United
States
Office
of
Air
Quality
Environmental
Protection
Planning
and
Standards
Agency
Research
Triangle
Park,
NC
27711
EPA­
454/
R­
OO­
036a
July
2000
AIR
&
ibEPA
Final
Report
Testing
a
2­
Stroke
Lean
Burn
Gas­
Fired
Reciprocating
Internal
Combustion
Engine
to
Determine
the
Effectiveness
of
an
Oxidation
Catalyst
System
for
Reduction
of
Hazardous
Air
Pollutants
Volume
1
of
2
n
­
­­
DISCLAIMER
Pacific
Environmental
Services,
Inc.
(
PES)
prepared
this
document
under
EPA
Contract
No.
68D98004,
Work
Assignment
No.
3­
01.
PES
reviewed
this
document
in
accordance
with
its
internal
quality
assurance
procedures
and
approved
it
for
distribution.
The
contents
of
this
document
do
not
necessarily
reflect
the
views
and
policies
of
the
U.
S.
EPA.
Mention
of
trade
names
does
not
constitute
endorsement
by
the
EPA
or
PES.

i
DISTRIBUTION
LIST
U.
S.
ENVIRONMENTAL
PROTECTION
AGENCY
Terry
Harrison,
Work
Assignment
Manager,
OAQPS,
EMC,
SCGA
Laura
P.
Autry,
Quality
Assurance
Manager,
OAQPS,
EMC
Kathy
Weant,
Contracting
Officer,
OAQPS,
EMAD
Sims
Roy,
Lead
Engineer,
OAQPS,
ESD,
CG
ENGINES
AND
ENERGY
CONVERSION
LABORATORY
COLORADO
STATE
UNIVERSITY
Dr.
Bryan
D.
Wilson,
Director,
EECL
PACIFIC
ENVIRONMENTAL
SERVICES,
INC.

John
T.
Chehaske,
Program
Manager,
Research
Triangle
Park,
NC
Dennis
A.
Falgout,
Project
Manager,
Herndon,
VA
PIPELINE
RESEARCH
COMMITTEE
INTERNATIONAL
Sam
L.
Clowney,
Chairman,
Compressor
Research
Supervisory
Committee
GAS
RESEARCH
INSTITUTE
James
M:
McCarthy,
Program
Team
Leader,
Air
Quality
ii
TABLE
OF'
CONTENTS
VOLUME
1
Page
1.0
INTRODUCTION..
................................................
l­
l
2.0
SUMMARY
OF
RESULTS
..........................................
2­
l
2.1
EMISSIONS
TEST
LOG
.......................................
2­
l
2.2
ENGINE
PARAMETERS
AND
EXHAUST
FLOW
RATES
..........
2­
6
2.3
FTIRS
AND
CEM
MEASUREMENTS
...........................
2­
6
2.4
GCMS
MEASUREMENTS
....................................
2­
7
2.5
POLYNUCLEAR
AROMATIC
HYDROCARBON
(
PAH)
MEASUREMENTS
..........................................
2­
12
2.6
DESTRUCTION
OF
ORGANIC
COMPOUNDS
BY
THE
CATALYST
2­
17
3.0
SOURCE
DESCRIPTION
AND
OPERATION
...........................
3­
1
3.1
ENGINE
DESCRIPTION
.......................................
3­
1
3.2
ENGINE
OPERATION
DURING
TESTING
.......................
3­
4
4.0
SAMPLING
LOCATIONS
...........................................
4­
l
5.0
SAMPLING
AND
ANALYSIS
METHODS.
.............................
5­
l
5.1
5.2
5.3
5.4
5.5
5.6
5.7
5.8
5.9
LOCATION
OF
MEASUREMENT
SITES
AND
SAMPLE/­
VELOCITY
TRAVERSE
POINTS
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
5­
l
DETERMINATION
OF
STACK
GAS
VOLUMETRIC
FLOW
RATE..
.
5­
3
DETERMINATION
OF
STACK
GAS
DRY
OXYGEN
AND
CARBON
DIOXIDE
CONTENT
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
5­
4
DETERMINATION
OF
STACK
GAS
MOISTURE
CONTENT
.
.
.
.
.
.
.
5­
4
DETERMINATION
OF
NITROGEN
OXIDES
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
5­
5
DETERMINATION
OF
CARBON
MONOXIDE
.
.
.
.
.
.
.
;
.
.
.
,
.
.
.
.
.
.
.
5­
5
DETERMINATION
OF
METHANE
AND
NON­
METHANE
HYDROCARBONS
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
5­
7
DETERMINATION
OF
GASEOUS
ORGANIC
HAPS
USINGFTIRS
.
.
.
.
.
.
.
.
.
.
.
.
.
..
t...............................
5­
7
DETERMINATION
OF
ORGANIC
HAPS
BY
DIRECT
INTERFACE
GCMS
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
5­
8
.
.
.
111
TABLE
OF
CONTENTS
(
Concluded)

Page
5.10
DETERMINATION
OF
POLYCYCLIC
AROMATIC
HYDROCARBONS
BYCARB429
.,............................................
5­
11
6.0
QUALITY
ASSURANCE/
QUALITY
CONTROL
PROCEDURES
ANDRESULTS....................................................
6­
l
6.1
FTIRS
QA/
QC
PROCEDURES
.................................
6­
l
6.2
CEMS
QA/
QC
PROCEDURES
.................................
6­
5
6.3
GCMS
QA/
QC
PROCEDURES
................................
6­
14
6.4
CARB
429
QA/
QC
CHECKS
..................................
6­
19
6.5
CORRECTIVE
ACTIONS
....................................
6­
26
6.6
DATA
QUALITY
ASSESSMENT
..............................
6­
30
APPENDIX
A
APPENDIX
B
VOLUME
2
APPENDIX
C
SUBCONTRACTOR
TEST
REPORT
­
COLORADO
STATE
UNIVERSITY
ENGINES
AND
ENERGY
CONVERSION
LABORATORY,
"
EMISSIONS
TESTING
OF
CONTROL
DEVICES
FOR
RECIPROCATING
INTERNAL
CQMBUSTION
ENGINES
IN
SUPPORT
OF
REGULATORY
DEVELOPMENT
BY
THE
U.
S.
ENVIRONMENTAL
PROTECTION
AGENCY
(
EPA)
PHASE
1:
TWO­
STROKE,
LEAN
BURN,
NATURAL
GAS
FIRED
INTERNAL
COMBUSTION
ENGINES"

SUBCONTRACTOR
TEST
REPORT
­
EMISSION
MONITORING,
INC.
"
RESULTS
OF
DIRECT
INTERFACE
GCMS
TESTING
CONDUCTED
ON
A
2­
STROKE
LEAD
BURN
ENGINE"

SUBCONTRACTOR
TEST
REPORT
­
EASTERN
RESEARCH
GROUP,
INC.
"
CARB
METHOD
429:
SAMPLE
ANALYSIS"

APPENDIX
D
CARB
METHOD
429
FIELD
DATA
iv
VOLUME
1
LIST
OF
TABLES
Table
2.1
Table
2.2
Table
2.3
Table
2.4
Table
2.5
Table
2.6
Table
2.7
Table
2.8
EmissionsTestLog
...........................................
2­
2
Summary
of
Exhaust
Gas
Flow
Rates
.............................
2­
4
Emission
Rates
of
Detected
FTIRS
and
CEM
Compounds
.............
2­
8
Emission
Rates
of
Detected
GCMS
Compounds
....................
2­
10
Summary
of
Stack
Gas
and
Sampling
Parameters
CARB
429
Catalyst
Inlet
and
Outlet.
......................................
2­
14
Emission
Rates
of
Detected
PAHS
at
Catalyst
Inlet
..................
2­
15
Emission
Rates
of
Detected
PAHS
at
Catalyst
Outlet
.....
;
..........
2­
16
Removal
Efficiencies
of
Detected
Organic
Compounds
.............
2­
l
8
Table
3.1
Engine
and
Catalyst
Specifications
...............................
3­
2
Table
3.2
Surnrnary
of
Nominal
Engine
Parameters
..........................
3­
3
Table
3.3
Target
Engine
Operating
Conditions
During
Testing
.................
3­
5
Table
3.4
Summary
of
Engine
Parameters
­
Cooper
Bessemer
GMV­
4­
TF
........
3­
7
Table
3.5
Summary
of
Engine
Parameters
During
Baseline
Runs.
...............
3­
9
Table
5.1
Summary
of
Sampling
and
Analysis
Methods
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
5­
2
Table
6.1
Table
6.2
Table
6.3
Table
6.4
Table
.6.5
Table
6.6
Table
6.7
Table
6.8
Table
6.9
Table
6.10
Table
6.11
Table
6.12
Table
6.13
Table
6.14
Table
6.15
Detection
Limits
of
FTIRS
and
CEMS
Compounds
..................
6­
7
Types
and
Frequencies
of
CEMS
Analyzer
Calibrations
..............
6­
l
0
Summary
of
Fuel
Factor
Values
................................
6­
13
Summary
of
CEMS
Analytical
Detection
Limits
...................
6­
14
Summary
of
GCMS
Continuing
Calibrations
And
Audit
Results
.......
6­
16
Detection
Limits
of
GCMS
Compounds
at
Catalyst
Inlet
.............
6­
17
Detection
Limits
of
GCMS
Compounds
at
Catalyst
Outlet
............
6­
l
8
CARB
429
Sample
Train
­
Summary
of
Temperature
Sensor
Calibration
Data
.............................................
6­
20
CARB
429
Sample
Train
­
Summary
of
Dry
Gas
Meter
and
Orifice
Calibration
Data
.............................................
6­
2
1
Summary
of
CARB
429
Blank
Results
...........................
6­
24
Summary
of
CARB
429
Surrogate
Recoveries
.....................
6­
25
Detection
Limits
of
PAH
Compounds
at
Catalyst
Inlet
...............
6­
27
Detection
Limits
of
PAH
Compounds
at
Catalyst
Outlet
.............
6­
28
Summary
of
Corrective
Actions
................................
6­
29
Summary
of
engine
and
Method
Performance
......................
6­
32
V
LIST
OF
FIGURES
VOLUME
1
Pag;
e
Figure
1.1
Test
Program
Organization
and
Major
Lines
of
Communication
.
.
.
.
.
.
.
.
l­
3
Figure
4.1
Figure
4.2
Sample
Port
Locations
for
Velocity,
CARE3
429,
FTIRS,
CEMS,
and
GCMS
Sampling
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
:.
.
.
.
.
.
.
.
4­
3
Sample
Point
Locations
for
Velocity
and
CARE3
429
Sampling
.
.
.
.
.
.
.
.
.
4­
4
Figure
5.1
Figure
5.2
Figure
5.3
Schematic
Diagram
of
EECL
CEMS/
FTIRS
Sampling
and
AnalysisSystem
..............................................
5­
5
Schematic
of
GCMS
Sampling
and
Analysis
System
................
5­
10
Schematic
Diagram
of
CARB
429
PAH
Sampling
Train
.............
5­
12
vi
1.0
INTRODUCTION
The
United
States
Environmental
Protection
Agency'(
EPA)
is
investigating
Reciprocating
Internal
Combustion
Engines
(
RICE)
to
characterize
engine
emissions
and
catalyst
control
efficiencies
of
hazardous
air
pollutants
(
HAPS).
This
document
describes
the
results
of
emissions
testing
conducted
on
a
Cooper­
Bessemer
GMV­
4­
TF
natural­
gas­
fired
2­
stroke,
lean
burn
(
2SLB)
engine.
Early
in
1998,
several
industry
and
EPA
representatives
agreed
that
the
Cooper­
Bessemer
GMV­
4­
TF
engine,
at
the
Colorado
State
University's
Engine
and
Energy
Conversion
Laboratory
(
CSU)
is
adequately
representative
of
existing
and
new
natural­
gas­
fired
2SLB
engines.
The
group
agreed
that
a
matrix
of
test
results
from
testing
conducted
at
the
EECL
could
be
used
to
develop
Maximum
Achievable
Control
Technology
(
MACT)
standards
for
RICE.
The
group
further
agreed
that
an
oxidation
catalyst
installed
on
the
Cooper
GW­
4­
TF
could
be
used
to
determine
the
effectiveness
of
oxidation
catalysts
for
these
engines,
and
that
the
EPA
could
use
the
results
from
testing
at
the
2SLB
matrix
conditions
at
CSU
as
the
basis
for
developing
the
MACT
standard
for
natural­
gas­
fired
2SLB
engines.

Emissions
testing
was
conducted
to
measure
pollutant
concentrations
in
the
exhaust
gas
both
up­
and
downstream
of
an
oxidation
catalyst.
Miratech
Corporation
manufactured
the
catalyst
and
CSU
personnel
installed
it
on
the
engine.
Several
sampling
and
analysis
methodologies
were
used
to
determine
HAP
emissions
before
and
after
the
oxidation
catalyst.
Fourier
transform
infrared
spectroscopy,
or
FTIRS,
was
used
to
measure
formaldehyde,
acetaldehyde,
and
acrolein.
Benzene,
toluene,
ethyl
benzene,
(
o,
m,
p)­
xylenes,
styrene,
hexane,
and
1,3­
butadiene,
were
measured
using
a
direct­
interface
gas
chromatograph
with
a
mass
spectrometer
detector,
or
GCMS.
Continuous
emission
monitors
(
CEMs)
were
used
to
measure
oxygen,
(
0,),
carbon
dioxide
(
CO,),
nitrogen
oxides
(
NO&
carbon
monoxide
(
CO),
total
hydrocarbons
(
THC),
and
methane.
Naphthalene
and
polycyclic
aromatic
hydrocarbons
(
PAHs)
[
acenaphthene,
acenapthylene,
anthracene,
benzo(
a)
amhracene,
benzo(
a)
pyrene,
benzo(
b)
fluoranthene,
benzo(
e)
pyrene,
benzo(
k)
fluoranthene,
benzo(
g,
h,
i)
perylene,
chrysene,
dibenzo(
a,
h)
anthracene,
fluoranthene,
fluorene,
indeno(
1,2,3­
cd)
pyrene,
2­
methylnapthalene,
perylene,
phenanthrene,
and
pyrene]
were
determined
using
California
Air
Resources
Board
(
CARB)
Method
429.

PES
used
three
subcontractors
for
this
effort.
The
CSU
EECL
provided
the
facility
and
the
engine
for
the
test
program,
operated
the
engine
at
predefined
conditions,
and
recorded
engine
operational
data
during
the
testing.
In
addition,
CSU
EECL
personnel
operated
two
FTIRS
sampling
and
analysis
systems
and
two
CEM
systems
that
measured
Final
Report
Cooper
Bessemer
GMY­
4­
TF
l­
l
July
2000
pollutants
and
diluents
in
the
exhaust
gas.
Emissions
Monitoring,
Inc.,
(
EMI)
of
Raleigh,
North
Carolina
provided
emissions
testing
services
and
two
direct­
interface
GCMS
sample
extraction
and
analysis
systems.
Eastern
Research
Group
(
ERG)
of
Morrisville,
North
Carolina,
prepared
filter
media
and
XAD­
2@
sorbent
resin
traps
and
analyzed
the
CARB
Method
429
samples
for
PAHs
using
Low
Resolution
Mass
Spectrometry
(
LRMS).
Under
a
separate
work
assignment,
ERG
personnel
operated
an
EPA­
owned
dynamic
spiking
system
for
the
validation
of
the
FTIRS
systems
for
formaldehyde,
acetaldehyde,
and
acrolein.

The
test
program
organization
and
major
lines
of
communication
employed
during
this
project
are
presented
in
Figure
1.1.
The
balance
of
this
report
contains
the
following
Sections:

Section
2.0
Summary
of
Results
Section
3
.
O
Source
Description
and
Operation
Section
4.0
Sampling
Locations
Section
5.0
Sampling
and
Analysis
Methods
Section
6.0
Quality
A
ssurance/
Quality
Control
Procedures
and
Results
Copies
of
raw
field
data,
quality
assurance
data,
subcontractor
reports,
and
example
calculations
are
included
in
the
appendices
to
this
document.

Final
Report
Cooper
Bessemer
Gh4V­
4­
TF
l­
2
July
2000
PES
Project
Manager
&
MiS
k
Falgout
(
703)
471­
8383
PES
QA/
QC
Officer
Jeff
Van
Atten
(
703)
471­
8383
Subcontractor
1
Emissions
Monitoring
Inc.
I
Subcontractor
Eastern
Research
Group,
Inc.
I
Figure
1.1.
Test
Program
Organization
and
Major
Lines
of
Communication
Draft
Final
Rvo*

PES
I
Subcontractor
Emissions
Monitoring,
Inc.
I
L
Subcontractor
1
CSU
EECL
1
{
El
I
Final
Report
Cooper­
Bessemer
GMV­
CTF
1­
3
July
2000
2.0
SUMMARY
OF
RESULTS
This
section
provides
summaries
of
the
stack
gas
parameters
and
HAP
emissions
during
the
test
program
conducted
on
the
Cooper­
Bessemer
GMV­
4­
TF
engine
March
3
1
through
April
2,
1999.
The
following
sub­
sections
present
the
test
times
and
durations,
engine
and
stack
gas
parameters,
and
HAP
concentrations
and
mass
flow
rates
before
and
after
the
oxidation
catalyst.
A
discussion
of
catalyst
removal
efficiencies
for
various
HAP
is
included
at
the
end
of
this
section.

2.1
EMISSIONS
TEST
LOG
The
test
team
conducted
sampling
at
the
EECL
starting
on
March
30
and
ending
on
April
2,
1999.
During
that
time
period
thirty­
one
test
runs
were
conducted.
These
test
runs
consisted
of
twelve
5­
minute
Quality
Control
(
QC)
runs,
twelve
33­
minute
sampling
runs
for
collection
of
FTIRS,
CEMS
and
GCMS
data,
three
2­
hour
CARB
Method
429
runs,
and
four
5­
minute
daily
baseline
runs.
Table
2.1
presents
the
emissions
test
log.
The
test
log
summarizes
the
date
and
time
that
each
run
was
conducted
and
the
sampling
methodologies
used
during
that
particular
run.
Additional
discussions
of
the
engine
operating
parameters
may
be
found
in
Section
3.0
of
this
document.

In
Table
2.1,
the
sampling
runs
are
presented
in
the
order
that
they
were
conducted.
In
the
tables
that
follow
Table
2.1,
the
sampling
runs
are
presented
in
numerical
order.
During
the
test
program,
engine
conditions
were
set
by
making
small
changes
in
engine
operation
from
run
to
run
rather
than
large
changes.
The
purpose
of
this
approach
was
to
minimize
both
the
time
between
test
runs
to
change
an
engine
condition
as
well
as
the
time
required
for
the
engine
to
stabilize
after
each
change..
The
effect
on
the
test
program
was
that
the
engine
load
conditions
for
which
emissions
data
were
sought
were
not
conducted
in
the
same
order
that
they
were
presented
in
the
Quality
Assurance
Project
Plan
(
QAPP).
To
maintain
consistency
with
the
QAPP,
the
numbers
denoting
the
engine
test
conditions
were
not
changed.

Final
Report
Cooper­
Bessemer
GMVATF
2­
l
July
2000
TABLE
2.1
EMISSIONS
TEST
LOG
Date
Run
Time
Run
ID
Sampling
Methodology
3/
30/
99
1256­
1301
Baseline
No.
1
CEMS,
FTIRS,
GCMS
313
l/
99
1207­
1212
Baseline
No.
2
CEMS,
FTIRS,
GCMS
3/
3
l/
99
1319­
1324
Run
1A
QC
CEMS,
FTIRS,
GCMS
313
l/
99
1340­
1413
Run
1A
CEMS,
FTIRS,
GCMS
313
l/
99
1539­
1544
Run
5
QC
CEMS,
FTIRS,
GCMS
313
I/
99
1600­
1633
Run5
CEMS,
FTIRS,
GCMS
313
l/
99
1741­
1746
Run
6
QC
CEMS,
FTIRS,
GCMS
3r3
l/
99
1805­
1838
Run
6
CEMS,
FTIRS,
GCMS
3/
3
l/
99
1943­
1948
Run
13
QC
CEMS,
FTIRS,
GCMS
313
l/
99
2305­
2338
Run
13
CEMS,
FTIRS,
GCMS
313
l/
99
2105­
2110
Run
14
QC
CEMS,
FTIRS,
GCMS
313
l/
99
2130­
2203
Run
14
CEMS,
FTIRS,
GCMS
313
l/
99
2310­
2315
RunSQC
CEMS,
FTIRS,
GCMS
3/
31199­
4/
l/
99
2335­
0008
Run
8
CEMS,
FTIRS,
GCMS
4/
l/
99
1135­
1140
Run
3
QC
CEMS,
FTIRS,
GCMS
4/
l/
99
1200­
1233
Run
3
CEMS,
FTIRS,
GCMS
4/
l/
99
13
19­
1324
Run
2f7
QC
CEMS,
FTIRS,
GCMS
4/
l/
99
1340­
1413
Run
217
CEMS,
FTIRS,
GCMS
4/
l/
99
1534­
1539
Baseline
No.
3
CEMS,
FTIRS,
GCMS
T
4/
l/
99
1627­
1632
Run
15
QC
CEMS,
FTIRS,
GCMS
411199
1650­
1723
Run
15
CEMS,
FTIRS,
GCMS
4/
l/
99
1817­
1822
Run
16
QC
CEMS,
FTIRS,
GCMS
4/
l/
99
1840­
1913
Run
16
CEMS,
FTIRS,
GCMS
Final
Report
Cooper­
Bessemer
GMY­
4­
TF
2­
2
July
2000
TABLE
2.1
(
Concluded)

EMISSIONS
TEST
LOG
Date
Run
Time
Run
ID
Sampling
Methodology
4/
l/
99
2025­
2030
Run
10
QC
CEMS,
FTIRS,
GCMS
411199
2050­
2123
Run
10
CEMS,
FTIRS,
GCMS
4/
l
J99
2340­
2345
Run
9A
QC
CEMS,
FTIRS,
GCMS
4/
l/
99­
412199
2355­
0028
Run
9A
CEMS,
FTlRS,
GCMS
412199
1204­
1404
PAH
1
(
Run
4)'
CEMS,
FTIRS,
GCMS,
CARB
Method
429
4l2l99
1625­
1825
PAH2
(
Run
SA)'
CEMS,
FTIRS,
GCMS,
CARB
Method
429
412199
4lil99
2000­
2
100
PAH3
(
Run
11)'
CEMS,
FTIRS,
GCMS,
2
100­
2200
PAH3
(
Run
12)'
CARB
Method
429
23
15­
2320
Baseline
No.
4
CEMS,
FTIRS,
GCMS
'
PAH
testing
was
conducted
at
multiple
load
conditions,
instead
of
one
load
condition
as
described
in
the
QAPP.
The
PAH
testing
was
conducted
in
this
fashion
to
make
up
for
field
delays.
A
discussion
of
this
issue
may
be
found
in
Section
5.10.

Final
Report
Cooper­
Bessemer
GMV­
4­
TF
2­
3
July
2000
TABLE
2.2
SUMMARY
OF
EXHAUST
GAS
FLOW
BATES
3
Run
ID
RunlA
1
Run%
7
Run3
Run4
Run5
Run5
Run8
RunSA
Run10
Ermine
SDeed
.
mm
300
I
299
269
270
300
300
270
299
299
Gas
Temoerature.
"
F
I
554
1
480
1
447
1
517
1
534
1
567
i
498
1
527
1
556
Oxygen,
%
vol
d.
b.
14.67
15.80
16.30
14.80
15.20
14.17
15.40
14.60
14.63
Dioxide.
%
vol
d.
b.
3.43
2.83
2.50
3.33
3.39
3.71
2.96
3.56
3.53
Gas
Volumetric
Flow
Rate,
dscfm
1
1907
1
1830
1
1782
1
1674
1
2077
1
1753
1
1840
1
1895
1
1929
rpm
­
revolutions
per
minute
MMBtu/
hr
­
million
British
Thermal
units
per
hour
ft­
lb
­
fast­
pounds
dscf/
MMEtu
­
dry
standard
cubic
feet
of
exhaust
products
par
million
Btu
of
heat
input
Q
0%
excess
air
bhp
­
brake
horsepower
`
F­
degrees
Fahrenheit
scth
­
standard
cubic
feat
per
hour
Q
68
°
F
and
29.92
in.
Hg
%
vol
d.
b.
­
%
volume
dry
basis
4
­
reciprocal
of
%
Excess
Air
dscfm
­
dry
standard
cubic
feet
per
minute
Q
68
"
F
and
29.92
in.
Hg
Btu/
cf
­
British
Themlal
Units
per
cubic
foot
of
natural
gas
Final
Report
Cooper­
Bessemer
GMVd­
TF
2­
4
July
2000
TABLE
2.2
(
CONCLUDED)

SUMMARY
OF
EXHAUST
FLOW
RATES
1
Run11
1
Run12
1
Run13
Run14
1
Run15
Engine
Speed
,
rpm
270
1
270
1
300
1
300
1
299
Engine
Torque,
ft­
lb
7356
7349
7727
7728
7729
Horsepower,
bhp
378
378
441
441
442
Fuel
Flow
Rate,
scfh
3277
3271
3727
3585
3715
Equivalence
Ratio,
I$
0.30
0.29
0.33
0.32
0.32
Higher
Heating
Value,
Btukf
1032
1032
1072
1072
1090
Heat
Rate,
MMBtulhr
3.38
3.38
3.99
3.84
4.05
Drv
Fuel
Factor.
F,+,
dscf/
MMBtu
8661
8661
8664
8664
8672
II
.
­.
I
I
I
I
I
Catalyst
Inlet
Run16
PAHI
PAH2
PAH3
299
270
270
270
7731
7326
7341
7353
442
377
377
378
3713
3277
3300
3274
0.33
0.33
0.29
0.29
I
I
I
1090
1
1032
1
1032
1
1032
II
4.05
1
3.38
1
3.41
1
3.38
11
8672
1
8661
1
8661
8661
11
ipm
­
revolutions
per
minute
ft­
lb
­
foot­
pounds
bhp
­
brake
horsepower
scfh
­
standard
cubic
feet
per
hour
Q
WF
and
29.92
in.
Hg
$
­
reciprocal
of
%
Excess
Air
Btulcf
­
British
Thermal
Units
per
cubic
foot
of
natural
gas
MMBtWhr
­
million
British
Thermal
units
per
hour
dscf/
MMBtu
­
dry
standard
cubic
feet
of
exhaust
products
per
million
Btu
of
heat
input
@
0%
excess
air
`
F
­
degrees
Fahrenheit
%
vol
d.
b.
­
%
volume
dry
basis
dscfm
­
dry
standard
cubic
feet
per
minute
Q
68
`
F
and
29.92
in.
Hg
Final
Report
Cooper­
Bessemer
GMV­
CTF
2­
5
July
2000
2.2
ENGINE
PARAMETERS
AND
EXHAUST
FLOW
RATES
Table
2.2
summarizes
some
of
the
engine
and
exhaust
gas
parameters
that
were
measured
and/
or
calculated
during
the
test
program.
The
EECL's
Data
Acquisition
System
@
AS),
monitored
and
recorded
approximately
200
engine
operating
parameters,
as
well
as
gas
temperatures,
and
concentrations
of
0,,
COa,
and
moisture
at
the
catalyst
inlet
and
exhaust.
(
The
test
report
generated
by
CSU
EECL
is
presented
in
Appendix
A).

The
exhaust
gas
volumetric
flow
rates
at
before
and
after
the
catalyst
are
presented
for
each
sample
run.
These
flow
rates
are
calculated
using
a
combustion
products,
or
Fd,
factor
and
correcting
for
excess
air
as
indicated
by
the
measurements
of
O2
concentration
at
each
location.
A
new
fuel
factor
was
calculated
daily
based
upon
daily
analysis
of
the
composition
of
the
natural
gas
fuel.

2.3
FTIRS
AND
CEM
MEASUREMENTS
Table
2.3
summarizes
the
mass
flow
rates
of
the
FTIRS
target
compounds
(
formaldehyde,
acetaldehyde,
and
acrolein)
and
the
CEM
target
compounds
(
carbon
monoxide,
nitrogen
oxides,
total
hydrocarbons,
or
THC,
methane,
and
non­
methane
hydrocarbons,
or
NMHC).

EECL
personnel
operated
two
FTIRS
sampling
and
analysis
systems
to
quantify
concentrations
of
the
FTIRS
target
compounds.
Exhaust
gas
samples
were
extracted
from
locations
before
and
after
the
oxidation
catalyst,
conditioned,
and
transported
to
a
Nicolet
Magna
560
FTIRS
tire­
catalyst
location)
and
a
Nicolet
Rega
7000
FTIRS
(
post­
catalyst
location).
The
outlet
FTIRS
was
also
used
to
measure
the
moisture
content
in
the
exhaust
gas.
Moisture
measurements
by
the
inlet
FTIRS
were
determined
by
EECL
to
be
inaccurate.
Therefore
a
carbon
balance
method
was
employed
to
calculate
the
moisture
concentration
at
the
pre­
catalyst
sampling
location.

Of
the
three
target
FTIRS
compounds,
only
formaldehyde
was
detected.
Formaldehyde
was
detected
before
and
after
the
catalyst
during
every
sampling
run
that
was
conducted.
Results
are
presented
for
each
of
the
18
test
runs
in
numerical
order.
Neither
acetaldehyde
nor
acrolein
were
detected
during
the
sampling
program.
The
final
column
of
Table
2.3
presents
the
average
mass
flow
rate
of
the
formaldehyde,
or
the
average
detection
limit
of
acetaldehyde
and
acrolein.
Run
by
run
detection
limits
for
the
FTIRS
compounds
are
presented
in
Table
5.2
of
this
document.

Table
2.3
also
presents
the
calculated
mass
flow
rates
of
the
CEMS
compounds.
EECL
personnel
operating
two
CEMS
sampling
and
analysis
systems.
Engine
exhaust
gas
samples
were
extracted
from
locations
before
and
after
the
catalyst,
conditioned,
and
transported
to
the
CEMS
analyzer
racks.
Moisture
was
removed
from
the
gas
sample
prior
to
Final
Report
Cooper­
Bessemer
GMV­
4­
TF
2­
6
July
2000
introduction
to
the
02,
CO,,
CO,
and
NO,
analyzers.
All
of
the
CEMS
target
compounds
were
detected
at
both
the
inlet
and
the
outlet
locations.
CEMS
detection
limits
are
presented
on
a
run
by
run
basis
in
Table
5.2.

2.4
GCMS
MEASUREMENTS
Table
2.4
presents
the
calculated
mass
flow
rates
of
the
GCMS
compounds
(
1,3­
butadiene,
hexane
benzene,
toluene,
ethyl
benzene,
(
o,
m,
p)­
xylenes,
and
styrene).
EM1
personnel
operated
two
Inficon
Portable
Gas
Chromatographs
with
Mass
Spectrometer
Detectors.
(
The
test
report
generated
by
EMI
is
presented
in
Appendix
B).
Gas
samples
for
GCMS
analysis
were
extracted
from
both
the
before
and
after
catalyst
locations
through
a
heated
probe
and
quartz
fiber
filter,
then
transported
via
a
heated
Teflon@
sample
line
to
a
Peltier
condenser
for
continuous
moisture
removal.
The
sample
was
then
co­
mixed
with
an
internal
standard
mixture
(
in
a
constant
ratio
of
10:
1)
in
the
GC
sampling
loop
for
1
minute
before
injection
into
the
GCMS.
Afier
purging
the
sample
loop
for
1
minute,
the
sample
was
injected
onto
the
separatory
column
to
resolve
the
target
compounds
for
quantification
by
the
detector.
Each
sample
run
consisted
on
4
injections.
Each
GCMS
was
supported
by
a
PC­
based
DAS
to
calculate
peak
areas
of
the
target
compounds.

The
only
target
analytes
that
the
GCMS
detected
at
the
catalyst
inlet
location
were
hexane,
benzene,
and
toluene.
Concentration
levels
of
hexane
reached
about
0.1
parts
per
million
(
ppm,
or
100
parts
per
billion
ppb)
which
is
approximately
the
instrument
detection
limit
for
hexane.
Concentration
levels
for
benzene
and
toluene
ranged
from
0.05
to
0.1
ppm
(
50
to
100
ppb),
and
0.01
to
0.23
ppm
(
10
to
230
ppb),
respectively,
for
the
16
engine
test
conditions.
Run
number
1A
had
the
lowest
concentration
levels
for
benzene
and
toluene
with
only
0.05
and
0.02
ppm
(
50
and
20
ppb)
detected,
respectively.
All
other
engine
test
conditions
produced
higher
concentration
results
for
these
compounds,
but
changes
in
engine
operation
had
little
effect
on
the
observed
results.
Benzene
and
toluene
concentration
levels
for
runs
2/
7,3,5,6,8,9A,
10,
13,
14,
15,
and
16
all
approximated
0.07
ppm
(
70
ppb)
for
benzene
and
0.22
ppm
(
220
ppb)
for
toluene.

A
gas
chromatograph
coupled
with
a
mass
spectrometer
(
GCMS)
detector
can
identify
compounds
that
are
not
contained
in
the
instrument
specific
calibration.
The
GCMS
identified
two
peaks
that
were
not
among
the
original
matrix
of
target
analytes.
The
compounds,
di­
methyl
ether
(
CAS#=
115­
10­
6,
MW=
46
AMU)
and
nitromethane
(
CAS+
75­
52­
5,
MW=
61
AMU),
were
tentatively
identified
in
nearly
every
run
at
the
inlet
location.
Neither
of
these
compounds
are
consider
HAPS
by
EPA.
We
could
not
quantify
the
compounds
because
we
had
no
calibration
analytes
that
are
chemically
similar,
and
therefore,
could
not
estimate
instrument
specific
response
factors
to
generate
estimated
concentrations.

Final
Report
Cooper­
Bessemer
GMV­
4­
TF
2­
7
July
2000
TABLE
2.3
EMISSION
BATES
OF
DETECTED
FTIB
AND
CEM
COMPOUkDS
(
Intalvst
ldet
Formaldehyde
mglbhp­
hr
mlblhr
mg/
bhp­
hr
Acetaidehyde
mlb/
hr
Acrolein
mg/
bhp­
hr
tilb/
hr
Nitrogen
Oxides
(
as
NO*)
g/
bhp­
hr
Ibihr
Carbon
Monoxide
g/
bhp­
hr
lb/
hr
g/
bhp­
hr
Methane
lblhr
Non­
methane
Hydrocarbons
glbhp­
hr
Ib/
hr
Total
Hydrocarbons
glbhp­
hr
lblhr
Formaldehyde
mg/
bhp­
hr
mlblhr
mglbhp­
hr
Acetaldehyde
mlblhr
mg/
bhp­
hr
Acrolein
mlb/
hr
Nitrogen
Oxides
(
as
NO3
glbhp­
hr
Ibfhr
Carbon
Monoxide
g/
bhp­
hr
lblhr
g/
bhp­
hr
Methane
lblhr
Non­
methane
Hydrocarbons
glbhp­
hr
lblhr
Total
Hydrocarbons
glbhp­
hr
lblhr
_­_­_
,­­
­­­­­­
i
181
287
280
182
180
188
198
172
187
158
191
158
135
175
184
185
188
182
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
1.5
0.14
0.14
4.6
0.80
3.0
0.49
1.8
2.2
1.5
0.092
0.085
3.8
0.58
2.9
0.41
1.8
2.1
0.75
2.7
2.5
0.88
1.0
0.87
1.2
0.70
0.72
0.73
1.8
1.5
0.57
1.0
0.86
1.0
0.89
0.70
3.7
10.8
10.3
5.9
4.5
3.5
5.9
4.1
4.1
3.8
7.2
6.1
4.9
4.4
3.4
4.9
4.0
4.0
0.81
1.8
2.2
0.71
0.93
0.71
1.2
0.44
0.42
0.79
1.1
1.3
0.59
0.91
0.89
0.98
0.43
0.41
4.8
12
13
7.2
5.4
4.3
7.5
4.8
4.8
4.5
8.2
7.9
8.0
5.2
4.2
8.2
4.5
4.8
Catalyst
Outlet
87
177
188
80
101
72
97
84
85
85
117
113
87
98
70
81
82
83
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
1.8
0.14
0.14
4.7
0.7
3.0
0.53
2.0
2.3
1.5
0.092
0.087
3.9
0.7
2.9
0.44
1.9
2.2
0.24
0.83
0.92
0.28
0.38
0.24
0.39
0.25
0.25
0.24
0.55
0.55
0.22
0.35
0.23
0.33
0.25
0.25
3.8
11
11
8.0
4.6
3.4
5.7
4.2
4.1
3.7
7.3
8.5
5.0
4.5
3.3
4.8
4.1
4.0
0.86
1.8
2.2
0.75
1.4
0.77
1.4
0.41
0.42
0.64
1.1
1.3
0.82
1.4
0.75
1.2
0.40
0.41
4.8
13
14
7.4
5.7
4.3
7.1
4.9
5.0
4.8
8.3
8.5
8.2
5.5
4.2
6.0
4.7
4.8
mgMphr
­
milligrams
per
brake
horsepower
hour
mlb/
bhphr
­
millipwnds
per
brake
horsepower
hwr
@
bhp­
hr
­
grams
per
brake
huaepower
hour
lbm
­
pounds
per
hour
ND
­
Not
Detected.
Refer
to
Table
6.1
for
run­
by­
iun
summary
of
detection
limits.

Final
Report
Cooper­
Bessemer
GMV­
4­
TF
2­
8
July
2000
TABLE
2.3
(
CONCLUDED)

EMISSION
RATES
OF
DETECTED
FTIR
AND
CEM
COMPOUNDS
Run
ID
~
Run11~
Runl2(
Runl3~
Run14~
Runl5)
Runl6~
PAHI
1
PAH2
1
PAH3
Catalyst
Inlet
­"
IL.%...
l.
r
1
188
192
193
162
198
194
157
194
190
131
161
158
mg/
biphr
­
milligrams
per
brake
horsepower
hour
mlblbhp­
hr
­
millipounds
per
brake
horsepower
hour
g!
bhp­
hr
­
grams
per
brake
horsepower
hour
lbhr
­
pounds
per
hour
ND
­
Not
Detected.
Refer
to
Table
6.1
for
rWby­
Iun
summaiy
of
detection
limits.

Final
Report
Cooper­
Bessemer
"
GMYV­
4­
TF
2­
9
July
2000
TABLE
2.4
EMISSION
RATES
OF
DETECTED
GCMS
COMPOUNDS
Run
ID
RunlA
RunZ­
7
Run3
Run4
Run5
Run6
Run8
RunSA
Run10
Catalvst
Inlet
I­
I
,3­
.
Butadiene
pg/
bhp­
hr
ND
ND
ND
ND
ND
ND
ND
ND
ND
plb/
hr
ND
ND
ND
ND
ND
ND
ND
ND
ND
pglbhp­
hr
ND
ND
ND
3000
ND
ND
ND
ND
ND
Hexane
ulb/
hr
ND
ND
ND
3000
ND
ND
ND
ND
ND
pglbhp­
hr
1000
2000
2000
2000
2000
2000
2000
2000
2000
Benzene
plblhr
1000
2000
1000
2000
2000
2000
1000
2000
2000
pglbhp­
hr
ND
8800
8800
2300
6600
6100
7200
8000
6200
Toluene
ylb/
hr
ND
5800
5300
1900
6400
5900
6000
5900
6100
pglbhp­
hr
ND
ND
ND
ND
ND
ND
ND
ND
ND
yl
ycI
dne
plblhr
ND
ND
ND
ND
ND
ND
ND
ND
ND
)­
Xylene
Styrene
o­
Xylene
pglbhp­
hr
ND
ND
plb/
hr
ND
ND
pg/
bhp­
hr
ND
ND
plblhr
ND
ND
pglbhp­
hr
ND
ND
plblhr
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND'

ND
ND
ND
ND
ND
ND
ND
Catalyst
Outlet
r
1,3.
pg/
bhp­
hr
ND
ND
ND
ND
ND
ND
ND
ND
ND
.
Butadiene
plb/
hr
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
Hexane
pglbhp­
hr
ND
plblhr
ND
ND
ND
ND
ND
ND'
ND
ND
ND
Benzene
ug/
bhp­
hr
ND
ND
ND
plblhr
ND
ND
ND
pglbhp­
hr
ND
ND
ND
Toluene
plblhr
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
Ethyl
Benzene
pglbhp­
hr
ND
ND
plbihr
ND
ND
ND
ND
ND
ND
ND
ND
ND
)
Ig/
bhp­
hr
ND
ND
ND
ND
ND
ND
ND
ND
ND
m/
p­
Xylene
plblhr
ND
ND
ND
ND
ND
ND
ND
ND
ND
pglbhp­
hr
ND
ND
ND
ND
ND
ND
ND
ND
ND
Styrene
plbkw
ND
ND
ND
ND
ND
ND
ND
ND
ND
pglbhp­
hr
ND
ND
ND
ND
ND
ND
ND
ND
ND
o­
Xylene
plblhr
ND
ND
ND
ND
ND
ND
ND
ND
ND
Cgmhphr
­
micrcgmms
&
brake
horsepower
hour
plbhr
­
micropounds
per
hour
ND
­
Refer
to
TaMe
6.6
far
nmby­
run
detection
Iknits
at
the
catalyst
inlet.
and
Table
6.7
for
run­
by­
run
detwtion
limits
at
the
catalyst
outlet
Final
Report
Cooper­
Bessemer
GMV­
4­
TF
2­
10
July
2000
TABLE
2.4
(
CONCLUDED)

EMISSION
BATES
OF
DETECTED
GCMS
COMPOUNDS
Run
ID
Run11
Run12
Run13
Run14
Run15
Caialyxt
Inlet
pglbhp­
hr
ND
ND
ND
ND
ND
1,3­
Butadiene
plb/
hr
ND
ND
ND
ND
ND
pglbhp­
hr
ND
ND
ND
ND
ND
Hexane
plb/
hr
ND
ND
ND
ND
ND
pglbhp­
hr
3000
2000
2000
1000
2000
Benzene
plblhr
2000
2000
2000
1000
2000
II
Toluene
Dglbhp­
hr
ulblhr
Ethyl
Benzene
pgLbhp­
hr
ND
ND
ND
ND
ND
plblhr
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
m/
p­
Xylene
pglbhp­
hr
l.
db/
hr
ND
ND
ND
ND
ND
H
Styrene
pglbhp­
hr
ulblhr
Run16
­

ND
ND
ND
ND
3000
3000
2000
2000
6300
6100
ND
ND
ND
ND
I
ND
~
ND
PAHI
PAHZ
1
1
PAH3
1400
t
ND
ND
ND
f
ND
ND
ND
o­
Xylene
pglbhp­
hr
ND
plblhr
ND
ND
ND
ND
ND
ND
ND
Catalyst
Outlet
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
pg/
bhp­
hr
ND
ND
ND
ND
ND
ND
ND
ND
ND
1,3­
Butadiene
plblhr
ND
ND
ND
ND
ND
ND
ND
ND
ND
pglbhp­
hr
ND
ND
ND
ND
ND'
ND
ND
ND
ND
Hexane
l.
db/
hr
ND
ND
ND
ND
ND
ND
ND
ND
ND
pglbhp­
hr
500
500
ND
ND
ND
500
ND
ND
500
Benzene
plblhr
400
400
ND
ND
ND
,500
ND
ND
400
pglbhp­
hr
ND
ND
ND
ND
ND
ND
ND
ND
ND
Toluene
plblhr
ND
ND
ND
ND
ND
ND
ND
ND
ND
pglbhp­
hr
ND
ND
ND
ND
ND
ND
ND
ND
ND
Ethyl
Benzene
plbkr
ND
ND
ND
ND
ND
ND
ND
ND
ND
pglbhphr
ND
ND
ND
ND
ND
ND
ND
ND
ND
rdpxylene
plb/
hr
ND
ND
ND
ND
ND
ND
ND
ND
ND
yglbhp­
hr
ND
ND
ND
ND
ND
ND
ND
ND
ND
Styrene
vlb/
hr
ND
ND
ND
ND
ND
ND
ND
ND
ND
o­
Xylene
pgibhp­
hr
plb/
hr
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
NO
­
Refer
to
TaMa
6.6
for
fw­
by­
wn
det&
on
limits
at
the
catalyst
inlet,
and
Table
6.7
for
rurrby­
run
d+
tecWn
limits
at
Me
catalyst
autle:

Final
Report
Cooper­
Bessemer
GMVATF
2­
11
July
2000
The
only
target
analyte
detected
at
the
catalyst
outlet
was
benzene.
The
di­
methyl
ether
and
nitromethane
peaks
were'
either
absent
or
at
very
low
concentrations
at
the
outlet
location.
The
highest
concentration
level
observed
for
benzene
at
the
outlet
was
0.03
ppm
(
30
ppb)
which
occurred
during
Run
4.

2.5
POLYNUCLEAR
AROMATIC
HYDROCARBON
(
PAH)
MEASUREMENTS
PES
used
CARB
Method
429
to
collect
samples
of
the
engine
exhaust
for
determination
of
PAHs.
A
sample
of
the
exhaust
gas
stream
was
extracted
through
a
glass
nozzle,
heated
glass­
lined
probe,
a
heated
quartz
filter,
and
a
chilled
sorbent
trap
containing
XAD­
2
sorbent
resin.
The
resin
was
extracted
and
combined
with
the
front­
half
train
rinses
and
the
filter
and
analyzed
for
PAH
content
by
ERG
using
Low
Resolution
Mass
Spectrometry.
(
The
analytical
reported
generated
by
ERG
is
contained
in
Appendix
C).

Table
2.5
presents
stack
gas
and
sample
train
parameters
for
the
CARB
429
testing.
Three
2­
hour
CARB
429
sample
runs
were
conducted
before
and
after
the
catalyst
by
PES
personnel.
The
first
PAH
run
was
conducted
at
Run
Condition
No.
4,
and
the
second
PAH
run
was
conducted
at
Run
Condition
8.
The
last
PAH
run
was
conducted
at
Run
Conditions
11
(
for
the
first
hour)
and
12
(
for
the
second
hour).
CARB
429
calls
for
testing
to
be
conducted
at
isokinetic
conditions.
The
isokinetic
sampling
ratios
should
be
100
%
f
10%.
The
(
3­
n­
m)
average
isokinetic
sampling
ratio
was
84.5
%
before
the
catalyst
and
84.8
%
after
the
catalyst.
PES
used
a
standard
pitot
tube
for
velocity
traverses
and
used
the
pitot
tube
coefficient
for
an
S­
type
pitot
tube
in
the
pre­
sampling
calculations.
This
error
resulted
in
sampling
at
a
velocity
approximately
15
%
less
than
the
exhaust
gas
velocity.

Isokinetic
sampling
is
used
to
ensure
that
the
distribution
of
large
versus
small
particles
in
the
collected
sample
is
representative
of
the
distribution
of
these
particles
in
the
exhaust
gas.
If
the
exhaust
is
composed
of
both
large
and
small
particles,
sampling
at
less
that
isokinetic
­
conditions
will
bias
the
particle
distribution
in
the
sample
towards
the
larger
particles.
The
larger
particles
will
be
collected,
but
not
the
smaller
particles.
The
effect
of
'
sampling
at
less
than
isokinetic
conditions
is
minimized
because
the
larger
particles
compose
most
of
the
mass
of
the
sample.
If
the
particles
in
the
exhaust
gas
are
composed
of
particles
that
are
the
same
size,
as
is
most
likely
the
case
for
an
engine
exhaust,
then
the
effect
of
anisokinetic
sampling
has
a
minimal
effect
on
the
particle
size
distribution
in
the
sample.

Table
2.6
presents
the
mass
emission
rates
of
detected
PAH
target
compounds
at
the
catalyst
inlet.
Napthalene
and
phenanthrene
were
the
only
PAHs
detected
during
every
run
before
the
catalyst.
Acenapthene
and
flurorene
were
detected
during
the
first
run,
and
acenapthylene
was
detected
during
the
second
run.
No
other
PAH
compounds
were
detected.
For
these
compounds,
(
3­
run)
average
detection
limit
is
presented
in
the
average
column.
Table
6.12
presents
the
in­
stack
detection
limits
at
the
catalyst
inlet
for
each
compound
on
a
Final
Report
Cooper­
Bessemer
GMV4­
TF
2­
12
July
2000
run­
by­
run
basis.
Table
2.7
presents
the
mass
emission
rates
of
detected
PAH
target
compounds
at
the
catalyst
outlet.
Acenapthene,
napthalene
and
phenanthrene
were
detected
during
all
three
sampling
runs.
Emission
rates
of
all
other
compounds
were
less
than
the
method
detection
limit.
For
these
compounds,
(
3­
run)
average
detection
limit
is
presented
in
the
average
column.
Table
6.13
presents
the
in­
stack
detection
limits
at
the
catalyst
outlet
for
each
compound
on
a
run­
by­
run
basis.

Final
Report
Cooper­
Bessemer
GMV­
bTF
2­
13
July
2000
TABLE
2.5
SUMMARY
OF
STACK
GAS
AND
SAMPLING
PARAMETERS
CARB
429
CATALYST
INLET
AND
OUTLET
Run
ID
Date
Time
PAH
1
PAH
2
PAH
3
4m99
412199
412199
Average
1204­
I
404
1625­
l
825
2000­
2200
Catalyst
Inlet
Sampling
Duration,
minutes
Average
Sampling
Rate,
dscfm
a
Sample
Volume,
dscf
Gas
Temperature,
"
F
Gas
Pressure,
in.
Hg
O2
Concentration,
%
by
Volume
CO2
Concentration,
%
by
Volumt
Moisture,
%
by
Volume
Gas
Volumetric
Flow
Rate:

acfm
b
dscfm
a
Gas
Velocity,
ft/
s
lsokinetic
Sampling
Ratio,
%

`

Sampling
Duration,
minutes
Average
Sampling
Rate,
dscfm
a
Sample
Volume,
dscf
Gas
Temperature,
"
F
Gas
Pressure,
in.
Hg
O2
Concentration,
%
by
Volume
COz.
Concentration,
%
by
Volum'

Moisture,
%
by
Volume
Gas
Volumetric
Flow
Rate:
e
4473
1738
95.0
82.8
Catalyst
120
0.661
79.366
586
25.35
14.7
3.4
8.1
120
120
120
0.708
0.677
0.682
84.998
81.198
81.854
582
582
583
25.35
25.35
25.35
15.5
15.4
15.2
2.9
3.0
3.1
7.7
7.4
7.7
acfm
'
4,420
4,420
4,460
4,433
dscfm
*
1,740
1,750
1,770
1,753
Gas
Velocity,
ft/
s
93.8
93.9
94.6
94.1
lsokinetic
Sampling
Ratio,
%
83.0
88.1
83.2
84.8
­
L
­
?

­
120
0.615
73.787
600
25.50
14.6
3.4
8.4
0
Wlet
120
0.648
77.723
616
25.50
15.4
3.0
7.6
4446
1717
94.4
88.3
120
0.612
73.475
620
25.50
15.3
3.1
7.4
4508
1740
95.7
82.4
a
Dry
standard
cubic
feet
per
minute
corrected
to
68"
F
(
20"
C)
and
1
atm.

b
Actual
cubic
feet
per
minute
at
exhaust
gas
conditions.
0.625
74.995
612
25.50
15.1
3.2
7.8
4476
1732
95.0
84.5
Final
Report
Cooper­
Bessemer
GMV­
4­
TF
2­
14
July
2000
TABLE
2.6
EMISSION
RATES
OF
DETECTED
PAFIS
AT
CATALYST
INLET
Run
ID
PAH
1
PAH
2
PAH
3
Date
412199
4/
2/
99
4lU99
Average
Time
12044404
16251825
2000­
2200
Acenaphthene
uglbhp­
hr
a
2.8
ND
ND
<
2.2
@
b/
hour
b
2.4
ND
ND
<
1.8
Acenaphthylene
pg/
bhp­
hr
ND
1.9
ND
<
1.9
@
b/
hour
ND
1.6
ND
<
1.6
Anthracene
uglbhp­
hr
ND
ND
ND
<
1.8
@
b/
hour
ND
ND
ND
<
1.5
Benzo(
a)
anthracene
uglbhp­
hr
ND
ND
ND
<
1.8
ulb/
hour
ND
ND
ND
<
1.5
Benzo(
b)
fluoranthene
uglbhp­
hr
ND
ND
ND
<
1.8
@
b/
hour
ND
ND
ND
<
1.5
Benzo(
k)
fluoranthene
uglbhp­
hr
ND
ND
ND
<
1.8
@
b/
hour
ND
ND
ND
C
1.5
Benzo(
g,
h,
i)
perylene
uglbhphr
ND
ND
ND
<
3.7
ylb/
hour
ND
ND
ND
<
3.1
Benzo(
a)
pyrene
pglbhp­
hr
ND
ND
ND
<
1.8
ulblhour
ND
ND
ND
C
1.5
Chrysene
uglbhp­
hr
ND
ND
ND
<
1.8
ulb/
hour
ND
ND
ND
<
1.5
Dibenz(
a,
h)
anthracene
ug/
bhp­
hr
ND
ND
ND
<
3.7
@
b/
hour
ND
ND
ND
<
3.1
Fluoranthene
pglbhp­
hr
ND
ND
ND
<
1.8
@
b/
hour
ND
ND
ND
C
1.5
pglbhp­
hr
4.7
ND
ND
<
Fluorene
2.8
@
b/
hour
3.9
ND
ND
C
2.3
Indeno(
l,
2,3cd)
pyrene
ug/
bhp­
hr
ND
ND
ND
<
3.7
ulblhour
ND
ND
ND
<
3.1
Naphthalene
pglbhp­
hr
64
38
42
48
ulb/
hour
53
32
35
PO
Phenanthrene
uglbhp­
hr
6.0
4.7
3.3
4.6
ulbkour
5.0
3.9
2.7
3.9
Pyrene
uglbhp­
hr
ND
ND
ND
C
1.8
ulblhour
ND
ND
ND
<
1.5
a
Micrograms
per
brake
horsepower
hour
b
Micropounds
per
hour
ND
indicates
that
the
compound
was
not
detected.
Averages
include
detection
limits.

Table
6.10
presents
run­
by­
run
detection
limits
for
all
PAHs.

Final
Report
Cooper­
Bessemer
GMV­
4­
TF
2­
15
July
2000
TABLE
2.7
EMISSION
RATES
OF
DETECTED
PAHS
AT
CATALYST
OUTLET
Run
ID
PAH
1
PAH
2
PAH
3
Date
4m99
4m99
4m99
Average
Time
1204­
1404
1625­
1825
2000­
2200
Acenaphthene
.
yg/
bhp­
hr
*
ND
ND
ND
<
1.7
plb/
hour
b
ND
ND
ND
<
1.4
Acenaphthylene
i.
Wfw­
hr
2.1
1.6
1.8
1.9
ulblhour
1
.
a
1.4
1.5
1.5
Anthracene
ug/
bhp­
hr
ND
ND
ND
<
1.7
ulb/
hour
ND
ND
ND
<
1.4
Benzo(
a)
anthracene
uglbhp­
hr
ND
ND
ND.
c
1.7
@
b/
hour
ND
ND
ND
<
1.4
Benzo(
b)
fluoranthene
ugibhp­
hr
ND
ND
ND
<
1.7
@
b/
hour
ND
ND
ND
<
1.4
Benzo(
k)
fluoranthene
WbW­
hr
ND
ND
ND
<
1.7
@
b/
hour
ND
ND
ND
<
1.4
Benzo(
g,
h,
i)
perylene
uglbhp­
hr
ND
ND
ND
c
3.4
ulblhour
ND
ND
ND
c
2.8
Benzo(
a)
pyrene
pglbhp­
hr
ND
ND
ND
<
1.7
@
b/
hour
ND
ND
ND
<
1.4
Chrysene
yglbhp­
hr
ND
ND
ND
<
1.7
Fib/
hour
`
ND
ND
ND
<
1.4
Dibenz(
a,
h)
anthracene
uglbhp­
hr
ND
ND
ND
<
3.4
@
b/
hour
ND
ND
ND
<
2.8
Fluoranthene
uglbhp­
hr
ND
ND
ND
<:
1.7
@
b/
hour
ND
ND
ND
c
1.4
Fluorene
pg/
bhp­
hr
ND
ND
ND
<
1.7
ylb/
hour
ND
ND
ND
<
1.4
Indeno(
l,
2,3­
cd)
pyrene
pglbhp­
hr
ND
ND
ND
c
3.4
plb/
hour
ND
ND
ND
c
2.8
Naphthalene
uglbhp­
hr
31
26
28
28
@
b/
hour
25
22
23
24
Phenanthrene
pglbhp­
hr
2.8
1.8
2.0
2.2
ulblhour
2.3
1.5
1.6
1.8
Pyrene
pglbhp­
hr
ND
ND
ND
<
1.7
@
b/
hour
ND
ND
ND
<
1.4
a
Micrograms
per
brake
horsepower
hour
b
Micropounds
per
hour
ND
indicates
that
the
compound
was
not
detected.
Averages
include
detection
limits.

Table
6.11
presents
run­
by­
run
detection
limits
for
all
PAH
at
the
catalyst
outlet.

Final
Report
Cooper­
Bessemer
GMV­
4­
TF
2­
16
July
2000
2.6
DESTRUCTION
OF
ORGANIC
COMPOUNDS
BY
THE
CATALYST
PES
calculated
the
catalyst
destruction
efficiency
of
several
of
the
target
compounds.
These
data
are
presented
in
Table
2.8.
Several
of
the
compounds
that
were
on
the
target
list
were
not
detected
at
the
inlet
or
the
outlet.
PES
did
not
attempt
to
calculate
destruction
efficiencies
for
these
compounds
(
acetaldehyde,
acrolein,
l­
3
butadiene,
hexane,
ethyl
benzene,
styrene,
xylenes,
acenaphthene,
acenapthylene,
anthracene,
benzo(
a)
anthracene,
benzo(
a)
pyrene,
benzo(
b)
fluoranthene,
benzo(
e)
pyrene,
benzo(
k)
fluoranthene,
benzo(
g,
h,
i)
perylene,
chrysene,
dibenzo(
a,
h)
anthracene,
fluoranthene,
fluorene,
indeno(
1,2,3
­
cd)
pyrene,
2­
methylnapthalene,
perylene,
and
pyrene).

Formaldehyde,
nitrogen
oxides,
carbon
monoxide,
methane,
non­
methane
hydrocarbons,
and
total
hydrocarbons
were
detected
on
every
run,
both
before
and
after
the
catalyst.
PES
calculated
the
destruction
efficiencies
of
these
compounds
for
every
run
using
the
calculated
mass
flow
data.

Benzene
was
detected
before
the
catalyst
during
every
sampling
run,
and
toluene
was
detected
before
the
catalyst
during
every
sampling
run
except
for
two.
PES
calculated
the
removal
efficiencies
of
these
compounds
when
they
were
detected
before
the
catalyst.
The
benzene
and
the
toluene
detection
limits
after
the
catalyst
were
used
to
estimate
destruction
efficiency
for
these
two
compounds.

At
the
direction
of
EPA,
PES
calculated
the
destruction
efficiency
of
the
PAH
compounds
only
when
the
compound
was
detected
on
two
of
three
PAH
sampling
runs
before
the
catalyst.
Napthalene
and
phenanthrene
were
detected
on
all
three
of
the
sampling
runs
before
and
after
the
catalyst.
PES
calculated
the
destruction
efficiencies
of
these
compounds
for
each
PAH
sampling
run
using
the
calculated
mass
flow
data.

Final
Report
Cooper­
Bessemer
GMV­
I­
TF
2­
17
July
2000
TABLE
2.8
REMOVAL
EFFICIENCIES
OF
DETECTED
ORGANIC
COMPOUNDS
Run
ID
Run
IA
Run2­
7
Run
3
Run
4
Formaldehyde
46%
39%
28%
56%

Nitrogen
Oxides
(
as
NO9
­
6%
­
1%
­
2%
­
3%

Carbon
Monoxide
67%
69%
63%
62%

Methane
­
2%
­
1%
­
5%
­
2%

Non­
methane
Hydrocarbons
­
6%
­
1%
3%
­
5%

Total
Hvdrocarbons
­
4%
­
2%
­
a%
­
3%

Benzene
Toluene
Napthalene
IiPhenanthrene
53%
75%
63%
79%
­
86%
85%
63%

I
­
I
­
I
­
I
­
Run
5
Run
6
44%
57%

­
13%
1%

65%
64%

­
2%
I
2%

­
49%

­
5%
­
8%
1%

69%
80%

86%
87%

+­
I­+
Run
6
I
Run
9A
Run
10
­
5%

65%

72%

87%

bun
ID
1
Run
11
1
Run
12
1
Run
13
1
Run
14
1
Run
15
[
Run
16
1
PAH
1
1
PAH
2
1
PAH
3
1
Average(

37%
40%

­
17%
­
14%
52%

­
5%
52%

­
4%
52%

­
6%
53%

­
7%
49%

­
4%
39%

­
15%
38%
1
46%

­
16%
1
­
7%

Carbon
Monoxide
Methane
58%

­
4%
60%
64%
67%
67%
66%
60%
60%
59%
64%

­
2%
1%
1%
0%
­
2%
­
3%
­
2%
­
3%
­
1%

II
I
I
I
I
I
I
I
Non­
methane
Hydrocarbons
1
­
3%
I
­
21%
I
0%
I
­
7%
I
1%
1
22%
1
­
a%
[
­
a%
1
­
11%
1
­
6%
1
I
Total
Hydrocarbons
­
7%
­
6%
­
1%
0%
­
3%
­
7%
­
3%
­
5%
­
7%
­
4%

Benzene
77%
79%
74%
70%
74%
75%
78%
73%
75%
73%
I
II
Toluene
I
I
I
I
I
I
1
58%
1
87%
1
86%
I
87%
I
87%
I
61%
I
60%
I
44%
I
77%
1
­
­
­
­
­
53%
31%
34%
39%

54%
62%
41%
52%

Final
Report
Cooper­
Bessemer
GMV­
CTF
July
2000
3.0
SOURCE
DESCRIPTION
AND
OPERATION
This
section
presents
discussions
of
the
candidate
engine
and
the
catalyst
that
was
used
for
the
test
program.
The
sections
that
follow
describe
the
engine
and
the
operation
of
the
engine
during
testing.

3.1
ENGINE
DESCRIPTION
The
Cooper­
Bessemer
GMV­
4­
TF
stationary
internal
combustion
engine
is
a
four­
cylinder,
2­
stroke
internal
combustion
engine
with
a
manufacturer's
sea
level
rating
of
440
brake­
horsepower
(
bhp)
at
300
rpm.
Due
to
the
elevation
of
the
EECL,
the
unit
is
site­
rated
at
378
bhp.
The
engine
was
originally
manufactured
in
1946,
but
was
rebuilt
and
installed
at
the
EECL
in
1993.
The
pistons
are
14
inches
in
diameter
with
a
14­
inch
stroke.
Air
is
delivered
to
the
engine
via
a
supercharged
air
delivery
system;
air
manifold
pressures
are
controlled
by
the
EECL
process
control
system.
Engine
loading
is
controlled
by
a
computer­
controlled
water
brake
dynamometer.
Before
the
test
program
EECL
installed
an
oxidation
catalyst,
manufactured
by
MiraTech
Corporation,
on
the
engine.
EECL
aged
the
catalyst
under
its
normal
operating
condition
(
i.
e.,
burned
in
the
catalyst)
before
the
test
program.
This
procedure
ensured
that
the
catalyst's
HAP
destruction
efficiency
approximated
the
HAP
destruction
efficiency
of
mature
catalysts
installed
on
2istroke
engines
in
industry.
Table
3.1
presents
specifications
of
the
engine
and
the
catalyst.
Table
3.2
presents
nominal
engine
operating
parameters.

The
2­
stroke
cycle
requires
only
one
revolution
of
the
engine
crankshaft
for
each
power
stroke,
compared
to
the
4­
stroke
cycle
which
requires
two
revolutions.
When
the
compressed,
air/
fuel
mixture
is
ignited,
the
piston
travels
down
the
chamber.
Near
the
end
of
the
stroke,
the
piston
uncovers
ports
in
the
wall
of
the
cylinder
chamber,
and
scavenging
air
is
introduced.
This
air
consists
of
fresh
air
mixed
with
fuel.
As
the
scavenging
air
enters
the
cylinder,
an
exhaust
valve
opens
which
allows
the
exhaust
products
to
escape.
When
the
piston
returns
up
the
cylinder,
the
ports
are
covered,
the
exhaust
valve
is
closed,
and
the
air/
fuel
mixture
is
compressed
in
preparation
for
the
next
power
stroke.

The
GMV­
4­
TF
engine
was
outfitted
with
lean­
burn
technology,
which
is
used
for
the
control
of
NO,
emissions.
The
lean­
burn
system
uses
pre­
combustion
chambers
to
ignite
a
lean
air/
fuel
mixture
in
the
main
combustion
chambers.
A
relatively
rich
mixture
of
air
and
fuel
is
drawn
into
the
pre­
combustion
chamber
and
is
ignited
by
a
spark
plug.
The
resulting
Final
Report
Cooper­
Bessemer
GMV4­
TF
3­
l
July
2000
TABLE
3.1
E:

L
ENGINE
AND
CATALYST
SPECIFICATIONS
Cooper­
Bessemer
GlUV­
4­
TF
(
2­
stroke
lean­
burn,
natural­
gas­
fired)

ngine
Classification
Two­
stroke,
lean­
burn,
natural­
gas­
fired
Manufacturer
and
Type:
Cooper­
Bessemer
GMV4­
TF
Number
of
Cylinders:
4
Bore
and
Stroke:
14
in.
x
14
in.

Nominal
Engine
Speed:
300
rpm
;
nition
System
Classification
Spark
Ignited
Pre­
combustion
Chamber
Ignition
System:
Altronic
CPU­
2000
Pre­
combustion
Chamber
Type:
Diesel
Supply
"
Screw­
In"
Chamber
Number
of
Pre­
combustion
Chambers:
1
per
cylinder
latalyst
Classification
Oxidation
Type
Manufacturer:
MiraTech
Corporation
Tulsa,
Oklahoma
Date
of
Manufacture:
December
1998
Model
Number:
None.
Custom­
designed
unit
Serial
Number:
None.
Custom­
designed
unit
Item
Number:
CSU­
1216
Platinum/
Palladium
on
Aluminum
Catalyst
Material:
Substrate.
Manufactured
in
Finland
by
Kemira.

Element
Size:
12
in.
x
16
in.
x
3
in.
Effective
Area:
11
in.
x
14
7/
8"

Number
of
Elements:
2
Final
Report
Cooper­
Bessemer
GMV­
4­
TF
3­
2
July
2000
TABLE
3.2
SUMMARY
OF
NOMINAL
ENGINE
PARAMETERS
Paramettir
Nominal
Value
Acceptable
Deviation
Designation
Torque
7720
ft­
lb
f
2%
of
value
Primary
Speed
Jacket
Water
Temp
(
Outlet)

Oil
Temperature
(
Outlet)

Air
Manifold
Temperature
Air
Manifold
Pressure
300
rpm
I
f
5%
of
value
I
Primarv
16.5
"
F
f
5%
of
value
Primary
155
"
F
f
5%
of
value
Primary
110
"
F
f
5%
of
value
Primary
Barometric
+
7.5
in.
Hg;
f
5%
of
value
Primarv
Exhaust
Manifold
Pressure
Air
Manifold
Pressure
­
2.5
in.
Hg
&
5%
of
value
Primary
Ignition
Timing,
LPP'

Overall
Air/
Fuel
Ratio
1
18"
ATDC
f
5%
of
value
Primary
I
f
5%
of
value
I
PrimaY
Inlet
Air
Humidity­
Absolute
0.015
lb.
H,
O/
lb
Air
Fuel
Flow
3650
scfh
Oil
Pressure
Inlet
28
psig
f
10%
of
value
Primary
f
5%
of
value
Primary
f
5%
of
value
Secondary
Inlet
Air
Flow
Average
Exhaust
Temp
1600­
1700
s&
n
700
"
F
f
5%
of
value
Secondary
f
5%
of
value'
Secondary
'
For
a
GMV­
4­
TP
engine
operated
in
normal,
i.
e.,
not
lean­
burn,
configuration,
the
manufacturer
calls
for
ignition
timing
to
be
set
at
10'
BTDC.
Because
the
pre­
combustion
chamber
spreads
the
flame
throughout
the
engine
much
faster
than
a
standard
spark
ignition,
the
ignition
timing
had
to
be
retarded.
Timing
was
retarded
so
that
the
Location
of
Peak
Pressure
(
LPP)
was
consistent
with
an
engine
in
the
normal
firing
configuration.
LPP
for
this
engine
is
18"
ATDC.

Final
Report
Cooper­
Bessemer
GMV­
4­
TF
3­
3
July
2000
flame
is
then
directed
into
the
main
combustion
chamber,
which
contains
a
lean
mixture
of
air
and
fuel.
The
jet
from
the
pre­
combustion
chamber
provides
a
sufficient
source
of
ignition
for
combustion
of
the
air
fuel
mixture
in
the
main
chamber.

3.2
ENGINE
OPERATION
DURING
TESTING
As
stated
in
Section
2
of
this
document,
there
were
four
types
of
test
runs
that
were
conducted
during
the
test
program:
quality
assurance
runs,
sampling
runs
for
FTIRS/
CEMS/
GCMS,
CARB
429
sampling
runs,
and
daily
baseline
runs.
The
operation
of
the
engine
during
these
various
runs
is
discussed
in
the
following
pages
and
tables.

Table
3.3
presents
the
test
matrix
for
the
Cooper­
Bessemer
engine.
The
test
matrix
was
originally
presented
in
the
Quality
Assurance
Project
Plan.
During
the
test
program,
the
six
engine
operating
parameters
that
were
expected
to
have
the
greatest
impact
on
pollutant
formation
were
varied.
These
parameters
were:
engine
speed
(
measured
in
revolutions
per
minute
or
rpm),
engine
torque
(
measured
in
foot­
pounds
or
ft­
lb),
air­
to­
fuel
ratio
(
calculated
as
an
equivalence
factor),
engine
timing
(
the
location
of
the
cylinder,
relative
to
top
dead
center,
at
the
time
of
peak
pressure
in
the
combustion
chamber),
air
manifold
temperature
(
measured
in
degrees
Fahrenheit),
and
jacket
water
outlet
temperature
(
also
measured
in
degrees
Fahrenheit).

Table
3.4
presents
engine
parameters
that
were
recorded
during
each
test
run
and
their
percent
deviation
from
the
target
values.
There
were
fifteen
sampling
runs
conducted
on
the
engine
during
the
four­
day
period.
Run
1A
was
a
make­
up
run
for
the
original
run
at
Condition
No.
1.
The
run
was
repeated
because
the
EECL
DAS
recorded
no
data
from
the
FTIRS
(
neither
up­
nor
downstream
of
the
catalyst).
Run
9A
is
a
repeat
of
Run
9.
Run
9
was
invalidated
because
the
engine
speed
at
the
completion
of
the
run
did
not
agree
with
the
target
engine
speed
for
that
condition.
In
the
original
test
plan,
the
sampling
runs
for
PAH
compounds
were
to
be
conducted
at
the
single
load
condition
at
the
end
of
the
16
test
points.
Because
of
time
constraints,
two
of
these
runs
(
PAH
1
and
PAH
3)
were
combined
with
the
CEMS,
FTIRS,
and
GCMS
sampling.
Run
PAH
1
was
conducted
simultaneously
with
Run
4,
and
Run
PAH
3
was
conducted
simultaneously
with
Runs
11
(
for
the
first
hour)
and
12
(
for
the
second
hour).
Sampling
Run
PAH
2
was
conducted
at
Condition
8
but
the
PAH
run
was
not
conducted
simultaneously
with
the
FTIRS,
CEMS,
and
GCMS
runs.
The
engine
was
set
up
at
parameters
prescribed
by
Condition
8
a
second
time
for
run
PAH
2.
Further
discussions
of
these
issues
may
be
found
in
the
report
submitted
by
CSU
EECL.
This
report
is
included
in
Appendix
A
of
this
docunient.

Conditions
2
and
7
were
combined
because
of
factors
surrounding
the
air/
fuel
ratio.
The
air
fuel
ratios
prescribed
in
the
QAPP
were
unrealistically
rich
for
the
2­
stroke,
lean­
burn
engine.
In
order
to
meet
the
air/
fuel
ratios
in
the
test
plan,
the
air
manifold
pressure
would
have
had
to
be
dropped
below
the
manufacturer's
minimum
recommendation.
Operating
the
Final
Report
Cooper­
Bessemer
GMV­
4­
TF
3­
4
July
2000
TABLE
3.3
TARGET
ENGINE
OPERATING
CONDITIONS
DURING
TESTING
Operating
Conditions
Tested:

Condition
1
Condition
2
Condition
3
Condition
4
Condition
5
Condition
6
Condition
7
Condition
8
Condition
9
Condition
10
Condition
11
Condition
12
Condition
13
Condition
14
Condition
15
Condition
16
Speed
Torque
Air/
Fuel
Air
Manifold
Jacket
Water
b­
pm)
(%
of
Ratio
Timing
("
BTDC)
Temperature
Temperature
maximum)
a0
("
0
(
OF)

H
H
N
S
S
S
H
L
N
S
S
S
L
L
N
S
s
S
L
H
N
S
S
S
H
H
L
S
S
S
H
H
H
S
S
S
H
L
H
S
S
S
L
H
L
S
S
S
H
H
N
S
L
S
H
H
N
S
H
s
H
H
N
S
S
L
.
H
H
N
S
S
H
H
H
N
L
S
S
H
H
N
H
S
S
H
H
N
S
S
S
H
H
N
S
S
S
L=
270
L=
70
N
=
0.33
s
=
2.5
s
=
110
S=
165
H=
300
H
=
100
L
=
0.30
L=
l
L=
90
L=
155
H
=
0.36
H=
6
H=
130
H=
175
Final
Report
Cooper­
Bessemer
GMV­
CTF
3­
5
July
2000
engine
this
way
could
have
damaged
the
engine.
CSU
set
the
air
manifold
over
pressure
to
7.75
inches
of
mercury,
which
resulted
in
an
air/
fuel
ratio
of
58.9,
which
corresponds
to
364%
excess
air,
or
an
equivalence
factor,
4,
of
0.27.
The
points
were
combined
because
this
air/
fuel
ratio
was
less
than
the
prescribed
air/
fuel
ratio
,
for
both
of
the
two
test
points.

Table
3.5
presents
engine
parameters
during
baseline
test
points.
The
testing
was
conducted
over
a
period
of
four
days.
During
that
period
the
engine
did
not
run
continuously,
but
was
shut
down
each
night.
Test
accuracy
required
that
the
overall
engine
operation
did
not
change
over
the
four­
day
period.
The
stability
of
the
engine
over
this
period
was
demonstrated
by
operating
the
engine
at
a
"
baseline"
condition
for
one
5minute
period
on
the
first
day
of
testing
and
for
one
5­
minute
period
on
each
subsequent
day
of
the
testing.
The
baseline
condition
was
corresponded
to
the
manufacturer's
recommended
settings.
Changes
to
the
baseline
parameters
would
have
indicated
a
change
in
the
overall
operating
characteristics
of
the
engine.
It
wouldnot
have
been
possible
to
distinguish
between
emission
rate
changes
attributable
to
changes
in
the
independent
variables
and
emission
rate
changes
attributable
to
random
changes
in
the
performance
of
the
engine.
Table
3.5
presents
values
of
the
stability
parameters
and
their
deviation
from
their
proscribed
values
(
see
Table
3.2)
for
the
engine
baseline
run
conducted
on
March
30,
1999.
The
table
presents
the
data
for
the
three
remaining
baseline
checks,
but
the
deviations
reported
are
from
the
measured
parameter
during
the
first
baseline
check.
We
present
the
data
in
this
fashion
because
stability
of
these
parameters
over
the
duration
of
the
test
program
is
more
important
than
the
deviation
of
the
parameters
from
the
engine
manufacturer's
nominal
values
for
them.

Final
Report
Cooper­
Bessemer
GMV­
4­
TF
3­
6
July
2000
TABLE
3.4
SUMMARY
OF
ENGINE
PARAMETERS
­
COOPER
BESSEMER
GM.
V­
6TF
Run
ID
Run
1A
Run
2­
7
Run
3
Run
4
Run
5
Actual
300
299
269
270
300
Engine
Speed,
rpm
Target
300
300
270
270
300
%
diff
0%
0%
0%
0%
0%

Actual
7723
5285
5286
7324
7731
Engine
Torque
ft­
lb
Target
7720
5404
5404
7720
7720
%
diff
0.0%
­
2.2%
­
2.2%
­
5.1%
0.1%

Actual
0.33
0.27
0.25
0.32
0.30
Equivalence
Ratio,
+
(=
l/%
EA)
Target
0.33
0.33
0.33
0.33
0.30
%
diff
­
1.4%
­
19.7%
­
23.6%
­
2.8%
­
0.4%

Actual
19.2
18.5
18.2
17.2
18.7
Timing,
Location
of
Peak
Pressure,
"
ATDC
Target
18.0
18.0
18.0
18.0
18.0
%
diff
7%
3%
1%
­
4%
4%

Actual
111
109
110
110
111
Air
Manifold
Temperature,
"
F
Target
110
110
110
110
110
%
diff
0.1%
­
0.1%
0.0%
0.0%
0.1%

Actual
164
165
164
165
164
Jacket
Water
Temperature,
"
F
Target
165
165
165
165
165
%
diff
­
0.5%
0.0%
­
0.5%
0.0%
­
0.4%

Horsepower
bhp
441
302
272.
377
441
Fuel
Flow
Rate
SCfh
3672
2835
2491
3279
3661
Higher
Heating
Value
Btulcf
1072
1090
1090
1032
1072
Heat
Rate
MMBtulhr
3.94
3.09
2.72
3.38
3.92
Dry
Fuel
Factor,
Fd
dscf/
MMBtu
8664
8672
8672
8661
8664
Run6
1
Run6
1
RunSA
1
RunlC
300
270
299
299
300
270
300
300
0%
0%
0%
0%

7727
7360
7728
7729
7720
7720
7720
7720
0.1%
­
4.7%
0.1%
0.1%

0.34
0.27
0.33
0.32
0.36
I
0.30
I
0.33
I
0.33
­
6.4%
­
8.9%
0.3%
­
2.0%

18.0
18.0
18.0
18.3
110
110
92
130
110
110
90
130
0.0%
0.0%
0.3%
0.0%

164
165
165
165
165
165
165
165
­
0.5%
­
0.3%
0.0%
­
0.1%

442
378
441
442
3646
3130
3626
3674
1072
1072
1090
1090
3.91
3.35
3.95
4.01
8664
8664
8672
8672
Final
Report
Cooper­
Bessemer
GMV­
4­
TF
3­
7
July
2000
TABLE
3.4
(
CONCLUDED)

SUMMARY
OF
ENGINE
PARAMETERS
­
COOPER
BESSEMER
GMV­
CTF
Run
ID
El
1
Run11
1
Run12
1
Run13
1
Run14
1
Run15
I
Run16
I
PAHI
1
PAH2
1
PAH3
El
El
Ti
PI
Actual
270
270
300
300
299
299
270
270
270
ngine
Speed,
rpm
Target
270
270
300
300
300
300
270
270
300
%
diff
0%
0%
0%
0%
0%
0%
0%
0%
­
10%

Actual
7356
7349
7727
7728
7729
7731
7326
7341
7353
ngine
Torque
ft­
lb
Target
7720
7720
7720
7720
7720
7720
7720
7720
7720
%
diff
­
4.7%
­
4.8%
0.1%
0.1%
0.1%
0.1%
­
5.1%
­
4.9%
­
4.8%

Actual
0.30
0.29
0.33
0.32
0.32
0.33
0.33
0.29
0.29
quivalence
Ratio,
0
(=
l/%
EA)
Target
0.33
0.33
0.33
0.33
0.33
0.33
0.33
0.33
0.33
%
diff
­
9.9%
­
11.6%
­
1.5%
­
1.7%
­
3.5%
­
1.4%
,
­
0.7%
­
13.1%
­
10.8%

Actual
18.9
18.7
21.3
16.9
19.0
19.0
17.1
18.9
18.8
ming,
Location
of
Peak
ressure,
"
ATDC
Target
18.0
18.0
21
.
o
16.9
18.0
18.0
18.0
18.0
18.0
%
diff
5%
4%
1%
0%
6%
6%
­
5%
5%
4%

Actual
110
110
110
110
111
111
110
110
110
ir
Manifold
Temperature,
"
F
Target
110
110
110
110
110
110
110
110
110
%
diff
0.1%
0.1%
0.1%
0.0%
0.1%
0.1%
0.0%
0.1%
0.1%

Actual
154
175
184
184
165
164.
165
165
164
x&
et
Water
Temperature,
"
F
Target
155
175
165
'
165
165
165
165
165
165
%
diff
­
0.5%
­
0.3%
­
0.6%
­
0.4%
0.0%
­
0.6%
0.0%
­
0.2%
­
0.4%

orsepower
bhp
378
378
441
441
442
442
377
377
378
Jel
Flow
Rate
SCtll
3277
3271
3727
3585
3715
3713
3277
3300
3274
igher
Heating
Value
Btulcf
1032
1032
1072
1072
1090
1090
1032
1032
1032
eat
Rate
MMBtulhr
3.38
3.38
3.99
3.84
4.05
4.05
3.38
3.41
3.38
ry
Fuel
Factor,
Fd
dscf/
MMBtu
8661
8661
8664
8664
8672
8672
8661
8661
8661
Ai
J:

Hi
FI
Hi
Final
Report
Cooper­
Bessemer
Gh4V­
6TF
3­
8
July
2000
TABLE
3.5
SUMMARY
OF
ENGINE
PARAMETERS
DURiNG
BASELINE
RUNS
Run
ID
Baseline
1
Baseline
2
Baseline
3
Baseline
4
Actual
299
300
Engine
Speed,
rpm
299
300
Deviation
'
­
0.33%
0.33%
0.00%
0.33%

Actual
7724
Engine
Torque,
ft­
lb
7723
7751
7724
Deviation
0.05%
­
0.01%
0.35%
0.00%

Actual
41.51
42.08
Air/
Fuel
Ratio
42.40
42.55
Deviation
­
1.2%
1.4%
2.1%
2.5%

Timing,
Location
of
Peak
Actual
18.5
18.4
18.7
18.9
Pressure,
"
ATDC
Deviation
2.7%
­
0.32%
1.4%
2.0%

Actual
110
110
Air
Manifold
Temperature,
"
F
110
110
Deviation
0.00%
0.00%
.
O.
OO%
0.00%

Actual
165
164
165
164
Jacket
Water
Temperature,
"
F
Deviation
0.00%
­
0.61%
0.00%
­
0.61%

Actual
155
155
154
155
Oil
Temperature,
"
F
Deviation
0.00%
0.00%
­
0.65%
0.00%

Actual
7.76
7.75
7.74
7.75
Air
Manifold
Pressure,
in.
Hg
Deviation
3.5%
­
0.13%
­
0.26%
­
0.13%

Exhaust
Manifold
Pressure,
in.
Actual
5.00
5.00
4.97
4.99
Hg
Deviation
0.00%
0.00%
­
0.60%
­
0.20%

Actual
0.01508
0.01497
Inlet
Air
Humidity,
lb
HzO/
lb
air
0.01532
0.01481
Deviation
0.53%
­
0.73%
1.6%
­
1.8%

Actual
3786
3770
3814
3873
Fuel
Flow,
scfh
Deviation
3.7%
­
0.42%
0.74%
2.3%

Actual
27.59
26.59
32.47
30.78
Oil
Pressure,
psig
Deviation
­
1.5%
­
3.6%
18%
12%

Actual
1716
1734
1767
1801
Inlet
Air
Flow,
scfh
Deviation
4.0%
1.0%
3.0%
5.0%

Actual
694
721
Exhaust
Temperature,
"
F
700
692
Deviation
­
0.86%
3.9%
0.86%
­
0.29%

'
Deviation
for
Baseline
Run
No.
1
is
calculated
with
respect
the
manufacturer's
recommended
engine
operating
parameten.

For.
Baseline
Run
Nos.
2.
3,
and
4,
deviation
is
calculated
with
respect
to
the
results
of
Baseline
Run
No.
1
Final
Report
Cooper­
Bessemer
GMV­
CTF
3­
9
July
2000
4.0
SAMPLING
LOCATIONS
Figure
4.1
presents
a
schematic
drawing
of
the
exhaust
gas
piping
on
the
GW­
4­
TF
engine.
The
exhaust
piping
consisted
of
a
12­
inch
internal
diameter
(
ID)
pipe
that
connected
the
engine
exhaust
manifold
to
the
catalyst.
A
second
section
of
12­
inch
pipe
connected
the
catalyst
outlet
to
the
exhaust
silencer.

The
sampling
location
before
the
catalyst
consisted
of
several
sets
of
sampling
ports
used
for
isokinetic
sampling,
velocity
traverses,
and
extraction
of
sample
gas
for
the
FTIRS,
CEM
and
GCMS
systems.
CARB
429
sampling
before
the
catalyst
was
conducted
using
two
sample
ports.
One
port,
which
was
a
3­
in
ID
port,
was
used
for
the
CARB
Method
429
sample
probe.
The
second
port
(
1
­
inch
ID)
was
used
for
velocity
traverses
before
and
after
each
test
run.
The
nearest
upstream
disturbance
from
the
isokinetic
sampling
port
was
56
inches
(
4.58
diameters)
upstream
of
the
port.
The
disturbance
consisted
of
a
90"
pipe
elbow
connecting
the
exhaust
pipe
to
the
engine's
exhaust
manifold.
A
flange,
which
connected
two
sections
of
the
exhaust
pipe,
was
23
inches
upstream
of
the
3­
inch
ports,
but
was
not
considered
a
flow
disturbance.
The
nearest
downstream
disturbance
from
the
3­
inch
port
was
another
90"
pipe
elbow,
which
was
52
inches
(
4.33
diameters)
downstream
of
the
3­
inch
port.
The
nearest
upstream
disturbance
from
the
l­
inch
port,
which
was
used
for
velocity
traverses,
was
the
90"
pipe
elbow
that
was
8
1.5
inches
(
6.79
diameters)
upstream.
The
nearest
disturbance
downstream
of
the
l­
inch
port
was
the
second
90"
pipe
elbow,
found
26.5
inches
(
2.21
diameters)
downstream
of
the
l­
inch
sample
port.
PES
conducted
isokinetic
and
velocity
traverses
through
the
horizontal
ports
using
a
twelve­
point
traverse
matrix.
The
sample
point
locations
are
presented
in
Figure
4.2
for
the
sampling
locations
before
and
after
the
catalyst.

Multiple
ports
were
also
installed
on
the
pipe
after
the
catalyst.
A
3­
inch
ID
sample
port,
located
9
inches
downstream
of
the
CARB
429
port,
was
used
for
the
CARR
429
sample
probe,
and
a
l­
inch
sample
port,
was
used
for
velocity
traverses.
The
nearest
upstream
disturbance
from
the
3­
inch
ID
sample
port
was
the
catalyst
outlet,
which
was
96.5
inches
(
8.04
diameters)
upstream
of
the
port.
The
nearest
downstream
disturbance
from
the
3­
inch
sample
port
was
a
90"
pipe
elbow
upstream
of
the
exhaust
silencer.
The
exact
distance
to
the
pipe
elbow
could
not
be
measured
because
the
elbow
was
not
accessible
from
the
roof.
The
elbow
was
greater
than
89
inches
(
7.4
diameters)
downstream
of
the
3­
inch
sample
port
used
for
CARD
sampling.
The
nearest
upstream
disturbance
from
the
l­
inch
port
used
for
velocity
traverses
was
the
catalyst
exit.
The
exit
was
105;
5
inches
(
8.79
diarkters)
upstream
of
the
port.
The
90"
pipe
elbow
was
greater
than
80
inches
Final
Report
Cooper­
Bessemer
GMV4­
TF
4­
l
July
2000
(
6.67
diameters)
downstream
of
the
1
­
inch
sample
port.
PES
used
a
four­
point
traverse
matrix
for
both
the
CAB
429
and
the
velocity
sampling
at
this
location.
All
four
sample
*
points
were
on
the
horizontal
traverse
line.

Final
Report
Cooper­
Bessemer
GMV­
4­
TF
4­
2
July
2000
4­­­

b
A
00
24"

uu
1"
ports
v
0
0
12"
Figure
1
1"
ports
1
3"
ports
Catalyst
L
53"
11"
32.5"
9"
34.5"
46"

g­

0
0
00
0
0
12"
u
l­
l
u
l­
l
Ll
Ll
u
wall
Sample
Port
Locations
for
Velocity,
CARB
429,
FTIRS,
CEMS,
AND
GCMS
Sampling
Final
Report
Cooper­
Bessemer
GMV4­
TF
j
4­
3
July
2000
121110
9
8
7
6543211
Catalyst
Inlet
Catalyst
Outlet
Traverse
Distance
Traverse
Distance
Point
from
inside
Point
from
inside
Number
wall
(
in.)
Number
wall
(
in.)

1
2
3
4
5
6
7
8
9
10
11
12
0
II2
0
13/
16
1
7/
16
2
II8
3
4
l/
4
7
314
!
7/
8
10
9116
11
3/
16
11
l/
2
1
0
3116
2
3
3
9
4
11
3/
16
Figure
4.2
Sample
Point
Locaths
for
Velocity
and
CARB
429
Sampling
Final
Report
Cooper­
Bessemer
GMV­
4­
TF
4­
4
July
2000
5.0
SAMPLING
AND
ANALYSIS
METHODS
This
section
discusses
the
various
sampling
and
analysis
methods
employed
by
PES,
EMI,
and
EECL
to
quantify
the
HAP
emission
rates
before
and
after
the
oxidation
catalyst.
PES
selected
the
sampling
and
analysis
procedures
that
would
provide
the
information
required
by
during
the
planning
stages
of
the
project.
The
methods
were
selected
to
provide
the
required
data
in
the
most
economical
fashion,
while
providing
the
quality
required
by
Emissions
Standards
Division
(
ESD).

PES
divided
these
methods
into
two
categories
based
upon
quality
control
procedures
employed.
Type
I
methods
were
typical
source
test
methods,
designed
by
EPA
to
be
portable,
field
test
procedures.
PES
and
the
subcontractors
followed
QA
and
calibration
procedures
described
in
40
CFR
60,
,
Appendix
A
(
or
other
references
as
appropriate)
for
these
methods.

Type
II
methods
were
those
that
used
permanently
installed
instruments
housed
in
a
temperature­
controlled
environment
and
operated
in
the
same
fashion
as
continuous
monitors
used
by
industry
to
show
compliance
with
emission
regulations.
Because
these
instruments
are
maintained
in
a
laboratory­
type
environment
(
the
control
room
at
EECL),
fewer
QA
activities
and
calibrations
adequately
show
their
continuing
accuracy.
The
only
significant
change
to
the
quality
assurance
activities
was
that
fewer
instrument
calibrations
were
done
to
quantify
instrument
drift.
Historical
calibration
data
for
the
instruments
shows
their
stable
operation
over
extended,
e.
g.,
24­
hour,
periods.
Multipoint
calibrations
were
conducted
(
including
the
sampling
system
bias
checks)
on
these
instruments
once
at
the
beginning
of
each
engine
test.

Table
5.1
summarizes
the
parameters
measured,
the
sampling
methods,
the
classification,
and
measurement
principle.
The
text
that
follows
presents
brief
descriptions
of
the
sampling
and
analysis
procedures
used.

5.1
LOCATION
OF
MEASUREMENT
SITES
AND
SAMPLENELOCETY
TRAVERSE
POINTS
PES
used
EPA
Method
1,
"
Sample
and
Velocity
Traverses
for
Stationary
Sources,"
to
select
the
measurement
sites
for
velocity
traverses
and
CARB
429
sampling
up­
and
Final
Report
Cooper­
Bessemer
GMY­
4­
TF
5­
l
July
2000
TABLE
5.1
SUMMARY
OF
SAMPLING
AND
ANALYSIS
METHODS
Parameter
Test
Method
QA
Category
Measurement
Principle
Sample
Point
Location
EPA
Method
1
Type
1
Linear
Measurement
Velocity
and
Volumetric
Flow
EPA
Method
2C
Type
I
Differential
,
Pressure
I
I
Volumetric
Flow
EPA
Method
19
Type
II
Stoichiometry
,
I
Oxygen
and
Carbon
Dioxide
EPA
Method
3A
Type
II
Paramagnetic
and
Non­
dispersive
Infrared
Analyzers
EPA
Method
4
Type
I
Gravimetric
Moisture
I
GRI
Protocol'
I
Type
I
I
FTIRS
Analyzer
I
Carbon
Balance2
I
Type
I
I
Stoichiometry
Nitrogen
dxides
Type
II
Chemiluminescent
Analyzer
Carbon
Monoxide
I
EPA
Method
10
I
Type
II
I
GFC/
NDIR
Analyzer
Formaldehyde,
Acetaldehyde,
Acrolein
I
GRI
Protocol
Type
II
FTIRS
Analyzer
1,3­
Butadiene,
Hexane,
Benzene,
Toluene,
Ethyl
benzene,
Xylenes,
Styrene
Alternate
Method
17
Type
I
Gas
Chromatograph
w/
Mass
Spectrometer
Detector
Methane
1
EPA
Method
25A
(
modified)
1
Type
II
I
GC­
FID
Analyzer
Non­
methane
hydrocarbons
1
EPA
Method
25A
(
modified)
1
Type
II
I
GC­
FID
Analyzer
Total
Hydrocarbons
I
EPA
Method
25A
I
Type
II
I
FID
Analyzer
Polycyclic
Aromatic
Hydrocarbons
CARB
429
Type
I
Low
Resolution
GCMS
'
Measurement
of
Select
Hazardous
Air
Pollutants,
Criteria
Pollutants,
and
Moisture
Using
Fourier
Transform
Infrared
(
FTIR)
Spectroscopy.
Presented
as
an
Appendix
to
Fourier
Transform
Infrared
Spectroscopy
(
FTIRS)
Method
Validation
at
a
Natural
Gas­
Fired
Internal
Combustion
Engine
(
GRI­
95/
0271),
Gas
Research
Institute,
December
1995.

*
Derivation
of
General
Equation
for
Obtaining
Engine
Exhaust
Emissions
on
a
Mass
Basis
Using
the
"
Total
Carbon"
Method.

Final
Report
Cooper­
Bessemer
GMV­
4­
TF
July
2000
downstream
of
the
catalyst.
PES
used
the
cyclonic
flow
check
procedure
outlined
in
Method
1
to
evaluate
the
suitability
of
the
inlet
location
for
isokinetic
sampling.
The
measurement
sites
are
discussed
in
Section
4.0.

5.2
DETERMINATION
OF
STACK
GAS
VOLUMETRIC
FLOW
RATE
PES
used
two
methods
to
calculate
the
volumetric
flow
of
the
stack
gas
up­
and
downstream
of
the
catalyst.
During
the
PAH
runs,
Method
2C
was
used
in
direct
support
of
the
CARJ3
429
sampling.
The
mass
flow
rates
of
the
PAH
compounds
and
the
run­
by­
run
detection
limits
are
calculated
using
the
results
of
these
velocity
traverses.
Method
19
was
used
to
calculate
the
volumetric
flow
rate
of
the
exhaust
gases
for
Runs
1A
through
16
and
during
the
PAH
runs.
The
mass
flow
rates
of
pollutants
measured
by
CEMS,
GCMS,
and
FTIRS
were
calculated
using
the
Method
19
flow
data.

PES
used
EPA
Method
2C,
"
Determination
of
Stack
Gas
Velocity
and
Volumetric
Flow
Rate
in
Small
Stach
or
Ducts
(
Standard
Pitot
Tube),"
to
determine
stack
gas
velocity
during
CARE3
429
sampling.
The
test
crew
used
a
standard
pitot
tube,
constructed
according
to
specifications
of
Section
2.7
of
Method
2
and
having
a
coefficient
CC,)
of
0.99.
The
pitot
tube
was
connected
to
an
inclined/
vertical
manometer
and
the
Ap
measured
at
each
traverse
point.
Stack
gas
temperature
was
measured
using
a
Type­
K
thermocouple.
The
average
stack
gas
velocity
was
calculated
from
the
average
of
the
square
roots
of
the
Ap
values,
the
average
stack
gas
temperature,
the
stack
gas
molecular
weight,
and
the
absolute
stack
pressure.
The
volumetric
flow
rate
is
the
product
of
velocity
and
the
stack
cross­
sectional
area
of
the
duct
at
the
sampling
location.
PES
conducted
a
velocity
traverse
using
the
standard
pitot
tube
before
each
run
and
adjusted
the
sampling
rate
of
the
CARB
429
train
based
upon
these
data.
PES
employed
this
approach
with
the
approval
of
the
WAM.
Access
to
the
sampling
locations
was
severely
restricted
due
to
the
short
runs
of
exhaust
piping
and
the
profusion
of
sampling
probes
required
during
each
sampling
run.

EPA
Method
19,
"
Determination
of
Sulfur
Dioxide
Removal
Eflciency
and
Particulate
Matter,
Sulfir
Dioxide,
and
Nitrogen
Oxides
Emissions
Rates,"
uses
a
fuel
factor
to
calculate
the
volume
of
combustion
products
generated
upon
combustion
of
specific
fuel
types.
EECL
personnel
analyzed
a
sample
of
the
natural
gas
fuel
during
each
day
of
testing.
The
results
of
the
compositional
analysis
were
used
to
calculate
the
higher
heating
value
and
oxygen­
based
F­
factor,
Fe
The
EECL
Engine
Control
and
Monitoring
System
recorded
stack
gas
O2
concentrations
and
the
fuel
consumption
rate
during
testing.
These
data
were
used
to
calculate
the
exhaust
gas
flow
rates
by
multiplying
the
fuel
consumption
by
the
fuel
factor
and
correcting
for
excess
air.
Exhaust
gas
flow
rates
were
calculated
before
and
after
the
catalyst
for
each
rtm.
The
natural
gas
heating
values
and
the
calculated
F­
factors
used
for
each
test
run
are
presented
in
Table
2.2.

Final
Report
Cooper­
Bessemer
GMV­
4­
TF
5­
3
July
2000
5.3
DETERMINATION
OF
STACK
GAS
DRY
OXYGEN
AND
CARBON
DIOXIDE
CONTENT
EECL
used
EPA
Method
3A,
"
Determination
of
Oxygen
and
Carbon
Dioxide
Concentrations
in
Emissionsfiom
Stationary
Sources
(
Instrumental
Analyzer
Procedure),"
to
measure
oxygen
and
carbon
dioxide
content
of
the
exhaust
gas
during
testing.
EECL's
sample
gas
extraction
and
transport
system
extracted
a
gas
sample
from
the
exhaust
gas
stream.
The
sample
was
conditioned
to
remove
moisture
and
entrained
particulate
matter
and
directed
the
Rosemount
NGA­
2000
gas
analysis
system.
Oxygen
was
measured
using
the
paramagnetic
detection
principle.
Carbon
dioxide
was
measured
using
and
non­
dispersive
infrared
(
NDIR)
analyzer.
The
oxygen
and
carbon
dioxide
monitors
were
calibrated
with
a
pre­
purified
zero
gas
and
three
upscale
gas
standards
corresponding
to
approximately
30,55,
and
85
percent
of
the
instruments'
analytical
ranges.
EECL
used
only
EPA
Protocol
gas
standards
certified
by
the
gas
manufacturer.
The
calibration
gases
that
were
used
and
the
calibration
responses
of
the
instruments
are
discussed
in
Section
6.0
of
this
document.
A
schematic
diagram
of
the
CEMSKTIRS
sampling
and
analysis
system
is
presented
in
Figure
5.1.

5.4
DETERMINATION
OF
STACK
GAS
MOISTURE
CONTENT
PES
and
EECL
used
three
methods
to
determine
the
moisture
concentration
in
the
exhaust
gas
before
and
after
the
catalyst.
During
the
PAH
runs,
Method
4
was
used
in
direct
support
of
the
CARB
429
sampling.
During
the
CEMS/
GCMS/
FTIRS
runs,
moisture
was
measured
using
the
FTIRS
upstream
of
the
catalyst,
and
by
a
carbon
balance
calculation
downstream
from
the
catalyst.
During
the
testing,
EECL
personnel
determined
that
the
moisture
concentrations
after
the
catalyst,
as
measured
by
the
Nicolet
Magna
560
FTIRS
analyzer,
were
about
6
percent
higher
than
actual.
EECL
calculated
the
moisture
concentration
after
the
catalyst
using
a
carbon
balance
method.

PES
used
EPA
Method
4,
"
Determination
of
Moisture
Content
in
Stack
Gases,"
to
measure
the
flue
gas
moisture
content
during
the
CARB
429
sampling.
The
gas`
sample
was
extracted
from
the
exhaust
pipe
and
pulled
through
an
impinger
train
chilled
by
an
ice
bath.
The
field
technicians
weighed
the
impinger
train
(
including
the
XAD@­
2
sorbent
trap)
before
and
after
sampling.
PES
then
calculated
the
quantity
of
water
collected
in
the
train
and
the
moisture
content
of
the
stack
gas.

EECL
used
methodology
described
in
the
document
measurement
of
Select
Hazardous
Air
Pollutants,
Criteria
Pollutants,
and
Moisture
Using
Fourier
Transform
Infrared
(
FTIRS)
Spectroscopy
"
to
measure
moisture
concentrations
upstream
of
the
catalyst
This
document
is
referred
to
in
this
report
as
the
GRI
Protocol,
and
is
presented
as
Appendix
B
of
a
report
published
by
the
Gas
Research
Institute:
`!
Fourier
Transform
Inpared
Spectroscopy
(
FTIRS)
Method
Validation
at
a
Natural
Gas­
Fired
Internal
Combustion
Final
Report
Cooper­
Bessemer
GMVATF
5­
4
July
2000
Engine."
A
sample
of
the
gas
was
extracted
from
the
exhaust
and
directed
to
a
Nicolet
Rega
7000
FTIRS
analyzer
to
measure
the
moisture
concentration
in
the
exhaust
gas.
The
gas
sample
was
filtered
to
remove
entrained
particulate
matter
and
transported
to
the
analyzer
via
a
heated
Teflon
sampling
line.
Further
discussion
of
the
FTIRS
sampling
and
analysis
method
may
be
found
in
the
report
generated
by
the
EECL
and
the
GRI
protocol,
which
is
contained
in
Appendix
D.

Because
the
FTIRS
analyzer
downstream
of
the
catalyst
did
not
measure
the
moisture
concentration
accurately,
EECL
used
a
carbon
balance
method
to
calculate
the
moisture
present
in
the
gas
stream
downstream
of
the
catalyst.
The
method
used
is
discussed
in
the
EECL
report
in
Appendix
A.

5.5
DETERMINATION
OF
NITROGEN
OXIDES
EPA
Method
7E,
"
Determination
of
Nitrogen
Oxide
EmissionsjFom
Stationary
Sources
(
Instrumental
Analyzer
Procedwe),
"
determined
nitrogen
oxide
content
of
the
exhaust
gases.
These
tests
also
provided
the
data
needed
to
do
the
EPA
Method
301
validation
of
the
FTIRS
for
NO,
emissions
from
this
source.
A
gas
sample
was
extracted
from
the
exhaust
gas
stream,
conditioned
to
remove
moisture,
and
the
nitrogen
oxide
concentration
determined
by
an
instrumental
analyzer.
The
measurement
principle
for
oxides
of
nitrogen
is
chemiluminescence.
The
NO,
monitor
was
calibrated
with
a
pre­
purified
zero
gas,
and
three
upscale
gas
standards
corresponding
to
approximately
30,55,
and
85
percent
of
the
instruments
analytical
ranges.
EECL
used
EPA
Protocol
gas
standards
certified
by
the
gas
manufacturer.
The
calibration
gases
that
were
used
and
the
calibration
responses
of
the
instruments
are
discussed
in
Section
6.0
of
this
document.
A
schematic
diagram
of
the
CEMWTIRS
sampling
and
analysis
system
is
presented
in
Figure
5.1.

5.6
DETERMINATION
OF
CARBON
MONOXIDE
EPA
Method
10,
`
IDetermination
of
Carbon
Monoxide
EmissionsJiom
Stationary
Sources,"
measured
CO
concentration
of
the
exhaust
gas
during
the
testing.
These
tests
also
provided
the
data
needed
to
do
the
EPA
Method
301
validation
of
the
FTIRS
sampling
and
analysis
system
for
CO
emissions
from
this
source.
A
gas
sample
was
extracted
from
the
exhaust
gas
stream,
conditioned
to
remove
moisture,
and
the
carbon
monoxide
concentration
determined
by
an
instrumental
analyzer.
The
measurement
principle
for
carbon
monoxide
is
GFCYNDIR.
The
CO
monitor
was
calibrated
using
a
pre­
purified
zero
gas
and
three
upscale
gas
standards
corresponding
to
approximately
30,55
and
85
percent
of
the
instrument's
analytical
range.
All
gas
standards
used
for
calibrations
were
prepared
according
to
EPA
Protocol
1
and
certified
by
the
gas
manufacturer.
The
calibration
gases
that
were
used
and
Final
Report
Cooper­
Bessemer
GMV­
4­
TF
5­
5
July
2000
Nicolet
Magna
560
a
FTIR
Analyzer
_
_
_
,

Heated
Sample
Line
I
I
I
CHjNMHC
Analyzer
­
­
­
i
I
Heated
Sample
Line
CO
Analyzer
t
­
­
f
r
Miratech
Oxidation
Catalyst
.
'

88
/
A;
qusIp
/

1
8
Calibration
Gas
Cylinders
:

1
NO,
Analyzer
i­­
j
Exhaust
Flow
Heated
Sample
Line
Nicolet
Rega
7000
'
FTIR
Analyzer
­
­
­
'

Figure
5.1.
.
Schematic
Diagram
of
EECL
CEMS/
FTIRS
Sampling
and
Analysis
System
Final
Report
Cooper­
Bessemer
GMVATF
5­
6
July
2000
the
calibration
responses
of
the
instruments
are
discussed
in
Section
6.0
of
this
document.
A
schematic
diagram
of
the
CEMS/
FTIRS
sampling
and
analysis
system
is
presented
in
Figure
5.1.

5.7
DETERMINATION
OF
METHANE
AND
NON­
METHANE
HYDROCARBONS
A
modification
of
EPA
Method
25A,
"
Determination
of
Total
Gaseous
Organic
Concentration
Using
a
Flame
Ionization
Analyzer,"
determined
the
methane
and
non­
methane
concentrations
at
the
inlet
and
the
outlet
of
the
catalyst.
Gas
samples
extracted
from
each
gas
stream
were
transported
to
MSA
103OH
Methane/
Non­
Methane
Analyzers.
These
analyzers
are
single­
purpose
gas
chromatographs
that
separate
methane
from
the
other
organic
compounds
in
the
sample
by
passing
the
sample
through
a
separation
column.
The
methane
elutes
from
the
column
first
and
is
directed
to
the
flame
ionization
detector.
Then,
the
analyzer
reverses
the
flow
through
the
column
and
the
remaining
organic
compounds
are
back
flushed
to
the
same
detector.
The
analyzer
sums
the
two
fractions
to
yield
the
concentration
of
total
organic
compounds.
Because
this
unit
is
a
gas
chromatograph,
it
cannot
measure
methane
and
non­
methane
concentrations
continuously.
During
testing,
each
analyzer
determined
concentrations
once
every
five
minutes.
This
frequency
is
sufficient
for
testing
on
RICE
because
the
operating
conditions
were
maintained
within
close
constraints.
Each
analyzer
was
calibrated
before
the
test
program
using
methane
and
propane
calibration
standards
corresponding
to
approximately
30,50,
and
85
percent
of
the
instrument
span.
The
calibration
gases
that
were
used
and
the
calibration
responses
of
the
instruments
are
discussed
in
Section
6.0
of
this
document.
A
schematic
diagram
of
the
CEMWFTIRS
sampling
and
analysis
system
is
presented
in
Figure
5.1.

5.8
DETERMINATION
OF
GASEOUS
ORGANIC
HAPS
USING
FTIRS
EECL
used
two
FTIRS
systems
that
met
the
sampling
and
analysis
requirements
set
forth
in
the
GRI
Protocol.
EPA
has
approved
the
methodology
outlined
in
the
GRI
protocol
for
use
on
natural­
gas­
fired
reciprocating
internal
combustion
engines
on
July
2
1,
1995.
The
extractive
FTIRS
continuously
extracts
a
sample
gas
from
the
stack,
transports
the
sample
to
the
FTIRS
system,
and
does
spectral
analysis
of
the
sample
gas.
The
computer
system
analyzes
sample
gas
spectra
for
target
analytes
continuously
and
archives
them
for
possible
later
reanalysis.

The
sampling
and
measurement
system
consists
of
the
following
components:

l
heated
probe;

l
heated
filter;

l
heat­
traced
Teflon@
sample
line;

Final
Report
Cooper­
Bessemer
GhN­
4­
TF
5­
7
July
2000
l
.
Teflon@
coated,
heated­
head
sample
pump;

l
FTIRS
spectrometer;
and
l
QA/
QC
apparatus.

EECL
validated
each
sample
extraction
and
analysis
system
for
formaldehyde,
acetaldehyde,
and
acrolein
before
testing.
The
results
of
the
FTIRS
validation
are
discussed
in
Section
6.
The
basic
sampling
procedure
consisted
of
EECL
taking
an
initial
interferogram
of
the
stack
gas
with
the
FTIRS
measurement
and
analysis
system
before
each
test
to
describe
the
sample
matrix.
This
measured
the
concentrations
of
moisture
and
the
target
pollutants
and
allowed
for
adjustments
to
the
cell
pathlength
and
the
spectral
analysis
regions
if
the
concentrations
differ
from
expectations.
Sample
conditioning
was
not
necessary
at
the
EECL
test
site.

After
QA/
QC
procedures
and
initial
adjustments
were
completed
for
a
given
test
day,
a
gas
sample
was
drawn
continuously
through
the
heated
FTIRS
cell
while
the
system
collected
spectral
data.
The
FTIRS.
systems
collected
data
simultaneously
with
the
other
continuous
monitors
and
with
the
manual
train
sampling
for
PAHs
during
CAREJ
429
runs.
The
spectrometer
collected
one
complete
spectrum
of
the
sarnple,
as
an
interferogram,
per
second
and
averaged
interferograms
over
1
­
minute
periods.
The
FTIRS
computer
converted
these
time­
integrated
interferogram
into
conventional
wave
number
spectra,
analyzed
for
the
target
compounds
and
archived
the
data.
Sample
collection
was
33
minutes
in
duration,
coinciding
with
the
test
runs.

5.9
DETERMINATION
OF
ORGANIC
HAPS
BY
DIRECT
INTERFACE
GCMS
EMI
conducted
direct
interface
GUMS
sampling
using
EPA
Alternate
Method
17,
"
Determination
of
Gaseous
Organic
Compounds
by
Direct
Interface
GUMS."
The
sampling
and
analytical
procedures
used
during
this
testing
program
followed
those
detailed
in
the
method,
which
is
presented
in
Appendix
D
of
this
document.
The
instrument
was
calibrated
specifically
for
this
test
project
using
a
manufacturer's
certified
compressed
gas
mixture
of
nine
target
analytes
(
benzene,
toluene,
o,
m,
p­
xylenes,
styrene,
ethyl
benzene,
1,3­
butadiene,
and
hexane).
The
instrument
was
also
calibrated
for
all
compounds
identified
in
Section
1
of
the
method
approximately
one
month
before
this
test;
this
calibration
was
used
to
identify
additional
potential
analytes
not
specific
to
this
test
program.
Run­
by­
run
detection
limits
for
the
GCMS
compounds
are
presented
in
Section
6
of
this
document.

Effluent
gas
samples
were
withdrawn
at
a
constant
flow
rate
from
a
single
point
near
the
center
of
the
duct.
Effluent
was
withdrawn
at
approximately
1
S
liters
per
minute
through
the
sampling
system
for
no
less
than
5
minutes
before
sample
acquisition.
This
conditioning
period
serves
to
equilibrate
fully
all
of
the
sampling
system
components.
EM1
estimated
that
the
gas
residence
time
through
the
sampling
system
at
this
flow
rate
is
less
than
1
minute.

Final
Report
Cooper­
Bessemer
GMV­
4­
TF
5­
8
July
2000
Figure
5.2
presents
a
schematic
of
the
GCMS
measurement
system(
s)
used
during
the
test
program.

Exhaust
gas
samples
were
acquired
simultaneously
from
the
catalyst
inlet
and
outlet
sampling
locations.
A
total
of
four
samples
was
acquired
from
each
location
for
each
of
the
designated
engine
test
runs.
The
run
duration
was
approximately
30
minutes.
For
the
test
runs
where
PAH
sampling
trains
were
run,
each
GCMS
measurement
system
acquired
as
many
samples
as
possible
during
the
run
duration.

The
GCMS
instrumentation
was
operated
using
a
non­
evaporative
getter
(
NEG)
pump
to
maintain
the
requisite
high
internal
vacuum
needed
to
generate
mass
spectra.
Internal
standards
were
co­
added
with
every
effluent
sample
in
the
GC
sample
loop
before
injection
into
the
GC.
The
internal
standards
used
were
1,3,5­
trifluoromethyl
benzene
(
TRIS)
and
bromopentafluoro
benzene
(
BPFB).
These
compounds
are
not
usually
found
in
industrial
processes.
They
were
used
to
tune
the
mass
spectrometer,
to
assess
the
stability
and
performance
of
the
GCMS
on
each
sample
run,
and
to
determine
adherence
to
the
method
QA/
QC.
The
GC
was
operated
isothermally
at
60
°
C
to
separate
and
detect
the
target
analytes.
The
mass
spectrometer
was
operated
in
a
limited
full
scan
mode
(
a
45­
125
amu).
All
internal
GCMS
components
were
maintained
at
60
°
C.
The
procedures
detailed
in
the
Alternate
Method
17
were
followed
for
this
testing
program.

Before
the
test
program,
instrument
calibrations
were
conducted
at
the
EMI
laboratory
using
a
limited
full
scan
mode
of
mass
spectrometer
operation
(
from
45
to
125
amu).
The
limited
full
scan
mode
of
operation
allowed
for
the
lowest
possible
detection
limits
for
the
specific
target
analytes
while
still
generating
all
of
the
fragments
in
each
target
compounds
mass
spectrum
in
every
run.
The
calibration
curve
prepared
in
the
EM1
laboratory
was
used
to
quantify
all
QA
and
effluent
samples
acquired
in
the
field.
Establishing
a
valid
calibration
curve
requires
a
20
percent
relative
standard
deviation
(%
RSD)
for
each
individual
analyte
over
the
calibration
range.

Calibrations
were
done
by
conducting
two
successive
GCMS
runs
at
each
of
4
concentration
levels:
10
ppm,
3
ppm,
1
ppm,
and
300
ppb.
Section
10
of
the
method
describes
the
procedures
used
to
calculate
the
%
RSD
for
each
analyte.
Four
calibration
points
were
used
instead
of
the
three
specified
by
the
method
in
order
to
obtain
a
wider
dynamic
calibration
range,
particularly
for
1,3­
butadiene
and
hexane
(
whose
Detection
Limits
are
higher
than
the
other
target
analytes).
The
calibration
and
internal
standards
used
for
this
testing
were
certified
by
the
manufacturers.

Final
Report
Cooper­
Bessemer
GMV%
TF
5­
9
July
2000
I
Sample
Line
Heated
to
300
°
F
Probe
Box
Heated
to
250'
F
Mass
Flow
Meter
Condenser
Flow
Control
Excess
Sample
Atmospheric
Vent
4
Connection
Line
s
stem
I
I­
YIIGU
Calibration
Calibration
h
I
P
50
cc/
mm
during
GC­
h&
S
sample
acquisition)

I
GC­
MS
Analyzer
Condenser
System
ondensate
Drain
Contr
1
Box
Heated
to
125
°
F
A
(
or
at
least
5OF
above
saturation
temperature
of
sample
gas
)

Figure
5.2
Schematic
of
GCMS
Sampling
and
Analysis
System
Final
Report
Cooper­
Bessemer­
GMV­
4­
TF
5­
10
July
2000
5.10
DETERMINATION
OF
POLYCYCLIC;
AROMATIC
HYDROCARBONS
BY
CARB
429
PES
used
CARB
Method
429,
"
Determination
of
Polycyclic
Aromatic
Hydrocarbon
(
PAH)
EmissionsJEom
Stationary
Sources,"
to
quantify
PAH
concentrations
and
emission
rates
before
and
after
the
catalyst.
Sample
run
times
were
120
minutes
in
duration.
The
test
plan
specified
that
the
PAH
tests
consist
of
three
sampling
runs
at
the
engine
operational
condition
that
exhibited
the
highest
emissions
of
BTEX
compounds
as
measured
by
the
GCMS
apparatus.
The
GCMS
data
collected
before
the
PAH
runs
showed
that
BTEX
emissions
were
close
to
the
analyte
detection
limits
for
most
of
the
conditions,
so
PAH
testing
was
conducted
at
four
different
operational
conditions.
The
first
CARB
429
run
was
conducted
while
the
engine
was
run
at
Test
Condition
4
and
the
second
run
was
conducted
while
the
engine
was
run
at
Test
Condition
8.
During
the
third
run,
the
engine
was
operated
at
Test
Condition
12
for
the
first
part
of
the
run
and
at
Test
Condition
13
for
the
second
part
of
the
run.
The
difference
in
conditions
was
target
temperature
of
the
jacket
cooling
water,
which
was
not
expected
to
affect
formation
of
PAHs.
The
PAH
testing
was
conducted
in
this
fashion
to
make
up
for
delays
during
earlier
phases
of
the
test
program.
Figure
5.3
presents
a
schematic
diagram
of
the
CARB
429
sample
train.

PES
field
technicians
recovered
the
CARB
Method
429
sample
train
as
described
by
CARB
Method
429.
Method
429
specifies
that
sample
recovery
rinses
be
done
with
acetone,
hexane,
and
methylene
chloride.
PES
collected
blank
samples
of
reagent
grade
water,
acetone,
hexane,
methylene
chloride,
unused
filters,
and
XAD@­
2
resin
cartridges
used
during
the
test
program.
The
sample
recovery
apparatus
consisted
of
pre­
cleaned
Teflon@
or
glass.
Field
technicians
performed
three
acetone
rinses,
three
hexane
rinses,
and
three
methylene
chloride
rinses
of
each
sample
train
component
from
the
nozzle
to
the
front
half
of
the
filter.
They
also
rinsed
the
back
half
of
the
filter
holder,
the
connector,
and
the
condenser
three
times
with
acetone.
They
soaked
the
back
half
of
the
filter
holder,
connector,
and
condenser
three
times
with
acetone,
hexane,
and
methylene
chloride,
for
five
minutes
each
time.
PES
provided
pre­
cleaned
amber
glass
sample
bottles
with
Teflon
seals
for
the
recovery
of
solvent
rinses.

After
sampling
and
recovery,
the
CARB
,429
sample
fractions
were
stored
on
ice
and
transported
by
PES
personnel
from
Fort
Collins,
Colorado
to
PES'
laboratory
facilities
in
Research
Triangle
Park,
North
Carolina.
The
sample
bottles
were
examined
for
breakage
and
sample
loss.
The
samples
were
then
transferred
by
PES
personnel
to
ERG
laboratory
facilities
in
Morrisville,
North
Carolina
for
sample
extraction
and
analysis.
ERG
extracted
the
sample
fractions
for
each
PAH
sampling
run~
with
methylene
chloride,
then
combined
the
extracts.
The
6
extracts
(
3
inlet
samples
and
3
outlet
samples)
were
concentrated
to
a
volume
of
about
15
ml
using
a
Kuderna­
Danish
flask,
then
evaporated
to
dryness
using
a
nitrogen
blowdown
apparatus.
The
extracts
were
each
reconstituted
with
1
ml
hexane
before
analysis
using
a
gas
chromatograph
with
a
low
resolution
mass
spectrometer.
ERG's
analytical
report
for
the
PAH
samples
is
attached
in
Appendix
C.

Final
Report
Cooper­
Bessemer
GMV­
4­
TF
5­
l
1
July
2000
Heated
Probe,
S­
type
Pitot
&
Temp.
Sensor
Ice
Water
Sorbent
Module
(
water
cooled)
Check
Pitot
Manometer
Impingers
in
Ice
Bath:

B&!
lavSes
Thermocouple
Main
w
Valve
Orifice
I
I
W
I
Pump
Figure
5.3.
Schematic
Diagram
of
CARB
429
PAH
Sampling
Train
Final
Report
Cooper­
Bessemer
GMV+
TF
5­
12
July
2000
6.0
QUALITY
ASSURANCE/
QUALITY
CONTROL
PROCEDURES
AND
RESULTS
Summarized
in
this
section
are
the
specific
QA/
QC
procedures
that
PES,
EECL,
EMI,
and
ERG
personnel
employed
during
the
performance
of
this
source
testing
program.
PES'
quality
assurance
program
was
based
upon
the
procedures
and
guidelines
contained
in
the
"
Quality
Assurance
Handbook
for
Air
Pollution
Measurement
Systems,
Volume
III,
Stationary
Source
Specific
Methods,"
EPA/
600/
R­
94/
038c,
as
well
as
in
the
test
methods.
These
procedures
ensure
the
collection,
analysis,
and
reporting
of
reliable
source
test
data.

6.1
FTIRS
QA/
QC
PROCEDURES
EECL
calibrated
the
FTIRS
instrurnents
before
each
engine
test
series
and
at
the
beginning
and
end
of
each
test
day.
The
calibration
procedures
employed
were
consistent
with
procedures
found
in
the
following
documents:

Gas
Research
Institute
Report
Number
GRI­
95/
0271
entitled,
"
Fourier
Transform
Infrared
(
FTIRS)
Method
Validation
at
a
Natural
Gas­
Fired
Internal
Combustion
Engine"

This
report
was
prepared
for
the
Gas
Research
Institute
by
Radian
Corporation.
Included
as
appendices
are
two
additional
documents,
which
also
have
relevance
in
the
test
program:

"
Measurement
of
Select
Hazardous
Air
Pollutants,
Criteria
Pollutants,
and
Moisture
Using
Fourier
Transform
Infrared
(
FTIRS)
Spectroscopy"
­
Prepared
by
Radian
International
for
the
Gas
Research
Institute.

"
Protocol
for
Performing
Extractive
FTIRS
Measurements
to
Characterize
Various
Gas
Industry
Sources
for
Air
Toxic?'
­
Prepared
by
Radian
International
for
the
Gas
Research
Institute.

Final
Report
Cooper­
Bessemer
GMY­
4­
TF
6­
l
July
2000
6.1.1
FTIRS
Svstem
Preparation
.
I
Both
FTIRS
sampling
systems
(
before
and
after
the
catalyst)
were
subjected
to
an
EPA
Method
301
validation
process
for
formaldehyde,
acetaldehyde,
and
acrolein.
The
validation
process
quantified
the
precision
and
accuracy
of
each
FTIRS
analyzer
for
these
compounds.
Besides
the
validation
program,
EECL
personnel
performed
the
following
calibration
procedures
before
each
engine
test
series.

1.
Source
Evaluation
­
Initial
source
data
were
acquired
to
verify
concentration
ranges
of
target
compounds
and
possible
interferences.
This
was
completed
before
and
during
the
Method
301
validation
process
for
formaldehyde,
acetaldehyde,
and
acrolein,
and
during
the
test
program
for
moisture.

2.

3.
Sample
System
Leak
Check
­
A
leak
check
was
done
on
the
portions
of
the
system
between
the
sample
filter
and
the
pump
outlet.
A
rotameter
was
connected
to
the
discharge
side
of
the
sample
pump.
The
indicated
sample
flow
rate
was
recorded
while
the
sample
system
operating
at
typical
temperatures
and
pressures
(
the
sample
pump
pulled
a
slight
vacuum
on
the
suction
side).
The
inlet
was
closed
off
just
downstream
of
the
sample
probe.
A
rotameter
monitored
the
flow
rate.
A
leak
rate
of
4%
or
less
of
the
standard
sampling
rate
of
500
ml/
min
indicated
an
acceptable
leak
check.

Analyzer
Leak
Check
­
Both
FTIRS
analyzers
were
checked
to
ensure
that
they
were
operating
at
normal
operating
temperatures
and
pressures.
The
operating
pressures
were
recorded.
The
automatic
pressure
control
device
was
disabled
and
the
inlet
to
the
FTIRS
was
closed.
The
cell
was
evacuated
to
20%
or
less
of
the
normal
operating
pressure.
After
the
cell
was
evacuated,
it
was
isolated
and
the
cell
pressure
was
monitored
with
a
dedicated
pressure
sensor.
The
leak
rate
of
the
measurement
cell
must
be
less
than
10
Torr
per
minute
for
1
minute
for
the
analyzer
leak
to
be
considered
acceptable.

4.
Cell
Pathlength
Determination
­
The
FTIRS
cell
pathlengths
were
to
be
determined
using
the
procedure
outlined
in
the
Field
Procedure
Section
the
document
entitled
"
Protocol
for
Performing
Extractive
FTIRS
Measurements
to
Characterize
Various
Gas
Industry
Sources
for
Air
Toxics.
"
Because
each
FTIRS
was
a
fixed
pathlength
unit
(
i.
e.,
the
pathlengths
were
not
adjustable)
measurements
of
the
cell
pathlengths
were
deemed
unnecessary.
The
cell
pathlengths
specified
by
the
manufacturer
were
used
in
the
measurement
algorithms.

Final
Report
Cooper­
Bessemer
GMV%
TF
6­
2
July
2000
6.1.2
FTIRS
Daily
Calibrations
and
OA
Procedures,

Before
each
day
of
testing,
EECL
personnel
calibrated
each
FAIRS
system
following
the
procedures
outlined
below.

1.
Instrument
Stabilization
­
Each
of
the
following
components
were
checked
for
proper
operation
to
ensure
the
stability
of
the
operation
of
the
FTIRS
instruments:

a)
Instrument
heaters
and
temperature
controllers.

b)
Pressure
sensors
and
pressure
controllers.

C>
Sample
system
(
pump,
filters,
flow
meters,
and
water
knockouts).

2.
The
FTIRS
analyzers
were
purged
with
conditioned
air
for
a
minimum
of
30
minutes
before
conducting
background
spectrum
analysis.
During
periods
when
the
instruments
were
in
stand­
by
mode
(
i.
e.,
between
sampling
runs
or
between
test
days),
they
were
maintained
at
normal
operating
temperatures
and
purged
with
conditioned
air.

3.
Background
Spectrum
Procedures
­
Each
instrument
was
allowed
to
stabilize
while
being
purged
with
Ultrahigh
Purity
(
UHP)
nitrogen
for
10
minutes.
The
FTIRS
spectra
were
monitored
during
this
time,
until
CO
and
HZ0
concentrations
reached
a
steady
state.
The
following
procedures
were
then
done:

a>
The
interferogram
signal
was
checked
using
signal
alignment
software.

b)
A
single
beam
spectrum
was
collected
and
inspected
for
irregularities.

c)
Using
the
single
beam
spectrum,
the
detector
was
checked
for
non­
linearity,
and
corrected
if
necessary.

4
The
instrument
alignment
procedure
was
done.

4
A
background
spectrum
consisting
of
256
scans
was
collected.

4.
Analyzer
Diagnostics
­
Analyzer
diagnostics
were
done
by
analyzing
a
diagnostic
standard.
The
standard
was
a
109
ppm
CO
EPA
Protocol
gas
standard.
EECL
uses
CO
because
it
has
distinct
spectral
features
that
are
sensitive
to
variations
in
system
operation
and
performance.
The
standard
was
introduced
directly
into
each
instrument,
and
instrument
readings
were
allowed
to
stabilize
a
5­
minute
period.
The
accuracy
and
precision
of
each
instrument
were
calculated.
The
pass/
fail
criterion
for
accuracy
and
precision
was
10%
of
the
concentration
of
the
Final
Report
Cooper­
Bessemer
GMV­
4­
TF
6­
3
July
2000
standard
gas.
A
second
diagnostic
standard
consisting
of
a
blend
of
C02,
CO,
CHa
and
NOx
was
analyzed
using
the
same
procedure.
Each
instrument
met
the
precision
and
accuracy
requirements.
Analyzer
diagnostic
data
is
presented
in
the
report
generated
by
EECL
5.
Indicator
Check
&
Sample
Integrity
Check
­
An
indicator
check
was
done
by
analyzing
an
indicator
standard.
A
10.66
ppm
formaldehyde
standard
was
introduced
directly
into
each
instrument.
The
instrument
readings
were
allowed
to
stabilize
and
a
5­
minute
data
set
was
collected.
The
indicator
standard
was
then
introduced
into
the
sample
system
at
the
sample
probe,
just
upstream
of
the
filter.
The
instrument
readings
were
allowed
to
stabilize
and
a
5­
minute
set
of
data
was
collected.
The
accuracy,
precision,
and
recovery
were
calculated
based
on
equations
in
the
document
entitled
"
Protocol
for
Performing
Extractive
FTIRS
Measurements
to
Characterize
Various
Gas
Industry
Sources
for
Air
Toxics",
prepared
by
Radian
International
for
the
Gas
Research
Institute.
The
pass/
fail
criterion
for
accuracy,
precision,
and
recovery
is
100
f
10%
of
the
known
standard
(
recovery
shall
be
100
f
10%
of
the
instrument
reading
when
the
indicator
gas
was
introduced
directly
into
the
instrument.)
Each
instrument
met
these
criteria.
Indicator
check
and
sample
integrity
data
sheets
are
included
with
the
EECL
report.

6.1.3
Backpround
Assessment
During
data
acquisition
procedures,
the
baseline
absorbance
was
continually
monitored.
If
at
any
time
the
baseline
spectrum
changed
by
more
than
0.1
absorbance
units,
the
instrument's
interferometer
was
realigned
and
a
new
background
spectrum
collected.

6.1.4
Post
Test
Checks
Upon
completion
of
the
daily
test
program
steps
4
and
5,
of
the
pre­
test
calibration
procedures
were
repeated.
Both
of
the
FTIRS
analyzers
met
all
of
the
acceptance
criteria
for
the
calibration
and
QA
procedures.
Post
test
calibration
data
sheets
are
included
in
the
EECL
report.

6.1.5
FTIRS
Validation
Before
the
initiation
of
testing
on
the
engine,
both
FTIRS
sampling
and
analysis
systems
were
validated
for
formaldehyde,
acrolein,
and
acetaldehyde.
The
validation
was
conducted
by
personnel
from
ERG,
using
procedures
outlined
in
EPA
Method
301
"
Field
Validation
of
Pollutant
Measurement
Methods
from
Various
Waste
Media."
The
validation
was
conducted
by
means
of
a
dynamic
spiking
the
sample
gas
with
known
concentrations
of
formaldehyde,
acrolein,
and
acetaldehyde.
The
spike
gas
consisted
of
a
compressed
gas
cylinder
containing
a
mixture
of
acrolein
and
acetaldehyde.
Formaldehyde
was
added
to
the
Final
Report
Cooper­
Bessemer
GMV­
CTF
6­
4
July
2000
mixture
by
injecting
a
stock
formalin
solution
onto
a
heated
block
at
a
fixed
rate.
The
acrolein/
acetaldehyde
gas
standard
was
used
as
a
carrier
gas
for
the
vaporized
formaldehyde.
The
three­
component
mixture
was
injected
into
each
FTIRS
sampling
system
at
a
point
upstream
of
each
system's
filter.
Further
discussions
of
the
validation
procedures
employed
may
be
found
in
the
report
generated
by
EECL.

6.1.6
FTIRS
Detection
Limits
Table
6.1
presents
the
in­
stack
detection
limits
for
formaldehyde,
acetaldehyde,
and
acrolein
as
reported
by
CSU
EECL.
These
detection
limits
have
been
used
to
calculate
the
run­
by­
run
mass
detection
limits
for
each
of
the
target
pollutants.

6.2
CEMS
QA/
QC
PROCEDURES
The
following
paragraphs
describe
the
CEMS
quality
assurance
procedures
that
EECL
personnel
used
during
the
test
program.
The
calibration
and
QC
frequencies
far
exceeded
those
required
for
permanently­
installed,
compliance
analyzers,
but
are
less
than
those
specified
for
compliance
tests
by
EPA
(
40
CFR
60,
Appendix
A).
EECL
operates
their
CEMS
in
a
way
that
is
more
similar
to
permanently­
installed
analyzers.

6.2.1
Analvzer
Calibration
Gaseq
EECL
used
EPA
Protocol
calibration
gases.
The
.
calibration
gases
were
manufactured
by
Scott
Specialty
Gases.
For
this
program,
EPA
Protocol
1
calibration
gases
(
RATA
Class)
were
used.
Formaldehyde
and
acetaldehyde/
acrolein
standards
with
concentration
ranges
between
5
­
20
ppm
were
obtained
for
FTIRS
calibrations.
These
gases
are
not
available
as
EPA
Protocol
Gases,
so
EECL
specified
the
highest
quality
available.
Scott
supplied
certification
sheets,
which
may
be
found
in
the
Appendices
of
EECL's
test
report.

6.2.2
Res­
oonse
Time
Tests
Response
time
tests
were
done
on
each
sample
system
before
initiation
of
the
engine
test
program.
The
response
time
tests
were
performed
before
the
FTIRS
validation
process
for
each
sampling
system.
The
response
time
of
the
slowest
responding
analyzer,(
Questar
Baseline)
was
determined.
Response
time
tests
conducted
at
the
EECL
indicated
sampling
system
response
times
of
1:
10
minutes.
This
is
the
time
for
the
Rosemount
Oxygen
Analyzer
(
the
slowest
responding
continuous
analyzer)
to
stabilize
to
response
output
of
the
analyzer.
The
Questar
Baseline
Industries
CH4/
Non­
CH4
analyzers
have
a
minimum
cycle
time
of
4:
50
minutes.
The
overall
response
time
for
these
analyzers
when
their
cycle
is
started
1:
10
minutes
after
a
sample
source
change
is
5:
50
minutes.
When
the
CHJNon­
CH,
analyzer
cycle
time
was
initiated
at
a
sample
source
change,
the
overall
response
time
was
Final
Report
Cooper­
Bessemer
GMV­
4­
TF
6­
5
July
2000
9:
00
minutes.
The
response
time
was
tested
to
assure
that
the
analyzers'
response
was
for
exhaust
gas
entering
the
sample
system
from
each
of
the
test
point
conditions.

6.2.3
Analvzer
Calibrations
Zero
and
mid­
level
span
calibration
procedures
were
done
on
the
CO,
CO,,
0,,
NO,,
and
THC
analyzers
before
each
test
day.
Zero
and
span
drift
checks
were
performed
upon
completion
of
each
data
point
and
upon
completion
of
each
test
day.
A
zero
and
mid­
level
gas
was
introduced
individually
directly
to
the
back
of
the
analyzers
before
testing
for
carbon
monoxide,
carbon
dioxide,
oxygen,
total
hydrocarbons,
Methane/
Non­
Methane,
and
oxides
of
nitrogen.
The
analyzers
output
response
was
set
to
the
appropriate
levels.
Each
analyzer's
stable
response
was
recorded.
From
this
data
a
linear
fit
was
developed
for
each
analyzer.
The
voltages
for
each
analyzer
were
recorded
and
used
in
the
following
formula:

Final
Report
Cooper­
Bessemer
GW­
CTF
6­
6
July
2000
TABLE
6.1
DETECTION
LIMITS
OF
FTIRS
AND
CEMS
COMPOUNDS
RunID
1
RunlAIRunZ­
71
Run3
1
Run4
I
Run5
I
Run6
I
Run6
I
RunSAl
Run10
Catalyst
Inlet
Formaldehyde
mglbhp­
hr
3.6
5.4
5.6
3.2
3.9
3.5
4.5
3.3
3.4
mlb/
hr
3.5
3.6
3.5
2.6
3.8
3.4
3.8
3.2
3.3
L
Acetaldehyde
mg/
bhp­
hr
18
27
29
16
19
18
23
16
17
mlb/
hr
18
Ia
17
13
19
17
19
16
17
Acrolein
mg/
bhp­
hr
18
25
25
19
18
18
20
15
17
mlb/
hr
17
16
15
16
18
17
16
15
16
Nitrogen
Oxides
(
as
N02)
glbhp­
hr
0.001
0.002
0.002
0.001
0.002
0.001
0.002
0.001
0.001
Ib,
hr
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
Carbon
Monoxide
glbhp­
hr
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
lb/
hr
0.02
0.02
0.01
0.01
0.02
0.02
0.02
0.02
0.02
Methane
glbhp­
hr
0.1
0.1
0.2
0.1
0.1
0.1
0.1
0.1
0.1
Iblhr
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
Non­
methane
Hydocarbons
glbhp­
hr
0.03
0.04
0.04
0.03
0.03
0.03
0.03
0.03
0.03
lblhr
0.03
0.03
0.03
0.02
0.03
0.03
0.03
0.03
0.03
Total
Hydrocarbons
glbhp­
hr
0.0002
0.0003
0.0003
0.0002
0.0002
0.0002
0.0002
0.0002
0.0002
lb/
hr
0.0002
0.0002
0.0002
0.0002
0.0002
0.0002
0.0002
0.0002
0.0002
Catalvst
Outlet
Total
Hydrocarbons
Final
Report
Cooper­
Bessemer
GMV­
CTF
6­
7
July
2000
TABLE
6.1
(
CONCLUDED)

DETECTION
LIMITS
OF
FTIRS
AND
CEMS
COMPOUNDS
Run
ID
Formaldehyde
IRun
Run12~
Run13~
Run14~
Run15~
Run16~
PAHI
1
PAH21
PAH3
ca*&
st
Inlet
­­­
ILL­
L_
mg,
D"
p"
r
3.6
3.7
3.9
3.6
3.5
3.5
3.0
3.4
3.3
mlb/
hr
3.0
3.0
3.8
3.5
3.4
3.4
2.5
2.6
2.7
Acataldehyde
lb
crolein
mglbhp­
hr
mlblhr
glbhp­
hr
Nitrogen
Oxides
(
as
N02)
,
Whr
16
16
0.002
0.002
0.001
0.001
II
Non­
methane
Hydocarbon
g/
bhp­
hr
lblhr
It
Carbon
Monoxide
glbhp­
hr
lb/
hr
otal
Hydrocarbons
g/
bhp­
hr
0.0002
0.0002
lblhr
0.0002
0.0002
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.01
0.02
0.02
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.02
0.03
0.03
0.0002
0.0002
0.0002
0.0002
0.0002
0.0002
0.0002
0.0002
0.0002
0.0002
0.0002
0.0002
0.0002
0.0002
Final
Report
Cooper­
Bessemer
GMV­
4­
TF
6­
8
July
2000
Y
=
MX+
B
Where:
B
=
Intercept
M
=
Slope
X
=
Analyzer
or
transducer
voltage
Y=
Engineering
Units
After
each
test
point
and
upon
completion
of
a
test
day,
calibrations
were
conducted
by
reintroducing
the
zero
and
span
gases
directly
to
the
back
of
the
analyzers.
The
analyzers'
stabilized
responses
were
recorded.
No
adjustments
were
made
during
testing
or
during
the
final
calibration
check.
Initial
calibration
values
and
all
calibration
checks
were
recorded
for
each
analyzer
during
the
daily
test
program.

The
before
and
after
calibrations
checks
were
used
to
determine
zero
and
span
drift
for
each
test
point
for
the
CO,
CO,,
02,
THC,
CHJNon­
CH4,
and
NO,
analyzers.
The
zero
and
span
drift
checks
for
all
test
points
and
all
test
days
were
less
than
&
2.0%
of
the
span
value
of
each
analyzer
used
during
the
daily
test
program.
The
calibration
data
sheets
are
presented
in
the
test
report
generated
by
EECL.
Table
6.2
presents
the
types
and
frequencies
of
the
analyzer
calibrations
conducted
by
EECL.

6.2.4
Analyzer
Linearity
Check
Analyzer
linearity
checks
were
done
before
beginning
the
test
program.
The
oxygen,
carbon
monoxide,
total
hydrocarbon,
methane/
non­
methane,
and
oxides
of
nitrogen
analyzers
were
"
zeroed"
using
either
zero
grade
nitrogen
or
hydrocarbon
free
air.
The
analyzers
were
allowed
to
stabilize
and
their
output
recorded.
The
analyzers
were
then
"
spanned"
using
the
mid­
level
calibration
gases.
The
analyzers
were
allowed
to
stabilize
and
their
output
recorded.
From
this
data
a
linear
fit
was
developed
for
each
analyzer.
The
voltage
for
each
analyzer
was
recorded
and
used
in
the
following
formula:

Y
=
MX+
B
Where:
B
=
Intercept
M
=
Slope
X
=
Analyzer
or
transducer
voltage
Y
=
Engineering
Units
Using
the
linear
fit,
the
linear
response
of
the
analyzer
was
calculated.
Low­
level
and
high­
level
calibration
gases
were
individually
introduced
to
the
analyzers.
For
each
calibration
gas,.
the
analyzers
were
allowed
to
stabilize
and
their
outputs
were
recorded.
Each
analyzer's
linearity
was
acceptable.
The
predicted
values
of
a
linear
curve
determined
from
the
zero
and
mid­
level
calibration
gas
responses
agreed
with
the
actual
responses
of
the
low­
level
and
high­
level
calibration
gases
within
&
2.0%
of
the
analyzer
span
value.
The
Final
Report
Cooper­
Bessemer
GIW­
4­
TF
6­
9
July
2000
TABLE
6.2
TYPES
AND
FREQUENCIES
OF
CEMS
ANALYZER
CALIBRATIONS
Calibration
Type
Gas
Calibration
Gas
Calibrant
Concentration
(
units
Frequency
Injection
Validation
of
%
of
span
(
I))
Point
Criterion
ACE
c2)
02,
co,,
co,
NO,

Methane/
Non­
Methane
Hydrocarbons
0
to
0.25,
40
to
60,
80
to
100
0
to
0.1,
25
to
35,
45
to
55,
80
to
90
<
2%
of
analyzer
span
for
each
gas
Before
each
Directly
into
engine
test
the
analyzer
<
5%
of
respective
cal.
gas
value
ZSD
0)
02,
co,,
co,
NO,

MethaneMon­
Methane
Hydrocarbons
0
to
0.25,
40
to
60
or
80
to
100
(`)

25
to
35,
45
to
55
Before
and
I
after
each
test
rnn
Directly
into
the
analyzer
All
errors
<
3%
of
span
All
errors
<
3%
of
span
SSB
t4)
NO,

Methane/
Non­
Methane
Hydrocarbons
0
to
0.25,
40
to
60
or
80
to
90
CS)

0
to
0.25,
25
to
35,
45
to
55
or
80
to
90
w
Before
and
after
each
Both
directly
test
day
into
the
Both
errors
analyzer
and
into
the
inlet
<
5%
of
Before
and
after
each
of
the
sample
analyzer
spar
line
test
day
(
l)
­
The
span
must
be
1.5
to
2.5
the
concentration
expected
for
each
pollutant
(*)
­
Analyzer
calibration
error
check
t3)
­
Zero
and
span
drift
check
c4)
­
Sampling
system
bias
check
c5)
­
Whichever
is
closer
to
the
exhaust
gas
concentration
Final
Report
Cooper­
Bessemer
GMV­
4­
TF
6­
10
July
2000
methane/
non­
methane
analyzers'
linearity
was
acceptable
as
the
predicted
valued
agreed
with
the
actual
response
of
the
low­
level
and
high­
level
calibration
gases
within
*
5.0%
of
the
actual
calibration
gas
value.
This
procedure
was
done
for
one
range
setting
for
each
analyzer.
The
Linearity
Check
data
sheets
are
presented
the
test
report
generated
by
EECL.

6.2.5
NO,
Converter
Check
EECL
did
NO2
converter
checks
before
the
test
program
began.
A
calibration
gas
mixture
of
known
concentration
between
240
and
270
ppm
nitrogen
dioxide
(
NO,)
and
160
to
190
ppm
nitric
oxide
(
NO)
with
a
balance
of
nitrogen
was
used.
The
calibration
gas
mixture
was
introduced
to
the
oxides
of
nitrogen
@
IOx)
analyzer
until
a
stable
response
was
recorded.
The
converter
was
considered
acceptable
if
the
instrument
response
indicated
a
90
percent
or
greater
NOz
to
NO
conversion.
The
NO,
Converter
Check
data
sheets
are
presented
in
the
test
report
generated
by
EECL.

6.2.6
Samde
Line
Leak
Check
The
sample
lines
were
leak­
checked
before
the
engine
test
program.
The
leak
check
procedure
was
performed
for
both
pre­
catalyst
and
post­
catalyst
sample
trains.
The
procedure
was
to
close
the
valve
on
the
inlet
to
the
sample
filter
found
just
downstream
of
the
exhaust
stack
probe.
With
the
sample
pump
operating,
a
vacuum
was
pulled
on
the
exhaust
sample
train.
Once
the
maximum
vacuum
was
reached,
the
valve
on
the
pressure
side
of
the
pump
was
closed,
thus
sealing
off
the
vacuum
section
of
the
sampling
system.
The
pump
was
turned
off
and
the
pressure
in
the
sample
system
was
monitored.
.
The
leak
test
was
acceptable
as
the
vacuum
gauge
reading
dropped
by
an
amount
less
than
1
inch
of
mercury
over
a
period
of
1
minute.
The
Sample
Line
Leak
Check
data
sheets
are
presented
the
test
report
generated
by
EECL.

6.2.7
SamDIe
Line
Intewitv
Check
A
sample
line
integrity
check
was
done
before
and
upon
completion
of
each
test
day.
The
analyzers'
response
was
tested
by
first
introducing
a
mid­
level
calibration
gas
directly
to
the
NOx
analyzer.
The
analyzer
was
allowed
to
stabilize
and
the
response
recorded.
The
same
mid­
level
calibration
gas
was
thenintroduced
to
the
analyzer
through
the
sampling
system.
The
calibration
gas
was
introduced
into
the
sample
line
at
the
stack,
upstream
of
the
inlet
sample
filter.
The
analyzer
was
allowed
to
stabilize
and
the
response
recorded.
The
analyzer
response
values
were
compared
and
the
percent
difference
did
not
to
exceed
rt5%
of
the
analyzer
span
value.

The
sample
line
integrity
check
was
to
be
done
for
both
the
NO,
and
methane/
non­
methane
analyzers.
Due
to
time
constraints,
EECL
performed
the
integrity
check
for
the
NO,
analyzers
only.
The
SSB
procedure
was
performed
for
the
methane/
non­
methane
analyzers
Final
Report
Cooper­
Bessemer
GMV­
4­
TF
6­
11
July
2000
before
and
upon
completion
of
the
test
program.
The
Sample
Line
Integrity
Check
data
sheets
are
presented
in
the
test
report
generated
by
EECL.

6.2.8
Carbon
Balance
Check
One
of
the
methods
used
to
calculate
mass
emissions
was
a
carbon
balance
calculation
developed
by
Southwest
Research
Institute
specifically
for
the
American
Gas
Association.
The
calculations
consist
of
a
theoretical
O2
calculation
based
upon
measured
exhaust
stack
constituents
and
fuel
gas
composition.
The
theoretical
exhaust
O2
is
then
.
compared
to
the
measured
exhaust
0,.
The
percent
difference
between
the
actual
and
theoretical
O2
measurements
was
within
h5
%
of
the
measured
O2
reading.
The
O2
balance
was
performed
for
every
l­
minute
average
and
the
33­
minute
averaged
valued
for
each
test
point.

6.2.9
Fuel
Gas
&
Fuel
Flow
Measurement
Engine
fuel
gas
was
analyzed
on
a
real
time
basis
with
a
dedicated,
Daniels
Industries
GC.
The
GC
was
calibrated
on
a
daily
basis
against
a
known
standard.
A
gas
analysis
was
done
on
each
test
day.
This
analysis
gave
the
actual
specific
gravity,
mole
fractions
of
specific
hydrocarbons,
and
BTU
content
so
that
fuel
flow
and
mass
emissions
could
be
accurately
calculated.
Fuel
flow
measurements
were
made
with
an
AGA/
PRCI­
specified
orifice
meter
equipped
with
dedicated
high
accuracy
pressure
and
temperature
transmitters.
All
fuel
flow
calculations
were
in
accordance
with
AGA/
PRCI
Report
#
3.
All
stoichiometric
air
to
fuel
ratios
were
calculated
using
the
fuel
gas
analysis.
From
this
information,
the
equivalence
ratios
for
each
day
of
testing
were
determined.
All
fuel
gas
calibrations
and
analysis,
stoichiometric
air
to
fuel
ratio
calculations,
and
fuel
specific
F
Factor
calculations
are
presented
in
the
test
report
generated
by
EECL.
In
addition,
a
blind
fuel
gas
sample
provided
by
PES
was
analyzed..
The
result
is
presented
in
the
test
report
generated
by
EECL.

6.2.10
Fuel
Factor
Oualitv
Assurance
Checks
Besides
the
CEM
calibration
and
QC
checks,
carbon
dioxide
and
oxygen
measurements
were
validated
by
calculating
the
fuel
factor,
F,,
using
the
following
equation:

20.9
­%
02
F"=
%
CO,

Final
Report
Cooper­
Bessemer
GMV­
4­
TF
6­
12
July
2000
The
vahtes
of
F,
at
the
inlet
and
the
outlet
for
each
sampling
run
are
presented
in
Table
6.3.
For
natural
gas
combustion,
the
value
of
F,
should
be
between
1.60
and
1.84.
The
F,
values
were
within
the
prescribed
ranges
for
32
of
the
sampling
runs
conducted.
There
were
four
runs
for
which
the
F,
values
were
outside
these
limits.
However,
the
maximum
exceedance
was
1.6
%
of
the
maximum
F,
value.
Based
upon
the
results,
the
integrity
of
the
CEM
sample
stream
was
not
compromised
due
to
leaks
in
the
sampling
system.

TABLE
6.3
SUMMARY
OF
FUEL
FACTOR
VALUES
Run
Number
I
1.78
I
1.83
­­­
II
5
I
1.65
I
1.68
II
6
1.68
1.81
8
1.63
1.86
II
10
I
1.71
I
1.78
11
II
11
I
1.87
I
1.83
11
II
12
I
1.84
I
1.84
II
16
I
1.75
I
1.76
II
II
PAHl
I
PAH2
1.83
1.85
PAH3
1.85
1.83
Final
Report
Cooper­
Bessemer
GMV­
4­
TF
6­
13
July
2000
6.2.11
CEMS
Detection
Limits
For
each
of
the
sample
runs,
the
mass
detection
limits
of
the
CEMS
were
presented
previously
in
Table
6.1.
For
each
run,
the
detection
limit
was
calculated
using
analytical
detection
limit
data
supplied
by
EECL.
Table
6.4
summarizes
these
values.

TABLE
6.4
SUMMARY
OF
CEMS
ANALYTICAL
DETECTION
LIMITS
Parameter
Oxygen
Inlet
Detection
Outlet
Detection
Limit
Limit
0.01
%
volume
0.01
%
volume
Carbon
Dioxide
Nitrogen
Oxides
Carbon
Monoxide
Methane
0.25
%
volume
0.1
ppm
2
PPm
20
PPm
0.1
%
volume
0.1
ppm
2
PPm
20
PPm
Non­
methane
Hydrocarbons
2
PPm
2PPm
'

Total
Hydrocarbons
0.04
ppm
0.04
ppm
6.3
.
GCMS
QA/
QC
PROCEDURES
Each
day
the
GCMS
measurement
system
was
tuned
according
to
the
criteria
identified
in
the
method.
Achieving
the
criteria
for
a
valid
mass
spectral
tune
and
achieving
the
internal
standard
relative
mass
abundances
during
each
GCMS
run
(
see
Tables
3
and
4
of
the
method)
verify
the
continuing
instrument
performance
and
ensure
that
the
QA/
QC
of
the
method
is
achieved.
Achieving
the
criteria
for
a
valid
tune
also
allows
searches
of
the
NIST
Mass
Spectral
library
for
compounds
that
are
not
contained
in
the
instrument
specific
calibration.

Daily
system
calibrations
were
conducted
to
check
both
the
validity
of
the
initial
instrument
calibration
and
the
effectiveness
of
the
sampling
system
to
transport
the
target
analytes.
Daily
system
calibration
check
procedures
were
conducted
after
accomplishing
a
successful
instrument
tune'
using
the
blended
mixture
of
the
internal
standards.
Immediately
following
the
system
continuing
calibration,
nitrogen
was
injected
into
the
GCMS
sampling
Final
Report
Cooper­
Bessemer
GMV­
CTF
6­
14
July
2000
system
and
a
system
blank
was
acquired.
No
analytes
were
detected
in
any
of
the
system
blank
analyses.

The
direct
interface
GCMS
test
method
requires
that
continuing
system
calibrations
be
conducted
using
a
blended
mixture
of
6
surrogate
compounds
at
1
ppm.
For
this
test
program,
all
of
the
target
analytes
were
checked
daily
at
the
1
ppm
concentration
level.
Besides
the
daily
calibration
check
procedures,
PES
gave
EMI
an
independent
audit
gas.
The
identity
of
the
compounds
contained
in
the
audit
gas
and
their
concentrations
were
not
revealed
to
EMI.
Analysis
of
this
audit
gas
was
conducted
using
both
GCMS
measurement
systems.
Table
6.5
presents
the
results
from
the
daily
system
continuing
calibrations
and
the
audit.

Additional
QA
procedures
conducted
during
this
testing
program
included
analyte
spiking.
Analyte
spiking
consists
of
adding
an
exact
amount
of
calibration
standard
to
the
effluent
stream
at
a
point
upstream
of
the
primary
particulate
matter
filter
within
the
sampling
system.
This
procedure
checks
the
ability
of
both
the
sampling
and
analytical
system
to
transport
and
quantify
effluent
samples.
Analyte
spiking
procedures
were
conducted
on
each
day
of
the
test
program
at
varying
concentration
levels.
Concentrations
of
100
ppb,
500
ppb,
and
1
ppm
were
used
for
the
spiking.
Spike
recoveries
of
between
79%
and
126%
were
achieved
at
the
100
ppb
concentration
level
for
the
target
analytes
detected
using
the
inlet
GCMS
measurement
system.
EM1
achieved
spike
recoveries
of
between
74%
and
136%
at
the
100
ppb
concentration
level,
64%
to
8
1%
at
the
500
ppb
concentration
level,
and
100%
­
105%
at
the
1
ppm
level,
for
the
target
analytes
detected
using
the
outlet
GCMS
measurement
system.

6.3.1
GCMS
Detection
Limits
Table
6.6
presents
the
GCMS
Detection
Limits
at
the
pre­
catalyst
and
the
post­
catalyst
sampling
locations.
PES
used
the
analytical
detection
limits
supplied
by
EMI
to
calculate
the
run­
by­
run
mass
detection
limits.

Final
Report
Cooper­
Bessemer
GMVhTF
6­
15
July
2000
TABLE
6.5
SUMMARY
OF
GCMS
CONTINUING
CALIBRATIONS
AND
AUDIT
RESULTS
3/
31/
99
4/
l/
99
412199
Audit
1
Compound
Result
c­
w
Result
(%
I
Result
("/
I
Result
(%
I
(
mm)
Diff.
(
wm)
Diff.
@
pm)
Diff.
(
PPm)
Diff.

Catalyst
Inlet
1,3­
Butadiene
1.23
19.4
1.03
0
1.06
2.9
­

Hexane
1.02
­
0.97
0.82
­
20.4
1.03
0
­

Benzene
1.02
­
1.9
0.86
­
17.3
1.02
­
1.9
0.52
­
3.7
Toluene
0.78
­
22.8
0.8
­
20.8
1.01
0
0.50
­
5.7
Ethyl
Benzene
1.07
2.9
1.04
0
1.11
6.7
0.52
­
2.0
m/
p­
Xylene
2.22
7.8
2.09
1.56
2.19
6.3
­

Styrene
0.86
­
17.3
0.93
­
10.6
1.11
6.7
­

o­
Xylene
1.04
0.97
1.06
2.9
1.08
4.9
0.48
­
7.7
Catalyst
Outlet
m/
p­
Xylene
2.13
3.4
2.17
5.3
2.15
4.4
­

Styrene
1.04
0
1.03
­
0.96
0.81
­
22.1
­

o­
Xylene
1.07
3.9
1.1
6.8
1.12
8.7
.
0.43
­
17.3
'
The
audit
cylinder
contained
540
ppb
benzene,
530
ppb
toluene,
510
ppb
ethyl
benzene,
and
520
ppb
o­
xylene.
The
analytical
accuracy
for
each
component
was
reported
to
be
f
5%
by
the
manufacturer.

Final
Report
Cooper­
Bessemer
GMV4­
TF
6­
16
July
2000
TABLE
6.6
DETECTION
LEtfITS
OF
GCMS
COMPOUNDS
AT
CATALYST
INLET
RunZD
wmvd
1
,
bButadiene
pg/
bhp­
hr
vlb/
hr
rwmvd
Hexane
pglbhp­
hr
Mlb/
hr
wmvd
Benzene
pglbhp­
hr
plblhr
wmvd
Tduene
pg/
bhp­
hr
plb/
hr
pwfd
Ethyl
Benzene
pglbhp­
hr
plb/
hr
ppmvd
m/
p­
Xylene
pglbhp­
hr
plb/
hr
wmvd
Styrene
pglbhp­
hr
plb/
hr
wmvd
o­
Xylene
uglbhp­
hr
vlb/
hr
0.50
0.50
0.50
0.50
0.50
0.50
0.50
0.50
8000
12000
12000
8000
9000
8000
10000
8000
8000
8000
7000
7000
9000
8000
8000
8000
0.09
0.11
0.10
o.
qs
0.10
0.09
0.10
0.09
0.10
0.09
0.09
0.09
0.09
0.11
0.09
0.09
0.09
0.09
2000
5000
3000
2000
3000
2000
4000
2000
3000
2000
2000
2000
2000
3000
2000
2000
2000
2000
2000
3000
2000
2000
3000
2000
3000
2000
3000
2000
2000
2000
2000
3000
2000
2000
2000
2000
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
200
300
300
200
200
200
200
200
200
200
200
200
200
200
200
200
200
200
200
200
200
200
200
200
200
200
200
200
200
200
200
200
200
200
200
200
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
500
800
800
600
600
500
600
500
600
600
600
500
500
600
600
600
600
600
500
500
500
500
600
500
500
500
600
500
500
500
500.
600
600
500
500
500
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
'
600
900
1000
600
'
700
600
700
600
600
700
700
600
600
700
600
600
700
700
600
600
600
500
700
600
600
600
600
600
600
600
600
700
600
500
600
600
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
6.01
0.01
0.01
0.01
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
600
900
1000
600
700
600
700
600
600
700
700
600
600
600
600
600
700
700
600
600
600
500
700
600
600
600
600
600
600
600
600
600
600
500
600
600
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
600
900
1000
600
700
600
700
600
600
700
700
600
600
700
600
600
700
700
600
600
600
500
700
600
600
600
600
600
600
600
600
700
600
500
600
600
Final
Report
Cooper­
Bessemer
GMV­
4­
TF
6­
17
July
2000
TABLE
6.7
DETECTION
LIMITS
OF
GCMS
COMPOUNDS
AT
CATALYST
OUTLET
Run
ID
wmvd
1,3­
Butadiene
lglbhp­
hr
plb/
hr
wmvd
Hexane
pglbhphr
plb/
hr
wmvd
Benzene
pglbhphr
plblhr
ppmvd
Toluene
pg/
bhphr
@
b/
hr
wmvd
Ethyl
Benzene
pg/
bhphr
plb/
hr
wmvd
m/
p­
Xylene
pglbhphr
plblhr
ppmvd
Styrene
pglbhphr
plblhr
ppmvd
o­
Xylene
pglbhp­
hr
plb/
hr
Final
Report
Cooper­
Bessemer
GMVd­
TF
6­
18
July
2000
6.4
CARB
429
QA/
QC
PROCEDURES
The
following
text
describes
the
QA/
QC
procedures
employed
by
PES
and
ERG
during
the
PAH
sampling
and
analysis.

6.4.1
Calibration
of
CARB
429
Samplin?
Apparatus
Because
no
mechanism
exists
for
an
independent
measurement
of
emissions
from
the
source,
careful
preparation,
checkout,
and
calibration
of
the
sampling
and
analysis
equipment
is
essential
to
ensure
collection
of
high
quality
data.
PES
maintains
a
comprehensive
schedule
for
preventive
maintenance,
calibration,
and
preparation
of
the
source
testing
equipment.

6.4.1.1
Barometers.
PES
used
aneroid
barometers
calibrated
against
a
station
pressure
value
reported
by
a
nearby
National
Weather
Service
Station
and
corrected
for
elevation.

6.4.1.2
Temperature
Sensors.
The
responses
of
the
Type
K
thermocouples
used
in
the
field
testing
program
were
checked
using
Calibration
Procedure
2e
as
described
in
the
Quality
Assurance
Handbook.
The
response
of
each
temperature
sensor
was
recorded
when
immersed
in
an
ice
water
bath,
at
ambient
temperature,
and
in
a
boiling
water
bath;
each
response
was
checked
against
an
ASTM
3F
reference
thermometer.
Table
6.8
summarizes
the
results
of
the
thermocouple
checks
and
the
acceptable
levels
of
variance.
Digital
temperature
readouts
were
checked
for
calibration
using
a
thermocouple
simulator
having
a
range
of
O­
2400
OF.

6.4.1.3
Pitot
PES
used
Type
S
Pitot
tubes
or
Standard
Pitot
tubes
constructed
according
to
EPA
Method
2
specifications.
Type
S
Pitot
tubes
were
calibrated
against
the
dimensional
criteria
described
in
Method
2
using
Calibration
Procedure
2a
as
described
in
the
Quality
Assurance
Handbook,
Volume
III,
1994.
Type
S
Pitot
tubes
meeting
these
criteria
are
assigned
a
pitot
coefficient
(
C,)
of
0.84.
Standard
Pitot
tubes
were
checked
for
dimensional
criteria
using
Calibration
Procedure
2b
as
described
in
the
Quality
Assurance
Handbook,
Volume
III,
1994.
Standard
Pitot
tubes
meeting
these
criteria
were
assigned
a
pitot
coefficient
(
C,)
of
0.99.

6.4.1.4
Differential
Pressure
Gauses.
PES
used
Dwyer
inclined/
vertical
manometers
to
measure
differential
pressures
including:
velocity
pressure,
static
pressure,
and
orifice
meter
pressure.
PES
chose
manometers
having
sufficient
sensitivity
to
accurately
measure
pressures
over
the
entire
range
of
expected
values.
Manometers
are
primary
standards
and
require
no
calibration.

Final
Report
Cooper­
Bessemer
GMV­
4­
TF
6­
19
July
2000
TABLE
6.8
CARB
429
SAMPLE
TRAIN
SUMMARY
OF
TEMPERATURE
SENSOR
CALIBRATION
DATA
Temp.
Sensor
I.
D.

RT­
14
RT­
15
Usage
Stack
Gas
Stack
Gas
Temperature,
OR
Reference
Sensor
492
492
532
529
670
671
492
493
530
530
670
670
Absolute
EPA
Difference
Criteria
%
%

0
e1.5
0.56
­=
I.
5
0.15
­
1.5
0.20
e1.5
0
e1.5
0
­
1.5
Dry
Gas
Meter
493
495
0.41
e1.5
RMB­
15
Inlet
534
534
0
­
1.5
668
670
0.30
­
1.5
Dry
Gas
Meter
493
493
0
­
1.5
RMB­
15
Outlet
534
535
0.19
­
1.5
668
668
0
e1.5
Dry
Gas
Meter
493
494
0.20
e1.5
MB­
10
Inlet
536
536
0
­
1.5
666
665
0.15
­
1.5
Dry
Gas
Meter
492
494
0.41
e1.5
MB­
IO
Outlet
536
537
0.19
e1.5
666
665
0.15
­
1.5
492
492
0
e1.5
SH­
1
Impinger
Exit
536
536
0
e1.5
668
668
0
e1.5
492
493
0.20
e1.5
SH­
5
Impinger
Exit
531
531
0
­
1.5
667
667
0
­
1.5
Final
Report
Cooper­
Bessemer
GMVATF
6­
20
July
2000
6.4.1.5
Drv
Gas
Meter
and
Orifice.
The
CARB
Method
429
dry
gas
meters
and
orifices
were
calibrated
according
to
Calibration
Procedure
5
in
the
Quality
Assurance
Handbook.
This
procedure
requires
direct
comparison
of
the
dry
gas
meter
to
a
reference
dry
test
meter.
PES
calibrates
its
reference
dry
test
meter
annually
against
a
wet
test
meter.
Before
its
initial
use
in
the
field,
the
metering
system
was
calibrated
at
several
flow
rates
over
the
normal
operating
range
of
the
metering
system.
Individual
meter
calibration
factors
(
y)
cannot
differ
from
the
average
by
more
than
0.02,
and
the
results
of
individual
meter
orifice
factors
(
AH@)
cannot
differ
from
the
average
by
more
than
0.20.
After
field
use,
the
metering
system
calibration
was
checked
at
the
average
flow
rate
and
highest
vacuum
observed
during
the
test
period.
The
results
of
the
post­
test
meter
correction
factor
check
cannot
differ
by
more
than
5%
from
the
average
meter
correction
factor
obtained
during
the
initial,
or
thereafter,
the
annual
calibration.
Table
6.9
presents
the
results
of
the
dry
gas
meter
and
orifice
calibrations.
All
dry
gas
meters
and
orifices
used
in
this
test
program
met
the
method
calibration
requirements.

TABLE
6.9
CARB
429
SAMPLE
TRAIN
SUMmY
OF
DRY
GAS
METER
AND
ORIFICE
CALIBRATION
DATA
Meter
Box
No.
Dry
Gas
Meter
Correction
Factor
(
y)
Meter
Orifice
Coefficient
(
AH&

Pre­
test
Post­
test
%
Diff.
EPA
Criteria
Average
Range
EPA
Criteria
MB­
10
1.015
1.013
­
0.27
<
5%
1.84
1.82
­
1.92
1.64­
2.04
RMB­
15
1.001
0.998
­
0.29
<
5%
1.87
1.79
­
1.98
1.67
­
2.07
6.4.2
ReaPents
and
Glassware
Preparation
Before
field
testing,
PES
pre­
cleaned
all
sample
train
glassware
following
the
procedures
in
CARB
Method
429.
Specifically,
the
glassware
was
cleaned
according
to
the
following
protocol.

1.
2.
3.
4.
5.
Wash
in
hot
soapy
water
with
Alconox.
Rinse
three
times
with
tap
water.
Rinse
three
times
with
reagent
(
i.
e.,
deionized)
water.
Soak
in
10%
(
v/
v)
nitric
acid
(
I­
NO,)
solution
for
a
minimum
of
4
hours.
Rinse
three
times
each
with
pesticide­
grade
acetone,
hexane,
and
methylene
chloride,
and
allow
to
air
dry.

Final
Report
Cooper­
Bessemer
GMV­
CTF
6­
21
July
2000
After
preparation
of
the
glassware,
the
openings
were
sealed
with
Teflon
tape
to
prevent
contamination,
and
the
glassware
wrapped
and
packed
for
transport
to
the
EECL.
ERG
prepared
the
W­
2@
sorbent
resin
traps.
ERG
then
pre­
spiked
the
traps
with
surrogates
and
capped
them
with
glass
balls
and
sockets.
Impinger
water
used
was
organic­
free,
reagent
grade.
Pesticide­
grade
acetone,
hexane,
and
methylene
chloride
were
used
as
recovery
solvents.

6.4.3
On­
site
Measurements
The
on­
site
QA/
QC
activities
included:

6.4.3.1
Measurement
Sites.
Before
sampling,
PES
checked
the
dimensions
of
the
exhaust
duct
to
assure
that
the
port
locations
complied
with
Method
1
criteria.
PES
confirmed
the
distances
to
upstream
and
downstream
disturbances
and
test
port
locations.
PES
also
measured
inside
stack
dimensions
through
perpendicular
ports
to
assure
uniformity
of
the
stack
cross
sectional
area.
PES
measured
the
inside
stack
dimensions,
stack
wall
thickness,
and
sample
port
lengths
to
the
nearest
0.1
inch.

6.4.3.2
Velocitv
Measurements.
PES
assembled,
leveled,
zeroed,
and
leak­
checked
all
velocity
measurement
apparatus
before
and
after
each
sampling
run.
The
stack
static
pressure
was
determined
at
a
single
point.
PES
selected
a
point
of
average
velocity
pressure
found
during
the
pre­
test
velocity
traverse.

6.4.3.3
Moisture.
During
sampling,
the
exit
gas
temperature
of
the
last
impinger
in
each
sampling
train
was
maintained
below
68
°
F
to
ensure
condensation
of
stack
gas
water
vapor.
The
moisture
gain
in
the
impinger
train
due
to
flue
gas
moisture
was
determined
gravimetrically
using
a
digital
top­
loading
electronic
balance
with
a
resolution
of
0.1
g.

6.4.4
Analvtical
Oualitv
Assurance
PES
and
ERG
personnel
employed
several
methods
to
ensure
the
quality
of
the
PAH
analytical
data.
These
methods
included
analysis
of
reagent
blanks,
a
laboratory
method
blank,
and
field
blanks.
In
addition,
the
XAD­
2
sorbent
traps
were
spiked
with
isotopically
labeled
internal
standards.
The
recovery
efficiency
of
the
internal
standards
is
used
to
evaluate
method
performance.
The
results
of
these
QA
checks
are
discussed
in
the
following
paragraphs.

6.4.4.1
Blank
Analvses.
During
the
field
testing,
PES
personnel
collected
blanks
of
the
CARB
429
sampling
train
reagents
to
quantify
contamination
levels.
Field
blank
trains
were
assembled,
transported
to
each
sampling
site,
and
leak
checked.
The
field
blank
trains
were
then
returned
to
the
PES
field
laboratory,
where
were
recovered
in
the
same
manner
as
the
trains
used
for
sampling.
The
field
blank
train
impingers
and
connecting
glassware
were
the
same
components
used
during
actual
sampling.
Since
the
sampling
glassware
is
cleaned
after
Final
Report
Cooper­
Bissemer
GMV­
CTF
6­
22
July
2000
each
run
and
reused,
analysis
of
field
blank
trains
is
used
to
find
out
if
poor
cleanup
technique
caused
cross­
contamination
between
sampling
runs.
Per
CARB
Method
429,
PES
did
not
correct
any
of
the
PAH
results
for
blank
results.
The
results
of
the
reagent
and
field
blank
analyses
are
presented
in
Table
6.10.
The
levels
of
any
unlabeled
analyte
quantified
in
the
blank
train
must
not
exceed
20
percent
of
the
level
of
that
analyte
in
the
sampling
train.

At
the
inlet
location,
naphthalene
was
present
in
the
blank
train
at
a
magnitude
approximately
30%
of
naphthalene
in
samples
collected
during
runs
PAH
.2
and
PAH
3.
Naphthalene
was
present
in
all
samples
and
the
XAD
laboratory
blank
and
the
field
blanks.
The
presence
of
naphthalene
is
due
to
the
ubiquitous
nature
of
this
compound.
For
chrysene,
the
inlet
field
blank
result
was
2
1.7
%
of
the
mass
of
chrysene
in
the
sample
for
run
PAH
1.
In
the
remaining
cases
at
the
inlet,
the
blank
levels
were
less
than
20%
of
the
levels
in
the
samples.

At
the
outlet,
naphthalene
was
also
present
at
levels
that
exceeded
the
acceptable
level.
For
runs
PAH
1,
PAH
2,
and
PAH
3
at
the
outlet,
the
blank
levels
were
38.9
%,
42.5
%,
and
42.4
%
of
the
levels
in
the
sample
trains.
The
only
other
compound
detected
in
the
outlet
blank
train
was
phenanthrene,
which
was
present
at
levels
well
below
20%
of
the
levels
in
the
sample
trains.

6.4.4.2
Internal
Standard
Recoveries.
Table
6.11
presents
the
recovery
efficiencies
of
isotopically
labeled
surrogate
compounds.
Recovery
efficiency
gives
a
measure
of
the
capture
efficiency
and
the
efficiency
of
the
solvent
extraction
for
specific
compounds.
Recoveries
for
each
of
the
internal
standards
must
be
greater
than
50
percent
and
less
than
150
percent
of
the
known
value.
This
criterion
is
used
to
assess
method
performance.
Because
this
is
an
isotope
dilution
technique,
it
should
be
independent
of
internal
standard
recovery.
Lower
recoveries
do
not
necessarily
invalidate
the
analytical
results
for
PAH,
but
they
may
result
in
higher
detection
limits.

Final
Report
Cooper­
Bessemer
GMV­
4­
TF
6­
23
July
2000
TABLE
6.10
SUMMARY
OF
CARB
429
BLANK
RESULTS
Compound
Laboratory
Reagent
Blank
Inlet
Field
Outlet
Blank
Blank
Field
Blank
Result
(
pg)
Result
(
P!
z)'
Result
@
g)
Result
(
pg)

I
I
I
I
Naphthalene
I
0.378
1
0.018
1
3.352
1
3.425
Acenaphthylene
I
I
ND
1
ND
1
ND
Acenaphthene
I
ND
I
ND
I
ml
ND
Fluorene
I
ND
1
NDI
N­
d
ND
Phenanthrene
I
ND
1
0.049
1
0.054
1
0.043
Anthracene
I
I
I
I
I
I
0.008
1
ND
1
ND
Fluoranthene
I
ND
I
0.006
1
ND
I
ND
P
yrene
I
ND
I
ND
I
0.029
1
ND
Benzo(
a)
anthracene
1
ND
I
ND
I
ND
I
ND
Chrysene
I
ND
I
ND
I
0.046
1
Benzo(
b)
fluoranthene
1
ml
ND
I
ND
I
ND
I
ND
I
ND
1
ND
I
I
ND
1
ND
1
ND
1
ND
Benzo(
k)
fluoranthene
I
Benzo(
a)
pyrene
Indeno(
1,2,3­
cd)
pyrene
I
Dibenz(
a,
h)
anthracene
I
Benzo(
g,
h,
i)
perylene
I
ND
1
ND
1
ND
1
ND
ND
I
ND
I
ND
I
ND
ND
I
ND
I
ND
I
ND
'
The
reagent
blank
value
is
the
sum
of
separate
analyses
hexane,
acetone,
methylene
chloride,
and
distilled
water
blank
samples.

Final
Report
Cooper­
Bessemer
GIvIV­
I­
TF
6­
24
July
2000
TABLE
6.11
iXJMMAFiY
OF
CARB
429
SURROGATE
RECOVERIES
Surrogate
Compound
Naphthalene­
d8
Acenaphthylene­
d8
Acenaphthene­
dl
0
Fluorene­
dl0
Phenanthrene­
d
10
Anthracene­
d
10
Fluoranthene­
dl0
Pyrene­
dl0
Benzo(
a)
anthracene­
d12
Chrysene­
d12
Benzo(
b)
fluoranthene­
d12
Benzo(
k)
fluoranthene­
d12
Benzo(
a)
pyrene­
dl2
[
ndeno(
1,2,3­
cd)
pyrene­
d12
Dibenz(
a,
h)
anthracene­
d14
3enzo(
g,
h,
i)
perylene­
d12
Lab
Blank
(%
I
86
45
64
80
73
1
73
39
10
68
74
73
ND
32
34
ND
Field
Blanks
PAH
Run
1
Inlet
Outlet
Inlet
Outlet
(%
I
(%
I
(%
I
VW
74
81
83
106
48
24
51
69
70
74
70
97
74
76
85
92
70
71
80
73
10
11
29
30
76
73
94
63
71
61
91
60
68
46
107
73
85
76
144
63
109
87
98
96
82
76
86
80
ND
ND
9
29
77
48
58
97
82
57
63
107
53
7
47
98
PAH
Run
2
Inlet
Outlet
W)
(%
I
96
111
58
69
83
93
97
99
116
69
30
34
137
67
133
65
111
61
125
7.1
143
0
127
0
13
32
70
49
80
54
57
46
PAH
Run
3
Inlet
Outlet
w
("
x3)

89
117
58
63
72
86
91
100
88
88
34
50.

101
109
100
107
110
113
113
115'

113
125
108
107
19
47
61
71
70
81
54
70
Final
Report
Cooper­
Bessemer
GMV­
4­
TF
6­
25
July
2000
6.4.5
CARB
429
Detection
Limits
Tables
6.12
and
6.13
present
the
in­
stack
detection
limits
of
each
PAH
compound
before
the
catalyst
and
after
the
catalyst.
The
volumes
of
the
CARB
429
samples
averaged
2.12
dry
standard
cubic
meters
(
dscm)
before
the
catalyst
and
2.32
dscm
after
the
catalyst.
The
expected
sample
volume
defined
in
the
QAPP
was
2.5
dscm,
and
the
in­
stack
detection
limits
on
the
PAH
compounds
were
based
on
this
volume.
Because
the
actual
sample
volumes
were
less
than
the
anticipated
volumes
by
approximately
15%
the
in­
stack
detection
limits
for
the
PAHs
are
approximately
15%
higher
than
those
presented
in
the
QAPP.

6.5
CORREXTIVE
ACTIONS
During
the
field
testing,
PES
and
EPA
made
several
changes
to
the
QAPP
describing
the
field
testing.
Field
and
engine
operating
conditions
mandated
these
changes.
These
changes
are
presented
in
Table
6.14.
'

Final
Report
Cooper­
Bessemer
GMV­
4­
TF
6­
26
July
2000
TABLE
6.12
DETECTION
LIMITS,
OF
PAH
COMPOUNDS
AT
CATALYST
INLET
Run
ID
PAH
1
PAH
2
Date
412199
M/
99
Time
12041404
1626­
1825
a
Acenaphthene
ugibhp­
hr
b
1.9
1.8
@
b/
hour
1.6
1.5
Acenaphthylene
uglbhp­
hr
1.9
1.8
@
b/
hour
1.6
1.5
Anthracene
yglbhp­
hr
1.9
1.8
@
b/
hour
1.6
1.5
Benzo(
a)
anthracene
ug/
bhp­
hr
1.9
1.8
@
b/
hour
1.6
1.5
Benzo(
b)
fluoranthene
uglbhp­
hr
1.9
1.8
ulb/
hour
1.6
1.5
Benzo(
k)
fluoranthene
uglbhphr
1.9
1.8
ulblhour
1.6
1.5
Benzo(
g,
h,
i)
perylene
pg/
bhp­
hr
3.7
3.5
@
b/
hour
3.1
2.9
Benzo(
a)
pyrene
ug/
bhp­
hr
1.9
1.8
ulb/
hour
1.6
1.5
Chrysene
uglbhphr
1.9
1.8
ulblhour
1.6
1.5
Dibenz(
a,
h)
anthracene
uglbhp­
hr
3.7
3.5
@
b/
hour
3.1
2.9
Fluoranthene
uglbhp­
hr
1.9
1.8
uib/
hour
1.6
1.5
Fluorene
uglbhphr
1.9
1.8
ul
b/
hour
1.6
1.5
Indeno(
l,
2,3cd)
pyrene
uglbhp­
hr
3.7
3.5
ulb/
hour
3.1
2.9
Naphthalene
uglbhphr
1.9
1.8
ulblhour
1.6
1.5
Phenanthrene
uglbhp­
hr
1.9
1.8
ulb/
hour
1.6
1.5
.
Pyrene
ug/
bhp­
hr
1.9
1.8
@
b/
hour
1.6
1.5
a
Micrograms
per
brake
horsepower
hour
b
Micropounds
per
hour
1.9
1.8
1.6
1.5
1.9
1.8
1.6
1.5
1.9
1.8
1.6
1.5
1.9
1.8
1.6
1.5
1.9
1.8
1.6
1.5
1.9
1.8
1.6
1.5
3.8
3.7
3.1
3.1
1.9
1.8
1.6
1.5
1.9
1.8
1.6
1.5
3.8
3.7
3.1
3.1
1.9
1.8
1.6
1.5
1.9
1.8
1.6
1.5
3.8
3.7
3.1
3.1
1.9
1.8
1.6
1.5
1.9
1.8
1.6
1.5
1.9
1.8
1.6
1.5
Final
Report
Cooper­
Bessemer
GMV­
4­
TF
6­
27
July
2000
TABLE
6.13
DETECTION
LIMITS
OF
PAH
COMPUNDS
AT
CATALYST
OUTLET
Run
ID
PAH
1
PAH
2
PAH
3
Date
412199
412199
412199
Average
Time
i204­
1404
16251825
2000­
2200
vg/
bhp­
hr
b
a
Acenaphthene
1.7
1.6
1.7
1.7
plb/
hour
1.4
1.4
1.4
1.4
Acenaphthylene
pg/
bhp­
hr
1.7
1.6
1.7
1.7
@
b/
hour
1.4
1.4
1.4
1.4
Anthracene
pg/
bhp­
hr
1.7
1.6
1.7
1.7
@
b/
hour
1.4
1.4
1.4
1.4
Benzo(
a)
anthracene
pg/
bhp­
hr
1.7
1.6
1.7
1.7
plb/
hour
1.4
1.4
1.4
1.4
Benzo(
b)
fluoranthene
pg/
bhp­
hr
1.7
1.6
1.7
1.7
@
b/
hour
1.4
1.4
1.4
1.4
Benzo(
k)
fluoranthene
pg/
bhp­
hr
1.7
1.6
1.7
1.7
@
b/
hour
1.4
1.4
1.4
1.4
Benzo(
g,
h,
i)
perylene
pglbhp­
hr
3.5
3.3
3.5
3.4
plb/
hour
2.9
2.7
2.9
2.8
Benzo(
a)
pyrene
pglbhp­
hr
1.7
1.6
1.7
1.7
@
b/
hour
1.4
1.4
1.4
1.4
Chrysene
vg/
bhp­
hr
1.7
1.6
1.7
1.7
@
b/
hour
1.4
1.4
1.4
1.4
Dibenz(
a,
h)
anthracene
pg/
bhp­
hr
3.5
3.3
3.5
3.4
@
b/
hour
.
2.9
2.7
2.9
2.8
Fluoranthene
pg/
bhp­
hr
1.7
1.6
1.7
1.7
@
b/
hour
1.4
1.4
1.4
1.4
Fluorene
vg/
bhp­
hr
1.7
1.6
1.7
1.7
plWho4.
H
1.4
1.4
1.4
1.4
Indeno(
l,
2,3­
cd)
pyrene
pg/
bhp­
hr
3.5
3.3
3.5
3.4
vlblhour
2.9
2.7
2.9
2.8
Naphthalene
vglbhp­
hr
1.7
1.6
1.7
1.7
plb/
hour
1.4
1.4
1.4
1.4
Phenanthrene
pglbhp­
hr
1.7
1.6
1.7
1.7
@
b/
hour
1.4
1.4
1.4
1.4
Pyrene
pg/
bhp­
hr
1.7
1.6
1.7
1.7
@
b/
hour
1.4
1.4
1.4
1.4
a
Micrograms
per
brake
horsepower
hour
b
Micropounds
per
hour
Final
Report
Cooper­
Bessemer
GMV4­
TF
6­
28
July
2000
TABLE
6.14
SUMMARY
OF
CORRECTIVE
ACTIONS
Corrective
Action
No.
Date
Time
Problem
Corrective
Action
Planned
1
3129199
1200
CARB
429
Traverse
points
on
Minimum
number
of
points
(
8)
used
outlet
(
12)
more
than
minimum
required.

2
3/
29/
99
1600
5­
min
spiking
regimen
will
take
5­
min
spikes
changed
to
2­
min
spikes
too
much
time
3
3/
29/
99
1600
Conventional
CARB
429
sample
Heated
flexible
sample
line
inserted
train
will
be
impossible
to
use,
into
sample
train
split
after
heated
given
engine
exhaust
geometry
filter
box
4
3129199
1600
Separate
FTIRS
validation
for
Acetaldehyde
I
acrolein
cylinder
used
formaldehyde
and
actetaldehye
I
as
make­
up
air
for
formalin
solution
acrolein
will
take
too
much
time
5
3130199
­

­
6
3130199
­
On
baseline
tests,
engine
ignition
New
baseline
set­
point
will
be
0.4"
set
to
0.4'
BTDC
rather
than
10'
BTDC,
which
reflects
lean­
burn
specified
in
QAPP
modification
to
engine
Inlet
air
humidity
target
value
in
New
value
of
0.015
lb.
water
/
lb.
air
QAPP
(
0.0015
lb.
water
I
lb.
air)
used
incorrect
7
3130199
­
Cylinder
values
acetaldehyde
Value
of
30
ppm
acetaldehyde
will
be
standards
seem
incorrect
used
instead
of
100
ppm
on
cylinder
8
3130199
­
Ignition
timing
on
Run
1
does
not
New
timing
values
defined
as
1.8"
fat
agree
with
value
in
QA
PP
normal,
0.3"
for
advanced,
5.3"
for
retarded
9
3130199
­
Oil
pressure
during
QA
check
No
corrective
action
planned
since
oil
outside
tolerance
level
pressure
is
a
secondary
parameter
10
3131199
­
Ignition
timing
for
Run
5
For
subsequent
runs,
ignition
timing
changed
to
2.8'
BTDC
will
be
set
to
yield
a
cylinder
peak
pressure
at
18O
ATDC
11
3/
31/
99
­
Analyzer
drift
checks
between
Runs
13
and
14
take
too
much
time
Checks
between
Runs
13
and
14
will
be
dropped,
and
observed
drift
over
the
entire
period
will
be
applied
to
Run
13
and
Run
14
data
sets
12
313
1199
­
Impinger
on
outlet
moisture
train
Outlet
impinger
train
for
moisture
wil
broken.
No
spare
is
available
use
3
impingers
instead
of
one
Final
Report
Cooper­
Bessemer
GMV­
4­
TF
6­
29
July
2000
6.6
DATA
QUALITY
ASSESSMENT
EPA
used
the
Data
Quality
Objective
(
DQO)
Process
to
plan
the
test
program.
The
DQO
Process
consists
of
seven
distinct
steps.

1.
State
the
problem.
2.
Identify
the
decision.
3.
Define
inputs
to
the
decision.
4.
Define
the
study
boundaries.
5.
Develop
the
decision
rule.
6.
Specify
tolerable
limits
on
decision
errors
7.
Optimize
the
design
for
obtaining
data.

The
DQO
outputs
for
this
test
program
were
presented
in
the
Quality
Assurance
Project
Plan.
The
problem
was
defined
in
the
QAPP
and
is
restated
below.

EPA
believes
that
there
is
a
need
to
conduct
emission
tests
on
a
subset
of
engines
of
differing
designs
to
evaluate
the
following
issues:

l
the
effectiveness
of
after­
combustion
control
systems
on
HAP
emissions,
and
l
the
effectiveness
of
combustion
modifications
(
engine
operating
parameters)
on
HAP
emissions.

EPA
then
developed
a
decision
statement.
The
decision
statement
defined
the
process
that
would
be
used
to
answer
the
stated
problem.
The
decision
statement
is
restated
below:

If
EPA
can
identtfi
a
range
of
engine
operating
conditions
for
a
defined
set
of
engines
with
spectjied
after­
combustion
treatment
systems
and
a
list
ofpollutants
of
interest,
and
EPA
collects
data
to
determine
emissions
of
those
pollutants
for
each
engine
operated
at
each
engine
operating
condition,
then
EPA
can
make
a
determination
of
the
control
eflectiveness
of
after­
combustion
and
combustion
modijications.
In
addition,
EPA
can
obtain
information
on
HAP
emissions
throughout
the
engine
operating
range.

PES,
EECL,
and
EMI
conducted
the
test
program
on
the
Cooper­
Bessemer
GMV­
4­
TF,
natural
gas­
fired,
2­
stroke,
lean­
burn,
reciprocating
internal
combustion
engine.
The
MiraTech
oxidation
catalyst
was
designed
to
provide
the
information
required
by
the
decision
statement.
Based
upon
the
inputs,
EPA
will
make
decisions
that
will
be
used
to
regulate
this
engine
subcategory.
Inputs
to
the
decision
were
defined,
agreed
to,
and
documented
in
the
QAPP.
These
inputs
consisted
of
agreement
on
a
finite
list
of
engines
to
test,
the
after­
combustion
control
systems
to
test,
the
range
of
engine
operating
conditions,
the
catalyst
conditioning
process,
the
target
list
of
pollutants,
and
the
sampling
and
analysis
methods,
and
sample
durations.

Final
Report
Cooper­
Bessemer
GMV­
4­
TF
6­
30
July
2000
During
conduct
of
the
test
program,
there
were
deviations
from
the
QAPP.
These
deviations
were
presented
in
the
previous
sub­
section,
as
well
as
in
Table
6.14.
Additional
deviations
to
the
QAPP
are
discussed
in
Section
3.0
for
deviations
in
engine
operation,
and
Section
5.0
for
deviations
in
Sampling
and
Analysis
procedures.

Table
6.15
presents
a
summary
of
engine
and
sample
method
performance
compared
to
the
QAPP
requirements.
Outlier
and
data
validation
issues
have
been
discussed
in
previous
sections.
Based
upon
the
engine
and
method
performance,
the
data
quality
is
evaluated
on
a
run­
by­
run
basis
for
suitability
in
the
assessment
of
pollutant
emissions
and
'
destruction
efficiency
of
HAPS
by
the
catalyst.

Six
engine
parameters
were
varied
over
the
course
of
the
test
program,
The
parameters
were
changed
so
that
emissions
data
and
HAP
destruction
efficiency
could
be
evaluated
at
a
range
of
engine
operating
conditions.
These
conditions
are
expected
to
simulate
the
range
of
engine
operating
conditions
in
industry.
Table
6.15
identifies
the
number
of
engine
parameters
that
were
within
the
tolerances
proscribed
in
the
QAPP.
The
target
engine
operating
conditions
were
estimates
based
upon
manufacturer's
recommendations.
There
are
differences
between
these
recommendations
and
the
nominal
engine
operating
parameters
of
the
GMV­
4­
TF
engine
located
at
the
EECL.
When
testing
was
conducted
some
of
the
proscribed
engine
parameters
could
not
be
met.
The
fact
that
a
pre­
set
engine
parameter
could
not
be
met
is
considered
to
be
minor.
The
testing
was
conducted
over
a
range
of
engine
operating
conditions,
and
these
operating
conditions
are
documented.

The
remainder
of
the
table
assesses
data
quality
using
a
three­
tiered
system.
A
(
J
+)
indicates
that
all
method
performance
parameters
defined
in
the
QAPP
and/
or
the
sampling
method
were
met.
A
(
J)
indicates
that
at
least
90
%
of
the
method
performance
parameters
were
met.
In
the
case
of
FTIRS
and
CEMS
detection
limits,
there
were
no
detection
limits
specified
in
the
QAPP..
The
calculated
detection
limits
are
reasonable
for
this
test
program.

A
(
J
­)
indicates
that
fewer
than
90
%
of
the
method
performance
parameters
were
met.
This
was
the
case
in
the
QA/
QC
requirement
for
GCMS
at
the
catalyst
inlet,
and
for
the
PAH
sampling
runs
at
the
catalyst
inlet
and
outlet
locations.
At
the
catalyst
inlet,
the
results
of
the
GCMS
continuing
calibrations
for
toluene
were
slightly
outside
of
the
requirement
of
f
20
%
for
three
of
the
four
days
of
testing.
The
continuing
calibration
for
hexane
was
slightly
outside
of
the
limit
on
one
day.
For
the
PAH
testing
the
isokinetic
sampling
ratios
were
not
met.
Because
the
sampling
ratios
were
low,
the
volume
collected
during
each
PAH
sampling
run
was
also
low,
which
resulted
in
PAH
in­
stack
detection
limits
approximately
15%
higher
than
those
proscribed
in
the
QAPP.

Final
Report
Cooper­
Bessemer
GMY­
bTF
6­
3
1
July
2000
=:

\
­

\

\
­

\

\

\

\
­

\
­

\
­

\

\
­

\
­

\
­

\
­

\

\
­

\
­

\
­
8
5
.­
0
ti
E
.­
5
3
I
a
a
9
=
I
I
I
I
I
,
APPENDIX
A
SUBCONTRACTOR
TEST
REPORT
COLORADO
STATE
UNIVERSITY
ENGINES
AND
ENERGY
CONVERSION
LABORATORY
"
EMISSIONS
TESTING
OF
CONTROL
DEVICES
FOR
RECIPROCATING
INTERNAL
COMBUSTION
ENGINES
IN
SUPPORT
OF
REGULATORY
DEVELOPMENT
BY
THE
U.
S.
ENVIRONMENTAL
PROTECTION
AGENCY
(
EPA)
PHASE
1:
TWO­
STROKE,
LEAN
BURN,
NATURAL
GAS
FIRED
INTERNAL
COMBUSTION
ENGINES"
CoLomDo
STATE
UNIVERSITY
EMISSIONS
TESTING
OF
CONTROL
DEVICES
FOR
RECIPROCATING
INTERNAL
COMBUSTION
ENGINES
IN
SUPPORT
OF
REGULATORY
DEVELOPMENT
BY
THE
U.
S.
ENVIRONMENTAL
PROTECTION
AGENCY
(
EPA)

PHASE
1:
TWO­
STROKE,
LEAN
BURN,
NATURAL
GAS
FIRED
INTERNAL'COMBUSTION
ENGINES
Prepared
for:

PACIFIC
ENVIRONMENTAL
SERVICES
Submitted
by:

Engines
&
Energy
Conversion
Laboratory
Colorado
State
University
Mechanical
Engineering
Department
MAY.
l&
l999
Stclrement
ofConfidentiality
This
repori
has
been
submittedfor
the
sole
and
exclusive
use
of
Pacific
Environmental
Services,
and
shaN
not
be
disclosed
or
provided
to
any
other
entity,
corporation,
or
third
partfor
pwposes
beyond
the
spec~
fc
scope
or
intent
of
this
document
without
the
express
written
consent
of
Colorado
State
University.
.)
I
>
