at Muscatine
Power and Water's Unit 9, an NSPS Subpart Da source burning high-
sulfur Southern Illinois coal (study results could be transferrable to
non-FGD facilities burning high sulfur coal by comparing CSA measure-
ments to effluent concentrations at the FGD inlet); and one at Iowa
Public Service Company's George Neal South Station, Unit 4, an NSPS
Subpart D source burning low-sulfur Powder River Basin coal.
In addition to CEMS and CSA data, EPA Method 6B data were col-
lected from the two plants, thus affording the opportunity to evaluate
the accuracy and precision of all three measurement systems. At
Muscatine, Method 6B also provided a means of determining the impact
of a scrubber bypass duct on the comparison of CEMS and CSA measure-
ments; and at Iowa Public Service Company, it provided an independent
check on CEMS data, and thus a potential basis for documenting the
impact of mid-course quality assurance program changes.
In order to ensure that CEMS data of acceptable accuracy would be
collected, both sources implemented quality assurance programs similar
to those developed during the Missouri SO2 CEMS pilot project. The
development and implementation of these quality assurance programs
were the principal elements of the Iowa S02 CEMS study. The primary
goals were: (1) to develop effective quality assurance programs for
the major types of SO2 CEMS's installed in Iowa; (2) to implement
these quality assurance plans to ensure that accurate CEMS data were
collected for the CSA demonstrations, and (3) to provide
documentation of the accuracy of S02 CEMS data. Unlike Missouri's
project, there was no comparable attempt to persuade Iowa to adopt a
CEMS program — such a program was already in place.
(2) Major Findings and Conclusions
(a) Missouri
o The SOy CEMS's achieved a high availability rate, and provided
accurate, precise and representative emissions data.
Initial CEMS performance audit results indicated acceptable
performance for all of the CEMS's in the study. The accuracy of the
data was controlled by criteria established for (1) the allowable
drift limits for daily calibration checks and (2) various auxiliary
monitoring parameters that could affect the relationship between
calibration check data, and analyzer accuracy (e.g., temperature
compensation circuitry, sample and calibration gas pressures, flow
rate, etc.).
Self-audits based on calibration gas injections were conducted as
part of the periodic QA procedures at two of the plants to verify data
accuracy. At the third plant, calibration gas injections were part of
the corrective action procedures, and served both to control and to
assess data at this plant. Final performance audits were conducted
-------
Page 21
for only two of the three GEMS's; both GEMS's demonstrated acceptable
performance at the conclusion of the study. (A catastrophic generator
failure forced an early termination of the study and prevented the
final GEMS performance audit at the third source.)
The average availability of the six Missouri SO2 CEMS's filing
EERs during the study was 94.3% for six quarters (ranging from 33.2%
to 99.6%); and audits showed a consistent accuracy within ±10% of
emission levels.
Stratification testing was conducted during all of the initial
and final GEMS performance audits. In all cases, the GEMS sampling
location was found to be non-stratified. The absence of stratifi-
cation indicates that the GEMS emission measurements were represen-
tative of the total emissions.
o Effective SO, GEMS QA programs can be developed and implemented
for the types of CEMS's evaluated in the Missouri study (i.e.,
extractive and point in situ type CEMS's).
An effective GEMS QA program may be defined as one that results
in data of sufficient quality to satisfy both source and agency
requirements without imposing an unnecessary burden on the source.
The results from the audits and self-audits conducted during the study
demonstrated that the QA programs implemented were sufficient to
maintain acceptable data quality. In all cases, the implementation of
the QA program did not increase the total time expenditure for CEMS-
related activities.
Source personnel involved with GEMS operation and repair from two
of the sources stated that QA program implementation resulted in more
time optimizing GEMS operation and less time repairing their CEMS's.
In both cases, the total time expenditure for CEMS-related activities
did not change significantly due to QA program implementation. At a
third source, QA program implementation resulted in significant
additional time for repairs and other corrective action; however, the
source had little successful CEMS operation experience prior to the
study.
Representatives from each of the sources also indicated that
implementation of the QA program resulted in increased confidence in
the CEMS data. CEMS data were used by source personnel in assessing
the performance of fuel blending and emission control equipment.
Decisions based on CEMS data may have significant economic impact, and
such decisions are sensitive to the accuracy and precision of the CEMS
data.
-------
Page 22
o The accuracy of data provided by SO, CEMS's that are capable of
accepting calibration gasses (i.e., extractive and point in situ
type CEMS) can be quantified by performing calibration gas
injection tests.
In the absence of interferents within the effluent stream,
calibration gasses properly injected into a CEMS will give an indica-
tion of accuracy that is equivalent to or better than that which can
be obtained from comparative reference method testing.
For the CEMS's in the Missouri SO2 CEMS pilot project, specific
calibration gas injection procedures were designed to ensure that the
gas flow rates did not produce adverse pressure or temperature effects
and, thus, biases in the CEMS responses. Comparisons between accuracy
indications based on calibration gas injections and reference method
testing were made during initial and final CEMS performance audits.
In all cases, both techniques gave essentially the same results.
Calibration gas injection testing may be used to indicate analyzer
accuracy at several concentration levels within the instrument's
measurement range, thus allowing for an assessment of analyzer linear-
ity. This additional information can be used by technical personnel
in optimizing CEMS performance and by operations personnel in inter-
preting CEMS data.
(b) Iowa
In general, the study results at Iowa Public Service Company were
consistent with the highly favorable results experienced in the
Missouri study. The average availability of the three Iowa S02 CEMS's
filing EERs during the study was 95.1% for six quarters (ranging from
87.3% to 99.6%).
The following additional findings and conclusions are based on
the Iowa S02 CEMS project:
o Contraves GEM-100 cross-stack in situ CEMS did not provide
acceptable accuracy during the project.
Regardless of daily calibration check values which consistently
indicated good CEMS accuracy, the outlet CEMS on Muscatine's Subpart Da
unit demonstrated unacceptable inaccuracy throughout the pilot project
study. Despite their best efforts, the manufacturer's field service
representatives were unable to obtain acceptable CEMS performance.
Several performance specification tests were attempted, but none was
successful.
The calibration check errors were generally less than ±2.5% of
scale, and on only two occasions did these errors exceed ±5% of scale.
However, audit test data and Method 6B measurements indicated that the
bias associated with the emissions data for the stack CEMS changed
from unacceptably high to unacceptably low during the project.
-------
Page 23
o The inability of cross-stack in situ type CEMS's to accept
calibration gasses for verification of GEMS performance is a
significant disadvantage from a quality assurance standpoint.
The overall calibration status of cross-stack in situ analyzers
cannot be assessed with calibration gasses. Therefore, the validity
of daily calibration check results can be checked on a periodic basis
only by means of independent etfluent measurements.
Periodic QA procedures developed for Muscatine were based on
comparisons between data produced by the cross-stack in situ type
CEMS's and data produced by a well calibrated (with gasses) extractive
CEMS's incorporated into the FGD control system. These procedures
were less useful and considerably more complex than the periodic
accuracy checks developed for the point in situ type CEMS's.
o Corrective action procedures should include verification of
CEMS accuracy using calibration gasses following repairs or
adjustments that could affect overall CEMS calibration.
An early version of the Iowa Public Service Company quality
assurance program did not clearly require the use of calibration
gasses for verification of CEMS performance following repairs. This
program was subsequently modified after discovery of a measurement
bias which did not show up in the daily calibration checks. The bias
is believed to have been introduced during repairs that affected the
temperature compensation circuitry.
o For optimum performance, in situ CEMS'S should be protected from
environmental extremes.
During the final audit at Iowa Public Service Company, the
effects of rapidly changing ambient temperatures on short term drift
were demonstrated. A 3% calibration drift was introduced in the
SM 810 S0a CEMS by opening the stack annulus door and allowing cool
ambient air to blow directly on the SM810 transceiver case.
o The QA procedures evaluated in this project required only a
relatively minimal amount of time to complete.
The daily QA check procedures generally required about 5 to 10
minutes per day at Iowa Public Service Company, plus an average of
7 hours of periodic procedures once every 2 months. Muscatine daily
check procedures generally required about 20 minutes per day to
complete.
-------
Page 24
C. Unavoidable or Excusable CEMS Downtime and Inaccuracy
Although CEMS downtime or inaccuracy may often be avoided,
certain downtime incidents seem to be warranted based on events
occurring during the study. The following are project team recommen-
dations relating to such incidents.
o Normal QA Activities (calibration, maintenance, etc.). In most
cases the down time should be minimal, less than 2% of the
operating time of the CEMS. However, effective QA will sometimes
require a more extensive effort, especially when accuracy prob-
lems are recurring. Infrequently, problems may arise which are
difficult to diagnose and may require unusual out-of-service
periods extending for several weeks. Diagnosis and corrective
action is an important priority, and adequate QA in such circum-
stances should not be discouraged by an arbitrary agency limit on
downtime.
o Emergency Unit Outages. These events necessarily preempt normal
operation and QA activities. When the emergency conditions are
resolved, normal checks and corrective action should be resumed.
o Unsafe Access Conditions. Floods, lightning storms, dangerously
cold weather, and chemical spills were events that occurred
during the pilot study that interfered with daily or periodic
monitor checks and were considered a reasonable justification for
postponed QA activities. Of these, continuous cold weather below
zero restricted on-stack monitor checks for an extended period of
time. While the CEMS location might be changed, this seemed
unwarranted in view of the infrequency of similar conditions and
the overall serviceability of CEMS's in this study.
o Catastrophic Failure of the CEMS. Periodically a CEMS will be
accidentally damaged or will fail for unpredictable reasons in
such a way that a reasonable inventory of spare parts will not
suffice to restore the CEMS to service. In view of the overall
serviceability of opacity CEMS's involved in the study, main-
taining redundant systems seems to be an unnecessary precaution.
o Vendor Software Problems and Repair Schedules interfered with
CEMS availability during the project. Operators made reasonable
efforts to identify and resolve these problems. Ultimately the
responsibility for operation of the CEMS's rests with each
facility; however, agency flexibility seems warranted when the
delays are attributable to vendors and not the operators.
-------
Page 25
IV. Excess Emission Report Review
A. Introduction
One hundred thirty-five excess emission reports covering 21
opacity CEMS's were evaluated to determine the status of EER practices
and to recommend appropriate EER review procedures and guidance for
Missouri.
B. Reporting Practices
(1) Findings
Most reporting practices conform to the requirements of Missouri's
regulation, and to the extent that there are deviations, most have
been accepted by the APCP. However, almost all of the reports vary
significantly from company to company and have some content or format
problem which makes agency review inefficient or which prevents the
reviewer from having an accurate picture of CEMS or emissions perfor-
mance during the quarter.
ILLUSTRATION 3
Summary of EER Reporting Practices
Type of Information
Number of Plants
Reporting (out of 13)
1. Information Identifying Quarter, 13
Monitor and Unit*
2. Exceedance Information
a. Nature and cause of excess emissions* 13
— Narrative explanations (5)
— Reason codes (5)
— Combination (3)
b. Maqnitude of excess emissions* 9
c. Specific periods of excess emissions 8
d. Information on corrective actions 5
e. Statement of no excess emissions* 13
3. CEMS Performance Information
a. Specific periods of CEMS downtime* 5
b. Causes of CEMS downtime 11
c. Information on corrective actions* 10
d. Statement of no CEMS downtime* 13
4. Source Operating Information 3
(Allows for normalization of data)
'Currently required by Missouri reporting regulations.
-------
Page 26
(2) Recommended EER Changes
The following are recommended changes based on the project team
review. They are separated into two groups: one having an impact on
all or most sources; the other having an impact on only one or a
limited number of sources.
No legal analysis was undertaken to determine which recommended
changes will require revisions to Missouri's reporting regulations.
Although some industry representatives indicated that proposed changes
would be accommodated if requested by the Missouri APCP, others felt
that changes establishing new requirements (e.g., uniform format,
summary, etc.) should require regulatory action. Missouri has expres-
sed its intent to confer with electric utility representatives regard-
ing all potential reporting changes resulting from the pilot study
prior to any actual change.
ILLUSTRATION 4
Changes Affecting All or Most Sources
Change
Rationale
(1) Report in a uniform format with
uniform criteria governing content.
(2) Provide a simple summary of exceed-
ances and CEMS downtime. The summary
should highlight a very limited selec-
tion of general reason categories to
aid in screening.
(3) Include specific corrective action
taken for exceedances where
corrective action is appropriate.
(4) Require a very brief narrative ex-
planation of the causes of exceed-
ances and CEMS downtime, along with
corrective action. (This explanation
may then be categorized within a
simple selection of reason codes for
the purpose of summarization.)
To ensure proper content; to facilitate a
more effective review? to put sources on an
equal reporting basis.
To facilitate EER screening and tracking
by agency personnel.
This substantially expands the EER eval-
uation capability of agency personnel.
This also will expand the EER evaluation
capability of agency personnel.
-------
Page 27
ILLUSTRATION 5
Changes Affecting One or a Limited Number of Sources
Change
Rationale
(1) Include specific causes of CEMS
downtime.
(2) Excess emissions "caused" by CEMS
problems should be included with
CEMS downtime and not reported as
exceedances.
(3) Do not report exceedances during a
source outage.
(4) Do not report CEMS downtime during
a source outage.
(5) Do not combine data from units that
are separately monitored.
(6) Use the strip chart recorder, if
available, when the computer data
acquisition system fails.
This is essential for an effective review
of CEMS performance.
To eliminate a bias in exceedances during
review.
To eliminate a bias in exceedances during
review.
To eliminate a bias in CEMS downtime during
review.
This unnecessarily prevents an evaluation of
problems on a single control system basis.
Use of this logical alternative data re-
cording system will reduce downtime due
to data recorder malfunction.
(3) Unresolved Issues for Agency Consideration
(a) Selection of Reason Categories for Summarization
During EER review,, the project team selected the following reason
categories for summarization: for exceedances — start-up/shutdown,
sootblowing, control equipment problems, process problems, fuel
problems, other known cause, and unknown cause; for CEMS downtime —
monitor equipment malfunction, non-monitor equipment malfunction,
calibration/QA, other known cause, and unknown cause.
During the EER evaluation and targeting phases, it was sometimes
useful to combine control equipment and process problems with unknown
causes, and to track separately certain additional monitor downtime
causes, including, for example, power supply problems, and operator
errors.
While the project team supports the continued use of the selected
reason categories, the State should establish categories that are most
meaningful in terms of its own objectives, while at the same time
ensuring that the categories ultimately selected are compatible with
those used by EPA in its own compliance tracking system (currently the
CDS CEM subset). Consideration should also be given to other indus-
tries that will be required to submit EERs; however, it should not be
-------
Page 28
necessary to make every reason category adaptable to every industry.
It should be possible,, instead, to designate certain reason categories
applicable to industries for which they logically apply.
(b) Normalization
EER data may be aidjusted to account for source and CEMS operating
time each quarter and thus allow for a more accurate comparison of the
performance of different sources. This may be accomplished for excess
emissions by comparing the total duration of exceedances in a quarter
to the CEMS operating time (excluding CEMS operation during a source
outage). To normalize CEMS downtime, a comparison to source operating
time would be made (again excluding CEMS operation during a source
outage).
In concept, the benefit of this approach is that a source with a
significant outage or significant periods of CEMS downtime will not
escape targeting if follow-up is deserved on the basis of its actual
period of operation. Because only two sources included in the pilot
study provided sufficient data to allow for normalization, the study
results are inconclusive on the actual value of this approach.
In order to normalize the data effectively, Missouri must obtain
source operating data each quarter in a form that allows for differen-
tiating periods of source downtime from CEMS downtime. This would
require a change in report content and data collection practices for
most facilities. The following are two factors for consideration by
Missouri.
o To what extent is normalization likely to result in targeting
facilities that would not have been targeted otherwise? In the
pilot study normalization would have resulted in targeting three
additional EERs (out of eight) if a 2% screening standard for
exceedances were applied, but no additional EERs if a 5% standard
were applied.
o Consider whether the targeting objective is focused on environ-
mental impact or on control system performance, or both. For
example, the air quality impact of reported exceedances from
Source A over less than a full quarter may not be significantly
different from the impact of the same duration of exceedances
from Source B over the entire quarter (depending on how the
exceedances in each case are actually distributed). However,
control system performance may well be worse for Source A.
C. Recommended Agency EER Review Procedures
Recommended Agency EER review procedures were influenced by EPA
guidance on EER review (along with supporting documents) published on
October 5, 1984; interviews with project participants; as well as the
-------
Page 29
review of approximately 135 EERs over a two-year period during the
project. The following is a summary of the recommended procedures:
(1) Select screening criteria for quarterly exceedances and downtime
review.
Screening criteria should be selected on the basis of what is
reasonable for workload planning, giving consideration to the trend of
improvement or degradation in performance. A 5% of the quarter
exceedance standard and a 10% of the quarter downtime standard would
have resulted in targeting 32 of 112 EERs for detailed review and
possible follow-up in 1986 based on experience during the project
timeframe.
(2) Expand the quarterly review to include the following elements;
o Utilization of an EER review checklist during review.
o Sending a follow-up letter if the EER is late.
o Review for obvious content problems.
o A detailed review of EERs failing the screening criteria.
o Some type of source contact if problems are identified.
Quarterly EER review would be expanded at a minimum to include a
more detailed review, additional computations, completion of an EER
review worksheet for each CEMS, and completion of a composite quar-
terly summary report. It may also include a reason code summary and
evaluation if the agency elects to do so.
The focus during review should be on any parameter which is
out-of-line in comparison to other EERs. The reviewer should deter-
mine whether the problems are currently occurring or are of a chronic
nature. The source should be contacted if problems are identified.
The time required to perform these tasks should not exceed 38
hours per quarter after start-up. This assumes that the agency will
perform a reason code summary and evaluation and that the average EER
will require 45 minutes to review. The estimated time is based on the
average time required by the project team to review EERs for the last
three quarters of the project. An average of approximately two hours
per EER were required for the first quarter.
-------
Page 30
(3) Establish an agency tracking system.
This should include a completed EER review checklist filed with
each EER, a composite quarterly summary of exceedances and downtime, a
quarterly highlight report of sources justifying follow-up and results
of the annual audit and review described below.
Establishing a new manual tracking system will require an addi-
tional effort at the outset; however, this effort should not be
significant after an initial break-in period.
(4) Conduct an annual audit and EER review.
This should include a sample audit of EERs to verify correct
application of EER review procedures. It should also include an
annual tabulation and simple analysis of the quarterly composite
summaries to target sources for special focus during the next year.
These activities are not intended to duplicate the quarterly
review. They provide a program management tool to consider the
overall effectiveness of the EER review program and a planning tool to
target certain sources for longer range consideration. Most impor-
tantly, they result in a yearly status report that a source may use to
evaluate its own performance in relation to others in the State.
Conducting an audit and preparation of an annual EER review
report should take no more staff time than a quarterly review.
However, additional management time will be required.
(5) Ensure management participation in the review process.
Most of the new EER review activities are designed in some way to
involve agency management more actively in the EER review program.
Each new activity includes some type of simple status report (which
can be easily prepared if the other review and evaluation activities
are adopted). These reports will serve to keep managers informed of
the results of each activity, and will also offer a convenient track-
ing mechanism for managers to gauge the progress of their pollution
control program for sources with GEMS's.
D. Opacity Emission Trends in Missouri
A review of opacity EER data during the project shows a trend of
improved emissions in Missouri from 1980 through mid-1984. The
following table illustrates this improvement.
-------
Page 31
ILLUSTRATION 6
Trend of Exceedances Reported
in Opaciry EERs from 1980-1984
(1) EERs Analyzed
(2) Average Duration* of
Exceedances for All ELFs
(3) Average Duration* of Exceed-
ances for All LEFcs Mir.'js Uie
Three Worst Performers
(4) Percentage of EERs Below
a 5* Duration* Threshold
(5) Percentage of EERs Be lew
a 2* Duration* Threshold
I960 1981 1982 1963 1984
(2 -2tr».)
41 43 57 78 41
4.1% 3.1% 2.0% 2.5% 1.0%
2.19% 1.03% .93% .54% .56%
73% 93% 68% 91% 98%
o3% 72% 74% 85% 68%
* Duration IB expressed a* a percentage of tfie total tine in each year.
During the project timeframe, the long term trend of improvement
is not as visible on a quarter to quarter basis. Nonetheless, there
is no significant degradation.
ILLUSTRATION 7
Quarterly Trend of Opacity Exceedances
Prior to and During the Project Timeframe
DURATION
(% OF THE
QUARTER)
,A11 OEMS 's
Six Study CEMS '3
JU.1 CEMS 'fc
J 3 Worst Pe.
l»t 4 ITC3 Zncl « IB§3 3rd ft 10B3 4U» 9 1M3 1st Q 18B4 Znd Q 19»i
Most exceedances during the project were directly related to
control and process problems at three facilities. The following three
graphs depict how these three sources compared to other sources during
the project timeframe. The first graph identifies the three sources
with the greatest duration of exceedances on the basis of an annual
average.
-------
Page 32
ILLUSTRATION 8
A Comparison of Individual Source Opacity Exceedances
During the One Year Project Timeframe
(Annual Average)
DURATION
(% OF
THE YEAR)
19 -
1O -
17 -
16 -
16 -
14 -
13 -
12 -
11 -
1O -
0 -
a -
7 -
c -
6 -
4 -
3 -
a -
1 -
o -
'/
.-•
r-
/'
>",•
'x-
|»- .
"
Hn
"7
'
/
n
^ ''
f
/
'
^
F
_ p~3 p~3 P~3 C*~5 P~l L ,
i i f 1 1 1 1 i n fn l m LI r
ABCOirCiBI JKLUNOPQR3TD MZ1N
The following graph shows how these three sources performed from
quarter to quarter, again in comparison to the 21 sources. It shows
that all three were over the mean for at least three of the four
quarters; but only one was over a 5% norm for three of the four
quarters.
ILLUSTRATION 9
Opacity Exceedance Trends for the Three Worst
Performers During the One Year Project Timeframe
DURATION
(% OF THE
QUARTER)
ZO -I
10 -
^
\
X\_.
a^J q 1083
MEAN
A
4U» Q 1OBJ
1st q
-------
Page 33
These exceedances are clearly attributable to control and process
problems. As the following graph indicates, the occurrence of these
problems at the three facilities is extraordinary in comparison to
similar occurrences at other facilities.
ILLUSTRATION 10
The Major Reasons for Opacity Exceedances
Among the Three Worst Performers
During the One Year Project Timeframe
PERCENT?IE
OF TOT;i
EXCEEDANCES
90 -
50 -
70 -
eo -
».
30 -
20 -
10 -
/ /' '" •
•x ••''''/
' .- -'' •-'
•• • ' -' X
^ -•"'/''/"
."'//'/'
ll
/•" •'' /' f
'/,-//
92%
yy'- •
' ••"'.-''' '"
'///'
'' / s / -
s .-' • .-
f .•' .'
/•' .-' ''
/'//',
•'/.•• •
'' S '' '
~'' / •' "
.->"'-.
96 5%
'///.
/ / ^ '
S •* s' *
.-'''/,-
/ :-'.''•-
'//'/.
s - .- x
^' .- " X
Q"7 Q%
a / • y%
'///.
'-'/,-'.•
39.8%
SOURC! A
SQURCI E
CONTROL EQUIPMENT FAILURE?
PROCESS PROBLEMS; UNKNOWN CAUSES
SOURCE G ALL OTHER SOUSCEJ
ALL OTHER REASONS
V. The Coal Sampling and Analysis (CSA) Project
The Iowa CSA project was not completed during the course of the
CEMS pilot study. The major goal was to develop criteria for use by
the Iowa agency in determining the acceptability of a CSA program as
an alternative to SOa CEMS's in coal-fired electric utility steam
generators. This goal was partially achieved by developing and field
testing CSA evaluation procedures at two utility locations in Iowa.
However, final acceptability criteria could not be established because
of unresolved policy issues at the State and EPA.
These issues involve primarily (1) the extent to which Iowa SO2
emission limits must be revised to satisfy EPA SIP requirements; (2)
the effect of possible NSPS regulation changes involving SOa moni-
toring that are currently under consideration; (3) the selection of an
-------
Page 34
emission averaging time for SO2 emissions (also an issue in the
pending NSPS rulemaking); (4) a determination of the confidence with
which the probability of short term SO2 exceedances must be estab-
lished for a CSA program to be approvable; and (5) a determination of
the extent to which potential short term exceedances must be avoided
for a CSA program to be approvable.
Unresolved questions were also raised by project team members and
industry consultants regarding the technical model used during the
project for predicting potential S02 emissions based on CSA data.
Additional comments from industry participants on the CSA portion of
the study are anticipated.
The following are major technical accomplishments of the CSA
project:
o The project has resulted in the first long term (and most exten-
sive) comparison of CSA, SO2 CEMS's, and Method 6B under control-
led conditions (i.e., thoroughly documented CSA procedures, CEMS
with an extensive QA program and external performance audits to
verify accuracy of the data, and long term Method 6B measure-
ments) .
o The concurrent S02 emission measurements provided by the three
measurement methods during the project allowed for evaluation of
model outputs and predictions derived from a sophisticated time
series/statistical model developed by EPA and already being used
to assist in the evaluation of CSA programs by some agency
offices. The evaluation showed a need for continued development
and refinement of the model. Principal problems identified
during the evaluation included:
specific user constraints,
some inaccurate and highly variable predictions/estimates,
excessive sensitivity of the model to missing data, and
assumptions regarding the impact of measurement error on the
likelihood of compliance which are inappropriate in some
circumstances.
o The project clearly demonstrated the relative precision of the
three measurement methods. The results showed that SO2 CEMS data
are much more precise than CSA data. The results also illustrate
the impact of imprecise measurement methods on determining the
likelihood of excess S02 emissions for various emission limits
and averaging periods. The results show that proper application
of Method 6B yields results less precise than SO2 CEMS's but more
precise than CSA, and that improper application of Method 6B may
yield data less precise than CSA data.
o The project demonstrated clearly that the level-of-effort and
time required to perform CSA/effluent S02 measurement comparisons
-------
Page 35
under controlled conditions were much greater than many people
expected. It is the opinion of the project team that a fundamen-
tal understanding of CSA principles and practices (rather than a
black-box performance evaluation approach) is necessary for the
practical evaluation of CSA programs.
The project allowed for an evaluation and a field test of a
protocol developed by the EPA Quality Assurance Division and UARG
consultants for the evaluation of alternate analytical methods.
This test ultimately resulted in modifications to the protocol.
Thus, the modified protocol was accepted (as was the alternate
analytical method). The project also allowed for the evaluation
of simpler, cheaper, and more expedient methods for the evalua-
tion of alternate analytical methods.
During the project, methods for characterizing fuel sulfur
depletion were evaluated. The project demonstrated that fuel
sulfur depletion can be very significant (i.e., greater than 20%)
for some coals and source conditions. Two methods of quantifying
fuel-sulfur depletion were evaluated: (1) comparison of CSA and
effluent SO2 measurements, and (2) ash sampling and analysis.
Together, these two methods provide a series of alternatives of
varying complexity and cost both for initial determination of the
fuel-sulfur depletion factor and for follow-up verification of
the depletion factors.
During the project, all major technical CSA issues were organized
into a conceptually consistent framework to facilitate an agency's
choice of options regarding:
CSA procedural requirements,
CSA precision,
CSA bias,
Use of CSA data as a surveillance method for short term
averaging periods, and
Fuel sulfur depletion.
-------
Page 36
Appendix A
Bibliography of Technical GEMS Pilot Project Reports
I. MISSOURI OPACITY CEMS PROJECT
Audit Report; LSI RM41 Opacity Monitoring System at Kansas City Power
and Light Company, latan Station, Unit #1, Entropy Environmen-
talists, Inc., (Draft Report, March 1983).
Performance Audit Report; LSI RM41 Opacity Monitoring System at
Kansas City Power and Light Company, latan Station, Unit #1,
Entropy Environmentalists,(Draft Report, June 1984).
Performance Audit Report; Dynatron Model 1100 Opacity Monitoring
Systems at City Utilities, James River Station, Units #4 and #5,
Springfield, Missouri, Entropy Environmentalists, Inc..(Draft
Report, April 1983).
Performance Audit Report; Dynatron Model 1100 Opacity Monitoring
Systems at City"utilities, James River Station, Units #4 and #5,
Springfield, Missouri, Entropy Environmentalists, Inc., (Draft
Report, June 1984).
Performance Audit Report: Contraves-Goerz Model 400 Opacity
Monitoring SysteitTat St. Joseph Light and Power Company, Lake
Road Station, Unit #5, St. Joseph, Missouri, Entropy Environ-
mentalists, Inc., (Draft Report, March 1983).
Performance Audit Report; Contraves-Goerz Model 400 Opacity
Monitoring SystenTat St. Joseph Light and Power Company, Lake
Road Station, Unit #5, St. Joseph, Missouri, Entropy
Environmentalists, Inc., (Draft Report, August 1984).
Audit Report: LSI RM41 Opacity and Monitoring System, Unit #1 and
Unit #2, Union Electric Power Company, Sioux Station, Entropy
Environmentalists^(Draft Report, March 1983).
Performance Audit Report; Lear Siegler Opacity Monitoring Systems at
Union Electric Company, Portage de Sioux Station, Unit #2,
Entropy Environmentalists, Inc.,(Draft Report, June 1984).
Final Opacity CEM Audit, Union Electric Company, Portage Des Sioux
Station, Unit #1, Entropy Environmentalists, (Draft Report,
February 25, 19135) .
An Analysis of Opacity CEMS Downtime in Excess Emission Reports, James
W. Peeler (Draft Report, March 1985).
-------
Page 37
EER Review Recommendations for Missouri's Air Pollution Control
Program; Opacity CEMS Installed at Electric Utilities, Perrin
Quarles, (Draft Report, May 23, 1985).
Evaluation of Opacity CEMS Reliability and QA Procedures: Volumes I
and II, James W. Peeler, (Draft Report, April 1985).
Recommended Quality Assurance Procedures for Opacity CEMS's, James W,
Peeler, (Draft Report, May 1985) .
II. MISSOURI S02 CEMS PROJECT
System Performance Audit: LSI Continuous SO2 Emission Monitoring
System: Union Electric Company, Labadie Generating Station, Unit
#2, Labadie, Missouri, Entropy Environmentalists, Inc., (Draft
Report, October 1983).
System Performance Audit; KVB Stack Test Gas Continuous Emission
Monitoring System; Sikeston Power Station, Unit #1, Sikeston,
Missouri, Entropy Environmentalists, Inc., (Draft Report, October
1983).
System Performance Audit: KVB Stack Test Gas Continuous Emission
Monitoring System; Sikeston Power Station, Unit #1, Sikeston,
Missouri, Entropy Environmentalists, Inc., (Draft Report, June
and August 1984).
System Performance Audit: Dupont Continuous Emission Monitoring
System: City Utilities, Southwest Power Station, Unit #1,
Springfield, Missouri, Entropy Environmentalists, Inc., (Draft
Report, January 1984).
System Performance Audit; Dupont Continuous Emission Monitoring
System; City Utilities Southwest Power Station, Unit #1,
Springfield, Missouri," Entropy Environmentalists, Inc., (Draft
Report, June 1984).
Missouri SOa CEM Pilot Project Report, Wayne E. Reynolds, (Draft
Report, August 1985).
III. IOWA S02 CEMS PROJECT
System Performance Audit; LSI SO, Continuous Emission Monitoring
System, Iowa Public Service Company, George Neal North Station,
Unit #4, Sioux City, Iowa, Entropy Environmentalists, Inc.,
(Draft Report, December 1983).
-------
Page 38
System Performance Audit; LSI S02 Continuous Emission Monitoring
System, Iowa Public Service Company, George Neal South Station,
Unit #4, Sioux City, Iowa, Entropy Environmentalists, Inc.,
(Draft Report, September 1984).
System Performance Audit: Dupont Continuous SO.,/0., Emission
Monitoring System, Muscatine Power and Water Unit #9, Muscatine,
Iowa, Entropy Environmentalists, Inc., (Draft Report, May 1984).
System Performance Audit: Continuous SO.,/0., Emission Monitoring
Systems, Muscatine Power and Water, Unit #9, Muscatine, Iowa,
Entropy Environmentalists, Inc., (Draft Report, September 1984).
Iowa SO, CEM Pilot Project; Evaluation of Quality Assurance
Procedures and Monitor Performance, Wayne E. Reynolds, (Draft
Report, September 1985).
IV. IOWA CSA PROJECT
Coal Sampling and Analysis Demonstration Project, Iowa Public Service
Company, George Neal, Unit #4, Draft Report, James W. Peeler and
Phillip J. Juneau, (Draft Report, July 22, 1985).
Coal Sampling and Analysis Demonstration Project: Muscatine Power and
Water, Unit #9, Entropy Environmentalists, Inc., (Draft Report,
July 1985) .
Memorandum: Evaluation of CSA Programs, James W. Peeler, (Draft
Report, August 6, 1985).
-------
Page 39
Appendix B
Bibliography of Written Comments on
CEMS Pilot Project Reports
"Comments on EPA Draft Report: An Analysis of Opacity CEMS Downtime
in Excess Emission Reports," prepared for the Utility Air
Regulatory Group by Kilkelly Environmental Associates, Inc.,
Report No. 85-I190-1-02F, (June 1985) .
Letter from Steven D. Brooks, St. Joseph Light & Power Company, to
Perrin Quarles, Perrin Quarles Associates, Inc., re comments on
draft report "EER Review Recommendations for Missouri's Air
Pollution Control Program: Opacity CEMS Installed at Electric
Utilities," (June 12, 1985).
Letter from Steven D. Brooks, St. Joseph Light & Power Company, to Jim
Peeler, Entropy Environmentalists, Inc., re comments on draft
report "Evaluation of Opacity CEMS Reliability and QA Procedures,
Volume 1," (June 13, 1985).
Letter from Steven D. Brooks, St. Joseph Light & Power Company, to Jim
Peeler, Entropy Environmentalists, Inc., re comments on draft
report "Recommended Quality Assurance Procedures for Opacity
CEM's," (June 14, 1985).
Letter from Steven D. Brooks, St. Joseph Light & Power Company, to Jim
Peeler, Entropy Environmentalists, Inc., re comments on draft
report "An Analysis of Opacity CEMS Downtime in Excess Emission
Reports," (June 14, 1985).
Letter from Walter C. Gray, Jr., Kilkelly Environmental Associates, to
Richard D. McRanie, Southern Company Services, Inc., and Gary L.
Huber, Union Electric Company, re review of EPA Region VII
Opacity Pilot Project Report "Opacity CEMS Audit Procedure
Guidelines," (June 28, 1985).
"Comments on EPA Draft Report: EER Review Recommendations for
Missouri's Air Pollution Control Program: Opacity CEMS Installed
at Electric Utilities," prepared for the Utility Air Regulatory
Group by Kilkelly Environmental Associates, Inc., Report No.
85-I190-2-02F, (July 1985).
Letter from David M. Fraley, Ph.D., City Utilities of Springfield, to
Perrin Quarles, Perrin Quarles Associates, Inc., re comments on
draft report "EER Review Recommendations for Missouri's Air
Pollution Control Program: Opacity CEMS Installed at Electric
Utilities," (July 11, 1985).
-------
Page 40
Letter from Gerald L. Bennett, Kansas City Power & Light Company to
Anthony P. Wayne, Air Branch, Region VII, U.S. Environmental
Protection Agency, re comments on draft report "EER Review
Recommendations for Missouri's Air Pollution Control Program:
Opacity CEMS Installed at Electric Utilities," (July 12, 1985).
Letter from Darrell McAllister, Iowa Department of Water, Air and
Waste Management, to James W. Peeler, Entropy Environmentalists,
Inc., re comments on draft reports "Coal Sampling and Analysis
Demonstration Project: Iowa Public Service COmpany, George Neal
Unit No. 4," and "Coal Sampling and Analysis Demonstration
Project: Muscatine Power and Water, Unit No. 9," (September 18,
1985).
"Comments on EPA Draft Report Entitled: 'Evaluation of Opacity CEMS
Reliability and QA Procedures'," prepared for the Utility Air
Regulatory Group by Kilkelly Environmental Associates, Inc.,
Report No. 85-I190-3-02F, (November 1985).
Letter from John Hardie, Iowa Public Service Company, to Wayne
Reynolds, Entropy Environmentalists, Inc., re comments on draft
report "Evaluation of QA Procedures and Monitor Performance,"
(November 20, 1985).
"Technical Memorandum: Comments on EPA-Contractor Draft Report
Entitled 'Coal Sampling and Analysis Demonstration Project, Iowa
Public Service Company, George Neal, Unit No. 4," prepared for
the Utility Air Regulatory Group by Kilkelly Environmental
Associates, Inc., (January 22, 1986).
Letter from Richard D. McRanie, Southern Company Services, Inc., to
Mark S. Siegler, Chief, Technical Support Branch, U.S.
Environmental Protection Agency, re comments on draft summary
report "A Pilot Project to Demonstrate the Feasibility of a State
Continuous Emission Monitoring System (CEMS) Regulatory Program,"
(February 19, 1986).
-------
Page 41
Appendix C
Project Reports to be Finalized
for General Distribution
(1) Summary Report: A Pilot Project to Demonstrate the Feasibility
of a State Continuous Emission Monitoring System Regulatory
Program
(2) Opacity Excess Emission Report Review Findings and Recommen-
dations
(3) Missouri Opacity CEMS Pilot Project: Evaluation of Opacity GEMS
Quality Assurance Procedures and Monitor Performance
(4) Missouri SO2 CEMS Pilot Project: Evaluation of Quality Assurance
Procedures and Monitor Performance
(5) Iowa SO2 CEMS Pilot Project: Evaluation of Quality Assurance
Procedures and Monitor Performance
(6) Recommended Quality Assurance Procedures for Opacity CEMS's
(7) Recommended Quality Assurance Procedures for S02 CEMS's
(8) Recommended Audit Procedures for Opacity CEMS's
(9) Recommended Audit Procedures for Gaseous CEMS's
-------
------- |