MEMORANDUM  SEQ CHAPTER \h \r 1    

TO:		Melanie King, EPA OAQPS/SPPD/ESG

FROM:	Bradley Nelson, EC/R, Inc.

DATE:		January 21, 2010

SUBJECT:	Summary of the October 14, 2009 Teleconference with the
American Petroleum Institute and the U.S. Environmental Protection
Agency to discuss the Proposed Test Plan for Existing Stationary Spark
Ignition Reciprocating Internal Combustion Engines

1.0	INTRODUCTION

The American Petroleum Institute (API) requested this teleconference
with EPA to discuss the proposed national emission standards for
hazardous air pollutants (NESHAP) for existing stationary reciprocating
internal combustion engines (RICE), and to propose a plan to gather
additional test data for spark ignition (SI) RICE.  This meeting summary
will be available in the docket for this rulemaking (Docket ID No.
EPA-HQ-OAR-2008-0708).  The proposed NESHAP for existing stationary RICE
was published in the Federal Register on March 5, 2009 (74 FR 9698).  

2.0	ATTENDEES

U.S. Environmental Protection Agency

Melanie King, OAQPS/SPPD/ESG

Michael Horowitz, OGC

Gary McAlister, OAQPS/AQAD/MTG

American Petroleum Institute

Jeff Panek, IES

Roger Zygmunt, Shell

Jenny Yang, Marathon

Tom McGrath, IES

Paul Little, Air Hygiene

Jeff Adams, BP

Tom Monahan, ExxonMobil

Jim McCarthy, IES

Lynn Reed, ONEOK

Janet Bounds, Chevron

Tim Helbig, ONEOK

Matt Todd, API 

EC/R Incorporated

Bradley Nelson 

3.0	SUMMARY OF MEETING

API started the teleconference by asking what the status was on the
extension for the SI engine portion of the final NESHAP.  EPA responded
that it is currently in discussions and expects a response on the
extension in the near future.  

API stated that it was in the process of developing “white papers”
on the following topics: above the floor cost analysis, work practices
for area sources, and maintenance intervals for work practices.  The
papers will be submitted to EPA when they are complete.  API stated that
the test data collection process for SI engines is ongoing and that it
is currently looking for test contractors to perform the testing.  API
stated that testing would not begin until November 2009 and API plans to
look at emission variability during the testing campaign.  EPA asked how
many tests would be provided from the testing, in which API responded
approximately 100 tests.  The test plan for the proposed testing for SI
engines is presented as Attachment A at the end of this meeting summary.

API asked EPA if there had been any additional discussions on the
urban/rural issue for existing stationary SI engines at area sources. 
API proposed only work practice standards for rural SI engines located
at area sources in comments on the proposed rule.  EPA indicated that no
final decisions have been made on this issue.  API also brought up the
issue of variability of emissions test data. EPA stated that it would
prefer to have variability data for all engines in the data set.  API
responded that it does not have the resources to test for variability
for so many engines and stated that it could perhaps provide test data
for the top 20 percent.  API stated it is finalizing a technical paper
on test data using Method 323 and will forward to EPA when the paper is
completed.

API asked EPA if it had received additional CI engine test data.  API
stated that the MACT floor for CI engines was based on test data from a
single engine tested at various loads, which was inappropriate in
API’s opinion.  EPA responded that no CI engine test reports were
received during the comment period, but emissions data was referenced in
a few of the comments.  EPA intends to follow up with the commenters to
see if the referenced emissions data is available as a test report and
will include in the MACT floor determinations.  API also asked EPA if
any decisions were made regarding the emission standards for cyclical
engines, and whether the urban/rural classification would be extended to
diesel engines.  EPA responded that the options are being looked at and
no final decisions have been made.

API stated that next steps were to wait for the decision on the final
rule extension for SI engines, and to send EPA the technical papers for
the Method 323 testing and the variability of emissions from SI
engines.

Attachment A

40 CFR 63 Subpart ZZZZ 

Natural Gas Fired Engine MACT Floor Test Plan 

& 

Equipment / Operations Information Gathering40 CFR 63 Subpart ZZZZ 

Natural Gas Fired Engine MACT Floor Test Plan 

& 

Equipment / Operations Information Gathering

PREPARED FOR: 

Stephanie Meadows

Upstream Senior Policy Analyst

American Petroleum Institute

1220 L Street, NW

Washington, D.C. 20005

PREPARED BY:

innovative environmental solutions, inc.

P.O. Box 177

Cary, IL 60013-0177

September 30, 2009							Version 1

MACT Floor Test Program Background and Objectives

The March 2009 proposed revisions to the Federal National Emission
Standards for Hazardous Air Pollutants (NESHAP) for reciprocating
internal combustion (IC) engines (74 FR 9698) include emission standards
based on insufficient and technically flawed emissions data.  The
resulting paucity of data makes it impractical to complete a MACT floor
analysis representative of the affected engine populations.  Flaws in
the existing data include emission test methods with known negative
biases and inappropriate engine operation during the testing.  Based on
industry comments, EPA has acknowledged that additional emissions data
and analysis is needed to define emission limits for the Final Rule.  To
address the data deficiency and inform the MACT floor analysis, API and
a consortium of engine rental companies are coordinating an emission
test project on existing reciprocating internal combustion engines at
oil and natural gas (O&G) facilities.  This document identifies the Test
Plan for that project to ensure accepted and consistent methods are used
for the testing.  The test results will be integrated into a single
emissions test report at the completion of the test program. 

For major source engines and most area source engines, the Proposed Rule
includes emission limits.  The emission limits are based on the “MACT
floor”, and the technology basis is an “uncontrolled” engine.  In
some cases, the emission limit is the MACT floor emission level, while
in other cases EPA assumed that “above the floor” catalytic control
is warranted and based the emission limit on the floor level less a 90%
reduction from catalytic control.  These emission limits will not be
achievable by the vast majority of uncontrolled existing engines; that
is, for an uncontrolled unit to meet the emission level associated with
the “best performing” MACT floor limit, catalytic control will be
required for nearly all units.  Refer to API Comments 5.1 and 6.2 for
additional detail.

This test program is designed to obtain uncontrolled “as found”
emission data to support revisions to the MACT floor limits in the
upcoming Final Rule that will amend the NESHAP for reciprocating IC
engines 40 CFR, Part 63, Subpart ZZZZ.  Data collected as part of this
intensive effort are needed to develop technically defensible emission
standards.

The MACT floor emissions test project objectives include:

Establish a test program to consistently obtain emission data for
developing MACT floor emission limits that address the size, type, and
operation of affected engines.  The focus of the MACT floor testing
program is primarily directed toward smaller natural gas fired engines 
   (< 500hp).   

Obtain operational variability data from engines that are expected to be
in the MACT floor data set (i.e. engines in the best performing 12
percent of engines in each category/subcategory) to allow determination
of emission limits that include normal operational variability. 
Variability parameters may include:

Excess O2 from 4SRB;

Percent load;

Multiple engines of identical make / model;

Fuel quality; and

Ambient affects including temperature, relative humidity, & elevation.

  

Variability is an important consideration because it allows
consideration of operating factors when establishing the MACT floor
emission limit.  Thus, these data are a key component of the test
program.  For example, when establishing the standard, our understanding
is that EPA is not required to use an average of the best performing
twelve percent of units as the MACT floor emission limit.  Rather, EPA
has the latitude to select the limits at the worst reasonable condition
for the best controlled similar unit. [“EPA would be justified in
setting the floors at a level that is a reasonable estimate of the
performance of the ‘best controlled similar unit’ under the worst
reasonably foreseeable circumstances.” Sierra Club v.EPA, 167 F.3d 658
(D.C. Cir. 1999)].  

In addition, in Mossville Environmental Action Now v. EPA, 370 F.3d 1232
(D.C. Cir. 2004), the Court held that floors may legitimately account
for variability because “each [source] must meet the [specified]
standard every day and under all operating conditions.”  Thus,
emission variability can be considered in setting emission limits, and
MACT floor emission limit can be based on the highest emissions for MACT
floor engines considering operational variation.  

Table 1 presents a draft test matrix template for the primary engine
categories.  The engine category priorities discussed in this plan are
based upon the RICE fleet used by natural gas production and gathering
facilities and test data reviewed in the RICE NESHAP proposal.  As
testing commitments, locations, schedules, and engine types are
finalized, Table 1 will be completed as a stand alone supplementary
document that identifies engine makes/models and counts.  At present,
industry and trade associations anticipate on the order of 100 tests
will be completed over a range of engine types, sizes, and operating
conditions.  

Additional subcategories or further refinement of the horsepower ranges
can be completed based on participant direction and engine types
available for testing.  Since the MACT floor is based on an uncontrolled
engine, NSCR and oxidation catalyst equipped engines are not preferred
for the purpose of establishing the MACT floor.  Should controlled
engines be included in the measurement program, pre-catalyst testing is
recommended.   In addition, proper consideration of air fuel ratio (AFR)
system test impacts should be made.  Potential test considerations
includes disconnecting the AFR and reverting to manual adjustments or
AFR system bypass (see 4SRB O2 considerations below).   

Similar to the proposed draft test matrix, Table 2 presents a table to
summarize emissions data that is compiled from previously completed
tests (i.e., available test results from O&G companies).  These data may
be used to establish or confirm MACT floor emission limits or further
the understanding of emissions variability.  These data are expected to
be very limited for smaller burn engines (i.e., where formaldehyde is
needed) and primarily encompass larger lean burn units.   

Size Category

 (HP)	Number of

4SRB 

Engines	Number of

2SLB 

Engines	Number of 

4SLB 

Engines	Subcategory-Specific

(e.g. cranes, air compressors, etc.)

50-100 hp











MACT Floor 

Engine Count	X	X	X	X

100-250 hp











MACT Floor 

Engine Count	X	X	X	X

250-500 hp

e.g. …5 (Ajax DPC 280)









MACT Floor 

Engine Count	X	X	X	X

>500 hp

















Total	X	Y	Z	AA

 Table 1.	Test Matrix and Test Engine Count (MACT floor basis is 12% of
the engines tested)

Table 2.	Existing Test Data to Support MACT Floor Determination

Size Category

 (HP)	Number of

4SRB 

Engines	Number of

2SLB 

Engines	Number of 

4SLB 

Engines	Subcategory-Specific

(e.g. cranes, air compressors, etc.)

50-100 hp











MACT Floor 

Engine Count	X	X	X	X

100-250 hp











MACT Floor 

Engine Count	X	X	X	X

250-500 hp











MACT Floor 

Engine Count	X	X	X	X

>500 hp

e.g. …1 - (Clark TLA-6) 



Total	X	Y	Z	AA



Reference Test Methods and Procedures

Testing should be conducted in accordance with EPA/ASTM reference test
methods and procedures; however, due to time constraints and limited
resources, data will typically not be specifically conform to
performance or compliance test criteria (e.g., three one-hour tests are
not recommended for the purpose of gathering MACT floor engine
performance data on multiple units when continuous test method (e.g.,
FTIR) is used).  Because these data will be submitted to EPA for
regulatory development, adherence to the test methods is necessary.  Any
exceptions or deviations to the reference method procedures must be
documented to assess the appropriateness, quality, adequacy, and
representativeness of the emission data.

The testing will include a single one hour test on an “as-found”
engine (i.e. no adjustment to the engine will be made – excluding 4SRB
O2 exhaust levels as discussed below).  For units identified as MACT
Floor engines, subsequent tests at varied load conditions, possible
varied ambient condition (e.g. early morning vs. mid-day operation),
and, for 4SRB units, varied exhaust O2 levels should be conducted to
gather data that characterizes emissions variability.

Preferred Test Methods and Notes

Although Fourier Transform Infrared spectroscopy (FTIR) – [EPA Method
320 or ASTM D6348] is more costly and resource intensive than the
alternative Method 323 acetylacetone method, it provides near real time
measurements and flexibility for assessing formaldehyde emissions from
rich burn MACT floor candidate units.  Alternatives to FTIR are
discussed below.  FTIR species include:

CO 

CO2

NO and NO2  

Formaldehyde

C1 – C6 speciated hydrocarbons (e.g., to determine VOCs)

Ammonia

Moisture

FTIR provides near real time instrument response and the ability to
judge test results (e.g., whether the unit is a MACT floor engine).
One-minute average data should be logged and archived for all FTIR
measured compounds.  If Method 323 formaldehyde data are obtained, care
must be taken to ensure that sample flow rates are within the prescribed
range (0.2 to 0.4 liters per minute) and three one hour test runs are
required.  A potential list of FTIR testing companies is provided as
Attachment 2.

O2 and CO2 by instrumentation – [EPA Method 3A] - Determination of
Oxygen and Carbon Dioxide Concentrations in Emissions from Stationary
Sources (Instrumental Analyzer Procedure)] or portable analyzer [ASTM
D6522-00]. 

Moisture by FTIR or impingers [EPA Method 4].

Exhaust gas flowrate by F-factor [EPA Method 19].

Sample Ports and Location [EPA Method 1 or 1A].  Two sample ports should
be located at 90 degrees apart and at least two, and ideally eight,
stack or duct diameters downstream from and at least one half, and
ideally two, diameters upstream of any process change or exhaust to
atmosphere.  One port will be used for a thermocouple measurement and
the second for the heated FTIR sample probe.

For 2SLB and 4SLB engines, the NESHAP emission limits are based on the
surrogate compound CO and a simplified gas measurement approach can be
used.  CO, NOx and O2 can be measured using [ASTM D6522] - Determination
of Nitrogen Oxides, Carbon Monoxide, and Oxygen Concentrations in
Emissions from Natural Gas-Fired Reciprocating Engines, Combustion
Turbines, Boilers, and Process Heaters Using Portable Analyzers.

Quality Assurance / Quality Control

All testing should be conducted in adherence with the reference test
methods.  Method deviations should be documented and explained (e.g.
sample port location).  Because formaldehyde is highly water soluble,
sample line temperature maintenance is required to avoid moisture
condensation. Recommended additional procedures to ensure data quality
include:

Low concentration O2 calibration gases for measuring oxygen in rich burn
engines exhaust.  From EPA Method 7E/3A “to the extent possible, the
measured emissions should be between 20 to 100 percent of the selected
calibration span.” Thus, to measure 0.5% O2, the calibrated range
should be about 0 – 2%.

Similarly, a range of CO2 calibration gases will be required to measure
CO2 from both rich burn and lean burn engines.

QA Spikes for the FTIR systems are recommended for each engine using
formaldehyde and the spiking compound.   Stack gas formaldehyde
concentrations should be measured immediately before and after the QA
spike.

CO2 measurements by both continuous analyzer (EPA Method 3A) and by FTIR
can be used to gauge the uncertainty in the FTIR emission measurements.

The operator must identify the test load and report the basis for
determining the operating load for the test.

Reporting

Emissions should be reported in units of ppmvd, ppmvd @ 15% O2, lb/hr,
and g/hp-hr.  Test reports should be provided to IES to facilitate
review and validation of MACT floor units.  Multiple test reports are
not required for test campaigns on multiple engines.  A single summary
report is preferred. 

Additional Test Considerations

Engines should be tested “as found”, and engine operation should not
be altered or adjusted prior to initial testing other than to adjust the
“baseline O2” for 4SRB engines as discussed below.  Operating
conditions should be logged (see Attachment 1).  Engine exhaust
temperature should be monitored for the duration of all tests.  For 4SRB
engines, emissions are dependent on the amount of excess oxygen (i.e.,
related to the air-to-fuel ratio).  Thus,   

Four-stroke rich burn engine considerations:

If the “as-found” excess O2 is higher than 0.5%, then a short term
test will be completed before adjusting to a “baseline O2 level”. 
In this case, test the 4SRB for approximately 10 minutes at the "as
found" oxygen measured value and record the data.  Stable operation of
the engine over this timeframe can be documented based on 1-minute test
data.  

Baseline O2 level:  If the as found excess O2 is greater than 0.5%,
after the step above is completed adjust the fuel regulator or
carburetor setting to obtain an exhaust O2 concentration of
approximately 0.5% or lower (baseline test condition for 4SRB engines). 
The manual adjustment is necessary because these engines are
unsophisticated and do not include active control of excess air. 
Complete a one hour test at this condition.  This test point is likely
to be the “best performing level” for that particular test condition
(i.e., load, ambient conditions, etc.).  If exhaust oxygen is at 0.5% or
lower "as found", the one hour test will be completed at the as found
condition.  As found units greater than 0.5% O2 are unlikely to be MACT
floor engines – thus the excess air adjust will provide more valuable
data for the analysis and the range of “as found” O2 levels is also
pertinent information.  

Subsequent fuel pressure adjustments or carburetor settings (e.g. for
variability testing) should result in exhaust O2 concentrations less
than 2 percent for consistency with the EPA definition of a rich burn
engine.  The “as found” conditions should be documented as indicated
above.

 

Two- and four-stroke lean burn engine considerations:

CO as surrogate for lean burn engines;

O2 calibrations should be within the expected range of operation (e.g.
approximately 10 to 18 % O2). 

Variability Testing

If a particular engine appears to be a candidate “MACT floor” engine
(i.e., within the best performing 12% based on the number of engines in
the test program), then additional testing will be completed on that
engine.  These tests will vary the excess O2 level (via carburetor or
fuel pressure adjustment) over multiple test points and also vary load,
as feasible.  If the “as found” test was at reduced load, load
adjustment to a higher load will be attempted.  Lower load tests will
also be attempted.  Multiple excess O2 levels will be tested at each
load.  The number of data points to conduct this testing introduces test
timing issues, because there is not time or resources available to spend
multiple days testing a single engines.  Thus, abbreviated run times
will be used for most test conditions – i.e., ten minutes of FTIR data
with stable operation and emissions stability ensured for that period.  

Excess O2 from 4SRB – Measure emissions for MACT floor units to
establish emission range.  Simplest matrix could be lowest excess O2
(nominally 0.5% or lower) and ~2% excess O2.  Additional excess air and
load conditions could be tested as time permits.

Percent Load – Engine load can be changed by adjusting the suction,
restricting the discharge, and/or changing the bypass valves.  Target
test loads are 50%, 75%, and 100% of maximum achievable load.  These
load levels are pertinent because the “maximum achievable” load in
practice for production engines will often be less than the rated load. 


When load is changed to a different test condition, obtain a ten minute
test run for most excess O2 tests, with the 0.5% excess O2 test and
interim excess O2 test runs lasting 10 minutes.  For each load, the
higher excess O2 test run at 2% O2 will acquire one hour of data.  The
higher O2 test condition may bound the upper end of the emissions
variability range, so a one hour test run is proposed for these tests. 
These data should provide insight on formaldehyde variability for
“MACT floor” 4SRB engines. Test data should be continuously recorded
during load transitions.

Identical Make / Model – Document information on identical or
“similar” MACT floor units.  “Location-based” issues that may
affect emissions such as altitude and ambient temperature may need to
consider data from an engine that is the same make and model as MACT
floor engines.  (For example, MACT floor data in Texas; altitude and
temperature data may be available for an identical engine in New Mexico
or Colorado.)

Fuel Variability – Obtain and document fuel analyses for evaluating
effect on emissions.

Ambient parameters temperature, barometric pressure, relative humidity,
and altitude – should be documented during testing for evaluating the
effect on emissions

Other

The dataset will be reviewed to evaluate whether ambient effects can be
discerned.  Depending upon that review, it may be necessary to consider
data from engines that are not MACT floor units, but identical make and
model to MACT floor units, to assess variability from external
environmental factors. Fuel analysis records will be collected and the
data analyzed to assess potential impacts of varied fuel quality on
emissions. 

Schedule

Pending a potential extension request, EPA has requested all supporting
MACT floor data by the end of October.  Industry responses on test
engines completing Table 1 and available test reports for Table 2 should
be accompanied with tentative schedule(s) for completing the test or
providing the data.

Attachment 1

Supplemental Engine Data

In addition to the source test report, the following information will
assist in the determination of whether an engine is a “best performing
12 percent MACT floor engine” [Clean Air Act mandated target from
§112(d)(2)]. All emissions data should be reported as ppmvd @ 15% O2.

Required Supplemental Information

Company

Contact Name __________	email address __________	Phone number __________

Engine and Driven Equipment

Engine Make __________	Engine Model ___________	Installation Date
__________

Engine serial number or unique identifier
__________________________________________

Engine Age ________________________

Nameplate hp _________________	Site Rated hp __________

Circle One 	4SRB		2SLB		4SLB	

Location _________________________________	Elevation __________(feet
above sea level)

Pollution Controls	Y / N		Description (including make and model):
_____________

________________________________________________________________________
______

Ignition system _______________ (e.g. Altronic)

Engine condition upon arrival:
__________________________________________________ 

Date and description of last preventative maintenance:
_______________________________

Maintenance interval and description:
_____________________________________________

Engine life cycle (overhaul interval (yr) remaining life of overhaul
(yr)) _________________

Aftermarket AFR __________	As Found Setting __________

Driven Equipment

Make __________	Model ___________ 

Compressor Cylinder Bore __________ Stroke __________

Generator Rating __________

Test Conditions & Murphy Panel Data

Percent Load __________	including parasitic hp __________ 

Description of how load was determined:
____________________________________________
________________________________________________________________________
________________________________________________________________________
________________

RPM __________	Suction Pressure __________ Suction Temperature
__________

Discharge Pressure __________	Discharge Temperature __________

Manifold Temperature __________	Pressure __________

Water Jacket Temperature __________		Compressor VVCP position:
_____________

Fuel Gas Analysis: include representative fuel gas analysis

Fuel Meter: make __________	model # __________	accuracy __________ ( %

Fuel Line:  pressure _____________ (psig)   temperature ____________ °F

EPA Test Method deviations

Describe all deviations:
________________________________________________________________________
________

________________________________________________________________________
________________________________________________________________________
________________________________________________________________________
________________________

Test Company ___________    email address ______________	Phone number
__________

Contact

Ambient Test Conditions

Temperature __________	Atmospheric Pressure __________ Relative Humidity
_________

Optional Information

Exhaust Temperature __________

Measured Fuel __________

Sample Port Location Description
__________________________________________________

________________________________________________________________________
________

________________________________________________________________________
________

Attachment 2

Selected FTIR Test Companies: Note that this list is for informational
purposes only and does not represent an endorsement of any source test
firm

ACCT a Division of CECO

5440 Alder, Houston, TX 77081-1798

(800) 836-7333

Fax (713) 664-6444

  HYPERLINK "mailto:actt@ceconet.com"  actt@ceconet.com 	Spectral
Insights LLC

214 Argonne Drive

Durham, NC 27704

(919) 620-6549

  HYPERLINK "http://www.spectralinsights.com"  www.spectralinsights.com 

Air Compliance Testing, Inc.

5525 Canal Road

Cleveland, Ohio 44125

216-525-0900

  HYPERLINK "http://www.aircomp.com"  www.aircomp.com 	TRC Environmental
Corporation 

Ross A. Thompson

Lead Project Manager - Operations Manager - Lubbock Air Measurements

7902 Oakridge Avenue Suite One

Lubbock, Texas 79424

806/ 744-8422 Office

806/ 744-8423 Fax

806/ 773-8851 Cell

rthompson@trcsolutions.com

Air Hygiene, Inc.

Paul Little

Director of Customer Service

5634 South 122nd East Avenue, Suite F

Tulsa OK 74146

(918) 289-6378

  HYPERLINK "http://www.airhygiene.com"  www.airhygiene.com 

	Clean Air Engineering, Inc. 

500 W. Wood Street 

Palatine, IL 60067, USA 

800-627-0033

  HYPERLINK "http://www.cleanair.com"  www.cleanair.com 

	Enthalpy Analytical, Inc.

Bryan Tyler 

2202 Ellis Road

Durham, NC 27703

919.850.4392 

  HYPERLINK "mailto:bryan.tyler@enthalpy.com"  bryan.tyler@enthalpy.com 
 



Pace Analytical

1700 Elm Street SE

Minneapolis, MN 55414

(612) 607-6398

  HYPERLINK "http://www.pacelabs.com"  www.pacelabs.com 

	

E C/R Incorporated	Providing Environmental Technical Support Since 1989



	

  PAGE   \* MERGEFORMAT  1 

October 2, 2009

Final Draft Test Plan 

  PAGE  1 

 

