United
States
Office
of
Air
Quality
Environmental
Protection
Planning
and
Standards
Agency
Research
Triangle
Park,
NC
27711
EPA­
454/
R­
00­
037
September
2001
AIR
Final
Report
­
Volume
I
of
I
Testing
of
a
4­
Stroke
Lean
Burn
Gas­
fired
Reciprocating
Internal
Combustion
Engine
to
Determine
the
Effectiveness
of
an
Oxidation
Reduction
Catalyst
System
for
Reduction
of
Hazardous
Air
Pollutant
Emissions
__
­­_
._..
____
_.
_"^­__
I.
I­_"~­
__
.­
­
­_.
__­­_
­­.­­_­­..~.­­
­­
­­­­­­
__
_­­_­­
­­­­­
ll_____
l_­­._
­..­
­
_­­_­~
FINAL
REPORT
TESTING
OF
A
4dTROKE
LEAN
BURN
GAS­
FIRED
RECIPROCATING
INTERNAL
COMBUSTION
ENGINE
TO
DETERMINE
THE
EFFECTIVENESS
OF
AN
OXIDATION
CATALYST
SYSTEM
FOR
REDUCTION
OF
HAZARDOUS
AIR
POLLUTANT
EMISSIONS
Prepared
for:

Terry
Harrison
(
MD­
19)
Work
Assignment
Manager
SMTG,
EMC,
EMAD,
OAQPS
U.
S
.
Environmental
Protection
Agency
Research
Triangle
Park,
NC
277
11
September
2001
Submitted
by:

PACIFIC
ENVIRONMENTAL
SERVICES,
INC.
5001
S.
Miami
Blvd.,
Suite
300
Research
Triangle
Park,
NC
27709­
2077
(
919)
941­
0333
FAX
(
919)
941­
0234
DISCLAIMER
Pacific
Environmental
Services,
Inc.
(
PES)
prepared
this
document
under
EPA
Contract
No.
68­
D­
01­
003,
Work
Assignment
No.
l­
04.
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.

ii
1.0
2.0
3.0
4.0
5.0
TABLE
OF
CONTENTS
INTRODUCTION..
...................................................
l­
l
SUMMARY
OF
RESULTS
.............................................
2­
l
2.1
EMISSIONS
TEST
LOG
..........................................
2­
l
2.2
ENGINE
PARAMETERS
AND
EXHAUST
GAS
FLOW
RATES
.........
2­
3
2.3
FTIRS
AND
CEM
MEASUREMENTS
..............................
2­
3
2.4
DESTRUCTION
OF
HAP
BY
THE
CATALYST
......................
2­
8
SOURCE
DESCRIPTION
AND
OPERATION
...............................
3­
l
3.1
ENGINE
DESCRIPTION
.........................................
3­
1
3.2
ENGINE
OPERATION
DURING
TESTING
..........................
3­
4
SAMPLING
LOCATIONS
..............................................
4­
l
SAMPLING
AND
ANALYSIS
METHODS.
................................
5­
l
5.1
5.2
5.3
5.4
5.5
5.6
5.7
`
5.8
5.9
DETERMINATION
OF
STACK
GAS
VOLUMETRIC
FLOW
RATE
......
5­
l
DETERMINATION
OF
STACK
GAS
OXYGEN
AND
CARBON
DIOXIDE
CONTENT
.....................................................
5­
3
DETERMINATION
OF
STACK
GAS
MOISTURE
CONTENT
..........
5­
3
DETERMINATION
OF
NITROGEN
OXIDES
........................
5­
3
DETERMINATION
OF
CARBON
MONOXIDE
......................
5­
5
DETERMINATION
OF
TOTAL
HYDROCARBONS
..................
5­
5
DETERMINATION
OF
METHANE
AND
NON­
METHANE
HYDROCARBONS
.............................................
5­
6
DETERMINATION
OF
GASEOUS
ORGANIC
HAPS
USING
FTIRS
.....
5­
6
DETERMINATION
OF
NATURAL
GAS
COMPOSITION
..............
5­
7
.
.
.
111
TABLE
OF
CONTENTS
(
Concluded)

Page
6.0
QUALITY
ASSURANCE/
QUALITY
CONTROL
PROCEDURES
~
RES~
TS.......................................................
6­
1
6.1
FTIRS
QA/
QC
PROCEDURES
....................................
6­
l
6.2
CEMS
QA/
QC
PROCEDURES
....................................
6­
5
6.3
DATA
QUALITY
ASSESSMENT
.................................
6­
12
APPENDIX
A
SUBCONTRWTOR
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
2:
FOUR­
STROKE,
LEAN
BURN,
NATURAL
GAS
FIRED
INTERNAL
COMBUSTION
ENGINES"

APPENDIXB
EXAMPLE
CALCULATIONS
RELATED
CORRESPONDENCE
iv
LIST
OF
TABLES
Page
Table
2.1
Emissions
Test
Log
..............................................
2­
2
Table
2.2
Summary
of
Exhaust
Gas
Flow
Rates
................................
2­
4
Table
2.3
Emission
Rates
of
Detected
FTIRS
and
CEMS
Compounds
..............
2­
6
Table
2.4
Mass
Flow
Scenarios
.............................................
2­
9
Table
2.5
Catalyst
HAP
Removal
Efficiencies
................................
2­
10
Table
3.1
Engine
and
Catalyst
Specifications
..................................
3­
2
Table
3.2
Summary
of
Nominal
Engine
Parameters
.............................
3­
3
Table
3.3
Target
Engine
Operating
Conditions
During
Testing
....................
3­
5
Table
3.4
Summary
of
Engine
Parameters
­
Waukesha
3521GL.
...................
3­
6
Table
3.5
Summary
of
Engine
Parameters
During
Baseline
Runs.
..................
3­
8
Table
5.1
Summary
of
Sampling
and
Analysis
Methods
..........................
5­
2
Table
5.2
FTIRS
Analyzer
Specifications
.....................................
5­
7
Table
6.1
Detection
Limits
of
FTIRS
and
CEMS
Compounds
.....................
6­
6
Table
6.2
Types
and
Frequencies
of
CEMS
Analyzer
Calibrations
.................
6­
8
Table
6.3
Summary
of
Fuel
Factor
Values
...................................
6­
l
1
Table
6.4
Summary
of
CEMS
Analytical
Detection
Limits
......................
6­
12
Table
6.5
Summary
of
Engine
and
Method
Performance
........................
6­
15
V
LIST
OF
FIGURES
Figure
1.1
Test
Program
Organization
and
Major
Lines
of
Communication
.
.
.
.
.
.
.
.
.
.
.
l­
3
Figure
4.1
Exhaust
Piping
Schematic
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
4­
2
Figure
5.1
Schematic
Diagram
of
EECL
FTIRWEMS
Sampling
and
AnalysisSystem.................................................
5­
4
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
(
HAP).
This
document
describes
the
results
of
emissions
testing
conducted
on
a
Waukesha
3521
GL
natural­
gas­
fired
4­
stroke,
lean
burn
(
4SLB)
engine.
Early
in
1998,
several
industry
and
EPA
representatives
agreed
that
the
Waukesha
3521GL
engine
at
the
Colorado
State
University's
Engine
and
Energy
Conversion
Laboratory
(
CSU
EECL)
is
adequately
representative
of
existing
and
new
natural­
gas­
fired
4SLB
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
Waukesha
3521
GL
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
4SLB
matrix
conditions
at
CSU
as
the
basis
for
developing
the
MACT
standard
for
natural­
gas­
fired
4SLB
engines.

Emissions
testing
was
conducted
to
estimate
HAP
emissions
before
and
after
the
oxidation
catalyst.
Miratech
Corporation
manufactured
the
catalyst
and
EECL
personnel
installed
it
on
the
engine.
Fourier
transform
infrared
spectroscopy
(
FTIRS)
was
used
to
measure
formaldehyde,
acetaldehyde,
and
acrolein.
Continuous
emission
monitoring
systems
(
CEMS)
were
used
to
measure
oxygen
(
0,),
carbon
dioxide
(
CO,),
nitrogen
oxides
(
NO,&
carbon
monoxide
(
CO),
total
hydrocarbons
(
THC),
methane,
and
non­
methane
hydrocarbons
WC)*

Operating
as
a
subcontractor
to
PES,
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
CEMS
that
measured
pollutants
and
diluents
in
the
exhaust
gas.
Under
a
separate
work
assignment,
Eastern
Research
Group
(
ERG)
of
Morrisville,
North
Carolina
operated
an
EPA­
owned
dynamic
spiking
system
for
the
validation
of
the
FTIRS
systems
for
formaldehyde,
acetaldehyde,
and
acrolein.

Final
Report
Waukesha
352
1
GL
l­
l
September
2001
The
test
program
organization
and
major
lines
of
communication
for
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
.
O
Sampling
and
Analysis
Methods
Section
6.0
Quality
A
ssurance/
Quality
Control
Procedures
and
Results
Appendix
A
presents
the
4SLB
report
issued
by
CSU
EECL
on
April
28,200O.
Appendix
B
contains
example
calculations
used
by
PES
to
calculate
results,
and
background
correspondence
pertaining
to
the
Waukesha
test
program.

Final
Report
Waukesha
352
1GL
l­
2
September
200
1
EPAIEMC
Quality
Assurance
Of&
xx
Lara
P.
Autry
(
919)
541­
5544
EPAIEMC
Work
Assignment
Manager
Terry
Harrison
(
919)
541­
5233
PES
Project
Manager
Dennis
A.
Falgout
(
703)
471­
8383
PES
QAIQC
Officer
Jeff
Van
Atten
(
703)
471­
8383
I
Pretest
Site
Survey
PES
I
Quality­=
Project
Plan
PES
L
EPAIESD
Lead
Engineer
Sims
Roy
(
919)
541­
5263
Site
Specific
Field
Test
Plan
Testing
PES
PES
I
Subcontractor
CSUEECL
Report
Preparation
PES
Subcontractor
r
CSUEECL
I
I
Figure
1.1
Test
Program
Organization
and
Major
Lines
of
Communication
Final
Report
Waukesha
3
52
1
GL
1­
3
September
200
1
2.0
SUMMARY
OF
RESULTS
This
section
provides
summaries
of
the
stack
gas
parameters
and
HAP
emissions
measured
during
the
test
program.
Testing
of
the
Waukesha
3
521
GL
engine
was
conducted
August
4
through
August
6,1999
at
CSU's
Engines
and
Energy
Conversion
Laboratory
in
Fort
Collins,
Colorado.
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.
The
end
of
this
section
contains
a
discussion
of
the
efficiencies
at
which
the
catalyst
removed
HAP.

CSU
submitted
a
report
documenting
the
results
of
the
test
program
to
PES.
This
report
is
reproduced
in
its
entirety
in
Appendix
A.
PES
discovered
errors
in
the
combustion
products
(
Fd)
factors
calculated
by
CSU,
which
resulted
in
errors
in
the
calculations
of
pollutant
mass
flow
rates.
PES
requested
that
CSU
correct
the
errors
and
re­
submit
the
emissions
calculations
so
that
the
final
report
could
be
completed.
The
corrected
results
submitted
by
CSU
showed
that
only
9
of
the
sixteen
runs
used
the
correct'Fd
values.
Slight
errors
still
exist
in
CSU's
calculated
results
for
Runs
2,4,7,8,10,11,
and
12.
In
the
tables
that
follow,
the
correct
Fd
values
are
used
for
each
run.
PES
believes
that
the
values
expressed
in
these
tables
are
correct
representations
of
pollutant
mass
flow
rates
during
the
test
program.

2.1
EMISSIONS
TElST
LOG
During
the
test
period,
the
test
team
conducted
thirty­
four
test
runs
using
FTIRS
and
CEMS.
These
test
runs
consisted
of
sixteen
5­
minute
Quality
Control
(
QC)
runs,
sixteen
33­
minute
sampling
runs
for
collection
of
FTIRS
and
CEMS
data,
and
two
S­
minute
baseline
runs.
Table
2.1
presents
the
emissions
test
log.
The
test
log
summarizes
the
date
and
time
that
each
run
was
conducted.
Additional
discussions
regarding
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
reduce
both
the
times
between
test
runs
to
change
an
engine
condition
and
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
Final
Report
Waukesha
3521GL
2­
l
September
2001
TABLE
2.1
EMISSIONS
TEST
LOG
c
Date
1
Run
Time
1
Run
ID
1
Date
1
RunTime
1
Run
ID
814199
1521­
1526
Run9QC
8/
5/
99
2230­
2235
Run6QC
814199
1552­
1625
Run9
8/
S/
99
2252­
2325
.
Run6
8/
4/
99
1
1747­
1752
1
RunlQC
1
8/
6/
99
1
004
l­
0046
1
Run7QC
814199
1812­
1845
Run1
816199
02
1
O­
0243
Run7
814199
2020­
2025
Run
14
QC
816199
03
15­
0320
Run2QC
8/
4/
99
2039­
2
112
Run
14
816199
0328­
040
1
Run2
815199
1302­
1307
Run3QC
816199
0520­
052s
Run4QC
8/
S/
99
1323­
1356
Run3
816199
0536­
0609
Run4
8/
S/
99
1511­
1516
Baseline
1
816199
0714­
0719
Run8QC
8/
S/
99
1637­
1642
Run
13
QC
816199
0727­
0800
Run8
8/
5/
99
1
1657­
1730
1
Run
13
1
8/
6/
99
1
0902­
0907
1
Run
12
QC
8/
5/
99
1
1805­
1810
1
Run
15
QC
1
8/
6/
99
1
0936­
1009
1
Run
12
8/
S/
99
1815­
1848
Run
15
816199
1119­
1124
Run
10
QC
8/
S/
99
1935­
1940
Run
16
QC
816199
1136­
1209
Run
10
815199
1952­
2025
Run
16
816199
1304­
1309
RunllQC
8/
5/
99
1
2101­
2106
1
RunSQC
1
8/
6/
99
1
1317­
1350
1
Run
11
8/
S/
99
2117­
2150
RUIl5
8/
6/
99
1416­
1421
Baseline
2
Final
Report
Waukesha
3
52
1
GL
2­
2
September
200
1
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.

2.2
ENGINE
PARAMETERS
AND
EXHAUST
GAS
FLOW
RATES
Table
2.2
summarizes
some
engine
and
exhaust
gas
parameters
measured
and
calculated
during
the
test
program.
The
EECL's
Data
Acquisition
System
@
AS),
monitored
and
recorded
approximately
200
engine
operating
parameters,
and
gas
temperatures,
and
concentrations
of
02,
CO,,
and
moisture
at
the
catalyst
inlet
and
exhaust.

The
exhaust
gas
volumetric
flow
rates
upstream
and
downstream
of
the
catalyst
are
presented
for
each
sample
run.
These
flow
rates
were
calculated
using
a
combustion
products
(
Fd)
factor.
The
flows
thus
calculated
were
corrected
for
excess
air
based
on
the
measurements
of
O2
concentration
at
each
location.
A
new
fuel
factor
was
calculated
each
day
based
upon
daily
analysis
of
the
composition
of
the
natural
gas
fuel.

On
August
$,
the
oxygen
monitor
at
the
catalyst
inlet
failed
during
Run
No.
3,
and
during
the
Baseline
Run
1.
The
oxygen
concentration
at
the
catalyst
inlet
was
reported
to
be
100%
during
these
runs.
Since
this
value
is
an
obvious
error,
the
oxygen
concentration
at
the
catalyst
outlet
was
substituted
for
the
catalyst
inlet
value,
and
used
to
calculate
volumetric
flow.
The
oxygen
value
at
the
catalyst
inlet
location
during
Run
No.
13
of
10.44%
appears
higher
than
the
other
oxygen
values.
The
monitor
and
the
DAS
data
were
examined.
Although
the
value
appears
higher
than
normal,
justification
for
invalidating
the
data
point
could
not
be
found.

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
CEMS
target
compounds
(
carbon
monoxide,
nitrogen
oxides,
THC,
methane,
and
NMHC).

EECL
operated
two
FTIRS
sampling
and
analysis
systems
to
quantify
concentrations
of
the
target
compounds.
Exhaust
gas
samples
were
extracted
from
locations
upstream
and
downstream
of
the
oxidation
catalyst,
conditioned,
and
transported
to
a
Nicolet
Rega
7000
FTIRS
(
upstream
location)
and
a
Nicolet
Magna
560
FTIRS
(
downstream
location).
The
upstream
FTIRS
also
measured
the
moisture
content
in
the
exhaust
gas.
Moisture
measurements
by
the
downstream
FTLRS
were
inaccurate.
EECL
used
a
carbon
balance
method
to
calculate
the
moisture
concentration
at
the
downstream
sampling
location.

Final
Report
Waukesha
3
52
1
GL
2­
3
September
2001
TABLE
2.2
SUMMARY
OFEXHAUST
GAS
FLOW
RATES
.
Run
ID
Run
1
Run
2
Run
3
Run
4
Run
5
Run
6
Run
7
Run
8
Engine
Speed,
rpm
1197
1197
1002
1002
1197
1197
1197
1001
Engine
Torque,
tt­
lb
3236
2263
3234
3235
3234
2263
3234
Horsepower,
bhp
737
516
432
617
737
737
516
617
Fuel
Flow
Rate,
sclh
5416
4114
322%
4401
5564
5377
4070
4457
Equiwlence
Ratio,
+
0.56
0.56
0.56
0.56
0.52
0.5%
0.5%
0.53
Lower
Heating
Value,
Btukf
1008
1024
1024
1024
1024
1024
1024
1024
Higher
Heating
Value,
Btulcf
1135
1095
1152
1095
1152
1152
1095
1095
Heat
Rate,
MMBtulhr
.
6.15
4.50
3.72
4.82
6.41
6.1%
4.46
4.66
Dry
Fuel
Factor,
Fd
dsdMMBtu
8654
6645
6655
6645
8655
8655
8645
8645
Catalyst
Inlet
Gas
Temperature,
`
F
735
706
677
665
711
75%
726
663
Oxygen,
%
d.
b.
9.80
9.62
9.61
9.60
10.5
9.10
9.20
10.50
Carbon
Dioxide,
%
d.
b.
6.2%
6.23
6.37
6.24
5.66
6.65
6.65
5.86
Moisture,
%
12.3
11.7
12.3
12.0
11.6
12.7
12.4
11.6
Gas
Flow
Rate,
dsclhr
1670
1224
1011
1307
1660
1583
1147
1413
Catalyst
Outlet
Gas
Temperature,
`
F
740
710
67%
688
718
760
731
66%

Oxygen,
%
d.
b.
9.60
9.63
9.61
9.80
10.4
9.01
9.0%
10.5
Carbon
Dioxide,
%
d.
b.
6.46
6.42
6.40
6.41
5.96
6.83
6.7%
5.9%

Moisture,
%
12.4
11.9
12.3
12.2
11.6
12.9
12.5
11.7
Gas
Flow
Rate,
dsckn
1670
1225
1011
1307
1646
1571
1136
1413
rpm­
rav­
pardllda
hlMBtulhr­
miUion&
iihlhemnlur&
perhaw
ft­
b
­
foot­
pounds
dscfMt3tu­
dyrtmdwdcubicfmto(
exhaustproductsper
bhp­
brdmtlorsapower
MCUl6tUdheethprt@
0%~
CS!
SS&

scfh­
standafdwbicfeatpahOur~~
ard~.
82h
kb
r­&
3(
pasFahmm
+­
mipmcatd966ccassAlr
Yvotd.
b.­%
vm6ybasb
BWcf­
BrkishIhwmtWspacubicfootdnatu~
gas
Lcfm­
dyrtandadcubicfestper~
e~
68rand29.92hHg
NOES:
oz
matyzar
fmrulctilm
at
C­
St
IIltst
binQ
tam
ND.
3.
VdJa
for
catalyst
altkt
was
rrrbstauted.
F,
v&
esforRms2,4,7,8,10,11,
end12diffarfromF,
vaksraportedbyCSU.
Seetexton~
2­
1
fordiscussbn.

Final
Report
Waukesha
3521GL
2­
4
September
200
1
rpn­
revcMicnsper~
0
bMtwhr­
mtlim6rlbhlhmmluntsparhour
ft­
b
­
foot­
pound6
dscfhemtlJ­
4yr­
crMefeatafex~
tpr~
ts~

bhp­
brdlehonepower
mBcn8tucfheethputg)
O%
excessdr
scfh­
stmdudcuMcfeetpahow@
86Tnd29.92htQ
r­
degrees­

+­
reciprcdd%
kcessAt
%
vofd.
b.­%
vohmm&
ybmts
B&&
f­
Eritish7?
mrmtWtsparcubkfoatdndurrlO#
dpdm­
6ys~
cubief~~
miatte~
68rand29.92h~

tS)
Tk
F,
vaWea
far
Runs
2,4,7,8,10,11,
and
12diffarfromF,
values
mportedby
CSU.
Seetext
m
­
2­
l
fordbcussim.
TABLE
2.2
(
CONCLUDED)

SUMMARY
OF
EXHAUST
GAS
FIBW
RATES
I
l­
feet
Rate,
MMBtuIhr
8.18
5.89
5.90
5.97
8.80
8.03
8.33
8.24
*
"­`­
Dry
Fuel
Factor,
F,,
dscffMMBtu
8854
8845
a845
8855
8854
8855
a855
Gas
Temperature,
`
F
,735
737
731
738
787
722
738
730
2
_
*
Oxygen,
%
d.
b.
9.89
9.80
9.81
9.90
10.4
9.81
9.90
9.82
Carbcm
Dioxide,
%
d.
b.
8.35
8.35
8.29
`
8.29
8.27
8.24
8.22
8.2.3
Moisture,
%
12.3
12.1
12.2
12.0
12.1
12.1
12.1
12.1
,
.
1
Gas
Fkw
Rate,
dscltn
1857
1598
1803
1834
1901
1839
1734
1897
Catlilyd
Outlei
Gas
Temperature,
`
F
740
741
738
740
771
718
742
733
Oxygen,
%
d­
b.
9.70
9.73
9.79
9.85
9.81
9.89
9.82
9.80
Carbon
Dioxide,
%
d.
b.
8.51
8.5
8.47
8.41
8.38
8.41
8.35
8.38
I,
..'
Moisture,
%
12.5
12.3
12.4
12.1
12.2
12.3
12.2
12.2
,
Gas
Flaw
Rate,
dsclhr
1859
1588
1800
1828
1793
1851
1722
1894
Final
Report
Waukesha
3521GL
2­
S
September
200
1
TABLE23
EMISSION
RATES
OF
DETECTED
FIIRS
AND
CEMS
COMPOUNDS
Run3
1
Run4
1
Run5
1
Run6
1
Run7
1
Run8
1
Catalyst
Outld
Formaldehyde
Acetaldehyde
Acrdein
Nitrogen
Oxides
(
as
NO3
mg/
bhp­
hr
99
.
mlWhr
160
100
mglbhp­
hr
ND
ND
mlb/
hr
ND
ib
mgbhp­
hr
ND
ND
mlWhr
ND
ND
Blmpru
O.
OlP
0.
W
I
IWhr
1.42
0.724
60.1
83.1
133
.
69.7
110
57.1
113
216
121
79.2
150
NU
NU
ND
NU
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
0.622
0.
/
alI
0.65/
1
AX
I
.
T"
I
in
I.
I"
0.469
0.592
1.06
1.03
2.41
1.25
0.665
Carbon
Monoxide
glbhp­
hr
0.165
.
0.101
0.127
0266
0.16/
0
115
0
*
163
IWhr
0.301
0.143
0.097
0.172
01433
0.271
0:
131
0.221
.

g/
bnp­
hr
2
­
I2
3.20
346
.
4.04
4.20
3.11
4
/
5
Methane
IWhr
5.23
4.59
4.00
5.04
7:
72
4.42
3.64
4:
70
glbhp­
hr
0.95
1.10
1.22
0.849
1.3/
0
­
692
0.662
1.23
Non­
methane
Hydrocarbons
IWhr
1.54
1.25
1.16
1.15
2.22
1.13
0.98
1.70
glbhp­
hr
4
/
9
563
6
51
8:
20
5
19
7:
05
f
02
ii.
4
364
4
59
6
74
Total
Hydrocarbons
IWhr
7:
78
8:
83
6:
24
5:
22
9:
1s
Final
Report
Waukesha
3
52
1
GL
2­
6
September
200
1
TABLE
23
(
CONCLUDED)

EMISSION
RATES
OF
DETECTED
FI'IRS
AND
CEMS
COMPOUNDS
Final
Report
Waukesha
3521GL
2­
7
September
200
1
Formaldehyde
was
detected
at
the
upstream
and
downstream
locations
during
every
sampling
run.
Neither
acetaldehyde
nor
acrolein
were
detected
on
any
sampling
run
either
before
or
after
the
catalyst.
Run
by
run
detection
limits
for
the
FTIRS
compounds
are
presented
in
Table
6.1.

EECL
personnel
operated
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
before
introduction
to
the
OS,
CO*,
CO,
and
NOx
analyzers.
All
of
the
CEMS
target
compounds
were
detected
at
the
catalyst
inlet
and
catalyst
outlet.

The
reported
concentration
of
NMHC
at
the
catalyst
outlet
for
Run
No.
8
was
284
ppmv
as
methane.
The
reported
concentration
at
the
catalyst
inlet
was
17
1
parts
per
million
by
volume
(
ppmv)
as
methane.
CSU
examine
the
NMHC
data
and
invalidated
the
data
for
this
run.
At
the
direction
of
the
WAM?
the
NMHC
values
obtained
during
the
5­
minute
QC
run
conducted
just
prior
to
Run
No.
8
were
substituted
at
both
locations.
The
NMHC
concentrations
were
168
ppmv
as
methane
and
155
ppmv
as
methane
at
the
catalyst
inlet
and
outlet
locations,
respectively.
These
values
are
presented
in
Appendix
C
of
the
CSU
test
report.

2.4
DESTRUCTION
OF
HAP
BY
THE
CATALYST
There
are
five
possible
HAP
mass
flow
rate
combinations
that
can
occur
across
the
oxidation
catalyst.
Table
2.4
presents
these
combinations,
and
notes
whether
a
destruction
efficiency
is
reported.
Out
of
the
five
possible
combinations,
there
are
two
instances
where
the
destruction
efficiency
of
the
target
pollutant
is
reported.
If
the
mass
flow
rate
of
a
pollutant
into
the
catalyst
(
Qi3
is
greater
than
the
mass
flow
rate
exiting
the
catalyst
(
QO,,
J,
%
DE
is
calcula
ted.
If
the
pollutant
is
detected
entering
the
catalyst,
but
is
not
detected
exiting
the
catalyst,
%
DE
is
estimated
using
the
measured
mass
flow
rate
at
the
inlet,
and
the
mass
flow
rate
corresponding
to
the
analytical
detection
limit
at
the
outlet.

The
removal
efficiency
of
HAP
for
various
target
compounds
is
presented
`
in
Table
2.5.
Formaldehyde
was
detected
on
every
run
that
was
attempted.
In
every
case,
the
mass
flow
rate
of
formaldehyde
into
the
catalyst
was
greater
than
the
mass
flow
rate
of
formaldehyde
leaving
the
catalyst.
Therefore
sixteen
formaIdehyde
removal
effkiencies
were
reported.
Neither
acetaldehyde
nor
acrolein
were
detected
on
any
run
at
either
location.
There
are
no
removal
efficiency
data
to
present
for
either
of
these
compounds.

The
removal
efficiency
of
NOx
is
not
presented
for
any
run.
In
every
case,
the
measured
concentration
of
NOx
at
the
catalyst
exit
was
greater
than
the
measured
concentration
of
NOx
at
the
catalyst
inlet.
The
difference
between
the
outlet
and
the
inlet
Final
Report
Waukesha
3521GL
2­
8
September
200
1
TABLE
2.4
MASS
FLOW
SCENARIOS
I
Scenario
No.
Result
DE
Reported?

1
Qin
'
0;
Qout
'
0;
Qin
'
Qout
YES
2
Qin'
0;
Q,=
l'+
JD
YES
3
Qin
<
Qout
NO
4
Qh=
ND;
Q,,>
O
NO
5
Qin=
m;
Qout=
m
NO
values
ranged
from
5
parts
per
million
by
volume,
dry
basis
(
ppmvd)
to
11
ppmvd,
which
is
1
.
O
to
2.2%
of
the
measurement
range
(
0
­
500
ppmvd)
of
the
NOx
analyzers.
The
apparent
increase
in
NOx
across
the
catalyst
may
be
due
to
uncertainty
inherent
in
using
two
analyzers,
and
not
due
to
any
increase
in
the
mass
flow
rate
of
NOx.
Carbon
monoxide
was
detected
at
the
catalyst
inlet
and
outlet
on
every
run.
The
mass
flow
rate
of
carbon
monoxide
showed
a
marked
decrease
across
the
catalyst.
Carbon
monoxide
destruction
efficiencies
are
presented
for
every
run.

Methane
and
NMHC
were
detected
by
each
methane/
NMHC
analyzer
at
both
locations
for
every
run.
There
are
sixteen
removal
efficiencies
calculated.
The
difference
in
methane
concentrations
across
the
catalyst
averaged
approximately
160
ppm
for
all
of
the
runs.
This
corresponds
to
about
3%
of
the
methane
analytical
range
of
0
­
5,000
ppmv.
The
small
difference
makes
it
difficult
to
determine
if
there
was
a
reduction
in
methane
across
the
catalyst,
or
if
the
differences
are
due
to
uncertainty
inherent
in
using
two
analyzers.
The
difference
between
the
NMHC
concentrations
measured
at
the
catalyst
inlet
and
the
catalyst
outlet
averaged
about
30
ppm
for
all
of
the
runs.
This
difference
is
about
6%
of
the
NMHC
analytical
range
of
0
­
500
ppmv.
The
data
indicates
that
some
NMHC
was
probably
removed
by
the
catalyst.

There
are
no
destruction
efficiencies
presented
for
total
hydocarbons.
In
fifteen
of
sixteen
cases,
the
mass
flow
rate
of
THC
exiting
the
catalyst
was
greater
than
the
mass
flow
rate
of
THC
entering
the
catalyst.
The
difference
in
the
concentration
measurements
at
these
locations
approached
1%
of
the
THC
analytical
range
of
0
­
5,000
ppmv.
Therefore,
there
was
most
likely
no
removal
of
THC
across
the
catalyst.
Since
THC
is
made
up
mostly
of
methane,
this
seems
to
indicate
that
the
difference
in
methane
values
was
most
likely
due
to
inherent
measurement
errors.

Final
Report
Waukesha
3521GL
2­
9
September
2001
TABLE
2.5
CATALYST
HAP
REMOVAL
EFFICIENCIES
Non­
methane
Hydrocarbons
12%
2%
4%
6%
14%
12%
11%
11%

21%
12%
8%
16%
38%
18%
29%
21%

Final
Report
Waukesha
352
1
GL
2­
10
September
200
1
3.0
SOURCE
DESCRIPTION
AND
OPERATION
This
section
presents
discussions
of
the
candidate
engine
and
the
catalyst
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
Waukesha
3521GL
is
a
4­
stroke
stationary
internal
combustion
engine.
The
engine
has
six
inline
cylinders;
the
total
piston
displacement
is
3520
cubic
inches.
Each
cylinder
is
9.375
inches
in
diameter,
and
has
an
8.5~
inch
stroke.
The
compression
ratio
is
10.5:
1.
Air
is
delivered
to
the
engine
via
the
EECL's
pressurized
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
4­
stroke
engines
in
industry.
Table
3.1
presents
specifications
of
the
engine
and
the
catalyst.
Table
3.2
presents
nominal
engine
operating
parameters.

The
4­
stroke
cycle
requires
two
revolutions
of
the
engine
crankshaft
for
each
power
stroke.
During
the
intake
stroke,
the
piston
moves
down
the
cylinder
and
an
air/
fuel
mixture
is
injected
into
the
piston
chamber.
On
the
compression
stroke,
the
piston
moves
back
up
the
chamber,
and
the
mixture
is
compressed
and
ignited.
The
expanding
gas
generated
upon
combustion
forces
the
piston
back
down
the
chamber.
This
stroke
is
the
power
stroke.
The
last
stroke
of
the
4­
stroke
cycle
is
the
exhaust
stroke.
The
piston
travels
back
up
the
chamber
and
the
combustion
products
are
vented
through
the
exhaust
manifold.

The
3
52
1
GL
engine
was
outfitted
with
lean­
burn
technology,
which
controls
NO,
emissions.
The
lean­
bum
system
uses
pre­
combustion
chambers
to
ignite
a
lean
air/
fuel
mixture
in
the
main
combustion
chambers.
A
rich
mixture
of
air
and
fuel
is
drawn
into
the
pre­
combustion
chamber
and
is
ignited
by
a
spark
plug.
The
resulting
flame
is
then
directed
into
the
main
combustion
chamber,
which
contains
a
lean
mixture
of
air
and
fuel.
The
flame
jet
from
the
pre­
combustion
chamber
ignites
the
air/
fuel
mixture
in
the
main
chamber.

Final
Report
Waukesha
3521GL
3­
l
September
2001
TABLE
3.1
ENGINE
AND
CATALYST
SPECIFICATIONS
hgine
Classification
Four­
stroke,
lean
burn,
natural­
gas­
fired
Manufacturer
and
Type
Waukesha
3521
GL
Number
of
Cylinders
6
Bore
and
Stroke
9.375"
x
8.5"

Nominal
Engine
Speed
1200
RPM
gnition
System
Classification
Spark
Ignited
Pre­
combustion
Chamber
Ignition
System
Altronic
Pre­
combustion
Chamber
Type
Standard
OEM
Product
Number
of
Pre­
combustion
Chambers
1
Per
Cylinder
Zatalyst
Classification
Oxidation
Type
Manufacturer
Miratech
Corporation
Date
of
Manufacture:
May
1999
Model
Number:
None.
Custom­
designed
unit
Serial
Number:
None.
Custom­
designed
unit
Item
Number:
CSU­
RE­
12160
Platinum/
Palladium
on
Stainless
Steel
Catalyst
Material:
Substrate.
Manufactured
in
Finland
by
Kemira
Element
Size:
12"
x
16"
x
3"
Effective
Area:
11"
x
14
7/
8"

Number
of
Elements
2
Final
Report
Waukesha
352
1
GL
3­
2
September
2001
TABLE
3.2
SUMMARY
OF
NOMINAL
ENGINE
PARAMETERS
.

Parameter
Torque
speed
lacket
Water
Temp
(
Outlet)

Oil
Temperature
Air
Manifold
Temperature
Air
Manifold
Pressure
Exhaust
Manifold
Pressure
Ignition
Timing
Overall
Air/
Fuel
Ratio
Inlet
Air
Humidity­
Absolute
Fuel
Flow
Oil
Pressure
Inlet
Inlet
Air
Flow
Average
Exhaust
Temp
Nominal
Value
3236
ft­
lb
1200
rpm
180
`
F
185
`
F
100
`
F
5
inches
Hg
over
ambient
pressure
(
i.
e.,
*
a
level)

5
inches
Hg
less
than
Air
Manifold
Pressure
10'
BTDC
28/
l
0.0
15
lb
HzO/
lb
Air
5460
scfh
52
psig
1730
scfin
700
`
F
Acceptable
Deviation
Designation
f
2%
of
value
­
w
f
5%
of
value
­=
Y
f
5%
of
value
*
str
f
5%
of
value
pri­
v
f
5%
of
value
primary
f
5%
of
value
primary
f
5%
of
value
pl­
imary
f
5%
of
value
*
ary
f
5%
of
value
*
ary
f
10%
of
value
primary
f
5%
of
value
F+
i­
Y
f
10%
of
value
Secondary
f
5%
of
value
secomialy
f
5%
of
value
secoxldaIy
­

&
lb
­
foot­
pounds
rpm
­
revolutions
per
minute
`
F
­
degrees
Fahrenheit
BTDC
­
Before
Top
Dead
Center
lb
H,
O/
lb
Air
­
pounds
water
vapor
per
pound
of
air
scfh
­
standard
cubic
feet
per
hour
psig
­
pounds
per
square
inch,
gauge
scfin
­
standard
cubic
feet
per
minute
Final
Report
Waukesha
352
1
GL
3­
3
September
200
1
3.2
ENGINE
OPERATION
DURING
TESTING
As
stated
in
Section
2
of
this
document,
three
types
of
test
runs
were
conducted
during
the
test
program:
quality
control
runs,
sampling
runs
for
FTIRS
and
CEMS,
and
baseline
runs.
The
operation
of
the
engine
during
these
various
runs
is
discussed
in
the
following
pages
and
tables.
The
four­
stroke
engine
test
matrix
described
in
the
QAPP
was
based
upon
estimated
operating
parameters
for
a
candidate
engine
to
be
installed
and
operated
at
the
EECL.
When
the
engine
was
received
and
first
operated
by
EECL
the
actual
operating
parameters
differed
fi=
om
the
estimates.
Table
3.3
presents
the
test
matrix
for
the
Waukesha
engine
based
upon
the
actual
engine
parameters.
During
the
test
program,
the
six
engine
operating
parameters
expected
to
have
the
greatest
impact
on
pollutant
formation
were
varied
from
their
baseline
values.
These
parameters
were:
engine
speed
(
measured
in
revolutions
per
minute
or
rpm),
engine
torque
(
measured
in
foot­
pounds
or
f&
lb),
air­
to­
fuel
ratio
(
calculated
as
an
equivalence
factor),
engine
timing
(
the
location
of
the
piston,
relative
to
top
dead
center,
at
the
time
of
spark
in
the
pre­
combustion
chamber,
measured
in
degrees),
air
manifold
temperature
(
measured
in
degrees
Fahrenheit),
and
jacket
water
outlet
temperature
(
measured
in
degrees
Fahrenheit).

Table
3.4
presents
engine
parameters
recorded
during
each
test
run
and
their
percent
deviation
from
the
target
values.
Sixteen
sampling
runs
were
conducted
on
the
engine
during
the
two­
day
period.
Except
for
air/
fuel
ratios,
the
actual
parameters
agreed
with
the
target
parameters
to
within
5%.
Although
the
calculated
air/
fuel
ratios
were
not
within
5%
of
the
target
air/
fuel
ratios,
testing
was
conducted
while
operating
at
rich
air/
fuel
ratios
(
Runs
5
and
8)
and
at
lean
air/
fuel
ratios
(
Runs
6
and
7).
The
air/
fuel
ratio
was
varied
to
simulate
the
range
of
air/
fuel
ratios
that
typcial
in
field
applications.
'

Before
starting
Run
7,
the
humidity
control
system
failed.
The
humidity
system
could
not
be
repaired
quickly
so
the
run
was
conducted
without
inlet
air
humidity
control.
Run
2
was
also
conducted
without
inlet
air
humidity
control.
The
set
point
for
the
humidity
ratio
for
all
test
points
was
0.015
lb.
water
/
lb.
air.
The
actual
humidity
ratios
for
Runs
2
and
7
were
0.0126
and
0.0127
lb.
water
/
lb.
air
respectively.
The
engine
emissions
for
Runs
2
and
7
should
be
similar
to
engine
emissions
at
the
specified
humidity
ratio.
The
most
dramatic
effect
will
be
on
NO,
emissions
as
can
be
seen
from
the
data
and
the
graphs
presented
in
Appendix
S
of
the
CSU
test
report.
At
a
constant
humidity
ratio,
it
would
be
expected
that
formaldehyde
emissions
would
either
remain
constant
or
increase
slightly
with
similar
changes
in
CO
and
THC
emissions.

Final
Report
Waukesha
3521GL
3­
4
September
2001
TABLE
3.3
TARGET
ENGINE
OPERATING
CONDITIONS
DURING
TESTING
N
=
Normal
Value
L
=
Low
value
H
=
High
Value
S
=
Set­
point
Value
Final
Report
Waukesha
352
1
GL
3­
5
September
2001
TABLE
3.4
SUMMARY
OF
ENGINE
PARAMETERS
­
WAUKESHA
3521
GL
RunO
1197
1200
­
0.26%

3234
3236
­
0.1%

0.56
0.61
­
7.6%

10
10
0.0%

119
120
­
0.1%

179
180
­
0.6%

737
5389
1024
5.52
Gx
1197
1200
­
0.26%

3237
3236
0.0%
Run
1197
1200
­
0.26%

3238
3236
0.1%
Run
xi7
1200
­
0.25%

2263
2265
­
0.1%
RunS
Xii­

1000
0.16%

2264
2265
­
0.1%
Run
TiiiZ­

1000
0.15%

3234
3236
­
0.1%
Run
1197
1200
­
0.26%

3235
3236
0.0%
Run
1197
1200
­
0.26%

3234
3236
0.0%
Runl
7r
1200
­
0.26%

2263
2265
4.1%
Run
1001
1000
0.15%

3234
3236
­
0.1%
Run
1197
1200
­
0.26%

3236
3236
0.0%
1197
1197
1197
1200
1200
1200
­
0.29%
­
0.26%
­
0.26Y
1197
1200
­
0.26%

3236
3236
0.0%
EngineToqueR­
tb
EquiwlenceRatio,+

IntercoolerWater
fpm
­
rmolutlons
per
minute
It­
lb­
foot
pounds
l
BIDC­
degrwsBeforeTopDeadCente8
3235
3236
0.0%

0.56
0.61
­
8.5%
3236
3236
0.0%

0.55
0.61
­
9.2%

10
10
0.0%

128
130
­
0.3%

179
180
­
0.7%

737
5452
1040
5.67
8655
3235
3236
0.0%

0.56
0.61
­
8.6%
0.56
0.56
0.52
0.59
0.59
0.53
0.61
0.61
0.56
0.62
0.62
0.56
­
8.6%
­
6.4%
8.3%
­
4.6%
­
5.4%
­
6.2%
0.56
0.61
­
8.5%
0.56
0.61
­
8.5%
0.55
0.61
­
9.3%
0.53
0.61
­
13.7%
0.56
0.61
8.6%

10
10
0.0%

128
130
­
0.4%
10
10
0.0%
6
6
0.0%
10
10
10
10
0.0%
0.0%

128
129
130
130
­
0.3%
­
0.2%

180
179
180
180
0.2%
4.5%

737
516
5337
4040
1040
1040
5.55
4.20
10
10
0.0%

130
130
0.1%

179
180
­
0.6%
10
10
0.0%

129
130
­
0.3%
10
10
0.0%

130
130
0.1%

179
180
­
0.3%
10
10
0.0%

130
130
0.0%

180
160
0.1%

737
5372
1040
5.58
8655
14
14
0.0%

132
130
0.3%

179
180
­
0.5%

737
5272
1024
5.40
141
140
0.2%

180
180
0.3%

738
5341
1040
5.55
8645
I
129
130
­
0.2%

179
180
­
0.7%

617
4424
1040
4.60
8645
130
130
0.0%

190
190
0.1%

738
5411
1040
5.63
8645
128
130
­
0.3%

163
180
1.7%

737
5683
1040
5.91
8655
179
180
­
0.5%
179
180
­
0.3%

432
617
737
3205
4368
5523
1040
1040
1040
3.33
4.54
5.74
665,
8645
8655
516
1040
4.24
8645
'
F­
degrwsFahnwheit
bhp­
brakehorsepower
scth­
standafdcubicfaetperhour
Btukf­
British'Ihermal
Units
percubic
M
ofnatwalgas
MMBtuIhr­
million
Btu
perhour
Final
Report
Waukesha
352
1
GL
3­
6
September
200
1
Table
3.5
presents
engine
parameters
measured
during
two
baseline
test
runs.
There
were
two
testing
periods.
On
August
4,
testing
was
conducted
over
a
six­
hour
period.
The
engine
was
shut
down
and
testing
resumed
on
August
5.
The
testing
was
completed
over
the
next
26
hours
of
continuous
engine
operation.
Test
accuracy
required
that
the
overall
engine
operation
did
not
change.
The
stability
of
the
engine
over
this
period
was
demonstrated
by
operating
the
engine
at
the
baseline
condition
for
one
Sminute
period
early
in
the
26­
hour
period
and
a
second
S­
minute
period
at
the
end
of
the
testing.
Changes
to
the
baseline
parameters
would
have
indicated
a
change
in
the
overall
operating
characteristics
of
the
engine.
It
would
not
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
compares
the
values
of
13
engine
parameters
measured
during
the
baseline
runs
to
the
manufacturer's
recommended
settings
that
were
presented
in
Table
3.2.
Deviations
are
calculated
in
percent.
Temperatures
were
converted
to
degrees
Rankine,
then
the
percent
deviation
was
calculated.

Final
Report
Waukesha
3521GL
3­
7
September
200
1
TABLE
3.5
SUMMARY
OF
ENGINE
PARAMETERS
DURING
BASELINE
RUNS
Engine
Speed,
rpm
Engine
Torque,
&
lb
Air/
Fuel
Ratio,
lb
air
/
lb
fuel
Ignition
Trming,
"
BTDC
Jacket
Water
Temperature,
"
F
Exhaust
Manifold
Pressure,
in.
Hg
Inlet
Air
Humidity,
lb
H,
O/
lb
air
Inlet
Air
Flow,
scth
Exhaust
Temperature,

rpm
­
rewlutions
per
minute
in.
Hg
­
inches
of
mercury
ft­
lb
­
foot­
pounds
lb
H20
/
lb
air
­
pounds
water
wpor
per
pound
of
air
lb
air
/
lb
fuel
­
pounds
air
per
pound
of
fuel
scfh
­
standard
cubic
feet
per
hour
"
BTDC
­
degrees
Before
Top
Dead
Center
psig
­
pounds
per
square
inch,
gauge
"
F
­
degrees
Fahrenheit
scfm
­
standard
cubic
feet
per
minute
Final
Report
Waukesha
352
1GL
3­
8
September
200
1
4.0
SAMPLING
LOCATIONS
A
schematic
drawing
of
the
exhaust
gas
piping
on
the
Waukesha
3521GL
engine
is
shown
in
Figure
4.1.
The
engine
exhaust
manifold
was
connected
to
the
inlet
of
the
catalyst
with
an
&
inch
internal
diameter
(
ID)
pipe.
The
pipe
extended
vertically
from
the
exhaust
manifold,
made
a
90'
bend,
and
was
connected
to
the
inlet
of
the
catalyst.
A
120inch
diameter
pipe
connected
the
outlet
of
the
catalyst
to
a
back
pressure
valve,
and
then
to
the
exhaust
header.
EECL
personnel
used
two
sampling
ports
to
extract
samples
for
analysis
by
FTIRS
and
CEMS.
One
port
was
located
before
the
catalyst
and
one
port
was
located
after
the
catalyst.

Final
Report
Waukesha
3
52
1
GL
4­
l
September
200
1
22'­
1
I"

­.,­
Back
Pressure
Valve
3
Post
Catalyst
Sampling
Port
'
1:
=
m
x
7
Pre
Catalyst
Sampling
Port
Expansion
Joint
Figure
4.1
Exhaust
Piping
Schematic
Final
Report
Waukesha
352
1
GL
4­
2
September
200
1
5.0
SAMPLING
AND
ANALYSIS
METHODS
This
section
discusses
the
various
sampling
and
analysis
methods
employed
by
PES
and
EECL
to
quantify
the
HAP
emission
rates
upstream
and
downstream
of
the
oxidation
catalyst.
PES
selected
the
sampling
and
analysis
procedures
that
would
provide
the
information
required
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
EPA's
Emissions
Standards
Division
(
ESD).

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

PES
and
EECL
used
QA
and
calibration
procedures
described
in
40
CFR
60,
Appendix
A
(
or
other
references
as
appropriate)
as
a
guideline
for
instrument
calibrations
and
drift
checks.
The
instrumental
methods
as
written
in
40
CFR
60
Appendix
A
are
designed
by
EPA
to
be
portable,
field
test
procedures.
Because
these
instruments
are
maintained
in
a
laboratory­
type
environment
(
the
control
room
at
EECL),
fewer
QA
activities
and
calibrations
are
required
to
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
series.

5.1
DETERMINATION
OF
STACK
GAS
VOLUMETRIC
FLOW
RATE
PES
used
EPA
Method
19
to
calculate
the
volumetric
flow
rate
of
the
exhaust
gases
for
Runs
1
through
16.
The
mass
flow
rates
of
pollutants
measured
by
FTLRS
and
CEMS
were
calculated
using
the
Method
19
flow
data.
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
upper
heating
value
and
oxygen­
based
F­
factor,
Fd.
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
upstream
and
downstream
of
the
catalyst
for
each
run.

Final
Report
Waukesha
352
1GL
5­
l
*
September
200
I
TABLE
5.1
SUMMARY
OF
SAMPLING
AND
ANALYSIS
METHODS
Parameter
Test
Method
I
Measurement
Principle
Volumetric
Flow
I
EPA
Method
19
I
Stoichiometry
Oxygen
and
Carbon
Dioxide
EPA
Method
3A
Paramagnetic
and
Non­
dispersive
Infrared
Analyzers
Moisture
GRI
Protocol'

Carbon
Balance2
FTIR
Analyzer
Stoichiometry
Nitrogen
Oxides
I
EPA
Method
7E
I
Chemihuninescent
Analyzer
~~
Carbon
Monoxide
I
EPA
Method
10
I
GFCMDIR
Analyzer
Formaldehyde,
Acetaldehyde,
Acrolein
Methane
GRI
Protocol
FTIRAnalyzer
'

EPA
Method
25A
(
modified)
GC­
FID
Analyzer
Non­
methane
hydrocarbons
1
EPA
Method
25A
(
modified)
1
GC­
FID
Analyzer
Total
Hvdrocarbons
I
EPA
Method
25A
I
FID
Analyzer
'
Measurement
of
Select
Hazardous
Air
Pollutants,
Criteria
Pollutants,
and
Moisture
Using
Fourier
Transform
Infrared
(
FTIR)
Spectroscopy.
Presented
as
an
Appendix
to
Fourier
Transform
Infrared
(
FTIR)
Method
Validation
at
a
Natural
Gas­
Fired
Internal
Combustion
Engine
(
GRI­
95/
027
l),
Gas
Research
Institute,
December
1995:

2
Derivation
of
General
Equation
for
Obtaining
Engine
Exhaust
Emissions
on
a
Mass
Basis
Using
the
"
Total
Carbon"
Method.

Final
Report
Waukesha
352
1
GL
5­
2
September
2001
5.2
DETERMINATION
OF
STACK
GAS
OXYGEN
AND
CARBON
DIOXIDE
CONTENT
EECL
used
EPA
Method
3A,
Determination
of
Oxygen
and
Carbon
Dioxide
Concentrations
in
Emissions
from
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
locations
upstream
and
downstream
of
the
catalyst.
Each
sample
was
conditioned
to
remove
moisture
and
entrained
particulate
matter
and
directed
to
the
CEMS..
Oxygen
was
measured
using
the
paramagnetic
detection
principle.
Carbon
dioxide
was
measured
using
non­
dispersive
infrared
(
NDIR)
analyzers.
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
FTIIWCEMS
sampling
and
analysis
system
is
presented
in
Figure
5.1.

5.3
DETERMINATION
OF
STACK
GAS
MOISTURE
CON'tENT
EECL
used
two
methods
to
determine
the
moisture
concentration
in
the
exhaust
gas
upstream
and
downstream
of
the
catalyst.
Moisture
content
downstream
of
the
catalyst
was
calculated
using
a
carbon
balance
method.
This
method
is
summarized
in
the
EECL
test
report,
which
may
be
found
in
Appendix
A.
During
the
testing,
EECL
personnel
determined
that
the
moisture
concentrations
downstream
of
the
catalyst,
as
measured
by
the
Nicolet
Magna
560
FTIR
analyzer,
were
about
6
percent
higher
than
actual.
Therefore
the
carbon
balance
method
was
used
to
calculate
moisture
content
at
this
location.

EECL
used
methodology
described
in
the
document
Measurement
of
Select
Hazardous
Air
Pollutants,
Criteria
Pollutants,
and
Moisture
Using
Fourier
Transform
Infrared
(
FTIR)
Spectroscopy
to
measure
moisture
concentrations
upstream
of
the
catalyst.
This
document,
called
the
GRI
Protocol
in
this
report,
is
presented
as
Appendix
B
of
a
report
published
by
the
Gas
Research
Institute:
Fourier
Transform
Infiared
(
FUR)
Method
Validation
at
a
Natural
Gas­
Fired
Internal
Combustion
Engine.
A
sample
of
the
gas
was
extracted
from
the
exhaust,
filtered
and
directed
to
a
Nicolet
Rega
7000
FTIR
analyzer
to
measure
the
moisture
concentration.
The
gas
sample
was
transported
to
the
analyzer
via
a
heated
Teflon@
sample
line.
Further
discussion
of
the
FTIRS
sampling
and
analysis
method
may
be
found
in
the
report
generated
by
the
EECL.

5.4
DETERMINATION
OF
NITROGEN
OXIDES
EPA
Method
7E,
Determination
of
Nitrogen
Oxide
Emissions
from
Stationary
Sources
(
Instrumental
Analyzer
Procedure),
determined
nitrogen
oxide
content
of
the
exhaust
gases.

Final
Report
Waukesha
352
1
GL
5­
3
September
200
1
5.7
DETERMINATION
OF
METHANE
AND
NON­
METHANE
HYDROCARBONS
A
modification
of
EPA
Method
25A
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
Questar
Baseline
1030H
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
each
week
of
testing
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
DETERMINATIOllJ
OF
GASEOUS
ORGANIC
HAPS
USING
FTIRS
EECL
used
two
FTIRS
systems
that
met
the
sampling
and
analysis
requirements
set
forth
in
the
GRI
Protocol.
Table
5.2
summarizes
the
specifications
of
each
FTIRS
analyzer.
GRI
validated
extractive
FTIRS
systems
successfully
for
an
analysis
of
emissions
from
natural
gas­
fired
RICE.
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
re­
analysis.

The
sampling
and
measurement
system
consists
of
the
following
components:

l
heated
probe;

l
heated
filter;

l
heat­
traced
Teflon@
sample
line;

l
Teflon@
coated,
heated­
head
sample
pump;

l
FTIRS
spectrometer;
and
l
QA/
QC
.
apparatus.

Final
Report
Waukesha
352
1
GL
5­
6
September
200
1
TABLE
5.2
FTIRS
ANALYZER
SPECIFICATIONS
I
Parameter
I
Pre­
catalyst
I
Post­
catalyst
II
Manufacturer
and
Type
I
Nicolet
Rega
7000
Nicolet
Magna
560
Spectral
Resolution
0.5
cm
­
l
0.5
cm
­
l
Detector
Type
MCT­
A
MCT­
A
Cell
Type
4.2
Meter
­
Fixed
Path
Length
2.0
Meter
­
Fixed
Path
Length
II
Cell
Temperature
I
185
"
C
I
165
`
C
II
Cell
Pressure
I
600
Torr
I
600
Ton:

11
Cell
Window
Material
1
Zinc
Cellinide
I
KBr
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
spectrometer
collected
one
complete
spectrum
of
the
sample,
as
an
interferogram,
per
second
and
averaged
interferograms
over
l­
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
NATURAL
GAS
COMPOSITION
The
composition
of
the
fuel
gas
combusted
in
the
engine
was
determined
daily
using
a
dedicated
Daniels
Industries
gas
chromatograph
(
GC).
The
GC
was
calibrated
each
day
using
a
Final
Report
Waukesha
352
1GL
5­
7
September
2001
known
standard.
From
the
results
of
the
daily
analyses,
the
specific
gravity,
mole
fractions,
and
BTU
content
of
the
fuel
were
calculated.
Fuel
flow
measurements
were
determined
using
an
American
Gas
Association
(
AGA)
specified
orifice
meter
run
equipped
with
high
accuracy
pressure
and
temperature
transmitters.
All
fuel
flow
calculations
were
in
accordance
with
AGA
Report
#
3.
Stoichiometric
air
to
fuel
ratio
calculations
were
also
made
using
the
results
of
the
fuel
gas
analysis.
The
results
of
fuel
gas
calibrations
and
analysis
are
presented
in
the
EECL
report,
which
is
attached
in
Appendix
A.
.

Final
Report
Waukesha
352
1
GL
5­
8
September
200
1
6.0
QUALITY
ASSURANCE/
QUALITY
CONTROL
PROCEDURES
AND
RESULTS
Summarized
in
this
section
are
the
specific
QA/
QC
procedures
that
PES
and
EECL
personnel
employed
during
the
performance
of
this
source
testing
program.
PIES'
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/
6OO/
R­
94/
03&
z,
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
instruments
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/
027
1
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
Corporation
for
the
Gas
Research
Institute.

"
Protocol
for
Performing
Extractive
FTIRS
Measurements
to
Characterize
Various
Gas
Industry
Sources
for
Air
Toxics"
­
Prepared
by
Radian
Corporation
for
the
Gas
Research
Institute.

Final
Report
Waukesha
352
1
GL
6­
1
September
200
1
6.1.1
FTIRS
Svstem
PreDaration
Both
FTIRS
sampling
systems
(
before
and
after
the
catalyst)
were
subjected
to
an
EPA
Method
3
0
1
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.
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
with
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.
The
rotameter
monitored
the
flow
rate.
If
the
leakage
rate
is
found
to
be
no
greater
than
500
ml/
min
or
4%
of
the
average
sampling
rate
(
whichever
is
less)
the
system
is
considered
to
be
acceptable
for
use.
The
leak
checks
conducted
by
EECL
personnel
indicated
that
the
system
was
acceptable
for
use.

3.
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
leakage
rate
of
the
measurement
cell
is
acceptable
if
it
is
less
than
10
Torr
per
minute.
All
analyzer
leak
checks
performed
by
EECL
were
within
the
acceptable
range.

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
Waukesha
352
1
GL
6­
2
September
200
1
6.1.2
FTIRS
Daily
Calibrations
and
OA
Procedures
Before
each
day
of
testing,
EECL
personnel
calibrated
each
FTIRS
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.

d)
The
instrument
alignment
procedure
was
done.

e>
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
for
5
minutes.
The
accuracy
and
precision
of
each
instrument
were
calculated.
The
pass/
fail
criterion
for
accuracy
and
precision
was
10%
of
the
concentration
of
the
standard
Final
Report
Waukesha
3
52
1
GL
6­
3
September
200
1
gas.
A
second
diagnostic
standard
consisting
of
a
blend
of
CO,,
CO,
CHI,
and
NOx
was
analyzed
using
the
same
procedure.
Each
instrument
met
the
precision
and
accuracy
requirements.
Analyzer
diagnostic
data
are
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
5minute
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
Corporation
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
is
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
Background
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
test
program,
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
30
1
"
Field
Validation
of
Pollutant
Measurement
Methods
from
Various
Waste
Media."
The
validation
was
conducted
by
spiking
the
sample
gas
with
known
concentrations
of
formaldehyde,
acrolein,
and
acetaldehyde.
The
carrier
gas
was
a
mixture
of
acrolein
and
acetaldehyde
and
was
introduced
into
the
spiking
system
from
a
compressed
gas
cylinder.
Formaldehyde
was
added
to
the
Final
Report
Waukesha
352
1GL
6­
4
September
200
1
carrier
gas
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
Analmer
Calibration
Gases
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
ResDonse
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
Methane/
Non­
methane
hydrocarbon
GC)
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
analyzers
have
a
minimum
cycle
time
of
450
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.

Final
Report
WauJcesha
352
1GL
6­
5
September
200
1
TABLE
6.1
DETECTION
LIMITS
OF
Fl'IRS
AND
CEMS
COMPOUNDS
mplbhphr­
miNlgmmsperbmkehonepovwrhou
6.2.3
Analvzer
Calibrations
Zero
and
mid­
level
span
calibration
procedures
were
done
on
the
CO,
CO*,
02,
NO,,
THC,
and
methane/
non­
methane
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
directly
to
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:

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
f5na.
l
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,
COZ,
02,
NO,,
THC,
and
methane/
non­
methane,
and
analyzers.
The
zero
and
span
drift
checks
for
all
test
points
and
all
test
days
were
less
than
=
t2.0%
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.

Final
Report
Waukesha
352
1GL
6­
7
September
2001
TABLE
6.1
DETECTION
LIMITS
OF
FIIRS
AND
CEMS
COMPOUNDS
Fonnaidehyde
Acetaldehyde
mgibhp­
hr
432
.
472
.
465
*
542
mlblhr
7.02
5.37
4.42
388
5:
27
s:
a1
406
.
463
5:
29
448
6109
421
.
404
.
4.08
412
6.59
8.70
482
6.65
6.57
6.64
7:
63
418
6:
60
442
'
432
7.19
7:
03
mglbhp­
hr
201
32:
7
219
24:
9
216
SO:
6
24:
6
181
250
40:
6
189
30:
6
214
24:
3
208
19/
188
191
192
mlb/
hr
28:
3
32:
0
30:
6
31:
o
31:
2
zb
36:
5
195
31:
7
206
33:
5
201
32:
7
Acrolein
mglbtphr
15.5
163
.
166
'
142
.
114
28:
3
15
­
/
164
'
149
20:
3
136
'
153
24:
9
155
25:
3
134
25:
1
169
153
*
166
163
*
mlb/
hr
25.2
18.5
15.8
19.3
25.5
18.7
25.7
30:
7
24.9
27:
l
26.4
NitrogenOxides
(
as
N02)
ww­
m
O.
Wl
0.001
0.001
0.001
0.001
0.001
0.001
O.
Wl
0.001
oxxn
ClsRn
0.001
0.001
O.
Wl
o.
ouT
lb/
hr
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
CarbonMonoxide
amhbnr
w­­­­
l­
.
.
.
.
0.05
004
.
004
.
0.05
004
004
004
004
005
.
otJ4
­
005
*
(
105
*
lblhr
0.07
0.05
0.04
0.08
0.06
0.07
0.05
0.06
0:
07
0:
07
0:
07
0:
07
0.08
0.07
0.08
0.07
Methane
WJp­
hr
.
.
0.2
.
.
.
.
.
0.1
0.1
0.1
0.1
.
.
.
.
IWhr
0.2
0.2
0.1
0.2
0.3
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.3
0.2
0.2
0.2
~
NorrmethanewdocaM
glbhp­
hr
0.04
0.04
0.04
0.04
0.04
0.
w
0.04
0.04
0.04
0.04
­­­
­
­*
­"­
­
0.04
Total
Hydrocarbons
Formaldehyde
Acetaldehyde
Mu
0.07
0.05
0.04
0.05
0.07
0.06
0.04
0.05
0.08
0.08
iii
;
I:;
i::
ii
KG
0.07
lhh­
Q­
w­~
Ib/
hr
0.0002
0.0001
0.0001
0.0001
0.0002
0.0002
0.0001
0.0001
0.0002
0.0002
0.0002
0.0002
0.0002
0.0002
0.0002
0.0002
Catalyst
Outlet
mg/
bhp­
hr
3.64
3
.
14
3
33
2
.
86
2.85
3
19
2
99
3
11
3.04
mlWhr
4.94
3.57
3:
17
2.91
342
.
2.80
2
88
3.20
3.03
2.86
3.95
5.57
4.56
3:
28
4.36
4.92
4.64
4.66
4.63
5119
4186
5:
05
4.95
mglbhphr
12.6
129
137
13:
l
120
*
14.2
116
18:
8
119
.
132
.
lz5
II8
mlb/
hr
20.4
14:
7
16.3
23.0
13.5
18.0
20:
3
19:
2
118
19:
3
`
I8
.
132
lz3
lz8
.
12­
6
19.1
21:
4
20:
o
20.8
20.4
Acrolein
mg/
bhp+
I
r
59.1
524
*
540
51:
4
490
66:
s
581
91:
1
510
mlb/
hr
97
59.6
82:
8
506
57:
s
50­
1
52l
.
50*
8
b0o
81:
4
502
81:
6
584
91:
7
523
68.9
84.6
82.6
85:
O
543
66:
3
534
66:
8
0001
0001
0001
0001
0001
owl
owl
.
0001
0001
owl
0001
0001
0001
*
0001
NittqenOxides(
as
NO2)
cmho­
tu
001
w­­­­
l­
­­
It&
0.001
0:
001
0:
001
Ok1
0:
OOl
0:
001
0:
001
0:
001
0.001
0:
001
0:
tm
0:
001
0:
001
0:
m
0.001
0:
OOl
001
0:
01
001
0:
01
001
0:
01
001
0:
Ol
001
io2
001
001
001
061
001
061
CarbonMonoxide
Whr
0:
01
0:
Ol
0:
01
0:
01
0:
01
0:
01
001
0:
01
001
0:
02
001
0:
01
001
0:
02
001
0:
01
glm+
r
014
0.16
0.14
014
015
015
*
O­
14
014
0:
23
O­
14
016
0:
25
014
0:
23
015
0:
24
013
Methane
Ib/
hr
0:
19
0.26
0.22
0:
16
0:
20
0.24
0.23
0.23
0:
24
wpty
004
0:
07
004
.
004
.
0.04
ocJ4
004
004
oofJ
.
004
.
004
.
0­
W
004
.
004
004
004
004
NonmethaneHydocarbons
IWhr
0.05
0.04
0.05
0:
07
0:
05
0:
w
0.05
0.08
0.08
0.08
0.06
0:
07
0:
08
0:
07
0:
07
TotalHydrocarbons
0morwwnm1
oum
oam
omm
oo?
m
Oa.
Kll
o?
Jm
OolJm
IJoiJol
lJ­
om1
l,
J­
mol
Owol
o­
aJm
o(
Jmi­
.
.
.
.
.
.
.
.
ii
o:
ooo2
.0:
wo,
.
.
.
.
o.
cHm
0.0001
0.0002
0.0002
0.0001
0.0001
0.0002
o.
wO2
0.0002
0.0002
0.0002
0.0002
0:
0002
0:
0002
mgb#
yhr­
miiigmmsperbrabhomepomrhou
6.2.3
Analvzer
Calibrations
Zero
and
mid­
level
span
calibration
procedures
were
done
on
the
CO,
CO*,
02,
NO,,
THC,
and
methane/
non­
methane
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
directly
to
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:

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,
COz,
OS,
NO,
THC,
and
methane/
non­
methane,
and
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.

Final
Report
Waukesha
352
1GL
6­
7
September
2001
TABLE
6.2
TYPES
AND
FREQUENCIES
OF
CEMS
ANALYZER
CALIBRATIONS
Calibration
Type
Gas
Calibration
Gas
Calibrant
Concentration
(
units
Frequency
Injection
Validation
of
%
of
span
(`
1)
Point
Criterion
ACE
(*
I
02,
co*,
co,
NO,

MethaneMon­
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
c2%
of
analyzer
span
for
each
gas
Before
each
Directly
into
engine
test
the
analyzer
<
5%
of
respective
cal.
gas
value
ZSD
(
3)
02,
co*,
co,
NO,

Methane/
Non­
Methane
Hydrocarbons
0
to
0.25,
40
to
60
or
80
to
100
w
25
to
35,
45
to
55
Before
and
after
each
test
run
Directly
into
the
analyzer
All
errors
<
3%
of
span
All
errors
<
3%
of
span
SSB
(`
I
NO,

Methane/
Non­
Methane
Hydrocarbons
0
to
0.25,
40
to
60
or
80
to
90
(`
1
0
to
0.25,
25
to
35,
45
to
55
or
80
to
90
(`)
Before
and
after
each
Both
directly
test
day
into
the
analyzer
and
Both
errors
into
the
inlet
<
5%
of
Before
and
analyzer
spar
after
each
of
the
sample
line
test
day
(*
I
­
The
span
must
be
1.5
to
2.5
the
concentration
expected
for
each
pollutant
(*
I
­
Analyzer
calibration
error
check
t3)
­
Zero
and
span
drift
check
t4)
­
Sampling
system
bias
check
@)
­
Whichever
is
closer
to
the
exhaust
gas
concentration
Final
Report
Waukesha
352
1
GL
6­
8
September
200
1
6.2.4
Analvzer
Linearitv
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
Tf52.0%
of
the
analyzer
span
value.
The
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
G.
O%
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
(
NO3
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
(
NO,
J
analyzer
until
a
stable
response
was
recorded.
The
converter
was
considered
acceptable
if
the
instrument
response
indicated
a
90
percent
or
greater
NO2
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
Final
Report
Waukesha
352
1GL
6­
9
September
2001
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
Samde
Line
Inteeritv
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
NO,
analyzer.
The
analyzer
was
allowed
to
stabilize
and
the
response
recorded.
The
same
mid­
level
calibration
gas
was
then
introduced
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
&
5%
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
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
OZ.
The
percent
difference
between
the
actual
and
theoretical
O2
measurements
was
within
*
5
%
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
Factor
Oualitv
Assurance
Checks
Besides
the
CEMS
calibration
and
QC
checks,
carbon
dioxide
and
oxygen
measurements
were
validated
by
calculating
the
fuel
factor,
F,,
using
the
following
equation:

Final
Report
Waukesha
352
1
GL
6­
10
September
200
1
F,'
20.9
­%
02
%
CO,

The
values
of
F,
at
the
inlet
and
the
outlet
for
each
sampling
run
are
presented
in
Table
6.3.
From
the
fuel
gas
analysis,
the
calculated
F,
was
1.69,
1.68,
and
1.70
for
August
4,5,
and
6
respectively.
The
F,
values
were
within
6%
of
the
calculated
F,
for
all
of
the
sampling
runs
conducted.
Based
upon
the
results,
the
integrity
of
the
CEMS
sample
stream
was
not
compromised
due
to
leaks
in
the
sampling
system.

TABLE
6.3
SUMMARY
OF
FUEL
FACTOR
VALUES
Run
Number
1
2
3
4
5
6
7
8
Inlet
F,
Outlet
F,

1.76
1.72
1.78
1.72
1.71
1.73
1.78
1.73
1.77
1.76
1.77
1.74
1.76
1.74
1.77
1.74
Run
Number
Inlet
F,
Outlet
F,

9
1.77
1.72
10
1.75
1.72
11
1.76
1.72
*
12
1.75
1.72
13
1.67.
1.74
1
14
1.78
1.72
15
1.77
1.74
16
1.78
1.75
Final
Report
Waukesha
352
1
GL
6­
11
September
2001
6.2.10
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
I
Inlet
Detection
I
Outlet
Detection
Limit
Limit
Oxygen
I
~~~
0.01
%
volume
I
0.01
%
volume
Carbon
Dioxide
I
0.25
%
volume
I
0.1
%
volume
Nitrogen
Oxides
I
0.1
ppm
I
0.1
ppm
Carbon
Monoxide
I
2
PPm
I
2
PPm
Methane
­
7
20
PPm
I
20
PPm
,
Non­
methane
Hydrocarbons
I
2
PPm
I
2
PPm
Total
Hydrocarbons
0.04
ppm
0.04
ppm
6.3
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.

Final
Report
Waukesha
352
1GL
6­
12
September
200
1
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:

0
the
effectiveness
of
after­
combustion
control
systems
on
hYP
emissions,
and
0
the
effectiveness
of
combustion
modifications
(
engine
operating
parameters)
on
h54P
emissions.

EPA
then
developed
a
decision
statement.
The
decision
statement
defined
the
process
that
would
be
used
to
answers&
he
stated
problem.
The
decision
statement
is
restated
below:
h*

IfEPA
can
ident@
a
range
of
engine
operating
conditions
for
a
defined
set
of
engines
with
specified
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
effectiveness
of
after­
combustion
and
combustion
modifications.
In
addition,
EPA
can
obtain
information
on
I­
UP
emissions
throughout
the
engine
operating
range.

PES
and
EECL
conducted
the
test
program
on
the
Waukesha
3521GL,
natural
gas­
fired,
4­
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.

During
conduct
of
the
test
program,
there
were
deviations
from
the
QAPP.
These
deviations
are
discussed
in
Section
3.0
for
deviations
in
engine
operation,
and
Section
5.0
for
deviations
in
Sampling
and
Analysis
procedures.

Table
6.5
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
HAP
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
Final
Report
Waukesha
3
52
1
GL
6­
13
September
2001
evaluated
at
a
range
of
engine
operating
conditions.
These
conditions
are
expected
to
simulate
the
range
of
engine
operating
conditions
in
industry.
Table
6.5
identifies
the
number
of
engine
parameters
that
were
within
the
tolerances
prescribed
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
Waukesha
3
52
1
GL
engine
located
at
the
EECL.
When
testing
was
conducted
some
of
the
prescribed
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
­)
would
indicate
that
fewer
than
90%
of
the
method
performance
parameters
were
met.
There
were
no
cases
where
less
than
90%
of
the
method
performance
parameters
were
met.

Final
Report
Waukesha
352
1
GL
6­
14
September
200
1
TABLE
6.5
SUMMARY
OF
ENGINE
AND
METHOD
PERFORMANCE
FTlR
QA
Requirements
FTfRQARequirements
J+
J+
J+
J+
J+
J+
J+
J+
J+
J+
J+
J+
J+
J+
J+
J+

FIR
Detection
Limits
a
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
CEMSQARequirements
J+
J+
J+
J+
J+
J+
J+
J+
J+
J+
J+
J+
J+
J+
J+
J+

CEMS
Detection
Limits
a
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
Assessment
of
Data
QuaI@

l
Neither
FFRS
nor
CBvlS
detection
hits
were
specified
in
the
WFP
Final
Report
Waukesha
352
1GL
6­
15
September
200
1
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
2:
FOUR­
STROKE,
LEAN
BURN,
NATURAL
GAS
FIRED
INTERNAL
COMBUSTION
ENGINES
EMISSIONS
TESTING
OF
CONTROL
DEVICES
FOR
RECIPROCATING
INTERNAL
COMBUSTION
ENGINES
IN
SUPPORT
OF
REGULATORY
DEVELOPMENT
BYTHE
U.
S.
ENVIRONMENTAL
PROTECTION
AGENCY
(
EPA)

PHASE
2:
FOUR­
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
April
28,200O
Statement
of
Confidntiaiity
7Iis
report
has
been
submitted
for
the
sole
and
exclusive
use
of
Pacific
Environmental
Services,
and
shall
not
be
disclosed
or
provided
to
any
other
entity,
corporation,
or
third
part
for
purposes
beyond
the
specific
scope
or
intent
of
this
document
without
the
express
written
consent
of
Colorado
State
University.
TABLE
OF
CONTENTS
1.0
INTRODUCTION
1.1
Overview
1.2
Background
2.0
TEST
PROGRAM
2.1
Objective
2.2
Incentives
2.3
Work
Plan
3.0
DEVIATIONS
TO
TEST
PROGRAM
3.1
FTIR
Validation
3.2
FTIR
Post­
catalyst
Water
Analysis
3.3
Baseline
Engine
Operating
Conditions
3.4
Four­
Stroke
engine
Test
Matrix
4.0
TEST
SAMPLING
PROCEDURES
4.1
General
Test
Procedures
4.2
Test
Specifics­
Data
Collection
4.3
Test
Specifics­
Engine
Stability
4.4
Test
Specifics­
Data
Collection
Hardware
4.5
Test
Specifics­
Data
Collection
Process
4.6
Test
Specifics­
Emissions
Analyzer
General
Test
Procedures
4.7
Test
Specifics­
Emissions
Analyzer
Checks
and
Calibrations
4.8
Test
Specifics­
m
Calibration
Procedures
4.9
Test
Specifics­
FTIR
Validation
Procedures
4.10
Test
Specifics­
General
Calibration
4.11
Test
Specifics­
Test
Bed
General
Description
Statement
of
Conjidurtiality
This
report
has
been
submitted
for
the
sole
and
exclusive
use
of
Pacific
Environmental
Services,
and
shall
not
be
disclosed
or
provided
to
any
other
entity,
corporation,
or
third
part
for
putposes
bqond
the
s~
ijk
scope
or
intent
of
this
document
without
the
exprus
written
consent
of
Colorado
State
University
I
COLORADO
STATE
UNIVERSITY
Appendix
A
Appendix
B
Appendix
C
Appendix
D
Appendix
E
Appendix
F
Appendix
G
Appendix
H
Appendix
I
Appendix
J
Appendix
K
Appendix
L
Appendix
M
Appendix
N
Appendix
0
Appendix
P
Appendix
Q
Appendix
R
Appendix
S
Appendix
T
Appendix
U
Appendix
V
Appendix
W
Appendix
X
APPENDIX
Engine
Test
Data
Daily
Baseline
Data
Points
Test
Point
QC
Checks
Test
Points
Reference
Method
Analyzers
Calibrations
FTIR
Calibration
FTIR
Validation
Calibration
Gas
Certification
Sheets
Baseline
Methane/
Non­
Methane
Analyzer
Pressure
and
Temperature
Calibrations
Equipment
Certification
Sheets
Dynamometer
Calibration
Dynamometer
Calibration
Procedure
Gas
Analysis
Gas
Analysis
Calibrations
Gas
Analysis
Calculations
­
Fuel
Specific
F
Factor
Stoichiometric
Air/
Fuel
Calculations
Computing
Air/
Fuel
Ratio
from
Exhaust
Composition
"
An
Investigation
on
Inlet
Air
Humidity
Effects
on
a
Large­
Bore,
Two
Stroke
Natural
Gas
Fired
Engine"
"
Derivation
of
General
Equation
for
Obtaining
Engine
Exhaust
Emissions
on
a
Mass
Basis
Using
the
"
Total
Carbon"
Method"
Annubar
Flow
Calculations
Additional
Calculations
Exhaust
Piping
Schematic
Catalyst
Schematic
Statement
of
Confidentiality
This
report
has
been
submitted
for
the
sole
and
exclusive
use
of
Pa@
Environmental
Services,
and
shall
not
be
disclosed
or
provided
to
any
other
entity,
corporation,
or
third
part
for
purposes
beyond
the
specific
scope
or
intent
of
this
document
without
the
express
written
consent
of
Colorado
State
University.
1.0
INTRODUCTION
1.1
OVERVIEW
Natural
gas
fueled
and
diesel
fueled
reciprocating
engines
represent
a
large
portion
of
the
horsepower
in
operation
within
the
oil
and
gas
industry
and
power
generation
markets.
With
stringent
emissions
regulations
being
required
by
federal,
state,
and
local
agencies,
accurate
data
on
current
engine
emission
levels
and
development
of
new
technologies
to
reduce
and
control
emissions
levels
has
become
essential
for
federal
agencies,
engine
manufacturers,
and
equipment
operators.
Criteria
pollutants
and
Hazardous
Air
Pollutants
(
HAPS)
issues
are
of
major
concern
for
operators
of
both
two­
stroke
and
four­
stroke
engines.
Current
Environmental
Protection
Agency
(
EPA)
and
natural
gas
industry
funded
test
programs
are
directed
toward
evaluating
emission
levels
from
existing
engines,
determining
formation
mechanisms
for
the
exhaust
gas
constituents
of
interest,
and
developing
new
technologies
to
reduce
the
emissions
levels
of
these
constituents.

The
investigation
of
the
application
of
commercially
available
techniques
designed
to
address
the
HAPS
emissions
from
reciprocating
internal
combustion
engines
(
RICES)
will
allow
the
EPA
to
quantify
the
effectiveness
of
current
commercially
available
control
devices.
These
devices
have
been
identified
as
having
the
potential
to
reduce
HAPS
emissions
from
stationary
RICE
sources.
Information
gained
through
this
program
will
assist
the
EPA
in
the
regulatory
development
effort.

Accurate
information
on
emission
levels
from
operational
facilities
is
difficult
to
obtain.
Based
upon
a
recommendation
from
the
Internal
Combustion
Coordinating
Rulemaking
Committee
(
ICCR)
to
the
EPA,
a
series
of
tests
were
conducted
on
industrial
class
engines
at
Colorado
State
University's
Engines
&
Energy
Conversion
Laboratory.
Testing
was
being
conducted
on
both
two­
stroke
and
four­
stroke
natural
gas
engines
and
a
four­
stroke
diesel
engine.
The
test
program
for
four­
stroke,
lean
bum,
natural
gas
fueled
internal
combustion
engine
was
performed
during
Phase
Two
of
this
test
program.
The
results
of
Phase
Two
testing
are
contained
within
this
document.

1.2
BACKGROUND
The
1990
Amendments
to
the
Clean
Air
Act
include
provisions
that
significantly
impact
the
operation
of
stationary
reciprocating
internal
combustion
engines.
Of
the
ten
titles
to
these
amendments,
four
have
direct
bearing.
They
are
as
follows:

Title
I
­
Attainment
of
Air
Quality
Standards
Defines
ambient
air
quality
standards,
defines
non­
attainment
areas
based,
imposes
emissions
reductions
to
achieve
attainment
per
specified
timeline
per
reasonably
available
control
technology
(
RACT).

Emissions
Testing
l­
l
Pacific
Environmental
Services
Of
Control
Devices
for
Reciprocating
Internal
Combustion
Engines
In
Support
of
Regulatory
Development
By
the
U.
S.
EPA.
Title
III
­
Hazardous
Air
Pollutants
Defines
189
pollutants
classified
as
hazardous
air
pollutants
(
HAPS),
specifies
thresholds
in
tons
per
year
(
TPY)
for
any
one
of
these
pollutants
or
a
combination
of
these
compounds,
introduces
maximum
achievable
control
technology
(
MACT)
for
sources
triggering
thresholds.

Title
V
­
Operating
Permits
Imposes
requirement
to
obtain
federal
operating
permits
for
major
sources,
imposes
requirement
to
provide
annual
certification
of
compliance,
defines
emissions
fees
based
on
actual
emissions.

Title
VII
­
Enforcement
Establish
mechanisms
to
enhance
and
strengthen
enforcement
of
CM,
establishes
criminal
penalties,
gives
authority
to
issue
administrative
orders
(
fines
/
penalties)
without
going
to
federal
court
for
certain
violations.

The
EPA
in
conjunction
with
the
RICE
Work
Group
of
the
ICCR
process
determined
that
additional
emissions
data
for
HAPS
exhaust
gas
constituents
is
required
to
support
the
regulatory
development
process.
In
a
RICE
Emissions
Test
Plan
Document
dated
November
1997,
a
five
component
test
plan
to
acquire
additional
HAPS
emissions
test
data
was
set
forth.
The
five
components
include
the
following:

Engines,
Fuels,
and
Emissions
Controls
to
be
tested
Matrix
of
Operating
Conditions
to
be
tested
Pollutants
to
be
Measured
During
Testing
Test
Methods
to
Quantify
Emissions
Prioritization
Eight
HAPS
pollutants
are
included
in
the
test
plan.
These
compounds
are:
formaldehyde,
acetaldehyde,
acrolein,
the
BTEX
compounds
(
benzene,
toluene,
ethylbenzene,
xylene)
and
1­
3
butadiene.
Polyaromatic
hydrocarbon
(
PAH.)
compounds
were
measured
on
the
two­
stroke
lean
burn
gas
engine,
but
were
not
measured
during
the
test
program
for
the
four­
stroke
lean
burn
gas
engine.

Insight
gained
through
the
test
program
will
provide
information
on
the
engine
operating
conditions
that
affect
the
formation
/
reduction
mechanisms
of
HAPS.
The
investigation
of
the
application
of
commercially
available
techniques
designed
to
address
the
HAPS
emissions
from
RICES
will
allow
the
EPA
to
quantify
the
effectiveness
of
current
commercially
available
control
devices.
These
devices
have
been
identified
as
having
the
potential
to
reduce
HAPS
emissions
from
stationary
RICE
sources.
Information
gained
through
this
program
will
assist
the
EPA
in
the
regulatory
development
effort.

Emissions
Testing
1­
2
Pacific
Environmental
Services
Of
Control
Devices
for
Reciprocating
Internal
Combustion
Engines
In
Support
of
Regulatory
Development
By
the
U.
S.
EPA.
2.0
TEST
PROGRAM
2.1
OBJECTIVE
The
objective
of
this
program
is
to
evaluate
commercially
available
catalyst
technologies
which
have
been
identified
as
having
the
potential
to
control
both
formaldehyde
and
other
Hazardous
Air
Pollutants
(
HAPS)
as
well
as
existing
criteria
pollutants
from
reciprocating
internal
combustion
engines
(
RICE).
The
specific
internal
combustion
engine
class
tested
under
the
Phase
Two
test
program
was
the
four­
stroke,
lean
burn,
natural
gas
fueled
internal
combustion
engines.
The
catalyst
hardware
was
evaluated
according
to
the
16­
point
test
matrix
developed
by
the
EPA,
and
the
Reciprocating
Internal
Combustion
Engine
(
RICE)
Work
Group
of
the
ICCR
process.
Investigation
of
catalyst
performance
during
operation
at
various
engine
operating
conditions
provides
insight
into
the
effectiveness
of
catalysts
at
various
conditions.
The
information
gained
through
the
test
program
will
assist
the
EPA
in
regulatory
development
efforts
for
control
of
HAPS
emissions
and
criteria
pollutants
from
RICE
sources.

2.2
INCENTIVES
Title
III
of
the
1990
Clean
Air
Act
Amendment
requires
the
development
of
Maximum
Achievable
Control
Technology
(
MACT)
standards
for
major
sources
of
Hazardous
Air
Pollutants
(
HAPS)
emissions.
A
MACT
major
source
is
defined
as
one
that
emits
greater
than
10
tons
per
year
of
any
single
HAP
or
25
tons
per
year
for
all
HAPS.
For
most
source
categories
(
RICE
included),
the
MACT
standards
will
require
existing
major
sources
apply
HAPS
emissions
control
technologies
that
reduce
emissions
to
a
level
achieved
by
the
best
performing
existing
sources.
In
some
cases,
depending
upon
the
cost
of
the
control
technology
and
the
amount
and
toxicity
of
the
HAPS
removed,
more
stringent
standards
may
be
set.
The
MACT
standards
for
RICES
are
scheduled
to
be
promulgated
by
the
year
2000.

Of
the
HAPS
listed,
the
EPA
in
conjunction
with
the
Internal
Combustion
Coordinating
Rulemaking
Committee
(
ICCR)
have
identified
compounds
which
may
be
present
in
the
exhaust
of
reciprocating
internal
combustion
engines.
Existing
test
data
indicates
that
the
only
HAP
present
in
the
exhaust
of
RICES
at
levels
approaching
10
tons
per
year
is
formaldehyde.
Commercially
available
after­
treatment
technologies
(
catalysts)
for
the
control
of
organic
compound
emissions
are
currently
in
operation
on
RICES.
These
technologies
have
demonstrated
performance
for
control
of
volatile
organic
compounds
(
VOCs)
and
products
of
incomplete
combustion.
However,
there
is
limited
information
on
the
effectiveness
of
these
technologies
for
reducing
organic
HAPS
emissions.
Determining
the
effectiveness
and
longevity
of
exhaust
catalyst
will
aid
the
EPA
in
evaluating
current
technologies
for
control
of
HAPS
emissions
from
RICE
sources
as
well
as
provide
information
in
support
of
regulatory
development
by
the
EPA
for
these
sources.

Emissions
Testing
2­
1
Pacific
Environmental
Services
Of
Control
Devices
for
Reciprocating
Internal
Combustion
Engines
In
Support
of
Regulatory
Development
By
the
U.
S.
EPA.
COLORADOSTATEUNIVERSITY
2.3
WORK
PLAN
Pacific
Environmental
Services
(
PES)
serves
as
the
prime
contractor
responsible
for
providing
information
to
the
EPA.
CSU
is
a
subcontractor
to
PES.
Testing
was
conducted
at
the
Colorado
State
University
`
s
Engines
and
Energy
Conversion
Laboratory.
The
engine
and
catalyst
type
tested
is
described
in
Table
1.

TABLE
1
ENGINE
AND
CATALYST
TYPE
Engine
Classification
Four­
Stroke,
Lean
Bum,
Natural
Gas
Fueled
Manufacturer
and
type
Waukesha
3521
GL
I
Number
of
Cylinders
I
6
I
Bore
and
Stroke
Engine
Speed
Ignition
System
Classification
9.375"
X
8.5"
1200
RPM
Spark
Ignited
Precombustion
Chamber
Ignition
System
Precombustion
Chamber
Type
­­
Number
of
Precombustion
Chambers
Catalyst
Classification
Altronic
Standard
OEM
Product
1
Per
Cylinder
Oxidation
Type
Manufacturer
Element
Size
Number
of
Elements
Miratech
Corporation
12"
xl6"
x3"
2
I
Substrate
Stainless
steel,
alternating
corrugated
&
flat
layers
I
The
test
matrix
as
originally
defined
is
presented
in
Table
2,
with
engine
baseline
conditions
shown
in
Table
3.
Deviations
from
the
described
test
conditions
are
detailed
in
Section
3
of
this
report.
Each
test
point
consisted
of
collecting
thirty­
three
minutes
of
data.
The
raw
data
was
averaged
into
thirty­
three
one­
minute
data
points.
The
data
points
were
then
averaged
to
provide
the
results
for
the
single
test
point.
The
results
are
presented
in
tabular
form
in
Appendix
A
of
this
report.

Emissions
Testing
2­
2
Pacific
Environmental
Services
Of
Control
Devices
for
Reciprocating
Internal
Combustion
Engines
In
Support
of
Regulatory
Development
By
the
U.
S.
EPA.
COLORADOSTATEUNIVERSITY
TABLE
2
ENGINE
OPERATING
CONDITIONS
DURING
TESTING
WAUKESHA
3521
GL
(
4­
STROKE
LEAN
BURN,
NATURAL­
GAS­
FIRED)
US
EPA
ICCR
RICE
HAP
EMISSIONS
TESTING
Operating
Torque
I
ntercooler
Jacket
Conditions
to
Speed
(%
of
Air­
to­
Fuel
Water
Water
be
Tested:
(
rpm)
baseline)
Ratio
Timing
Temp.
Temp.

?
un
1
H
H
N
S
S
S
tun
2
H
L
N
S
S
S
?
un
3
L
L
/
N
S
S
S
?
un
4
L
H
N
S
S
S
?
un
5
H
H
L
S
S
S
?
un
6
H
H
H
S
S
S
Qn
7
H
L
H
S
S
S
?
un
a
L
H
L
S
S
S
?
un
9
H
H
N
S
L
S
Xun
10
H
H
N
S
H
S
Run
II
H
,
H
N
S
S
L
Run
12
H
H
N
S
S
H
,

Run
13
H
H
N
L
S
S
Run
14
H
H
N
H
S
S
Run
15
H
H
N
S
S
S
Run
16
H
H
N
S
S
S
L=
1000
L
=
70
N
=
0.61
s=
IO
s=
130
s=
lao
H
=
1200
H=
lOO
(
9.8
%
02)
L
=
6
L=
12Q
L=
17u
L
=
0.56
H=
14
H
=
140
H=
19C
(
10.7
%
02)
H
=
0.62
(
a.
9
%
02)

l
Note:
Air/
he1
ratio
calculations
based
on
Appendix
Q
and
Appendix
R.

Emissions
Testing
2­
3
Pacific
Environmental
Services
Of
Control
Devices
for
Reciprocating
Internal
Combustion
Engines
In
Support
of
Regulatory
Development
By
the
U.
S.
EPA.
COLOMDOSTATEUNIVERSITY
TABLE
3
WAUKESHA
3521
GL
BASELINE
CONDITIONS
1
Engine
Operating
Parameters
Nominal
Value
Acceptable
Range
Designation
1
4
Engine
Torque
3383
ft­
lb.
to
3230
+
2%
of
value
Primary
ft­
lb.

1
Engine
Speed
I
1000
RPM
I
*
5%
of
value
I
Primary
I
1
Jacket
Water
Temperature
Outlet
(
180
°
F
I
*
5%
of
value
1
Primary
I
1
Engine
Oil
Temperature
Header
I
180
°
F
I
*
5%
of
value
I
Primary
I
I­­
~~~~
Arr
Manifold
Temperature
I
85
°
F
to
130
°
F
I
f
5%
of
value
I
Primary
I
30.0"
Hg
above
Atm.
+
5%
of
value
Primary
I
Exhaust
Manifold
Pressure
I
5.0"
Hg
below
AMP
I
f
5%
of
value
I
Primary
I
i­

I
Overall
Air:
Fuel
Ratio
Ignition
Timing
I
I
28:
l
1O"
BTDC
I
I
+
5%
of
value
+
5%
of
value
I
I
Primary
Primary
I
I
Inlet
Air
Humidity­
Absolute
Engine
Fuel
Flow
SCFH
.0015
lb
HzO/
lb
Air
+
10%
of
value
Primary
4460
to
4360
f
5%
of
value
Primary
SCFH
Engine
Oil
Pressure
Inlet
Inlet
Air
Flow
.

Average
Engine
Exhaust
Temperature
45
lb.
+,
10%
of
value
Secondary
1400­
1500
SCFM
f
5%
of
value
Secondary
660"
F
*
5%
of
value
Secondary
Emissions
Testing
2­
4
Pacific
Environmental
Services
Of
Control
Devices
for
Reciprocating
Internal
Combustion
Engines
In
Support
of
Regulatory
Development
By
the
U.
S.
EPA.
Cotou~
o
STATE
WNERTITY
3.0
DEVIATIONS
TO
TEST
PROGRAM
Testing
on
the
four­
stroke,
lean
burn,
natural
gas
fired
IC
engine
was
originally
scheduled
for
mid­
June,
but
a
main
bearing
failure
occurred
on
June
16.
The
engine
was
rebuilt
by
Pamco
/
Stewart
&
Stevenson
under
Waukesha's
direction.
The
actual
test
program
was
conducted
between
August
4,
1999
and
August
6,
1999.
Prior
to
initiation
of
the
16
point
test
matrix,
a
validation
procedure
was
performed
on
the
two
FTIR
analyzers.
The
analyzers
were
validated
for
formaldehyde,
acrolein,
and
acetaldehyde.
Modifications
to
the
baseline
engine
operating
conditions
were
made
prior
to
the
beginning
of
the
test
matrix.
The
variances
from
the
original
test
program
are
described
below:

3.1
FTIR
VALIDATION
A
validation
procedure
was
performed
on
the
two
FTIR
analyzers.
Eastern
Research
Group
(
ERG)
performed
the
validation
procedures
on
March
30,1999.
The
validation
procedure
was
conducted
in
basic
accordance
with
procedures
outlined
in
EPA
Method
301­"
Field
Validation
of
Pollutant
Measurement
Methods
from
Various
Waste
Media".
Validation
procedures
for
aldehydes
utilized
an
analyte
spiking
technique
as
specified
in
Method
301.
Deviations
from
the
described
procedures
are
as
follows:

Analyte
Spiking:
The
validation
for
the
target
aldehyde
compounds
was
carried
out
by
means
of
dynamic
analyte
spiking
of
the
sample
gas.
The
sample
stream
of
the
exhaust
gas
was
spiked
with
all
of
the
specific
analytes
simultaneously.
This
change
had
no
impact
on
the
test
procedure
or
results.

Formaldehyde:

Formaldehyde
spike
gas
was
generated
by
volatilization
of
a
formalin
solution
prepared
from
a
stock
formalin
solution
of
37%
formaldehyde
by
weight.
The
solution
was
vaporized
by
means
of
a
heated
vaporization
block.
The
vaporized
formalin
solution
then
mixed
with
a
carrier
gas
and
flowed
into
the
sample
exhaust
stream.
Carrier
gas
flow
rate
was
measured
by
a
mass
flow
meter
equipped
with
readout.
The
carrier
gas
was
to
be
Nitrogen;
however,
since
it
was
determined
to
perform
the
validation
process
for
all
aldehyde
compounds
simultaneously,
the
Acetaldehyde/
Acrolein
blend
calibration
gas
was
used
as
a
carrier
gas
for
the
vaporized
Formaldehyde.
No
impact
on
the
quality
of
the
validation
data
resulted
from
this
deviation
in
procedure.

Acetlyaldehyde/
Acrolein
Acetlyaldehyde
and
acrolein
spiked
samples
were
generated
from
a
certified
gas
standard
(
Scott
Specialty
Gases,
ti%
analytical
accuracy)
which
contain
both
analyte
species
and
Emissions
Testing
3­
1
Pacific
Environmental
Services
Of
Control
Devices
for
Reciprocating
Internal
Combustion
Engines
In
Support
of
Regulatory
Development
By
the
U.
S.
EPA.
COLORADOSTATEUNIVERSITY
a
sulfur
hexaflouride
(
SFa)
tracer
gas.
The
gas
flow
rate
was
measured
by
a
mass
flow
meter
equipped
with
readout.
The
validation
of
Acetaldehyde/
Acrolein
was
conducted
in
conjunction
with
the
Formaldehyde
validation.
No
impact
on
the
quality
of
the
validation
data
resulted
from
this
deviation
in
procedure.

Upon
investigation,
it
was
determined
that
the
Acetaldehyde
calibration
gas
standards
supplied
by
Scott
Specialty
Gases
were
inaccurate.
The
Nicolet
Rega
7000
FTIR,
which
had
previously
been
validated
for
Acetaldehyde,
showed
that
the
calibration
gases
for
Acetaldehyde
were
reading
lower
ppm
values
than
the
certification
indicated.
The
spectra
for
the
Acetaldehyde
were
analyzed
and
the
calibration
gases
were
found
to
have
an
impurity
in
the
standard.
A
method
was
developed
to
compensate
for
the
impurity
so
that
the
Acetaldehyde
standards
could
be
analyzed
on
the
Nicolet
Magna
560
FTIR.

The
two
component
standard
for
Acetaldehyde/
Acrolein
was
used
to
perform
the
validation
process.
The
concentration
of
the
Acetaldehyde
in
the
two
component
standard
was
determined
by
analyzing
the
spectra
with
both
FTIRs.
Both
units
were
in
agreement
on
the
value
of
the
Acetaldehyde
concentration
in
the
calibration
standard.
The
validation
process
was
conducted,
and
upon
completion,
the
calibration
gas
standard
was
shipped
to
PES
for
evaluation.
Scott
Specialty
Gases
was
contacted
and
informed
of
the
situation.
The
other
calibration
standards
for
Acetaldehyde
were
returned
to
Scott
Specialty
Gases
for
analysis.
PES
will
provide
information
on
the
two
component
standard
used
for
the
validation
process.

Imnact
on
the
validation
process:

1.
The
calibration
gas
standard
is
analyzed
and
found
to
be
in'
agreement
with
the
field
evaluation.
The
validation
for
the
Acetaldehyde
will
be
complete.
2.
The
calibration
gas
standard
is
analyzed
and
found
to
not
be
in
agreement
with
the
field
evaluation.
In
this
event,
several
options
are
available:

­
The
validation
for
the
Acetaldehyde
will
need
to
be
performed
and
data
adjusted
if
the
initial
validation
is
deemed
inaccurate.
­
Third
party
analysis
of
the
spectra
to
determine
if
the
Acetaldehyde
concentration
values
are
correct.
It
is
anticipated
that
this
would
be
performed
by
ERG.
­
Other
options
may
exist
and
should
be
explored
with
the
assistance
of
EPA,
PES,
ERG,
and
CSU.
­
Acceptance
of
data
based
upon
analysis
of
the
calibration
gas
by
the
Nicolet
Rega
7000,
which
has
previously
been
validated
for
Acetaldehyde,
Acrolein,
and
Formaldehyde.

Emissions
Testing
3­
2
Pacific
Environmental
Sexvices
Of
Control
Devices
for
Reciprocating
Internal
Combustion
Engines
In
Support
of
Regulatory
Development
By
the
U.
S.
EPA.
COLORADOSTATEWIVERSITY
3.2
FTIR
POST­
CATALYST
WATER
ANALYSIS
Analysis
method
on
the
Nicolet
Magna
560
FTIR
analyzer
gave
water
measurements
that
were
excessively
high
for
post­
catalyst
emissions
measurements.
The
spectra
for
HZO,
provided
by
Nicolet,
on
the
Magna
560
calculated
water
content
to
be
approximately
6%
higher
than
actual
exhaust
gas
concentrations.
Carbon
balance
calculations
for
each
one­
minute
data
point,
at
all
test
conditions,
agreed
with
the
Hz0
readings
from
the
Rega
7000
FT.
IR
analyzer,
precatalyst
emissions
measurement.
The
measurements
agreed
within
*
OS%
to
kl%
water
content.
The
carbon
balance
calculations
for
the
post­
catalyst
water
content
agreed
with
the
precatalyst
measurements
within
&
OS%
to
kl%
water
content
at
all
test
conditions.
The
carbon
balance
measurements
are
based
upon
the
precatalyst
and
post­
catalyst
reference
method
analyzers.
Since
the
pre­
catalyst
and
post­
catalyst
measurements
were
made
with
separate
analyzers,
the
variability
in
the
H20
calculation
could
be
caused
by
variability
in
emissions
analyzers.

The
changes
in
the
water
content
were
calculated
by
the
carbon
balance
method
and
detected
by
the
FTIR
analyzer.
Based
on
the
agreement
between
the
pre­
catalyst
FMR
measurements
and
the
carbon
balance
calculation
for
water
content
at
every
test
condition,
and
between
the
precatalyst
and
post­
catalyst
calculations,
the
water
content
from
the
post­
catalyst
FTIR
measurements
were
used
to
convert
the
wet
FTIR
measurements
to
dry
measurements.
As
both
FTTR
analyzers
passed
the
validation
process
and
passed
all
QC
checks,
the
variation
in
water
readings
from
the
Nicolet
Magna
560
analyzer
has
no
impact
on
the
results
of
the
testing
conducted
during
Phase
Two
of
the
overall
test
program.

Emissions
Testing
3­
3
Pacific
Environmental
Services
Of
Control
Devices
for
Reciprocating
Internal
Combustion
Engines
In
Support
of
Regulatory
Development
By
the
U.
S.
EPA.
3.3
BASELINE
ENGINE
OPERATING
CONDITIONS
Baseline
engine
operating
conditions
as
described
in
the
Scope
of
Work
are
presented
in
Table
3
of
this
report.
These
conditions
were
estimates.
After
running
the
engines,
these
values
were
found
to
be
inaccurate.
Deviations
from
the
Baseline
engine
operating
conditions
as
presented
are
as
follows:

TABLE
4
WAUKESHA
3521GL
BASELINE
CONDITIONS
Acceptable
Engine
Operating
Parameters
Nominal
Value
Range
Designation
Engine
Torque
3236
ft­
lb.
+
2%
of
value
Primary
Engine
Speed
1200
RPM
+
5%
of
value
Primary
Jacket
Water
Temperature
Outlet
180
°
F
f
5%
of
value
Primary
Engine
Oil
Temperature
Header
185
°
F
+,
5%
of
value
Primary
Air
Manifold
Temperature
100
°
F
+
5%
of
value
Primary
Air
Manifold
Pressure
5.0"
Hg
above
Atm.
+,
5%
of
value
Primary
Exhaust
Manifold
Pressure
5.0"
Hg
below
AMP
+
5%
of
value
Primary
Ignition
Timing
1O"
BTDC
+,
5%
of
value
Primary
Overall
Air:
Fuel
Ratio
28:
l
+
5%
of
value
Primary
Inlet
Air
Humidity­
Absolute
.015
lb
H*
O/
lb
Air
*
10%
of
value
Primary
Engine
Fuel
Flow
SCFH
5460
SCFH
*
5%
of
value
Primary
Engine
Oil
Pressure
Inlet
52
lb.
+
IO%
of
value
Secondary
Inlet
Air
Flow
1730
SCFM
+
5%
of
value
Secondary
Average
Engine
Exhaust
700
°
F
f
5%
of
value
Secondary
Temperature
Humidity
Ratio:
Baseline
humidity
ratio
is
0.015lb.
HzO/
lb.
air.
The
baseline
humidity
ratio
was
stated
as
0.0015lb.
HzO/
lb.
air
.
This
is
a
misprint.
Documentation
should
be
corrected
to
show
0.015lb.
H*
O/
lb.
air
as
baseline
humidity
ratio.

August
6,1999
Baseline:
The
air/
fuel
ratio
appears
to
be
wrong
because
it
is
calculated
from
the
output
of
the
pre­
catalyst
O2
monitor.
The
monitor
failed
during
this
baseline.
The
air/
fuel
ratio
is
28:
l
when
the
O2
is
9.8%.
The
catalyst
outlet
O2
is
9.9%,
so
the
air/
fuel
ratio
is
correct.

Emissions
Testing
3­
4
Pacifk
Environmental
Services
Of
Control
Devices
for
Reciprocating
Internal
Combustion
Engines
In
Support
of
Regulatory
Development
By
the
U.
S.
EPA.
COLORADO
STATE~
IVER~~

3.4
FOUR­
STROKE
ENGINE
TEST
MATRIX
The
four­
stroke
engine
sixteen
point
test
matrix
and
associated
engine
operating
conditions
as
described
in
the
Scope
of
Work
are
presented
in
Table
2
of
this
report.
During
testing
discrepancies
between
the
CSU
"
Scope
of
Work"
and
the
QAPP
in
relation
to
engine
operating
conditions
were
identified.
The
QAPP
referenced
engine
operating
data
in
relation
to
field
engines
originally
proposed
in
the
ICCR
process
and
not
the
engines
at
Colorado
State
University.
Deviations
Tom
the
engine
operating
conditions
described
in
the
sixteen­
point
test
matrix
are
referenced
to
the
CSU
"
Scope
of
Work".
Deviation
from
the
described
engine
operating
conditions
are
as
follows:

Global
Deviation
in
Engine
Operating
Conditions
Speed:
The
baseline
speed
condition
was
changed
to
1200
rpm
as
indicated.
This
value
was
used
for
the
high
speed
points.
The
value
used
for
the
low
speed
points
was
1000
rpm.

Air/
Fuel
Ratio:
The
baseline
air/
fuel
ratio
condition
is
28:
1.
This
corresponds
to
an
equivalence
ratio
of
0.57
and
an
oxygen
concentration
of
9.8%
in
the
exhaust.
This
value
was
used
for
the
normal
points.
The
low
condition
corresponds
to
an
equivalence
ratio
of
0.53
and
an
oxygen
concentration
of
10.7%.
The
high
condition
corresponds
to
an
equivalence
ratio
of
0.62
and
an
oxygen
concentration
of
8.9%.

Test
Point
Specific
Variances
Only
deviations
which
were
not
previously
described
in
the
"
Global
Deviation"
section
will
be
described.

Test
Point
3:
The
air/
fuel
ratio
on
test
point
3
appears
to
be
wrong
because
it
is
calculated
from
the
output
of
the
pre­
catalyst
O2
monitor.
The
monitor
falied
during
this
test
point.
The
air/
fuel
ratio
is
28:
1
when
the
02
is
9.8%.
The
catalyst
outlet
02
is
9.81%,
so
the
air/
fuel
ratio
is
correct.

Test
Points
2
and
7:

The
humidity
system
experienced
a
failure
prior
to
initiation
of
the
test
point
and
it
was
determined
to
conduct
the
test
points
without
inlet
air
humidity
control.
The
set
point
for
humidity
ratio
for
all
test
points
is
0.015
lbs.
water
/
lbs.
dry
air.
The
actual
humidity
ratios
for
Test
Point
2
and
Test
Point
7
were
0.0
126
and
0.0
127
lbs.
water
/
lbs.
dry
air
respectively.

Emissions
Testing
3­
5
Pacifx
Environmental
Services
Of
Control
Devices
for
Reciprocating
inter&
Combustion
Engines
In
Support
of
Regulatory
Development
By
the
U.
S.
EPA.
Appendix
S
contains
a
paper
entitled
"
An
Investigation
on
Inlet
Air
Humidity
Effects
on
a
Large­
Bore,
Two­
Stroke
Natural
Gas
Fired
Engine"
presented
at
the
1998
Gas
Machinery
Conference.
The
paper
presents
work
funded
by
the
PRCI
and
GRI.
The
paper
details
the
effects
of
variations
in
humidity
on
engine
performance
and
emissions.
Results
from
investigation
into
the
effects
of
humidity
on
engine
emissions
show
the
following
(
Appendix
S:
Figure
27
­
Figure
30):
­
With
increasing
humidity
ratio,
NO,
emissions
decrease.
­
With
increasing
humidity
ratio
formaldehyde
production
increases.
­
With
increasing
humidity
ratio,
CO
emissions
decrease
slightly
while
THC
emissions
remain
fairly
constant.
­
With
increasing
humidity
ratio
exhaust
temperatures
increase
slightly,
approximately
5
°
F
over
the
range
of
humidity
ratios
at
the
air
manifold
boost
pressure
for
Test
Points
2
and
7(
Appendix
S:
Figure
9).

Over
the
range
which
the
humidity
ratio
deviated
from
the
test
matrix
for
Test
Points
2
and
7,
the
engine
emissions
should
be
similar
to
engine
emissions
at
the
specified
humidity
ratio.
The
most
dramatic
effect
will
be
on
NO,
emissions
as
can
be
seen
from
the
data
and
the
graphs
presented
in
Appendix
S.
At
reduced
air
manifold
temperatures
(
with
engine
operating
parameters
remaining
constant),
reduction
in
NO,
emissions
would
be
the
most
noticeable
change.
NO,
emissions
would
be
reduced
due
to
the
lower
inlet
air
temperature
and
increased
inlet
air
density.
At
a
constant
humidity
ratio,
it
would
be
expected
that
CHZO
emissions
would
either
remain
constant
or
increase
slightly
with
similar
changes
in
CO
and
THC
emissions.

The
data
collected
at
Test
Point
2
and
Test
Point
7
is
indicative
of
engine
field
data
under
similar
operating
conditions.
The
variation
in
humidity
ratio
represents
minimal
impact
on
the
overall
emissions
obtained
for
this
data
point.
The
most
noticeable
impact
would
be
increased
NO,
emissions
due
to
changes
in
ambient
conditions,
which
would
result
in
elevated
in­
cylinder
temperatures
and
reducde
heat
capacity
of
the
inlet
air
charge.

3.4
OTHER
DEVIATIONS
The
Annubar
mass
flowmeter
used
for
the
exhaust
flow
measurement
did
not
perform
correctly
during
the
test
program.
Analysis
by
the
manufacturer
showed
the
unit
to
be
properly
calibrated
and
the
operation
of
the
unit
was
confirmed
in
another
location
on
our
piping
system.
Although
the
flowmeter
was
installed
well
downstream
from
flow
obstacles,
the
unit
shows
signs
of
operating
in
disturbed
flow.
The
fuel
flow
measurement
is
the
primary
mass
flow
measurement
on
the
system.
The
exhaust
flow
system
was
not
required
for
the
test
program
and
was
disconnected,
showing
a
negative
value
on
the
data
sheets.

Emissions
Testing
3­
6
Pacific
Environmental
Services
Of
Control
Devices
for
Reciprocating
Internal
Combustion
Engines
In
Support
of
Regulatory
Development
By
the
U.
S.
EPA.
4.0
TEST
SAMPLING
PROCEDURES
Engines
&
Energy
Conversion
Laboratory
Industrial
Engine
Test
Facility
Colorado
State
University
To
aid
industrial
engine
research,
Colorado
State
University
was
commissioned
to
design
and
install
a
dedicated
test
facility
for
industrial
class,
reciprocating
internal
combustion
engines.
The
Industrial
Engine
Test
Facility
was
installed
at
the
Engines
&
Energy
Conversion
Laboratory
to
provide
a
mechanism
by
which
environmental
and
technological
issues
related
to
industrial
class
engines
could
be
evaluated
in
an
independent,
economical
and
efficient
manner.
The
facility
would
also
provide
a
level
of
expertise
and
understanding
not
obtainable
from
field­
testing.

4.1
GENERAL
TEST
PROCEDURES
As
with
any
viable
testing
program,
a
procedure
has
been
established
which
affords
accurate
and
repeatable
results.
The
test
program
developed
for
the
Industrial
Engine
Test
Facility
located
at
the
Colorado
State
University's
Engines
&
Energy
Conversion
Laboratory
is
no
exception
to
this
rule.
Testing
criteria
established
for
the
test
facility
ensures
that
the
data
collected
has
a
high
degree
of
accuracy
and
can
be
repeated
if
warranted.
However,
since
the
Industrial
Engine
Test
Facility
was
designed
to
allow
for
several
different
industrial
engine
types
to
be
tested
in
a
laboratory
environment,
testing
procedures
differ
somewhat
from
field
test
procedures
and
are
unique
to
this
facility.
The
sampling
procedure
and
calibration
procedures
are
described
under
their
respective
sections
of
the
TEST
SPECIFICS
portion
of
this
report.

4.2
TEST
SPECIFICS
­
DATA
COLLECTION
The
data
collection
process
has
been
standardized
to
afford
accurate
and
repeatable
results
throughout
a
test
program.
The
high
degree
of
accuracy,
which
can
be
obtained
at
the
Industrial
Engine
Test
Facility,
is
due
to
the
sophisticated
level
of
instrumentation
utilized
at
the
facility.
However,
without
proper
implementation
no
amount
of
instrumentation
can
assure
accurate
or
repeatable
results.
Therefore
a
specific
outline
of
the
data
collection
process
has
been
developed
for
the
Industrial
Engine
Test
Facility.

Data
Point
Definition
A
typical
data
point
consisted
of
engine
operating
data
taken
over
a
specified
time
period
and
averaged.
During
normal
field
operations,
engine­
operating
parameters
will
fluctuate.
Variations
in
facility
process
conditions
can
effect
engine
speed
and
load.
Minimal
control
equipment
or
equipment
which
is
not
specialized
to
provide
precision
control
required
for
engine
research,
can
also
generate
unstable
operation.
Changes
in
environmental
Emissions
Testing
4­
1
Pacific
Environmental
Services
Of
Control
Devices
for
Reciprocating
Internal
Combustion
Engines
In
Support
of
Regulatory
Development
By
the
U.
S.
EPA.
CoLom~
o
STATEUNIVERSITY
conditions
during
the
course
of
a
test
program
will
introduce
additional
unlaiowns
into
typical
emissions
field
data.
The
Industrial
Engine
Test
Facility
was
developed
through
an
initiative
to
provide
a
facility
which
would
provide
accurate
and
repeatable
data
by
reducing
variations
in
engine
operation.
Under
controlled
conditions
at
the
EECL,
fluctuations
of
engine
load,
speed,
environmental
conditions,
etc.
have
been
minimized.
This
effort
allows
more
accurate
and
repeatable
engine
data
than
possible
with
field
research
programs.

A
standard
data
point
collected
at
the
EECL
consists
of
engine
operating
data
being
gathered
over
either
a
three­
minute
or
five­
minute
period
and
averaged.
It
has
been
determined,
based
on
previous
tests,
that
3­
5
minutes
provides
an
acceptable
time
period
required
for
an
appropriate
data
set
to
be
collected
and
an
average
for
each
parameter
calculated.

The
Large
Bore,
Engine
Testbed,
which
has
been
functional
since
1993,
was
used
as
a
reference
for
the
other
test
beds
at
the
EECL.
A
data
point
at
the
LBET
consists
of
101
engine­
operating
parameters,
which
are
collected
and
averaged
for
each
data
point.
The
data
point
consists
of
30
parameters
which
provide
basic
engine
operating
information,
twenty
parameters
which
are
received
from
the
emissions
computer
and
the
remaining
51
parameters
are
engine
combustion
parameters
calculated
with
a
combustion
analysis
system.
For
each
data
point
an
average
value,
minimum
value,
maximum
value,
and
standard
deviation
are
obtained
for
all
engine
operation
and
emissions
parameters
collected.
The
combustion
analyses
system
was
not
used
for
this
test
due
to
the
lack
of
sensor
access
ports
on
the
Waukesha
3521
Engine.

For
the
work
conducted
under
this
test
program,
a
test
point
consisted
of
a
series
of
data
points
taken
in
succession
and
averaged.
The
data
were
gathered
in
l­
minute
averages
over
a
33minute
test
period.
Using
a
data
set
consisting
of
thirty­
three,
one­
minute
data
points
would
highlight
any
large
fluctuations
in
load
and
other
parameters
that
would
have
a
significant
effect
on
emissions
data.
No
fluctuations
in
data
occurred
during
any
test
points.
This
demonstrated
that
the
engine
was
operating
at
a
steady
condition
and
the
data
recorded
in
the
individual
data
points
was
repeatable.

Table
5
provides
information
on
the
nominal
number
of
samples
collected
under
each
data
point
/
test
run
scenario
for
the
LBET.

Emissions
Testing
4­
2
Pacific
Environmental
Services
Of
Control
Devices
for
Reciprocating
Internal
Combustion
Engines
In
Support
of
Regulatory
Development
By
the
U.
S.
EPA.
COLORADOSTATEVNIVERSITY
TABLE
5
SAMPLING
SPECIFICATIONS
Measured
Number
of
Samples
Collected
Parameters
1
Minute
30
Minute
Data
Point
Test
Run
Engine
.
30­
60
900­
1800
Operation
43
TEST
SPECIFICS
­
ENGINE
STABILITY
For
data
taken
during
testing
to
be
reliable,
the
engine
was
operated
in
a
state
of
equilibrium
at
each
test
point.
The
engine
control
system
allowed
for
engine
operation
data
to
be
monitored
so
that
engine
stability
could
be
easily
recognized..
The
stability
of
each
specific
engine's
operation
was
not
only
determined
on
a
test
point
by
test
point
basis,
but
also
on
a
daily
basis.
Since
combustion
parameters
for
each
engine
type
will
vary,
engine­
operating
parameters
were
used
to
determine
engine
stability.
Procedures
used
for
determining
acceptable
engine
stability
are
as
follows:

Engine
Stability:
Engine
Start
Up
Procedures
Prior
to
the
beginning
of
data
collection
each
day,
the
engine
was
"
warmed
up"
and
a
thermal
equilibrium
state
established.
This
was
nominally
determined
when
the
engine
coolant
water
systems
and
lubricating
oil
reached
a
steady
state
temperature.
Once
steady
state
operation
was
achieved,
a
daily
"
baseline"
data
point
was
gathered.
The
length
of
time
required
to
obtain
steady
state
operation
was
highly
dependent
upon
the
ambient
temperature
and
the
temperature
of
the
engine
when
started.
Due
to
the
dependence
on
these
factors,
there
was
no
predetermined
warm­
up
time.

Engine
Stability:
Daily
Baseline
Data
Point
The
Scope
of
Work
for
the
project
required
that
a
specified
number
of
test
points
be
collected
on
the
engine.
The
data
collection
process
encompassed
multiple
days
of
testing.
To
ensure
that
the
engine
was
operating
in
a
similar
manner
on
each
test
day,
a
set
of
engine
"
baseline"
data
was
collected.
An
initial
set
of
engine
"
baseline"
data
(
one
five­

Emissions
Testing
4­
3
Pacific
Environmental
Services
Of
Control
Devices
for
Reciprocating
Internal
Combustion
Engines
In
Support
of
Regulatory
Development
By
the
U.
S.
EPA.
COLORADOSTATEUNIVERUTY
minute
data
point)
was
collected
prior
to
the
first
data
point.
On
the
ensuing
test
days,
a
"
baseline"
data
point
was
collected
to
verify
the
data
collection
for
that
day.
The
primary
engine
operating
parameters
of
the
data
point
must
compare
to
within
a
specified
acceptable
range
of
the
values
of
the
primary
engine
operating
parameters
on
the
original
."
baseline"
data
set
for
engine
stability
and
to
the
baseline
operating
conditions
specified
in
Table
4.
If
primary
engine
operating
parameters
did
not
compare
to
within
the
predetermined
range,
corrective
measures
were
taken
to
isolate
and
correct
the
cause
of
the
unacceptable
values
for
the
primary
engine
operating
parameters.
Both
CSU
and
PES
representatives
initialized
the
daily
"
baseline"
data
set.
All
baseline
data
points
were
acceptable
during
the
test
program.
The
primary/
secondary
engine
operating
parameters,
acceptable
ranges,
and
their
nominal
values
for
a
"
baseline"
data
set
are
presented
above
in
Table
4.

Engine
Stability:
Pm­
Data
Point
Test
Procedures
As
with
the
daily
engine
"
baseline"
data
point,
the
engine
must
maintain
a
stable
mode
of
operation
prior
to
and
during
a
test
run.
Changing
various
operating
parameters
to
achieve
the
desired
test
condition
will
cause
the
engine
to
operate
in
an
unstable
mode
during
the
transition
period
from
one
condition
to
the
next.
The
engine
parameter
which
has
the
most
effect
on
engine
equilibrium
is
engine
load.
Fluctuations
in
load
will
result
in
erratic
and
inaccurate
emissions
data
and
for
this
reason
load
was
closely
monitored
during
testing.
Changes
in
load
will
also
affect
the
engine's
thermal
equilibrium
and
will
require
the
longest
time
for
the
engine
to
return
to
a
thermal
equilibrium
state.

Although
the
effects
are
not
as
significant
as
those
of
changing
engine
load,
any
changes
in
air
manifold
pressure,
temperature,
exhaust
back­
pressure,
or
ignition
timing
also
affected
the
engine's
equilibrium.
As
with
load
changes,
the
engine
must
be
closely
monitored
for
return
to
an
equilibrium
state
after
any
changes
are
made.
Typically,
the
engine
will
return
to
equilibrium,
steady­
state
condition
within
30­
45
minutes.
Prior
to
initiating
a
test
run,
a
pre­
test
run
data
point
was
gathered.
The
data
point
was
five­
minutes
in
length.
For
each
pre­
test
run
data
point,
an
average
value,
minimum
value,
maximum
value,
and
standard
deviation
were
obtained
for
all
engine
operation
and
emissions
parameters
collected.
Primary
engine
operating
parameters
specified
at
a
test
condition
must
agree
with
the
test
condition
value
within
+/­
2%
to
+/­
lo%
of
the
requested
value
dependent
upon
the
engine
parameter.
The
relative
standard
deviations
of
the
primary
operating
variables
were
below
1
.
O%
for
engine
operating
parameters
and
below
3.0%
for
the
engine
emissions
parameters.
The
primary
engine
operating
parameters
and
their
nominal
values
for
a
"
pre­
test
run"
data
point
are
presented
below
in
Table
6.

Emissions
Testing
4­
4
Pacific
Environmental
Services
Of
Control
Devices
for
Reciprocating
Internal
Combustion
Engines
In
Support
of
Regulatory
Development
By
the
U.
S.
EPA.
COLORADOSTATEIRVNERITY
If
primary
engine
operating
parameters
did
not
agree
with
the
requested
test
condition
values
within
the
predetermined
range,
corrective
measures
was
taken
to
isolate
and
correct
the
cause
of
the
unacceptable
values
for
the
primary
engine
operating
parameters.
All
pre­
test
run
data
points
were
acceptable
for
the
test
program.
Both
CSU
and
PES
representatives
initialized
each
"
pre­
test
run"
data
point.

Engine
Stability:
Test
Run
Stability
A
test
run
consisted
of
a
set
of
one­
minute
averaged
data
points
taken
consecutively
over
a
33minute
time
period.
For
each
data
point,
the
average
value
for
each
primary
engine
operating
parameter
must
compare
to
within
the
acceptable
range
of
the
specified
target
value
at
the
test
condition
for
engine
stability
and
the
data
collection
process
to
be
valid
for
the
specific
test
condition.
If
primary
engine
operating
parameters
did
not
compare
to
within
the
predetermined
range,
the
data
point
was
invalid,
and
corrective
measures
were
taken
to
isolate
and
correct
the
cause
of
the
unacceptable
values
for
the
primary
engine
operating
parameters.

Engine
stability
was
maintained
throughout
the
data
collection
process
for
each
test
run.
The
relative
standard
deviation
of
the
primary
operating
variables
was
below
1.0%
for
engine
operating
parameters
and
below
3.0%
for
the
engine
emissions
parameters
at
each
data
point.

Both
CSU
and
PES
representatives
initialized
each
data
point
of
a
test
run.
The
tabular
format
of
the
primary
engine
operating
parameters,
designation,
and
the
acceptance
criteria
is
presented
in
Table
6:

Emissions
Testing
4­
5
Pacific
Environmental
Services
Of
Control
Devices
for
Reciprocating
Internal
Combustion
Engines
In
Support
of
Regulatory
Development
By
the
U.
S.
EPA.
COLORADO
STATE
UNIVERSITY
TABLE
6
TEST
POINT
­
ENGINE
STABILITY
Engine
Operating
Parameters
Acceptable
Range
Standard
Deviation
Designation
Engine
Torque
f
2%
of
value
Il.
0
Primary
Engine
Speed
Jacket
Water
Temperature
Outlet
f
5%
of
value
f
5%
of
value
6;
1.0
Il.
0
PliIllaIy
Primary
I
Engine
Oil
Temperature
Outlet
1
l
5%
of
value
1
Il.
0
I
primary
I
I
Air
Manifold
Temperature
1
f
5%
of
value
1
I
1.0
I
firnary
I
I
Air
Manifold
Pressure
I
f
5%
of
value
I
Il.
0
I
pl+=
Y
I
I
Exhaust
Manifold
Pressure
I
*
5%
of
value
I
Il.
0
I
primary
I
I
Ignition
Timing
I
f
5%
of
value
I
51.0
I
primary
I
I
Overall
Air/
Fuel
Ratio
I
f
5%
of
value
I
Il.
0
I
primary
1
I
Inlet
Air
Humidity­
Absolute
I
f
10%
of
value
I
Il.
0
I
Em=­
Y
I
Engine
Fuel
Flow
SCFH
/
Gal./
Hr.
Engine
Oil
Pressure
Inlet
Inlet
Air
Flow
f
5%
of
value
f
5%
of
value
f
5%
of
value
Il.
0
13.0
s3.0
w­­
Y
Secondary
Secondary
I
Average
Engine
Exhaust
Temperature
I
f
5%
of
value
I
13.0
1
Secondaxy
I
I
I
NO,
Emissions
(
PPM)
CO
Emissions
(
PPM)
THC
Emissions
(
PPM)
co2
(%)
02
(%)
Exhaust
Air
Flow
I
f
5%
of
value
I
5
3.0
1
Secondary
I
I
f
5%
of
value
1
I
3.0
1
Secondary
I
l
5%
of
value
f
5%
of
value
f
5%
of
value
f
5%
of
value
13.0
53.0
s3.0
s3.0
Secondary
Secondary
Secondary
Secondary
Emissions
Testing
4­
6
Pacific
Environmental
Services
Of
Control
Devices
for
Reciprocating
Internal
Combustion
Engines
In
Support
of
Regulatory
Development
By
the
U.
S.
EPA.
4.4
TEST
SPECIFICS
­
DATA
COLLECTION
HARDWARE
The
design
of
the
test
facility
provides
a
platform
for
accurate
and
versatile
performance
and
emission
research
on
industrial
engines.
Control
and
measurement
systems
installed
on
the
Industrial
Engine
Test­
Beds
are
as
follows:

Waukesha
3521
GL:
Two­
Stroke
Lean
Burn
Engine
Control
and
Monitoring:
Bristol
Babcock
Control
and
Monitoring
system
Emission
Analysis
Systems:
Pre­
catalyst
Emissions
Rosemount
NGA­
2000
Five
Gas
Analyzer
Rack
for
NO,,
CO,
COz,
02,
&
THC
Emission
Analysis
System:
Precatalyst
Emissions
Nicolet
Rega
7000
Fourier
Transform
Infrared
(
FTIR)
Exhaust
Gas
Analyzer
Emission
Analysis
Systems:
Post­
catalyst
Emissions
Five
Gas
Analyzer
Rack
TECO
NO,
CO,
&
THC
Servomex
CO*
&
O2
Emission
Analysis
System:
Post­
catalyst
Emissions
Nicolet
Magna
560
Fourier
Transform
Infrared
(
FTIR)
Exhaust
Gas
Analyzer
Ignition
Analysis
System:
Altronic
Diagnostic
Module
.
\

4.5
TEST
SPECIFICS
­
DATA
COLLECTION
PROCESS
The
data
collection
process
consisted
of
acquiring
information
from
the
various
control
and
monitoring
systems.
The
engine
control
and
monitoring
system
(
ECMS)
collected
all
engine
operating
and
emissions
parameters
(
criteria
pollutants
only).
All
engine
operating
parameters
were
direct
measurements
of
the
ECMS,
while
emissions
parameters
(
criteria
pollutants)
were
passed
by
communication
link
from
a
computer
dedicated
to
emissions
hardware
control
and
monitoring.
All
emissions
parameters
measured
with
an
FTIR
were
collected
and
stored
on
a
computer
dedicated
to
individual
FTIR
operation.

After
engine
stability
had
been
confirmed,
the
data
collection
process
for
a
test
run
condition
commenced.
The
data
collection
process
was
performed
as
follows:

Emissions
Testing
4­
7
Pacific
Environmental
Services
Of
Control
Devices
for
Reciprocating
Internal
Combustion
Engines
In
Support
of
Regulatory
Development
By
the
U.
S.
EPA.
COLORADO
STATE~
IVERSITY
Data
Collection
Process:
1.)
Verification
of
engine
stability
confirmed,
accepted,
and
initialized
by
PES
and
CSU
representatives.
2.)
Proper
file
names
are
assigned
to
all
data
acquisition
hardware.
3.)
Commence
acquisition
of
data
point
for
specified
test
condition
4.)
At
completion
of
data
point,
electronic
files
are
saved
and
hard
copies
are
printed
out.
5.)
PES
and
CSU
representatives
initialize
hard
copies
verifying
acceptable
data
point.
6.)
Move
engine
operation
to
next
test
condition.

4.6
TEST
SPECIFICS
­
EMISSION
ANALYZER
GENERAL
TEST
PROCEDURES
Introduction
The
following
general
test
procedures
and
calibration
checks
guaranteed
the
integrity
of
our
sampling
system
and
the
accuracy
of
our
data.
The
testing
was
conducted
in
basic
accordance
with
approved
Environmental
Protection
Agency
(
EPA)
test
methods
as
described
in
the
Code
of
Federal
Regulations,
Title
40,
Part
60,
Appendix
A.

General
Procedure
Exhaust
oxygen
and
oxides
of
nitrogen
concentrations
from
the
engine
were
determined
in
basic
compliance
with
EPA
Method
20,
"
Determination
of
Nitrogen
Oxides,
Sulfur
Dioxide,
and
Diluent
Emissions
From
Stationary
Gas
Turbines'knd
EPA
Method
7E,
"
Determination
of
Nitrogen
Oxides
Emissions
From
Stationary
Sources
(
Instrumental
Analyzer
Procedure)".
The
sampling
procedure
for
CO
concentrations
was
based
on
EPA
Method
10,
"
Determination
of
Carbon
Monoxide
Emissions
from
Stationary
Sources."
EPA
Method
25A,
"
Determination
of
Total
Gaseous
Organic
Concentration
Using
a
Flame
Ionization
Analyzer"
was
the
sample
procedure
used
to
determine
THC
emission
concentrations.
A
modified
EPA
Method
18A
was
used
for
the
sampling
procedures
for
Methane/
Non­
Methane
Analysis.
The
method
for
calculating
mass
emissions
levels
was
based
upon
an
EPA
Method
19
"
Determination
of
Sulfur
Dioxide
Removal
Efficiency
and
Particulate
Matter,
Sulfur
Dioxide,
and
Nitrogen
Oxides
Emission
Rates"
calculation.
Mass
emissions
were
also
calculated
using
carbon
balance
calculations
developed
by
Southwest
Research
Institute
specifically
for
the
American
Gas
Association.
Calibration
and
test
procedures
are
detailed
under
their
respective
sections
of
the
TEST
SPECIFICS
portion
of
this
report.

Emissions
Testing
4­
8
Pacific
Environmental
Services
Of
Control
Devices
for
Reciprocating
Internal
Combustion
Engines
In
Support
of
Regulatory
Development
By
the
U.
S.
EPA.
Sampling
System
Dedicated
analyzers
were
used
to
determine
the
NO,,
CO,
THC,
CO2,
and
02
emissions
level
on
a
dry
basis
for
both
pre
and
post­
catalyst
emissions.
Dedicated
analyzers
were
used
to
determine
the
Methane/
Non­
Methane
emissions
on
a
wet
basis
for
both
pre­
and
post­
catalyst
emissions.
FTIR
analyzers
were
used
to
determine
aldehyde
emissions
on
a
wet
basis
for
both
pre
and
post­
catalyst
emissions.
Refer
to
Table
7
for
the
analyzers
and
the
methods
of
analysis.

Exhaust
gas
entered
the
system
through
a
318"
stainless
steel
multi­
point
probe.
Sample
points
were
located
in
accordance
with
procedures
described
in
Method
1.
Exhaust
gas
then
passed
through
a
heated
3­
way
sample
valve
and
glass
wool
filter
assembly.
The
sample
was
transported
via
a
heat­
traced
Teflon
sample
lines
and
heated
sample
distribution
manifold.
Sample
for
the
"
dry"
gas
analyzers
then
passed
through
a
4­
pass
minimum
contact
condenser
specifically
designed
to
dry
the
sample.
The
"
dry"
sample
then
entered
a
stainless
steel
sample
pump.
The
discharge
of
the
pump
passed
through
318"
Teflon
tubing
to
a
Balston
Microfibre
coalescing
filter,
moisture
sensor,
and
then
to
the
sample
manifold.
The
sample
manifold
was
maintained
at
a'constant
pressure
by
means
of
a
pressure
bypass
regulator.
A
flow
meter,
placed
in
line
at
the
exhaust
of
each
analyzer,
monitored
exact
sample
flows.
Heated
sample
flow
for
all
"
wet"
measurement
analyzers
will
be
provided
by
means
of
a
heated
sample
distribution
manifold
prior
to
sample
gas
entering
the
"
dry"
gas
analyzer
platform.
Each
heated
analyzer
had
a
dedicated
sample
pump
and
heat
traced
line
from
the
main
sample
train
to
the
analyzer.

Emissions
Testing
4­
9
Pacific
Environmental
Services
Of
Control
Devices
for
Reciprocating
Internal
Combustion
Engines
In
Support
of
Regulatory
Development
By
the
U.
S.
EPA.
COLORADOSTATEWIVERSITY
TABLE
7
CURRENT,
INSTRIJMENT.
ATION
­

n
­
F
(

E
r
I
r
I
z
I
1
(
1
I
1
I
,

(

I
Post­
catalyst
Emissions
danufacturer
and
Model
Parameters
losemount
NGA­
2000
NO
or
NO,
ZLD
Analyzer
<
osemount
NGA­
2000
CO
DIR
Analyzer
<
osemount
NGA­
2000
COZ
WIR
Analyzer
<
osemount
NGA­
2000
THC
TD
Analyzer
Detection
Principle
Thermal
reduction
of
NO2
to
NO.
Chemiluminescent
reaction
NO
with
03.
NDIR
with
Gas
Filter
Correlation
NDIR
Flame
Ionization
RwF
Variable
to
10000
PPM
Variable
to
2000
PPM
Variable
to
20%

Variable
to
10,000
PPM
$
osemount
NGA­
2000
02
Paramagnetic
Variable
to
?
MD
Analyzer
100%
&
estar
Baseline
1030H
CH4
Gas
Chromatograph
Variable
to
HeatedGC
/
FID
Non­
CH4
Flame
Ionization
5000
PPM
qicolet
Magna
560
Multiple
FTIR
analysis
utilizing
a
Variable
See
Attached
medium
range
IR
source.
Pre­
catalyst
Emissions
Manufacturer
and
Model
1
Parameters
1
Detection
Principle
I
Range
DECO
Model
42H
CLD
Analyzer
TECO
Model
48H
NDIR
Analyzer
NO
or
NO,

co
Thermal
reduction
of
NO2
to
Variable
to
NO.
Chemiluminescent
5000
PPM
reaction
NO
with
03.
NDR
with
Gas
Filter
Variable
to
Correlation
20000
PPM
I
Servomex
NDIR
Analyzer
CO2
NDIR
O­
25%
TECO
Model
5
1
THC
Flame
Ionization
Variable
to
FID
Analyzer
10000
PPM
Servomex
02
Paramagnetic
O­
5%
PMD
Analyzer
O­
25%
Questar
Baseline
103
OH
CH4
Gas
Chromatograph
Variable
to
HeatedGC
/
FID
Non­
CH4
Flame
Ionization
50000
PPM
Nicolet
Rega­
7000
Multiple
FTIR
analysis
utilizing
a
Variable
See
Attached
medium
range
IR
source.

Emissions
Testing
4­
10
Pacific
Environmentd
Services
Of
Control
Devices
for
Reciprocating
Internal
Combustion
Engines
In
Support
of
Regulatory
Development
By
the
U.
S.
EPA.
TABLE
8
COMPONENTS
MEASURED
BY
NICOLET
FTIR
Component
Formula
Component
Name
H20
CO
co2
NO
NO2
NO
NH3
NOX
C&

C2H2
c2Hs
C2&

c3Hs
H2C0
CH30H
C3Ha
w&
o
N­
G&
o
CH&
HO
so2
TIE
Water
Carbon
Monoxide
Carbon
Dioxide
Nitric
Oxide
Nitrogen
Dioxide
Nitrous
Oxide
Ammonia
Oxides
of
Nitrogen
Methane
Acetylene
Ethylene
Ethane
Propane
Formaldehyde
Methanol
Propane
Iso­
Butylene
.
Normal­
Butane
Acetaldehyde
Sulti
Dioxide
Total
Hydrocarbons
Emissions
Testing
4­
11
Pacific
Environmental
Services
Of
Control
Devices
for
Reciprocating
Internal
Combustion
Engines
In
Support
of
Regulatory
Development
By
the
U.
S.
EPA.
4.7
TEST
SPECIFICS
­
EMISSION
ANALYZERS
CHECKS
AND
CALIBRATIONS
The
following
instrument
checks
and
calibrations
guaranteed
the
integrity
of
our
sampling
system
and
the
accuracy
of
our
data.

Analyzer
Calibration
Gases
Standard
calibration
gases
used
at
the
facility
are
Scott
Specialty
Gases
EPA
Protocol
Gas
Standard
calibration
gases
with
a
kl
.
O%
or
k2.0%
accuracy.
For
this
program,
EPA
Protocol
1
calibration
gases
(
RATA
Class)
were
used.
Manufacturer
supplied
certification
sheets
were
available
during
the
testing
procedure
and
copies
of
the
current
inventory
of
gases,
which
were
used
for
calibration
and
integrity
checks
on
the
reference
method
and
FTIR
analyzers,
are
provided
within
this
document.

EPA
Protocol
1
gases
(
Rata
Class)
were
used
to
calibrate
the
reference
method
analyzers
and
FTIR
analyzers.
Formaldehyde
standards
with
a
concentration
range
between
5
­
20
PPM
were
obtained.
Acetylaldehyde/
acrolein
standards
were
also
acquired.
Any
calibration
standards
which
were
not
EPA
Protocol
1
gases
were
the
highest
quality
standard
available.

Analyzer
Specifications
Vendor
instrument
data
concerning
interference
response
and
analyzer
specifications
are
available
if
requested.
Information
supplied
by
the
manufacturer
on
the
factory
specification
sheets
will
be
furnished
if
requested.

Response
Time
Tests
(
Prior
to
initiation
of
engine
test
program)

Response
time
tests
were
performed
on
each
sample
system.
The
response
time
tests
were
performed
prior
to
the
FTIR
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
(
slowest
responding
analyzer
which
continuously
monitors)
to
stabilize
to
response
output
of
the
analyzer.
The
Questar
Baseline
Industries
CH4/
Non­
CH4
analyzers
have
a
minimum
cycle
time
of
450
minutes.
The
overall
response
time
for
these
analyzers
when
their
cycle
is
started
1:
10
minutes
after
a
sample
source
change
is
550
minutes.
When
the
CH4/
Non­
CH4
analyzer
cycle
time
was
initiated
at
a
sample
source
change,
the
overall
response
time
is
9:
OO
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.

Emissions
Testing
4­
12
Pacific
Environmental
Services
Of
Control
Devices
for
Reciprocating
Internal
Combustion
Engines
In
Support
of
Regulatory
Development
By
the
U.
S.
EPA.
COLORADOSTATEUAWERUTY
Calibration
(
Daily)

Zero
and
mid­
level
span
calibration
procedures
were
performed
on
the
reference
method
analyzer
prior
to
each
test
day.
Zero
and
span
drift
checks
were
performed
upon
completion
of
each
data
point
and
upon
completion
of
each
test
day.
This
procedure
is
referenced
as
ZSD
(
zero
and
span
drift
check)
in
the
CSU
"
Scope
of
Work".
A
zero
and
a
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
voltage
for
each
analyzer
were
recorded
and
used
in
the
following
formula:

Y=
MX+
B
Where:
B
=
Intercept
N=
Slope
X=
Analyzer
or
transducer
voltage
Y
=
Engineering
Units
After
each
test
point
and
upon
completion
of
a
test
day,
calibration
checks
were
conducted
by
re­
introducing
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
will
be
used
to
determine
a
zero
and
span
drift
for
each
test
point
for
the
CO,
COZ,
02,
THC,
CH4/
Non­
CH4,
and
NO,
analyzers.
The
zero
and
span
drift
checks
for
each
test
point
and
each
test
day
were
less
than
X2.0%
of
the
span
value
(
specific
range
setting)
of
each
analyzer
used
during
the
daily
test
program.
The
calibration
data
sheets
are
presented
in
Appendix
E
of
this
document.

Linearity
Check
(
Prior
to
initiation
of
engine
test
program)

Prior
to
initiation
of
the
test
program,
analyzer
linearity
checks
were
performed.
This
procedure
is
referenced
as
ACE
(
analyzer
calibration
error
check)
in
the
CSU
"
Scope
of
Work".
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
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
voltages
for
each
analyzer
were
recorded
and
used
in
the
following
formula:

Emissions
Testing
4­
13
Pacific
Environmental
Services
Of
Control
Devices
for
Reciprocating
Internal
Combustion
Engines
In
Support
of
Regulatory
Development
By
the
U.
S.
EPA.
COLORADOSTATEUNIVERSITY
.

Y=
hW+
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
analyzers'
linearity
was
acceptable
as
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
methane/
non­
methane
analyzers'
linearity
was
acceptable
as
the
predicted
values
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
performed
for
one
range
setting
for
each
analyzer.
The
Linearity
Check
data
sheets
are
presented
in
Appendix
E
of
this
document.

NO2
Converter
Check
(
Prior
to
initiation
of
engine
test
program)

Prior
to
initiation
of
the
test
program,
NO2
converter
checks
were
performed.
A
calibration
gas
mixture
of
known
concentrations
between
240
and
270
PPM
nitrogen
dioxide
(
NO2)
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
(
No,)
analyzer
until
a
stable
response
was
recorded.

The
converter
is
considered
acceptable
if
the
instrument
response
indicated
a
90
percent
or
greater
NO2
to
NO
conversion.
The
NOz
Converter
Check
data
sheets
are
presented
in
Appendix
E
of
this
document.

Sample
Line
Leak
Check
(
Prior
to
initiation
of
engine
test
program)

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
involved
closing
the
valve
on
the
inlet
to
the
sample
filter
located
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
in
Appendix
E
of
this
document.

Ihissions
Testing
4­
14
Pacific
Environmental
Services
Of
Control
Devices
for
Reciprocating
Internal
­
Combustion
Engines
In
Support
of
Regulatory
Development
By
the
U.
S.
EPA.
COLORADO
STATE
WIVEM~

Sample
Line
Integrity
Check
(
Daily)

A
sample
line
integrity
check
was
performed
prior
to
and
upon
completion
of
each
test
day.
This
procedure
is
referenced
as
SSB
(
Sampling
System
Bias
Check)
in
the
CSU
"
Scope
of
Work".
The
analyzer's
response
was
tested
by
first
introducing
the
mid
level
calibration
gas
directly
to
the
NO,
analyzer.
The
analyzer
was
allowed
to
stabilize
and
the
response
recorded.
The
same
mid
level
calibration
gas
was
then
introduced
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
+
5
%
of
the
analyzer
span
value
(
range
setting).

The
SSB
procedure
was
to
be
performed
for
both
the
NO,
and
methane/
non­
methane
analyzers.
It
was
determined
to
perform
the
integrity
check
for
the
NO,
analyzers
only.
The
SSB
procedure
was
performed
for
the
methane/
non­
methane
analyzers
prior
to
and
upon
completion
of
the
test
program.
The
Sample
Line
Integrity
Check
data
sheets
are
presented
in
Appendix
E
of
this
document.

Carbon
Balance
Check
(
Continuous)

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.
As
part
of
a
QC
check,
the
calculations
involve
performing
a
theoretical
02
calculation
based
upon
measured
exhaust
stack
constituents
and
fuel
gas
composition.
The
theoretical
exhaust
O2
is
then
compared
to
the
measured
exhaust
Oz.
The
percent
difference
between
the
actual
and
theoretical
O2
measurements
was
within
&
5
%
of
the
measured
02
reading.
The
O2
balance
was
performed
for
every
one­
minute
average
and
the
thirty­
three
minute
averaged
value
for
each
test
point.
The
averaged
value
for
each
test
point
is
included
in
the
test
point
data
in
Appendix
A.

Fuel
Gas
Analysis
&
Fuel
Flow
Measurements
Natural
Gas
Fuel
Gas:
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
daily
gas
analysis
was
acquired
for
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
using
an
AGA
specified
orifice
meter
run
equipped
with
dedicated
high
accuracy
pressure
and
temperature
transmitters.
All
fuel
flow
calculations
were
in
accordance
with
AGA
Emissions
Testing
4­
15
Pacific
Environmental
Services
Of
Control
Devices
for
Reciprocating
Internal
Combustion
Engines
In
Support
of
Regulatory
Development
By
the
U.
S.
EPA.
CoLok4~
0
STATE
uiwmsm
Report
#
3.
Additionally,
stoichiometric
air
to
fuel
ratio
calculations
were
made
using
the
fuel
gas
analysis.
From
this
information,
the
equivalence
ratios
for
each
day
of
testing
were
determined.
All
fuel
gas
calibrations
and
analysis
are
presented
in
Appendix
0
and
Appendix
N,
respectively.
Stoichiometric
air
to
fuel
ratio
calculations
are
presented
in
Appendix
Q.
Calculations
for
fuel
flow,
stoichiometric
air­
to­
fuel
ratio
calculations,
and
fuel
specific
F
Factor
are
presented
in
Appendix
V,
Appendix
Q,
and
Appendix
P,
respectively.

A
blind
sample
provided
by
PES
was
analyzed.
The
results
are
included
in
Appendix
N
of
this
report.

4.8
TEST
SPECIFICS:
FTIR
CALIBRATION
PROCEDURES
Calibration
was
performed
on
the
FTIR
instrument
prior
to
each
phase
of
the
test
program
and
at
the
beginning
and
end
of
each
test
day.
The
calibration
procedures
described
within
this
document
are
consistent
with
procedures
found
in
the
following
documents:

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

"
Protocol
for
Performing
Extractive
FTIR
Measurements
to
Characterize
Various
Gas
Industry
Sources
for
Air
Toxics"
­
Prepared
by
Radian
International
for
the
Gas
Research
Institute.

Both
documents
are
contained
with
the
Gas
Research
Institute
Report
Number
GRI­
9510271
entitled,
"
Fourier
Transform
Infrared
(
FTIR)
Method
Validation
at
a
Natural
Gas­
Fired
Internal
Combustion
Engine"
­
Prepared
by
Radian
International
for
the
Gas
Research
Institute.

Instrument
Description
Dedicated
FTIR
analyzers
and
sampling
conditioning
systems
were
used
to
measure
pre­
catalyst
and
post­
catalyst
exhaust
emissions.
A
description
of
each
unit
is
presented
in
Table
9:

Emissions
Testing
4­
16
Pacific
Environmental
Services
Of
Control
Devices
for
Reciprocating
Internal
Combustion
Engines
In
Support
of
Regulatory
Development
By
the
U.
S.
EPA.
COLORADOSTATEUNIVERSITY
TABLE
9
FTIR
EQUIPMENT
DESCRIPTION
1
Pre­
catalyst
Analyzer
I
Manufacturer
and
Type
1
Nicolet
Rega
7000
Spectral
Resolution
OScm"
Detector
Type
MCT­
A
Cell
Type
4.2
Meter
­
Fixed
Path
Length
Cell
Temperature
185OC
Cell
Pressure
600
Ton:
Cell
Window
Material
Zinc
Cellmide
Post­
catalyst
Analyzer
Manufacturer
and
Type
Nicolet
Magna
560
I
Spectral
Resolution
I
0.5cm'
1
Detector
Type
Cell
Type
MCT­
A
2.0
Meter
­
Fixed
Path
Length
Cell
Temperature
Cell
Pressure
Cell
Window
Material
165
°
C
600
Torr
KBr
Each
unit
and
the
associated
test
method
have
been
designed
for
measurement
of
raw
exhaust
gases
from
internal
combustion
engines.
Dedicated
temperature
controllers
maintained
cell
temperature
and
associated
sample
lines
at
the
appropriate
the
design
temperature.
Pressure
was
controlled
by
means
of
an
MKS
pressure
controller
for
each
system.
Sample
flow
to
each
analyzer
was
between
8
­
15
liters/
minute.
The
units
utilized
a
high­
energy
mid­
range
IR
source
and
are
equipped
with
modulating,
potassium
bromide
beamsplitters
with
MCT­
A
liquid
nitrogen
cooled
detectors.
The
cells
have
been
equipped
with
specific
optical
windows
to
prevent
signal
degradation
from
damaged
optics
due
to
moisture
and
corrosive
gases
present
in
the
exhaust
stream.

Pre
Engine
Test
Calibration
Prior
to
initiation
of
an
engine
specific
test
program,
the
FTIR
sampling
systems,
both
pre
and
post­
catalyst
sample
trains
underwent
an
EPA
Method
301
validation
process.
The
validation
process
was
to
verify
the
sample
and
analytical
system
performance
in
relation
to
precision
and
accuracy
of
data
collected.
Additional
calibration
procedures
prior
to
testing
of
the
engine
were
as
follows:

Emissions
Testing
4­
17
Pacific
Environmental
Services
Of
Control
Devices
for
Reciprocating
Internal
Combustion
Engines
In
Support
of
Regulatory
Development
By
the
U.
S.
EPA.
COLORADOSTATELNWERSITY
1.)
Source
Evaluation
­
Acquired
initial
source
data
to
verify
concentration
ranges
of
target
compounds
and
possible
interfits.
This
was
accomplished
prior
to
and
during
the
Method
30
1
validation
process
2.)
Sample
System
Leak
Check
­
Sample
system
leak
checks
were
performed.
The
leak
check
procedure
encompassed
the
sample
train
from
the
sample
filter
to
the
pump
outlet.
A
dedicated
rotameter
was
installed
on
the
discharge
side
of
the
sample
pump.
With
the
sample
system
operating
at
typical
temperatures
and
pressures
(
sample
pump
will
pull
a
slight
vacuum
on
the
suction
side),
the
sample
flow
rate
from
the
rotameter
was
recorded.
The
inlet
to
the
sample
filter
located
just
downstream
of
the
sample
probe
was
closed
and
the
flow
rate
through
the
rotameter
was
monitored.
The
flow
rate
through
the
rotameter
went
to
zero.
The
leak
checks
were
determined
to
be
acceptable,
as
the
leak
rate
was
less
than
4%
of
the
standard
sampling
rate
or
SOOml/
min,
whichever
is
less.
Sample
system
leak
check
data
sheets
are
provided
in
Appendix
F
of
this
document.
3.)
Analyzer
Leak
Check
­
With
the
FTIR
analyzers
operating
at
normal
operating
temperatures
and
pressures,
the
operating
pressures
were
recorded.
The
automatic
pressure
controllers
were
then
disabled,
and
the
inlet
valves
to
the
FTIR
analyzers
were
then
closed.
The
measurement
cells
were
then
evacuated
to
20%
or
less
of
their
normal
operating
pressure.
After
the
measurement
cells
were
evacuated,
each
measurement
cell
was
then
isolated
and
the
cell
pressure
monitored
with
a
dedicated
pressure
sensor.
The
leak
rate
of
each
measurement
cell
was
less
than
10
Tour
per
minute
for
a
one­
minute
period.
The
analyzer
leak
rate
was
determined
to
be
acceptable.
Analyzer
leak
check
data
sheets
are
provided
in
Appendix
F
of
this
document.
4.)
Cell
Pathlength
Determination
­
The
cell
pathlength
was
to
be
determined
using
the
measurement
procedures
as
outlined
in
the
Field
Procedure
Section
of
the
document
entitled
"
Protocol
For
Performing
Extractive
FTIR
Measurements
To
Characterize
Various
Gas
Industry
Sources
For
Air
Toxics",
prepared
by
Radian
International
for
the
.
Gas
Research
Institute.
Because
the
units
are
fixed
pathlength
(
non­
adjustable)
measurement
cells
which
are
stationary
units
dedicated
to
a
specific
task,
the
pathlength
determination
process
was
determined
not
to
be
necessary.
The
units
are
"
as
specified"
from
the
manufacturer,
and
have
passed
all
validation
and
calibration
procedures
at
this
fixed
pathlength.

Emissions
Testing
4­
18
Pacific
Environmental
Services
Of
Control
Devices
for
Reciprocating
Internal
Combustion
Engines
In
Support
of
Regulatory
Development
By
the
U.
S.
EPA.
Daily
Calibration
Procedures
­
Pre
Test
The
following
daily
calibration
procedures
were
performed
prior
to
the
initiation
of
each
day's
testing.

I.>
Instrument
Stabilization
­
To
ensure
the
FTIR
instruments
were
operating
in
a
stable
manner,
verification
of
the
operation
of
the
following
components
at
the
beginning
of
each
day
was
performed:
a.)
All
instrument
heated
devices
and
temperature
controller
were
at
operating
temperature
and
performing
properly.
b.)
Pressure
sensor
and
pressure
controllers
were
at
operating
conditions
and
performing
properly.
c.)
Sample
systems
(
pumps,
filters,
flow
meters,
and
water
knockouts)
were
functioning
properly.
2.)
Instruments
were
operated
on
a
conditioned
air
source
for
a
minimum
of
30
minutes
prior
to
conducting
background
spectrum
procedures.
When
the
instruments
were
in
standby
mode,
between
test
days,
the
analyzers
and
all
components
were
kept
at
normal
operating
temperatures.
The
analyzers
operated
on
conditioned
air
at
all
times
when
not
involved
with
data
acquisition.
3
.>
Background
spectrum
procedures
­
After
purging
with
a
conditioned
air
source
for
a
minimum
of
30
minutes,
the
instruments
were
allowed
to
stabilize
by
flowing
an
ultra
high
purity
Nz
gas
through
the
measurement
cell
for
a
minimum
of
ten
minutes.
During
the
stabilization
process,
the
PTIR
spectra
were
monitored
until
the
concentrations
of
CO
and
Hz0
were
reduced
and
normal
steady
state
background
levels
had
been
achieved.
The
following
procedures
were
then
performed:
a.)
Check
for
proper
interferogram
signal
using
alignment
software
b.)
Collect
a
single
beam
spectrum
and
inspect
for
irregularities
c.)
Check
the
single
beam
spectrum
for
detector
non­
linearity
and
correct
if
necessary
d.)
Perform
an
instrument
alignment
procedure
e.)
Collect
a
background
spectrum
­
The
background
spectrum
was
comprised
256
scans,
which
was
equal
to
or
greater
than
the
number
of
scans
used
for
sample
analysis.
4.)
Analyzer
Diagnostics
­
Perform
an
analyzer
diagnostic
procedure
by
analyzing
a
diagnostic
standard.
The
standard
was
a
EPA
Protocol
1
CO
gas
standard
at
concentration
levels
indicative
of
the
emissions
source,
109
ppm.
A
CO
standard
was
recommended
due
to
the
distinct
spectral
features,
which
are
sensitive
to
variations
in
system
operation
and
performance.
The
standard
was
introduced
directly
into
the
instrument.
The
instrument
readings
were
allowed
to
stabilize
and
a
Emissions
Testing
4­
19
Pacific
Environmental
Services
Of
Control
Devices
for
Reciprocating
Internal
Combustion
Engines
In
Support
of
Regulatory
Development
By
the
U.
S.
EPA.
COLORADOSTA.~
UNIVERSITY
five­
minute
set
of
data
was
acquired.
The
calculated
accuracy
and
precision
based
on
equations
from
the
document
entitled
"
Protocol
for
Performing
Extractive
FTIR
Measurements
To
Characterize
Various
Gas
Industry
Sources
for
Air
Toxics",
prepared
by
Radian
International
for
the
Gas
Research
Institute,
was
acceptable.
The
pass/
fail
criteria
for
accuracy
and
precision
was
f
10%
of
the
known
standard
for
the
instrument
to
be.
acceptable.
Each
instrument
meets
this
criteria
for
all
daily
calibrations.
Analyzer
diagnostic
data
sheets
are
provided
in
Appendix
F
of
this
document.
5.)
Additional
Analyzer
Diagnostic
­
An
additional
diagnostic
check
was
performed
to
ensure
system
operation
and
performance.
A
second
diagnostic
standard
comprised
of
a
multi­
gas
composition
was
analyzed
by
the
same
procedure.
The
gas
consisted
of
COZ,
CO,
C&,
and
NO,
in
concentrations
similar
to
exhaust
gas
composition.
The
same
pass/
fail
criteria
was
used
to
evaluate
each
analyzer's
performance
when
analyzing
the
multi­
gas
standard.
Each
instrument
meets
this
criteria
for
all
daily
calibrations.
Analyzer
diagnostic
data
sheets
are
provided
in
Appendix
F
of
this
document.
6.)
Indicator
Check
&
Sample
Integrity
Check
­
An
indicator
check
procedure
was
performed
on
each
analyzer
by
analyzing
a
certified
indicator
standard.
The
standard
was
either
a
NIST
traceable,
EPA
Protocol
1
gas
standard,
or
highest
grade
standard
available
of
a
surrogate/
analyte
gas
concentration
at
levels
indicative
of
the
emissions
source.
A
formaldehyde
standard
(
concentration
of
10.66
ppm)
was
used
due
to
the
fact
that
formaldehyde
represents
a
sampling
challenge
because
of
its
solubility
in
water.
The
standard
was
introduced
directly
into
the
instrument.
The
instrument
readings
were
allowed
to
stabilize
and
a
five­
minute
set
of
data
was
acquired.
Next,
the
indicator
standard
was
introduced
into
the
sample
system
at
the
sample
filter
located
just
downstream
of
the
sample
probe.
The
instrument
readings
were
allowed
to
stabilize
and
a
five­
minute
set
of
data
was
acquired.
The
calculated
accuracy
and
precision
based
on
equations
from
the
document
entitled
"
Protocol
For
Performing
Extractive
FTIR
Measurements
To
Characterize
Various
Gas
Industry
Sources
For
Air
Toxics",
prepared
by
Radian
International
for
the
Gas
Research
Institute.
The
pass/
fail
criteria
for
accuracy,
precision,
and
recovery
was
f
10%
of
the
known
standard
(
recovery
was
f
10%
of
the
instrument
reading
with
the
indicator
gas
introduced
directly
into
the
instrument.)
for
the
instrument
to
be
acceptable.
Each
instrument
meets
this
criteria
for
all
daily
calibrations.
Indicator
check
and
sample
integrity
check
data
sheets
are
provided
in
Appendix
F
of
this
document.

Emissions
Testing
4­
20
Pacific
Environmental
Services
Of
Control
Devices
for
Reciprocating
Internal
Combustion
Engines
In
Support
of
Regulatory
Development
By
the
U.
S.
EPA.
Daily
Calibration
Procedures
­
Background
assessment
The
baseline
absorbance
was
continually
monitored
during
da&
acquisition
procedures.
If
it
was
determined
by
PES,
ERG,
and
CSU
personnel
that
the
baseline
had
changed
by
more
than
0.1
absorbance
units,
the
instrument
interferometer
was
realigned
and
a
background
spectrum
collected.

Daily
Calibration
Procedures
­
Post
Test
Upon
completion
of
the
daily
test
program
steps
4
­
6
of
the
pre
test
calibration
procedures
were
repeated.
Both
analyzers
meet
all
acceptance
criteria
for
calibration
procedures.
All
post
test
calibration
data
sheets
are
presented
in
Appendix
F
of
this
document.

4.9
TEST
SPECIFIC
­
FTIR
VALIDATION
PROCEDURES
To
ensure
the
accuracy
of
data
collected
during
testing,
the
test
program
required
procedures
to
evaluate
instrument
performance.
Prior
to
collecting
test
data,
a
validation
procedure
was
performed
on
each
FTIR
sample
train,
both
pre­
catalyst
and
post­
catalyst,
for
the
natural
gas
fueled
engine
classification.
The
specific
sample
trains
are
as
follows:

1.)
Precatalyst
emissions
sample
trains
from
the
exhaust
of
natural
gas
fueled
engines.
This
comprises
the
two­
stroke
lean
burn
engine
class
and
four­
stroke
lean
burn
engine.
2.)
Post­
catalyst
emissions
sample
trains
from
the
exhaust
of
natural
gas
fueled
engines.
This
comprises
the
two­
stroke
lean
burn
engine
and
four­
stroke
lean
burn
engine.

Each
sample
train
was
validated
for
the
following
target
compounds:

1.)
Formaldehyde
2.)
Acetaldehyde
3.)
Acrolein
Instrument
Description
Refer
to
FTR
calibration
procedures
for
FTR
instrument
description.

Emissions
Testing
4­
2
1
Pacific
Environmental
Services
Of
Control
Devices
for
Reciprocating
Internal
Combustion
Engines
In
Support
of
Regulatory
Development
By
the
U.
S.
EPA.
COLORADO
STATE
CRvIvERsm
Procedures
Eastern
Research
Group,
ERG,
performed
the
validation
for
the
target
aldehyde
compounds.
The
validation
procedure
was
conducted
in
basic
accordance
with
procedures
outlined
in
Method
30
1­"
Field
Validation
of
Pollutant
Measurement
Methods
from
Various
Waste
Media".
Validation
procedures
for
aldehydes
utilized
an
analyte
spiking
technique
as
specified
in
Method
'
301.
The
procedures
for
the
validation
process
are
as
follows:

Analyte
Spiking:
The
process
was
carried
out
by
means
of
dynamic
analyte
spiking
of
the
sample
gas.
The
sample
stream
of
the
exhaust
gas
was
spiked
with
the
specific
analyte
after
the
sample
probe,
and
before
the
sample
filter.
Spike
levels
for
the
specific
aldehydes
were
determined
and
the
spike
gas
concentrations
were
generated
for
the
specific
aldehydes
using
the
following
methods:

Formaldehyde:
Formaldehyde
spike
gas
was
generated
by
volatilization
of
a
formalin
solution
prepared
from
a
stock
formalin
solution
of
37%
formaldehyde
by
weight.
The
solution
was
injected
into
a
heated
vaporization
block.
The
vaporized
formalin
solution
was
mixed
with
a
acetylaldehyde/
acrolein
carrier
gas
and
carried
into
the
sample
exhaust
stream.
Carrier
gas
flow
rate
was
measured
by
a
mass
flow
meter
equipped
with
readout
Acetlyaldehyde/
Acrolein:
Acetlyaldehyde
and
acrolein
spike
samples
were
generated
from
a
certified
gas
standard
(
Scott
Specialty
Gases,
ti%
analytical
accuracy)
which
contained
both
analyte
species
and
a
sulfur
hexaflouride
(
SF6)
tracer
gas.
Carrier
gas
flow
rate
was
measured
by
a
mass
flow
meter
equipped
with
readout.

Analyte
specific
spike
gas
was
introduced
to
the
FTTR
sample
train
upstream
of
the
sample
system
filter.
The
spike
gas
was
introduced
at
a
known
flow
rate.
The
spike
gas
flow
was
controlled
by
a
three­
way
solenoid
valve,
which
directed
gas
either
into
the
sample
stream
or
diverted
the
spike
gas
to
the
atmosphere.
This
allowed
for
uninterrupted
flow
of
the
analyte
spike
gas
source
during
the
validation
procedures.

The
formaldehyde
and
acetylaldehyde/
acrolein
validation
runs
were
conducted
simultaneously.
The
validation
test
runs
consisted
of
24
test
runs,
12
spiked
and
12
unspiked
runs,
which
were
paired
and
grouped
further
into
s&
groups
of
2
spiked/
unspiked
pairs
to
simulate
the
"
quad
train"
approach
used
for
Method
301
Emissions
Testing
4­
22
Pacific
Environmental
Services
Of
Control
Devices
for
Reciprocating
Internal
Combustion
Engines
In
Support
of
Regulatory
Development
By
the
U.
S.
EPA.
CoLoh4~
0
STATE:
UN~
ERSI~
Y
statistical
calculations.
Samples
were
one
minute
in
duration.
Measurement
procedures
for
acquiring
the
spiked/
unspiked
pairs
are
as
follows:

1.)
Verify
stable
engine
operation
2.)
Begin
measurement
of
the
unspiked
native
exhaust
stack
gas.
3.)
Upon
completion
of
acquiring
the
unspiked
sample,
initiate
spike
gas
flow
into
sample
stream.
4.)
Let
system
equilibrate.
5.)
Begin
measurement
of
the
spiked
exhaust
gas
sample.
6.)
Upon
completion
of
acquiring
the
spiked
sample,
divert
spike
gas
flow
to
atmosphere.
7.)
Let
system
equilibrate.
8.)
Repeat
items
2
through
7.

This
procedure
was
performed
twelve
times
to
acquire
the
appropriate
number
`
of
spiked/
unspiked
pairs.
To
ensure
stable
engine
operation
during
the
validation
procedure,
engine
operating
data
was
collected
during
the
spiking
process.

4.10
TEST
SPECIFIC
­
GENERAL
CALIBRATION
To
ensure
the
accuracy
of
data
collected
during
testing,
the
test
procedure
required
that
all
instrumentation
be
routinely
calibrated.
Calibrations
and/
or
calibration
checks
were
dormed
within
one
week
before
initiation
of
testing,
and
upon
completion
of
the
entire
test
program
to
ensure
that
no
"
drift"
has
occurred.
The
devices
calibrated
included
the
dynamometer
5000­
lb.
load
cell
and
amplifier,
all
thermocouples,
pressure
transducers,
and
all
pressure
transmitters.

Dynamometer
Load
Cell
and
Amplifier
(
Daily)

The
5000
pound
load
cell
and
amplifier
was
calibrated
prior
to
the
engine
test
section.
The
calibration
procedure
is
outlined
in
a
document
contained
in
Appendix
M
of
this
document.
Calibration
of
the
load
cell
and
amplifier
were
then
be
verified
by
applying
the
full
range
of
load
without
any
adjustments
to
the
offset
or
gain
of
the
instrumentation.
Calibration
checks
were
performed
on
a
daily
basis
prior
to
starting
the
engine
to
identify
and
correct
any
drift
in
the
load
cell
or
amplifier.
These
checks
used
the
same
procedure
as
the
calibration
verification.
Jf
the
daily
calibration
check
showed
an
indicated
load
that
exceeded
fl.
O%
of
the
torque
applied
by
the
standard
weights,
the
full
calibration
procedure
was
performed.
The
dynamometer
was
within
acceptable
limits
during
the
test
program.
Dynamometer
calibration
data
sheets
are
*
provided
in
Appendix
L
of
this
document.

Emissions
Testing
4­
23
Pacific
Environmental
Services
Of
Control
Devices
for
Reciprocating
Internal
Combustion
Engines
In
Support
of
Regulatory
Development
By
the
U.
S.
EPA.
COLORADOSTATEUNIVERSITY
Thermocouples
(
Within
one
week
prior
to
initiation
of
each
engine
test
program)

K­
type
insertion
thermocouples
are
used
throughout
the
Large
Bore
Engine
Testbed
with
compensation
performed
through
the
engine
control
and
data
acquisition
hardware.
The
thermocouples
were
calibrated
using
a
Ronan
X88
portable
calibrator
calibrated
within
kl.
O
°
F
of
N.
I.
S.
T.
standard
by
an
independent
laboratory.
The
thermocouple
signal
was
zeroed
and
the
gain
adjusted
at
full
span
until
the
value
displayed
by
the
NetCon
5000
matched
the
setting
of
the
Ronan
X88
within
&
2.0
°
F.
Once
the
zero
and
gain
have
been
set
a
minimum
of
two
mid­
point
temperatures
were
checked
to
verify
the
calibration.
Thermocouple
calibration
data
sheets
are
provided
in
Appendix
J
of
this
report.

Pressure
Transducers
(
Within
one
week
prior
to
initiation
of
each
engine
test
program)

A
3­
way
valve
has
been
installed
to
allow
pressure
transducer
calibration
without
removing
the
sensor
from
the
system.
The
Model
320
Beta
calibrator
used
for
transducers
calibration
provides
an
accuracy
of
0.05%
of
reading
or
0.02%
of
full
span
and
is
calibrated
to
N.
I.
S.
T.
standards
by
an
independent
laboratory.
The
transducer
was
zeroed
and
the
gain
adjusted
at
full
span
until
the
value
displayed
by
the
NetCon
5000
was
within
*
l
.
O
psig
of
the
pressure
supplied
by
the
pressure
calibration
standard.
A
minimum
of
two
midpoints
was
checked
to
verify
calibration.
Pressure
transducer
calibration
data
sheets
are
provided
in
Appendix
J
of
this
report.

Pressure
Transmitters
(
Within
one
week
prior
to
initiation
of
each
engine
test
program)

Pressures,
which
were
critical
to
control,
and
emissions
calculations
were
measured
using
Rosemount@
305
1C
transmitters.
The
calibration
was
performed
at
the
transmitter
and
no
adjustments
are
made
to
the
current
loop.
A
known
pressure
was
supplied
to
the
sensing
port
of
the
transmitter
using
the
Model
320
Beta
calibrator.
The
transmitter
was
zeroed
and
then
spanned
at
the
full
range
value
of
the
system.
Once
spanned,
the
value
displayed
by
the
NetCon
5000
within
ItO.
5%
of
the
full
range
value.
A
minimum
of
two
mid­
span
points
was
checked
to
verify
calibration.
Pressure
transmitter
calibration
data
sheets
are
provided
in
Appendix
J
of
this
report.

4.11
TEST
SPECIFICS
­
TEST
BED
GENERAL
DESCRIPTION
Colorado
State
University's
Engines
&
Energy
Conversion
~
Laboratory
The
continued
operation
of
stationary
reciprocating
internal
combustion
engines
is
faced
with
tremendous
challenges
in
meeting
ever
tightening
restrictions
on
air
borne
pollutants.
The
regulatory
environment
continues
to
evolve
toward
lower
allowable
limits
for
criteria
pollutants,
including
new
limitations
on
hazardous
air
pollutants
(
HAPS),
even
as
current
statutes
are
being
Emissions
Testing
4­
24
Pacific
Environmental
Services
Of
Control
Devices
for
Reciprocating
Internal
Combustion
Engines
In
Support
of
Regulatory
Development
By
the
U.
S.
EPA.
implemented.
Although
ominous
the
task
of
meeting
compliance,
difficulties
involved
in
complying
with
tightening
emissions
regulations
have
advanced
the
knowledge
and
understanding
of
engine
emissions
and
performance.
The
mechanism,
which
has
elevated
the
understanding
of
exhaust
emissions,
is
research
and
development.
To
aid
in
this
effort
the
Engines
&
Energy
Conversion
Laboratory
was
established
at
Colorado
State
University.
The
engines
located
at
the
Engines
&
Energy
Conversion
Laboratory
(
EECL)
located
at
Colorado
State
University,
and
are
representative
of
the
types
used
by
the
oil
and
gas
industries
as
well
as
power
generation
markets.
The
CSU
facility
currently
comprises
the
only
independent
large­
bore
industrial
engine
test
facility
in
North
America.
Engines
that
are
located
at
the
facility
are
as
follows:

l
Cooper
­
Bessemer
GMV+
TF
,
Two­
Stroke
Lean
Burn
Natural
Gas
Fired
Engine
l
Waukesha
3
52
1
GL,
Four­
Stroke
Lean
Burn
Natural
Gas
Fired
Engine
l
White
Superior
66825,
Four
Stroke,
Rich
Burn
Natural
Gas
Fired
Engine
l
Caterpillar
3508,
Four
Stroke,
Lean
Burn
Diesel
Fueled
Engine
The
natural
gas
pipeline'
industry
has
supported
the
installation
of
three
four­
stroke
engines
in
the
same
manner
as
the
original
engine
installation.
The
program
sponsor
for
the
installation
of
the
engines
is
the
Gas
Research
Institute
(
GRJ).
The
additional
engines
have
been
installed
at
the
facility
to
assist
research
efforts
in
addressing
needs,
both
emissions
and
performance
related,
on
multiple
engine
types.
The
high­
speed,
four­
cycle,
industrial
engines
(
approximately
1000­
l
800
r­
pm)
represent
a
large
portion
of
the
current
horsepower
in
operation
within
the
oil
and
gas
industry
and
power
generation
markets.

Emissions
Testing
4­
25
Pacific
Environmental
Services
Of
Control
Devices
for
Reciprocating
Internal
Combustion
Engines
In
Support
of
Regulatory
Development
By
the
U.
S.
EPA.
