United States
Environmental Protection Air and Radiation EPA/400/1-91/005.C
Agency (ANR-445) April 1991
EPA Acid Rain Advisory
Committee Meeting:
December 13-14,1990
Emissions Monitoring
Issue Papers
-------
Emissions Monitoring Issue Paper
Index
E-l Background Paper on Continuous
Emission Monitoring
-------
El
Background Paper
Continuous Emissions Monitoring
Continuous emissions monitoring (CEM) is pivotal to the
successful implementation of Title IV of the 1990 Amendments to
the Clean Air Act (CAA) and to the achievement of mandated
reductions in the principal acid deposition precursors—emissions
of sulfur dioxide and nitrogen oxides from the combustion of
fossil fuels. Further, the Title's market-based allowance
trading program will greatly expand both the uses and users of
CEM data. Under Title IV, CEM will evolve from a technical
instrument for assessing affected sources' operations into, as
suggested in the allowances background paper, the gold standard
backing up the paper currency of emissions allowances, enabling
the market to work. New CEM data users will include financial
brokers, ordinary citizens, and others who have little knowledge
of CEM technology. CEM will become the yardstick people will use
for measuring EPA's progress towards the emissions reduction
goals of Title IV.
WHAT IS CEM?
CEM is the monitoring on a continuous basis by utility and
industrial sources of pollutant concentrations emitted into the
atmosphere by exhaust gases from combustion processes. Monitored
gas concentrations include:
• Sulfur dioxide (S02) ,
• Nitrogen oxides (NOX), and
• Other pollutants (e.g., CO, H2S) and diluents (CO2, O2).
When used in its broadest sense (as in Title IV), CEM also refers
to the continuous opacity monitoring (COM) of obscuration caused
by source particulate emissions.
Prior to 1978, EPA used CEM only as a relative indicator of
proper operation and maintenance of pollution control equipment.
EPA's September 19, 1978 Subpart Da New Source Performance
Standards (NSPS) proposal, however, launched a new era for CEM by
specifying that CEM data will be used to assess affected sources'
compliance with SO2 and NO, emission standards. Today, almost
half of U.S. electric utility plants are already subject to
federal and/or state CEM regulations, promulgated under State
Implementation Plans (SIPs), NSPS, National Emission Standards
for Hazardous Air Pollutants, Prevention of Significant Deterior-
ation (PSD), and other programs. About three-quarters of the
large Phase I sources have CEM systems (CEMSs) installed on their
affected units, whereas very few of the smaller Phase II sources
have CEMSs.
-------
Three basic types of CEMSs exist: extractive, in-situ, and
remote. Extractive systems extract samples of exhaust gases from
the stack, condition them, and analyze them for pollutant con-
centrations. In-situ systems, as the name implies, analyze
pollutant concentrations in exhaust gases directly in the stack.
They typically use optical methods, including infrared
absorption. Remote systems determine pollutant concentrations in
exhaust gases after they have exited from the stack, generally
from a distant location. They include sophisticated light
detection and ranging (lidar) systems that use various types of
laser to determine pollutant concentrations near the stack tip.
Although remote systems are evolutionary and few, if any, sources
use them on a permanent basis, they are being used by EPA
enforcement divisions to assess compliance. However, we do not
expect remote systems to play a major role in the Acid Rain
program implementation.
At a recent Air & Waste Management Association specialty
conference on CEM, the keynote speaker stated that a review of
available information indicates a serious failing in the area of
uniform application of current CEM requirements for compliance
determination. Part of the reason for the uneven implementation
of CEM programs, according to this spokesperson, is the
widespread (but unfounded) concern that CEM equipment may be
inherently unreliable.
Actual CEM data in a national 1988 database of excess
emission reports (EERs) submitted by utility sources required to
have CEM does not support this belief. Reported failures for
systems that meet EPA's minimum performance specifications,
quality assurance, and quality control requirements are on the
order of 4-8 percent or less. These reports indicate that
expected CEM data capture levels of 95 percent would not be
unreasonable.
MAJOR CEM PROVISIONS IN TITJ-*? Tv
Require CEMS on Affected Units. Section 412 (Monitoring,
Reporting, and Recordkeeping) of Title IV requires each affected
unit to install and operate a CEMS, or an alternative monitoring
system which provides information with the same precision,
reliability, accessibility, and timeliness as CEM. The source
must quality assure, record and report to EPA the data for, so2,
NOX, opacity, and volumetric flow at each of its affected units.
Where two or more units share a single stack, a separate CEM for
each unit is not required, but sufficient information must be
provided to permit reliable compliance determinations for each
unit.
Require Flow Monitors to Compare Unit's Annual SO, Emissions
With Allowances. Traditionally, CEM data has been recorded in
pounds of pollutant per unit of heat input (Ibs/mmBtu), parts per
-------
oillion, or micrograms per cubic meter. Under Title IV, however,
volumetric flow»data is required to compute each affected unit's
SO, emissions in tons/year for comparison with the allowances it
holds. Such data is also needed for levying the Title's excess
emissions fee and offset against units who do not have "balanced
books" at the end of the year. Flow monitors (or acceptable
alternatives) will be needed to measure so, mass emitted per unit
time.
Several types of flow monitors are commercially available,
although only a handful of sources currently have them installed.
Flow monitors determine volumetric stack gas flow using techni-
ques such as sound wave velocity differential, temperature
differential, and pressure differential. The suitability of at
least some of these systems (which are typically used in process
industries for measuring the flow of steam, natural gas, and
other gaseous and liquid materials) for precise CEM applications
may be questionable.
Restrict Use of Alternative Monitoring Systems. Section 412
contains explicit and stringent criteria for assessing the
equivalency of alternative monitoring systems ("same precision,
reliability, accessibility, and timeliness as CEMS"). Alterna-
tive monitoring systems for SO2 emissions determinations which
EPA has accepted under some circumstances in other CEM
regulations include: fuel sampling and analysis (FSA), and a
continuous wet-sampling technique (Method 6B). EPA proposed
these alternative systems primarily as a means to supplement CEM
data during periods of monitor failure and to provide (smaller)
plants with more flexibility in complying with emissions
monitoring regulations. No appropriate alternatives have been
identified for nitrogen oxides emissions*
Section 412 affords considerable discretion to EPA in
restricting the use of alternative monitoring systems—even those
that satisfy the criteria for equivalency. The CEM regulation
which EPA must promulgate may include limitations on alternative
systems "as necessary to preserve the orderly functioning of the
allowance system or [to] ensure the [mandated] emissions
reductions.*1
Use Standard Method for Calculating Emissions When CEM Data
is Unavailable. If data from a CEMS or an approved alternative
monitoring system is unavailable, and the owner or operator
cannot provide satisfactory emissions information, the statute
requires EPA to deem the unit "to be operating in an uncontrolled
manner." It further states that EPA must prescribe a standard
method in the CEM regulation to calculate emissions for that
period. Publication of this universal method more than a year
before Phase I begins should contribute to the certainty needed
for a robust allowance trading market and create incentives for
CEM compliance.
-------
Rulemakina and Compliance Schedules. EPA must promulgate a
CEM regulation within 18 months of enactment (by May 15, 1992)
that contains:
• CEMS performance specifications (certification tests,
QA/QC and audit procedures) for SO2, NOX, opacity and
volumetric flow monitoring;
• Definition and use of alternative monitoring systems;
• Method for calculating emissions when CEM data is
unavailable; and
• Requirements for recordkeeping and reporting various
types of CEM data.
A fundamental issue is whether existing NSPS performance specifi-
cations are adequate for acid deposition control, or should
improvements be incorporated into the Acid Rain CEM program? In
particular, additional specificity may be required in existing
flow monitor performance specifications given their importance to
the proper functioning of the allowance trading program.
Within 36 months of enactment (by November 15, 1993), Phase
I affected units, including substitution units (Sections 404 (b) &
(c) , and eligible Phase I extension units (Section 404 (d)) must
install and operate CEMS, quality assure the data, keep records
and report the data in accordance with the CEM regulation.
Not later that January 1, 1995, Phase II affected units must
comply with the regulation. New utility units must comply upon
commercial operation. Each industrial or process source that
elects to become an affected unit under Section 410 must also
comply with the CEM regulation.
RELATIONSHIP OF SECTION 412 CEM REQUIREMENTS TO OTHER
SECTIONS OF T
CEM is Essential to Enforcement. Particularly Under the
Allowance Trading Program. As stated previously, explicit and
certain CEM requirements are essential to implementation of the
enforcement provision (Section 414) and the excess emissions
penalty (Section 411) within the context of the Title's market-
based allowance trading program (Section 403). They are also
necessary for enforcement of the nitrogen oxides emissions
reductions mandated by Sections 401 and 407. An important
purpose of Section 412 is to ensure that the Phase I and II
allowances (Sections 404-406) are not exceeded and that the SO2
emissions cap (Section 403) holds. Because interstate trading of
allowances is likely, the Section 412 accounting system will be a
federal program, implemented primarily from EPA Headquarters.
-------
CEM will be Integral to the Permitting Process. The CEM
monitoring, recordkeeping, and reporting requirements under
Section 412 will become part of the source permit requirements
and specifications (Section 408). The intent is to communicate
unequivocally what is required and its application to each
affected facility. Like Section 412, the permitting process for
Phase I sources will be a federal program implemented by EPA
Headquarters.
CEM wj_ll Support Conaressionally Mandated Research and
Emissions Inventories. Section 412 could be used to provide
valuable data in support of the industrial source sulfur dioxide
emissions inventory and a twenty-year industrial emissions
projection'required every five years under Title IV. The acid
deposition research by the U.S. Fish and Wildlife Service will
also be supported by Section 412.
KEY IMPLEMENTATION ISSUES
There are a number of important issues that will need to be
addressed before an effective monitoring regulation can be
promulgated. These issues include: required accuracy for CEMS,
method for calculating emissions when CEM data is unavailable,
criteria for alternative monitoring systems, averaging time for
pollutant concentration data, data integrity/security, emissions
tracking system, format and frequency of data reporting, multiple
units with common stack, current source population with CEMS and
flow monitors, auditing program, and cost effective improvements
to existing CEMS.
What should be the accuracy of emissions data? What ac-
curacy can/should be required of CEMS, including flow monitors?
This issue involves the need to accurately determine the amount
of allowances consumed at the end of the year so that the books
can be balanced to the nearest ton of SO2. EPA expects that the
CEM and flow monitoring systems together would produce results
better than + 10 percent. If the system meets EPA quality
control and quality assurance (QC/QA) requirements, it would be
equally probable over time that errors in monitored data would be
plus or minus; and EPA could treat the monitored data values as
accurate to any required significant figure.
Are there any methods equivalent to CEM? Another issue is
whether there are any alternative methods that provide informa-
tion with the same precision, reliability, accessibility, and
timeliness as CEM; and if so, what criteria should be used to
judge equivalence? No acceptable alternative methods have as yet
been developed for nitrogen oxides emissions. What percentage of
emissions data, if any, should be allowed from an alternative
method?
-------
What should be the averaging time for pollutant concentra-
tion data? CEMS can provide pollutant averaging times as short
as a minute or less. Should the averaging times vary by pol-
lutant? Should the averaging times be the same as for NSPS?
Ultimately, a cumulative annual emissions mass per unit time
value is needed for sulfur dioxide emitted by each affected unit.
Do we need short-term emissions data, say, hourly emissions, for
quality assurance procedures, or to satisfy other Acid Rain
program requirements? A basic question is how can EPA
efficiently integrate the acid rain monitoring requirements with
the other air program monitoring requirements to be promulgated
under the amended CAA?
How will accurate and timely emissions tracking information
be assured? Such information is vital to the functioning of the
allowance trading program and to the conduct of credible program
evaluations. How will data reporting be handled? At what
frequency will data be reported? How should data integrity/
security be addressed? Should electronic reporting be required
or at least strongly encouraged? How much data should EPA
require the source to retain on site, and how much should the
source be required to send to EPA? What EPA data system should
be used to record emissions data initially, and in the longer
term? Should a different system be used to record opacity or
nitrogen oxides?
How will emissions be calculated for periods of missing or
invalid CEM data? The statute requires uncontrolled emissions to
be used for missing or invalid CEM data when no acceptable
substitute can be provided, but what about data gaps caused by
calibration periods or due to events such as planned outages for
maintenance?
How should multiple units with a common stack be handled?
While the statute does not require that each such unit have a
GEM, it does require that data be provided to allow compliance
determinations for each unit. What types of information would
facilitate reliable compliance determinations for each affected
unit using a common stack?
What are current CEMS/source population characteristics?
How many, and what kinds of sources currently have CEM equipment;
what type of CEM equipment? Will vendors be able to meet ex-
pected increases in demand for new and/or improved CEM equipment?
What kind of auditing program should be established, con-
sidering possible resource constraints? Proper auditing of
emissions monitoring systems is crucial in achieving consistently
accurate emissions data.
How can cost-effective improvements be made to existing
CEMSs? It could be burdensome if units were required to replace
-------
their existing CEMS with a new one. Can the accuracy of existing
CEMSs be increased by improving operation and maintenance proce-
dures?
We hope this background document on continuous emissions
monitoring under Title IV of the 1990 CAA Amendments will be
useful to the Acid Rain Advisory Committee (ARAC) as ARAC begins
discussion on this Title. We look forward to working with the
ARAC throughout the rulemaking process, and to benefiting from
ARAC opinions. Our hope is that a better, less litigated,
emissions monitoring regulation will be the result.
------- |