LOW NOX EMISSION
COMBUSTOR DEVELOPMENT
FOR AUTOMOBILE
GAS TURBINE ENGINES
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Air and Water Programs
Office of Mobile Source Air Pollution Control
Alternative Automotive Power Systems Division
Ann Arbor, Michigan 48105
-------
-------
APTD-1374
LOW NOx EMISSION
COMBUSTOR DEVELOPMENT
FOR AUTOMOBILE
GAS TURBINE ENGINES
Prepared By
D.W. Dawson, K.A. Hanson, R.C. Holder
AiResearch Manufacturing Company of Arizona
A Division of The Garrett Corporation
402 S. 36th Street
Phoenix, Arizona 85034
Contract No. 68-04-0014
EPA Project Officer:
R.B. Schulz
Prepared For
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Air and Water Programs
Office of Mobile Source Air Pollution Control
Alternative Automotive Power Systems Division
Ann Arbor, Michigan 48105
February 1973
-------
The APTD (Air Pollution Technical Data) series of reports is issued by the
Office of Air and Water Programs, U.S. Environmental Protection Agency,
to report technical data of interest to a limited number of readers. Copies
of APTD reports are available free cf charge to Federal employees, current
contractors and grantees, and non-profit organizations - as supplies permit
from the Air Pollution Technical Information Center, U.S. Environmental
Protection Agency, Research Triangle Park, North Carolina 27711 or may be
obtained, for a nominal cost, from the National Technical Information Service,
U. S. Department of Commerce, 5285 Port Royal Road, Springfield, Virginia
22151.
This report was furnished to the U.S. Environmental Protection Agency
by AiResearch Manufacturing Company of Arizona, Phoenix, Arizona, in ful-
fillment of Contract Number 68-04-0014. The contents of this report are repro-
duced herein as received from the AiResearch Manufacturing Company of Arizona.
The opinions, findings, and conclusions expressed are those of the author
and not necessarily those of the Environmental Protection Agency.
Publication Number APTD-1374
11
-------
TABLE OF CONTENTS
Page
1, INTRODUCTION AND SUMMARY 1-1
1.1 Introduction 1-1
1.2 Acknowledgments 1-2
1.3 Summary 1-2
2. COMBUSTOR PRELIMINARY DESIGN 2-1
2.1 Design Criteria 2-1
2.2 Engine Design/Off-Design Cycle Analysis 2-7
2.3 Preliminary Sizing 2-21
2.4 Analytical Design Techniques 2-27
i. i
2.4.1 Flov; Pattern Numerical Analysis
2.4.2 Chemical Kinetic Analysis 2-29
2.4.3 Flow Pattern and Chemical Kinetic Analysis 2-35
2.5 Experimental Flow Visualization Model 2-46
3. COMBUSTOR TEST RIG AND INSTRUMENTATION 3_1
3.1 Rig Design and Fabrication 3-1
3.2 Instrumentation 3-2
3.2.1 Corabustor Performance 3-6
3.2.2 Emissions Analyzing Equipment 3-9
4. DATA REDUCTION METHODS AND PRESENTATION 4-1
4.1 Coiubustor Performance Data Reduction 4-1
4.2 Gaseous Emissions Data Reduction 4-3
4.3 Humidity Corrections to Emissions Results 4-8
5. COMBUSTOP. DEVELOPMENT AND EVALUATION 5-1
5.1 Test Period (5-12-71 to 12-10-71) 5-1
5.2 Test Period (12-11-71 to 1-31-72) 5-14
5.2.1 Emissions Performance 5-14
5.2.2 Emission Pickup Probe Conversion 5-30
5.2.3 Conventional Performance (Non-Emissions) 5-31
5.3 Test Period (February 1, 1972 to.August 10, 1972) 5-35
5.3.1 Test and Analysis Activity 5-35
5.3.2 Test Results 5-36
5.3.3 Analytical Effort 5-65
AT-S037-P.1?
Page iii
-------
TABLE OF CONTENTS (Contd)
Page
5.4 Test Period (8-10-72 to 11-15-72) 5-74
5.4.1 Test Results 5-74
5.4.2 Discussion of Test Results 5-98
5.5 Effect of Inlet Temperature on NO Over FDC 5-108
5.6 Development Test Summary x 5-113
6. CONCLUSIONS AND RECOMMENDATIONS 6-1
6.1 Conclusions 6-1
6.2 Recommendations 6-5
6.2.1 Recommendations for Future Programs 6-8
APPENDICES I THROUGH VI following
Page 6-8
AT-6097-R12
Page iv
-------
FINAL REPORT
LOW NO EMISSION COMBUSTOR
DEVELOPMENT FOR AUTOMOBILE GAS
TURBINE ENGINES
1. INTRODUCTION AND SUMMARY
1.1 Introduction
The investigation reported herein was performed by AiResearch
Manufacturing Company of Arizona, A Division of The Garrett Corpora-
tion, to satisfy the Low HO Emission Combustor Study for Automobile
X
Gas Turbine Engines under Contract 68-04-0014 for the Environmental
Protection Agency, Office of Air and Water Programs, Advanced Automo-
tive Power Systems Development Division.
The purpose of this program was to perform an analytical and
experimental study of gas turbine combustors suitable for automotive
engines. A chemical kinetics analysis was formulated and performed.
Combustors representative of regenerated and nonregenerated automotive
gas turbines were analyzed, designed, tested, and evaluated relative
to reducing exhaust emissions. The program goal was to establish com-
bustor design data and emission criteria that through test demonstra-
tion would aid in achieving the 1976 Federal Emissions Standards tabu-
lated below. The program included emphasis on the reduction of NO
A
emissions which for gas turbine combustors is the most difficult emis-
sion species to reduce to the 1976 Standards.
1976 Federal Emission Standards
NO (as NO9) 0.40 gm/mi
X £•
CO 3.40
HC (as CH, Q1-v 0.41
1 . OD)
AT-6097-R12
Page 1-1
-------
A number of combustor types were fabricated and evaluated under
specified test conditions. When the estimated program cost was
expended at 8 months, an extension was negotiated, after a four month
hold period to cover 8 tests on two specified combustor configurations.
The specified configurations were a vaporizer combustor and a
pneumatic-impact injector combustor.
1.2 Acknowledgments
The authors would like to acknowledge major technical contribu-
tions given throughout the program by the following individuals:
S. C. Hunter, Principal Investigator; C. G. Mackay, Project Engineer;
K. W. Benn, J. T. Irwin, Program Managers; Dr's. J. G. Sotter,
V. Quan, and C. A. Bodeen of KVB Engineering Inc., Consultant Firm.
1.3 Summary
This document reports on analytical and experimental work per-
formed for the Environmental Protection Agency, Office of Air and
Water Programs (formerly Office of Air Programs) under Contract 68-04-
0014. The total program, including the four month hold, covered the
period from May 11, 1971 to November 30, 1972.
During the contract period 35 combustor configurations were tested
to determine emissions characteristics. Six combustor types were
checked for emissions characteristics but not extensively investigated.
The goal of demonstrating emissions (NO , HC, and CO) lower than
X
the 1976 Federal Emission Standards for Light Duty Vehicles was met,
except for NO , when utilizing simulated Federal Driving Cycle proced-
X
ures. A design technique that achieved significant NO reductions in
A
a gas turbine combustor was demonstrated. This technique involved the
application of recuperator (or regenerator; bypass air directly into
the combustor primary zone.
AT-6097-R12
Page 1-2
-------
Some difficulties were encountered during the program. One of
these was the difficulty of achieving reliable extrapolations of emis-
sions at combustor inlet temperatures higher than the maximum available
test temperature. This revealed the desirability of performing combus-
tor emissions tests in facilities capable of providing full-scale air
flows with combustor inlet temperatures of at least 1400°F.
Another difficulty related to the method of predicting or simula-
ting the Federal Driving Cycle (FDC) emissions from steady-state test
conditions. However, during the course of the program, AiResearch
developed a 5-point simulation procedure to predict emissions from
steady-state combustor tests. For the vaporizer combustor compared
by both procedures with bypass operation, this procedure predicts NO
2\
emissions almost three times as high as the EPA 6-point procedure. It
is also higher than procedures suggested by other EPA sub-contractors.
Differences in simulation procedures used by AiResearch and others,
including EPA, typically predict orders-of-magnitude differences in
the individual species of pollutant. Therefore, comparisons of pre-
dicted FDC emissions should be made with the same procedure utilized
by all subcontractors. The need for additional study directed toward
improvement of the simulation procedure is apparent.
Out of the configurations tested, the vaporizer combustor resulted
in the most significant improvement by the use of bypass flow. For
combustors with a fixed bypass flow quantity, a bypass of 10 percent
appears to be the best selection. Further improvements can be attained
by using a combustor with a variable bypass flow. The best results
obtained for simulated FDC emissions in gm/mi are as follows:
(Procedure-*) AiR 5 Pt AiR 5 Pt AiR 1 Pt EPA 6 Pt EPA 6 Pt
76 Std 0% BP 10% BP 10% BP 10% BP Variable BP
no
X
HC
CO
0.4
0.41
3.4
6.38
0.012
0.13
1.73
0.005
0.071
0.45
0.70
1.60
0.66
0.12
1.90
0.78
0.09
0.90
AT-6097-R12
Page 1-3
-------
The results of the program have revealed that the optimum low
emissions engine would utilize an engine cycle and variable bypass flow
that have been matched to provide the best balance between fuel econ-
omy and related emissions.
Significant conclusions of the program include the following:
• Recuperator bypass, correctly applied, is an effective
110 control technique
• An 82 percent 110 reduction has been demonstrated with
a vaporizing combustor (SKP26489SD) over the AiResearch
5-point FDC simulation at constant 10 percent bypass.
• A 73 percent NO reduction was demonstrated on the
X
AiResearch 5-point FDC simulation between the vaporizer
combustor with the lowest zero-bypass NO emissions
X
(SKP26489 M_) and the vaporizer combustor with the
the highest zero-bypass NO emissions (SKP26489SD) when
X
operated at a constant 10 percent bypass.
• Recuperator bypass as a NO control technique is applicable
X
to a variety of combustor concepts including:
• Vaporizer
• Pneumatic impact
• Atomizer (with air-assist atomization)
• Premix
• An engine cycle can be optimized to provide the best balance
of the emission constituents and fuel economy.
AT-6097-R12
Page 1-4
-------
• Variable recuperator bypass is a simple and convenient
alternative to variable combustor geometry. The required
control system is simpler and has the potential of:
• Lower cost
• Higher reliability
• Better maintainability
• Recuperator bypass does result in higher vehicle fuel con-
sumption. Based on limited testing of non-optimized com-
bustion and without the benefit of cycle optimization con-
siderations, the fuel consumption penalty was about 15 to
20 percent.
AT-6097-R12
Page 1-5
-------
-------
2. COMBUSTOR PRELIMINARY DESIGN
2.1 Design Criteria
The program was originally directed to two types or classes of
gas turbine engines that exhibit the following characteristics:
Class A Class B
Low pressure ratio High pressure ratio
Low turbine inlet temperature High turbine inlet temperature
Regenerated or recuperated No form of waste heat recovery
The combustor emission goals originally specified in the "Request-
For Proposal" (established before the 1976 Federal Emissions Standards
were announced) are shown together with the preliminary design criteria
for Class A and Class B combustors in Table 2-1.
Conceptual design studies of combustion systems for both the
regenerated and nonregenerated applications were conducted on single-
can configurations. Conceptual designs proposed for this program were
evaluated based on the following criteria:
(a) Capable of operation at low primary zone equivalence ratio
to minimize flame temperature
(b) Amenable in itself or by the addition of mixing devices to
the establishment of a homogeneous primary zone devoid of
local areas of high equivalence ratio
(c) Minimum primary zone residence time
Based on the above criteria, three configurations were selected for
further study.
AT-6097-R12
Page 2-1
-------
TABLE 2-1
EMISSION LEVEL CRITERIA
Hydrocarbons**
Carbon Monoxide
Oxides of Nitrogen***
Particulates
Vehicle Emission
Goals
grams/mile
0.14
6.16
0.40
0.03
Combustion Emission
Goals*
mgyg fuel
0.48
21.3
1.38
0.10
Combustor Emissions
Levels, ppm
Class A**** Class B****
0.5 to 2. 4 1.0 to 4. 8
66 to 325 132 to 650
2.6 to 13 5.2 to 26
-
*For 10.0 miles/gal fuel economy and JP-4 fuel (specific gravity = 0.763)
**Total hydrocarbons plus total aldehydes expressed as hexane (Cg*1]*)
***0xides of nitrogen computed as NO-
****Parts per million by volume, wet basis
DESIGN CRITERIA FOR GAS TURBINE COMBUSTORS
Heating Rate, Btu/hr
Inlet Air Temperature, °F
Inlet Air Pressure, atm
Overall Fuel-Air Ratio
Outlet Temperature, °F
Class A
Design Point
1.386 x 106
1100
4.0
0.01
1700
Testing Range
2.10 x 10 min
200 to 1200
2.0 to 6.0
0.003 to 0.015
1000 to 2200
Class B
Design Point Testing Range
1.880 x 106 2.82 x 105 min
760 400 to 900
12.0 6.0 to 16.0
0.020 0.006 to 0.030
1900 1200 to 2400
AT-6097-R12
Page 2-2
-------
(a) External Prevaporizing Combustor (Pre-Mix) - The external
prevaporizing design (Figure 2-1) incorporates an external
chamber into which the fuel is sprayed, vaporized, and
mixed with air. The mixture is delivered to the combus-
tion section primary zone through a large number of small
connecting tubes. More air is introduced into the primary
zone around the tubes to further lean out the mixture prior
to combustion. Small recirculating combustion zones are
established at the exits of the fuel delivery tubes result-
ing in a homogeneous primary zone with a very compact
flame. Minimizing the combustion zone volume in this man-
ner reduces the length of time the post-flame gases are
exposed to high temperature conditions which encourage
nitric oxide formation.
(b) Airblast Pneumatic Impact Combustor - A configuration util-
izing an air blast injector (Figure 2-2) was designed for
the Class B application. By introducing most of the primary
air through the injector, maximum atomization and mixing
occurs prior to reaching the injector exit.
The airblast injector was sized with the aid of an analyti-
cal model developed at AiResearch for the USAAVLABS Advanced,
Small, High-Temperature-Rise Combustor Program. The model
calculates fuel velocities through the venturi, atomization
quality at the injector outlet, and droplet trajectories for
five classes of droplet size at the injector exit. At 4.5
percent pressure drop (isothermal conditions) with the Class
B fuel flow, an injector geometry consisting of a 3/16-in.
diameter fuel delivery tube discharging into the 0.38-in.
diameter throat of a 0.65-in. long venturi gave a Sauter
mean droplet diameter at the injector exit of 7.5 microns,
which is more than adequate atomization for successful com-
bustion operation.
AT-6097-R12
Page 2-3
-------
HEAT
FIGURE 2-1 EXTERNAL PREVAPORIZING COMBUSTOR
(PRE-MIX)
FIGURE 2-2 AIRBLAST PREMIXING COMBUSTOR
(PNEUMATIC IMPACT)
FIGURE 2-3 FILM VAPORIZING COMBUSTOR
AT-6097-R12
Page 2-4
-------
(c) Film-Vaporizer Combustor - One film-vaporizing type
combustor (Figure 2-3) was designed for each combustor class.
Liner open area was sized for a pressure loss of 4.5 percent
of inlet total pressure .(isothermal) in both cases.
Features of the design incorporated specifically for
emission control include a ?ean primary zone (equivalence
ratio = 0.6) to minimize nitric oxide formation rates by
reducing flame temperature. Aerodynamic mixing devices
are included within the fuel delivery primary pipe to
ensure uniform fuel distribution at the pipe exit. Uniform
fuel distribution is required to avoid local areas of high
equivalence ratio in the primary zone which contribute to
high nitric oxide formation rates. In addition, a solid
boundary has been introduced between the primary pipe exit
and the combustor dome to allow vaporization to begin at an
earlier time. The effect is to reduce primary zone resi-
dence time by minimizing the effective primary zone length
and to allow earlier dilution air introduction to quench
emission formation reactions. Figure 2-4 illustrates these
effects.
AT-6097-R12
Page 2-5
-------
FUEL DELIVERY TUBE
VAPORIZATION
INITIATED HERE"
,
f
\
f
f
0 w 0.9
IKDIAL!
u
ALL JET|
/ PRIMARY
Y r PIPE
\ \
DILUTION
P_x
SfftiCTlVE A
PRIMARY — ^T VAPORIZATION
LENGTH COMPLETED HERE
a. CONVENTIONAL FILM-VAPORIZER
FUEL DELIVERY TUBE
EFFECTIVE
PRIMARY
ZONE
LENGTH
VAPORIZATION
— INITIATED HERE
H_ VAPORIZATION
COMPLETED HERE
b. LOW NO FILM-VAPORIZER
COMPARISON OF CONVENTIONAL AND LOW
FILM-VAPORIZING COMBUSTOR DESIGN FEATURES
FIGURE 2-4
AT-6097-R12
Page 2-6
-------
2.2 Engine Design/Off-Design Cycle Analysis
Cycle studies, including design and off-design performance, were
conducted to integrate the Class A and Class B combustor preliminary
design criteria into representative automotive gas turbine engine
applications. The data derived from the cycle analysis were used to
establish combustor operating conditions for use in the test program.
(a) Design-Point Cycle Analysis - A compressor-turbine flow
match was generated for the Class A and Class B preliminary
design criteria utilizing a variable-geometry free-power-
turbine engine cycle. The matching studies resulted in a
corrected power output for the Class A regenerated engine of
182.5 shp and 124.2 shp for the nonregenerated Class B
engine at the design conditions specified. In addition to
the design-point calculations, off-design steady-state and
transient performance data were also generated for both
engine classes.
Because of the wide variation in power output between the two
combustor classes, it was recommended and approved that the
design criteria be modified to bring the power output of both
engines to 150 shp. Consequently, the heat release rate of
the Class A combustor was reduced from 1.5 x 10 Btu/hr to
1.396 x 10 Btu/hr, and the Class B heat release was
increased from 1.5 x 10 Btu/hr to 1.880 x 10 Btu/hr.
The new design criteria for Class A and Class B are shown in
Tables 2-2 and 2-3, respectively.
AT-6097-R12
Page 2-7
-------
at o *-
1/1 a _i or a o at
»- VI UJ U U/UJUIUJXUI
z uj w acz«nz 1/1
I
O
Z
«
X «-!»•«
UJ
a >-
o a o- t*
u. _i •»••*•
u « • «i
CD I -* a-
t- f- "
z •» •»>
uJ
UJ _l >-
z ^ o
i-l O 2 O O
a 2 1-1 o- a-
3 < u. . ..
K- x ->
in u u.
z uJ 3: j
»-« ^
cr o
or a o
D uj -c o
»- UJ K * •
x a. " - x 3 -o in
o M -<
I/I UJ
' •- a
o
o
u
.1 in
•-t O
o o
I
a
i o
V) O
O 0.0
in u. •
-i o
(U
a
«r
i—
UJ
uj
O O
a
i o
o m i~«
u >
a
100
• •
t~ c o
u- in
•a. -i
I
O CO
»-t O ^
t- X -O
a. I -<
a o
UJ o
a
o
o
o
1/1
T
o
-J
I o
_J in
03
K IX
or «-•
o <
3c or _lo
o uj o
in ?• o
I/) UJO
UJ
a:
a.
i: -e
o *
o a. •
T ^-
in
! IV)
O UJ
2 2
u. •
m
ct
o
CD
UJ {
19 o m mo o ou. o
i^oocvioooru •* o
V ooooooo u,o'
4 •• ••• • • •-••'
Ul O 000 U, •* !
-» u/«. ,
°-
a. inoor^ooooor^o t» i
^^O^-tlNJ^-OIMOLn^O^^ I
_J ooooooooooo
UJ ...........
o o o o I
; o
UJ '
OIB ooo o o ^-'^i'
fL OO« OO>I> O O O*''
N. Or^ OOO o CO >O ^ '
U- o cc ir» ^ cr cp t
'UJ ...... . i
i *^ O. i
i '*- • i
ixj o :
, OOOOOOOOOOOO Zi/>
ooooooooooo o —•
^OOOOOOOOOOOO
X.OOOOOOOOOOOO j
i OOOOOOOOOOOO j
' I
a ,
j oomtniAinoo u i
' 0^^5-^0^0'ff' * !
'«t O OOOOOOO O '
•V. o ooooooo a. i
li. ••«••••.* '
! ° !
'o 1
3rooa>o»-4-— «rn'r)(*in<*inrrifr)(*i ,
SJ-j:^^jj;^^j:jj: !•
- , !
tLoo^-tr-im— «»£)*DJjror>^^--i '
4 •*••»«•»*•»• ;
Tr**Mfu*inaiaO(ntcaicD i
^-fXJMoooein\CxT«-*— » •*» m
,?-««-tf\j'\ii'ntn*<»*-*oa(xj
'ui ,'(
, ^<*^CT'<>vOiniAininL'>ift '
^-*-»-»-*^)^>^)j3o»oa3
; rnron-nrnfin^rornnrnf*) |
o:». ..*..•.... i
, mnnmnnmnmnnn
mininijO inui'.n'.r> Lnm^nin
i • ' « !
,^oomnf^«*^ ooftJ(\/OsO'i)>O'-»^^O5* U.l/>
ujooiO^o*-tmiriV(vjcoco (/>* ,
r...... ......
H- *— * •— « »^^^(ri-^ n^i'ifnro *"^ •-* '
;
'i
--»-oo o '
UJ. ».........* UJ ;
: • o . .
UJ « '
f^t^»T'Cinr>-»o%p«*>(*)^^ orh-
o... .......... a: ;
'T^r'C'-*'- '
^inmccrcin^ccfrussoo^a' ,
^«cv ^ ^ — • •-• i
UJ I
• rro^io— r-*^(0-oo-n* ._IIP :
,o»r--o-of-nj-rj»o*» o i-in
in............ or~
U-'-»-*oc^>n»O)— r-ino'coo i i
orovoo^-^^'jn-r— *oo*i*i o i
jnnr^i^o-«Mivr)t^-«iM ,_j
o:
Ci
I
w
o
s
u !
SI
S
§
s
jJD IT)3'OD 3
XU.3:tDKU-Qru.ZU,
au.2ajou.Du.ccu,
AT-6097-R12
Page 2-8
-------
ui
z
<9
e
v) o
O. o
Ul
O
•o
CO
*
I/I
o
D
l/v
o
u
a
t
O
T
r
o
To
j in
.
a i K ~i
X UIM «
; ui.
or >•
oo. K o>
; u. -JOB f>'
UJ < • •>
• a> i u in
' z * <•>
: Ul
.
z « o
M O 20 o
CD (-* Ul CO CD
ec z «» o
3 < U •• •'
i- r -•
u u.
U CC Ul U.
Z UJ Z J
•->*
OD O
a: a. o
a uim *•
K Ul - o o
u. m
a o
uio
« a:
ct M
o <
U o
(O o
>- 2 o
o-« .
Of ft _)0
O UJO
VI Z O
I/I Ul U
Ul
•X
a.
•r- o
o m
u a. .
o
§
u
Ul
o
<
if
<
UJ
_)
OL
U>
O
a
M ec (e
I- VI UIUIUI Ul
z ui a z 1/1 z v>
ui>- a uj-• ;> «D
*-*uia. ZAU.CQU.
?z'§5,5ii5ii
< — UCD»-Ol-0
r-
o
o
O <-• G O
O (\| O O
o o o o
tf> e o o o o o
«—-.-* O MO ^D
o a o o o o o
O -O O —•
O O O .6
on o **
O
OOOOOOOO
4OOOOQOOO
,XOOOOOOOO
X.. ......
in f- »-in in
OQ r- r^ r~ ^
o
Ul
a
o i
tf«. i
o •• !
uj (u !
x in ;
g- !
O I
,N
ooooo
UIO
z in
ut
X
o
a
< -« •-> co -* ec o ~ r*
IS
ujoonmPjfMcoa?
X ........
«JOO*O1/1CV»-«OO
OC
ui
.
r*s>ceo-.o.D:>a'
•aj^-^Hotnoo^ix
*-u>m*\i UJ
Z Z
» Ol "
•-•ac
a 3
2
a
f/)
uj-
a
a.
O*f-Cfflor>J-»o-»
_j>o.o--<(vmm.nf-
u.
o:
U
_
Ul
D
u.
ZUI!/IUJ?liJ?UJ
AT-6097-R12
Page 2-9
-------
(b) Off-Design Cycle Analysis - The following discussion briefly
describes the development of combustor test conditions as
they evolved during this study program. Also, the method
for obtaining mass emissions in grams per mile from combus-
tor rig test data is discussed.
Within the constraints of Table 2-1, an engine cycle design
point (Tables 2-2 and 2-3) was generated for each class
engine. These data were used to calculate preliminary com-
bustor size. Simultaneous to the preliminary sizing of the
two combustor types and the effort to acquire off-design
engine performance, an effort was made to adequately repre-
sent and simulate the Federal Driving Cycle (FDC) for auto-
mobiles.
A mission analysis computer program, formulated during a
previous company-sponsored effort, was modified for the
automobile gas turbine optimization study (Contract 68-04-
0012). The program matched the engine characteristics to
the route profile, with due consideration for engine oper-
ating speed range, varying transmission efficiency, and
vehicle tire adhesion limits. Two versions of the program,
one for a single-shaft engine and the other for a free-
turbine type, were created to facilitate mathematical model-
ing of engine dynamic characteristics. The program versions
were sufficiently flexible to subject each candidate engine
to the various performance tests and driving cycles speci-
fied for this study.
A block diagram of the mission analysis program logic appli-
cable to both types of engines is depicted in Figure 2-5.
Route profile data was introduced by means of cards desig-
nating length, grade, and speed limit for individual route
segments. In this manner, it was possible to simulate an
AT-6097-R12
Page 2-10
-------
READ ROUTE AND
ENGINE DATA
SET SPEED EQUAL
TO SPEED LIMIT
SELECT GEAR RATIO
CALCULATE VEHICLE
POWER REQUIRED
CALCULATE DRIVE-
LINE LOSSES
CALCULATE MAXIMUM
ENGINE POWER
AVAILABLE
REDUCE SPEED IF
NECESSARY TO MATCH
SELECT GEAR RATIO
YIELDING HIGHEST
SPEED, LOWEST SFC
ACCUMULATED
PRINTOUT
SEGMENT
PRINTOUT
ACCELERATION
SUBROUTINE
BRAKING
SUBROUTINE
COMPARE WITH PRECEDING
AND FOLLOWING SEGMENTS
MISSION ANALYSIS PROGRAM 'LOGIC DIAGRAM
FIGURE 2-5
AT-6097-R12
Page 2-11
-------
actual trip with a high degree of accuracy, provided that
a sufficient number of route subdivisions were used. The
required computation time was a function of both the number
of segments and the number of accelerations.
Engine map data consisted of net engine horsepower, fuel
flow, gas generator and power turbine speed (where appli-
cable) , turbine inlet temperature, and emission indices
(gin/kg fuel) , all as a function of output speed and throttle
setting. Net engine horsepower was defined as gross power
output minus the load imposed by engine accessories and the
engine-mounted speed reduction gearbox. A separate set of
engine map cards was required for each ambient temperature
of interest. Fuel consumption for all gas turbine engines
in this study was based on a specific gravity of 0.748 (6.25
Ib/gal) and a lower heating value of 18,500 Btu/lb. Other
input data included; engine-to-transmission speed reduction,
maximum vehicle speed at maximum engine speed, engine design
speed, gas generator and power turbine inertia, and vehicle
total weight.
A 2-dimensional interpolation subroutine determined any
desired engine parameter as a function of two other vari-
ables. Consequently, the engine was not restricted to a
single speed/horsepower operating line, b> t was free to seek
the best match point for a given load condition in accord-
ance with gear-shift logic built into the program.
The initial calculation for a given route segment involved
identification of transmission gear ratios that would permit
the speed limit to be achieved on a steady-state basis. If
the engine power was insufficient to reach the speed limit,
a lower match-point vehicle speed was determined by itera-
tion. If it was not possible to attain the speed limit in
AT-6097-R12
Page 2-12
-------
any gear ratio, the gear yielding the highest car speed was
selected. Conversely, when the speed limit was attainable in
more than one gear ratio, the latter was optimized with respect
to fuel consumption. Infinitely variable speed mechanical
transmissions were simulated by means of a large number of
discrete gear splits.
Engine and vehicle inertia were ignored when a computer run was
designated as steady-state on the appropriate control card. If
such was the case, the calculation proceeded from segment to
segment, as described above, until the route was completed.
Individual segment printout information included:
Length
Grade
Speed Limit
Required Engine
Horsepower
Actual Vehicle
Speed
Gear
Turbine Inlet
Temperature
Power Turbine
Speed
Gas Generator
Speed
Torque Converter
Efficiency
Drive Line
Efficiency
Segment Time
SFC
Fuel Economy
Cumulative Distance
and Time
Weight of Fuel
Consumed
Grams of Individual
Pollutants Emitted
BTU/Mile
The hydrocarbon and NO constituents were expressed as equiv-
alent CH, oc and N0~, -respectively.
1 . o_> Z
Segment printouts were optional and could be restricted to a
few segments or eliminated altogether. An accumulated data
printout containing the following information was displayed
after completion of a run:
Total number of segments
Total distance, miles
Total fuel consumed, Ib
• Elapsed time, min
• Average fuel consumption, mpg
AT-6097-R12
Page 2-13
-------
• Average speed, mph
• Number of gear changes
• Average fuel heat release, Btu/mile
• HC emission, gm/mile
• CO emission, gm/mile
• NO emission, gm/mile
X
In September 1971, the Office of Air and Water Programs suggested
the following six point test approach to FDC simulation by the mission
analysis program:
Wf, Ib/hr
6
8
10
11
12
20
P, psig
18
13
13
13
18
13
T. °F
in.
1380
980
1000
1000
1380
1000
Wa Ib/sec
0.59
0.44
0.44
0.44
0.59
0.44
Time-Seconds
41
466
302
302
247
14
The emissions were to be computed by averaging weighting time
with grams/mile to be based on: a total distance of 7.5 miles, a
0.763 fuel specific gravity, and 10.0 miles per gallon fuel economy.
However, note that the corresponding output power levels were unknown
Subsequently, through engine part-load analysis and combustor test
experience AiResearch evolved the following three sets of test condi-
tions/procedures to simulate the FDC:
AiResearch Multiple Extrapolation (ME) 4-Point - December 71
(Based on Free-Turbine Cycle)
(a) Test at 2.18 Ib/sec
4 atmospheres pressure
Highest available inlet temperature
0.0096 f/a
AT-6097-R12
Page 2-14
-------
In one group of tests, vary f/a
In second group of tests, vary T. , .
In third group of tests, vary AP/P
Use data and graphical extrapolation of three variables to
obtain emission index at 5.5, 9.5, 18 and 27 hp. Compute
grams/mile from grams/mile = /JEI x K
HP
K
5.5
0.03606
9.5
0.07094
18
0.04503
27
0.3031
AiResearch Temperature Extrapolation 4-Point - January 72 -
(Based on Free-Turbine Cycle)
(b) Test at reduced hp conditions over a range of inlet temper-
atures as follows:
HP
5.5
9.5
18
27
Wa, Ib/sec
0.47
0.50
0.635
0.763
P., ATMS
1.25
1.32
1.50
1.C3
Tr°R
1900
1897
1860
1809
Wf, Ib/hr
4.4
7.6
12.2
16.2
These horsepowers were obtained from a time weighted Federal
Driving Cycle simulation analysis.
The grams/mile were obtained from
Grams/mile = 2-/EI x
K
Using the K values listed for (a), and obtaining the emis-
sion index (El) from extrapolation of El vs Tinie+. to tne
correct inlet temperature. This method involves only one
extrapolation on temperature for .each test point.
AT-6097-R12
Page 2-15
-------
The following final modification to the test procedure was
made to conform to the latest optimized engine cycle.* The
recommended cycle was a recuperated single-shaft engine with
variable inlet guide vanes (labeled NII2V).
AiResearch Temperature Extrapolation (TE) 5-Point - April 72
(c) The gas turbine optimization study (Contract No. 68-04-0012)
recommended an engine designated, NII2V, and having charac-
teristics shown on Table 2-4 for a sea level, standard day.
The Federal Driving Cycle (PDC) was simulated by a mission anal-
ysis computer program*, as previously described, with each route seg-
ment represented by a speed change phase followed by a sustained speed
phase to achieve the correct segment average speed and end speed of
the automobile. Then the complete mission (FDC) was surveyed to ob-
tain the total time spent within each horsepower range. All horse-
power levels were covered, using 1-hp intervals to 31 hp, and 3-hp
intervals from 30 hp to 91 hp. For the Federal Driving Cycle, the
Nil2V Engine does not operate at more than 91 hp at any time on a
sea level, 85°F day. Under these conditions this gas turbine engine
meets the performance requirements specified by the Environmental
Protection Agency document "Prototype Vehicle Performance Specifica-
tion" 3 January 1972.
From the mission analysis program output, a set of test condi-
tions can be chosen that satisfactorily represents the ranges in fuel
flow, pressure, and temperature over which the engine combustion
system must operate. Table 2-5 presents each of five test conditions
selected to represent a range of operating variables over the range
*Refer to "Automobile Gas Turbine Optimization Study," Final Report
(AT-6100-R7), Contract 68-04-0012.
AT-6097-R12
Page 2-16
-------
Of >-
*.-£-_! -. r
j uj r
, .... _. . . . . . TABLE 2-4 — - - - -
C. MoniFiF.n NII * v F-ioiNE FOR COMBUSTION STUDY Fen 7 "C
" - - - ----- - UHv • H/C ATTITUDE FPS
TURBINE
)HP. HP MF.CHANICAU RFFORE MIXINO
>in "WATER" SHAFT~HP ACC~HP"~ REQUIRED "FFFICIENCY ENTHALPY " TCMP
0.000 155,0 6.0 350.6 .980 **6.93 1767.9
. FLOW
1>
COMPRESSOR
>«I«INO • E'IFROY C<
'ORPINE ~~ MU COOLINO 1
\ 166.35 0.000
RURNER WATER HlEl
Z u. ^"! (i .• u» w
IZC— UIDJ—UK- '
O
' = = £
O 0 0 ' '
1 O • • ! J
GO4OOOO' U i
ULC COCO O ••
1 -1 Ul 0
o. -«
O 0 0
c
Ul
i &. O O C C ^+ O U •
X CC C*CO C U>P*~
C '
: U. c — iT C^ GO ;
U. C 0 Ct 9-
UJ 1"* 1
•Kr*-trtr-«^-^-r»-oo or*
u.
^ „_____. ^
_• ^V-l-^^-Tlf — -• _| !
, ' — — IV .. _ • ^ AT-6097-I
- r . Page 2-1"
! •
i
*
j
i
j
i
i
t
t
'.
\
t
i
ol
3
0 ^2
ni S!T
i
U C\)
1 "
"3 1
2 4
^ *j
s «
01 c9
112 i
J
'
\
I
1
I
3
c/1
^
0
O
1*^
oO
5
i^rf
^^
0
O-
i
I I
C (
-------
TABLE 2-5
TABLE OF 5-POINT TEST EVALUATION
Test
Points
Min
1 Av
Max | Min
2 Av
- Max 1 Min
3 Av
Max Min
4 Av
Max Min
5 Av
Max
HP
1.3
1.3
6.5
8.5
18. "5
29.5
38.5
53.5
62.5
80.5
121.0
Airflow,
Ib/sec
0.331
0.331
0.392
0.412
0.506
0.615
0.705
0.850
0.935
1.080
1.380
Temperature ,
Tr °R
1960
1960
1935
1915
1850
1780
1722
1660
1632
1580
1485
Pressure,
P,, psia
20.5
20.5
21.9
22.5
25.8
29.3
32.8
38.2
42.0
49.4
65.0
Fuel Flow,
Ib/hr
3.64
3.64
5.30
5.85
9.30
13.40
17.00
23.50
27.00
35.00
54.00
Fuel/Air
0.00305
0.00305
0.00373
0.00394
0.00500
0.00605
0.00705
0.00768
0.00835
0.00900
0.01050
AT-6097-R12
Page 2-18
-------
of engine operation during the Federal Driving Cycle simulation
analysis. Likewise, one or more of these conditions were applied to
the combustor development tests during this program. Note that the
test points selected minimize the cycle parameter variation between
the output power extremes, thus ensuring maximum accuracy in the
conversion from measured emmission index to grams of pollutant per
mile. Table 2-6 summarizes these data.
The use of the 5-point evaluation accounts for all steady-state
conditions, including a detailed integration of horsepower versus time
during engine accelerations and decelerations. It does not account
for the exhaust emissions that would be generated during the one cold
and one hot engine start in the 1975 Federal Driving Cycle nor the
variation in emissions associated with any other engine transient
operation.
The effect of the new driving cycle simulation is to increase the
predicted emission levels in grams-per-mile. This is illustrated by
values calculated according to the OAP-suggested procedure compared
with values from the two AiResearch procedures (original 4-point
simulation versus revised 5-point simulation). Calculations for the
vaporizer combustor (SKP26489-M2) yield the following values:
NO (as NO,)
X ^
FDC Simulation gm/mi Percent
AiR 5-pt 6.38 137
AiR 4-pt 5.46 117
OAP 4.67 100
AT-6097-R12
Page 2-19
-------
UJ
-I
o
o
o
o
QC
UJ
O
UJ
u.
Q
UJ
CO
(O
CM
UJ
_l
00
o
I
UJ
H
in
§
S
s
R
o
in
00
°
*~
Q.
X
cc
UJ
2
CM
7
5
i!
•^
7
00
00
(O
P r-
s 2 § • §
O 00 f- CM O
00
r-
S« 8 ? i
. en S co 9
0 CM i- t- 0
i- in J2 g g
6 S £ S 6
0
v—
s g 1 s i
C> CM t- CO O
0 rr
UJ Cu
«5 < 5
3 1 S- 3
i - ^ £• ,«
3 a. t- S ^
o
o
CM
o
(O
CM
cn
o
i
§
T—
6
m
_T
UJ
D
u.
8
6
(0
CM
o
6
LO
in
o
o
00
00
o
6
CO
o
r*
O
O
*
I
2
LU
H
LU
LL
UJ
X
CC
0
LL
X
X
UJ
Q
Z
^"*
o
3
s
UJ
w
n
UJ
CO
cc
o
ro b
CO
-j
C3
(O
CO
O o ^
Q ^ X
II
s
cc x ct
LU CW
O
in 0) II
i s s
*""
cn
o x i- N
S Z H II
i»; co < z .
coX X -
CQ
H
LU
O
LU
CO
Z
Q
Z
QC
CQ
_J
LU
"^
LL
<
>_
D
_i _i
re O — '
•^ n Q. O
+ ^ o ^ i
" < a
0
_, C 1 I
I s = o S
? S. W > g
CO ^ -J ££
_j § < 5
i £E o ^
° < H 0
^ J 0 UJ
CO m < _J
-J 3 CC O
X LL LL
LU
Q
II "Z. « «
^ z
0
— CD
S
n
i
X
X
LU
LL
O
1-
X
UJ
§
CC
_l
D
0
LU
_l
O
2
II
co 5: oo S S
CO
AT-6097-R12
Page 2-20
-------
2.3 Preliminary Sizing
Based on the combustor operating data derived from the engine
cycle analysis, preliminary designs of each of the candidate combus-
tion systems were accomplished using existing design techniques. Com-
bustor preliminary sizing was accomplished by conventional methods for
both combustors in the following manner:
(a) A combustor jet velocity was calculated assuming one velocity
head loss across the combustor at a pressure drop of 3 per-
cent of inlet total pressure.
(b) An outside annulus area was computed assuming a maximum
annulus velocity head equal to one-half of the jet velocity
head from (a). The minimum outside annulus area can then be
calculated from the volume flow rate and the maximum velocity,
(c) Combustor volume was calculated based on an assumed heat
release rate limit of 5 x 10 Btu/hr-ft -atm. The combustor
cross-sectional area was then determined using assumed values
of length-to-diameter ratio for each case and setting a prac-
tical limit of 250 fps for the combustor discharge velocity.
(d) Minimum combustor reference area is then the sum of the
annulus area from (b) and the combustor area from (c).
The above computational method is applicable to turbine engines
operating at essentially constant speed and constant combustor through-
flow. When engine performance for the variable-geometry free-turbine
cycle was calculated over the Federal Driving Cycle, however, it was
determined that most engine operating time was spent at part-power con-
ditions less than 40 shp with correspondingly reduced throughflow,
pressure ratio, and engine speed. At these conditions combustor load-
ing, Q, is increased and, therefore, a correction to increase combustor
AT-6097-R12
Page 2-21
-------
volume was made to reduce Q to ensure that a reasonable combustion
efficiency could be attained at these more severe operating conditions.
Therefore, the volume was based on.idle output power condition which
corresponds to a combustor loading, Q, range of values between 0 and
0.1. Figure 2-5 was the design basis used for combustor sizing. The
resultant vaporizer combustor geometry for Class A and Class B com-
bustors are shown in Figures 2-6 and 2-7. Table 2-7 presents a sum-
mary of preliminary vaporizer combustor design parameters.
During the test program, it was determined that the measured com-
bustion efficiencies were higher at a given aerodynamic loading param-
eter Q than .shown on the design curve. The measured efficiencies indi-
cated that a design line passing through a combustion efficiency of
92.5 percent at Q = 0.8 could have been conservatively used. This
compares to the design line predicting 80 percent combustion effi-
ciency at Q = 0.8. This indicated that a Class A combustor volume of
3 3
100 in. would have been adequate rather than the 247 in. used. If
the combustor were reduced in size, lower NO emissions (and higher CO
X
and HC emissions) than those reported in the following test results
should be achieved.
AT-6097-R12
Page 2-22
-------
1UU
EH
Z
W 0
U o
OH ^H
w
&< X
£« 90
H W
U Q
H H
W .01
WL
Q- a K1/2
VOL P2T1/2
\
\
\
\
0 0.2 0.5 0.6 0.8 1.0 1
AERODYNAMIC LOADING, Q, LB/SEC FT3 ATM2 °R1/2
COMBUSTION EFFICIENCY VERSUS AERODYNAMIC LOADING
FIGURE 2-5
AT-6097-R12
Page 2-23
-------
TABLE 2-7
SUMMARY OF PRELIMINARY VAPORIZER COMBUSTOR DESIGN PARAMETERS
Heat rate, Btu/hr
Inlet air temperature, °F
Inlet air pressure, atm
Fuel air ratio
Outlet temperature, °F
Fuel flow rate, lb/hr(1)
Airflow rate, Ib/sec
Volume flow rate, cu ft/sec
Reference diameter, in.
Liner diameter, in.
Liner length, in.
Length/diameter ratio
Reference velocity, fps
Liner discharge velocity, fps
(2)
Characteristic residenct time, ms
Combustor volume, cu ft
Heat intensity, Btu/hr/ft /atm
Class A
1.396 x 106
1100
4.0
0.0096
1700
75.5
2.178
21.4
6.1
4.8
13.7
1.5
106
250
2.40
0.143
4.6 x 106
Class B
1.88 x 106
760
12.0
0.0185
1900
101.6
1.525
3.86
3.7
3.25
10.0
2.0
52.3
138
3.92
0.481
5.0 x 106
NOTES:
(1)
(2)
Fuel lower heating value = 18,500 Btu/lb
Characteristic residence time - liner length/liner discharge
velocity
AT-6097-R12
Page 2-24
-------
•FUEL INJECTION TUBE
COMBUSTION CHAMBER LINER
OUTER CASING
PRELIMINARY VAPORIZER COMBUSTOR DESIGN - CLASS A
(NO SCALE)
FIGURE 2-6
AT-6097-R12
Page 2-25
-------
PRIMARY PIPE
FUEL INJECTION TUBES
COMBUSTION CHAMBER LINER
3.7
OUTER CASING
E>^
\
\
i^<:
\
I
\
r
\
0
O
r\
^J
o
1 °
3.25
i
10
PRELIMINARY VAPORIZER COMBUSTOR DESIGN - CLASS B
(NO SCALE)
FIGURE 2-7
AT-6097-R12
Page 2-26
-------
2.4 Analytical Design Techniques
Internal flow fields of the preliminary combustor designs were
predicted with the use of an existing two-dimensional finite-element
gas flow computation program from Gosman ^ '. Figure 2-8 presents the
basis used for flow-field calculations. The results of this analysis
provided velocities, temperatures, and mixture concentrations over the
range of combustor operating conditions defined in the cycle analysis.
An existing Gosman computer program provides time, temperature,
and mixture maps and equations for nitric oxide kinetics which were
incorporated into the Gosman computer program. Results from this
program were compared on a limited basis with results from an exist-
ing one-dimensional method of analysis.
The combustion of fuel and the formation of carbon monoxide were
computed from overall rates based on stirred reactor experimental data.
2.4.1 Flow Pattern Numerical Analysis
Flow pattern analysis was conducted on the preliminary combustor
designs. The analysis initially concentrated on the primary zone
region where nitric oxide formation predominates. Grid node patterns
and boundary conditions were established. Computations were conducted
for both cold flow (no combustion) and hot flow conditions for compari-
son with the flow visualization experimental analysis. Hot flow con-
ditions can be computed assuming local chemical equilibrium for which
combustion rates are limited only by mixing processes or by a two-step
global kinetic process utilizing stirred reactor combustor rates. The
Gosman, A. D., W. M. Pun, A. K. Runchal, D. B. Spalding, and
M. Wolfshtein, Heat and Mass Transfer in Recirculating Flows,
Academic Press, 1969.
AT-6097-R12
Page 2-27
-------
z
<
o
o
Q £
O
CO
CO
o
p
<
_J
to
LU
D
o
Z
9
a:
0
UJ
CO
I
0-4
D
a
H
AT-6097-R12
Page 2-28
-------
latter process allows an assessment of the relative effects of mixing
and kinetics and was selected for use in this program. The global rate
source terms were incorporated into the finite difference computer pro-
gram to calculate rates of fuel conversion to carbon monoxide and sub-
sequent conversion to carbon dioxide. Resulting combustion efficiency
calculations for a simplified configuration compared favorably with
both stirred reactor and practical combustor data.
Primary zone isothermal flow patterns were computed for the film
vaporizing combustor for the Class B engine. Figure 2-9 shows the
flow streamlines and velocity profiles. Air and fuel from the primary
pipe form a radial wall jet which is entrained into the cooling air
wall jet. The ejecting action of the primary pipe air results in for-
mation of a recirculation zone. High velocities are maintained along
the combustor walls by judicious introduction of high energy air to
delay combustion until mixing to a lean fuel-air ratio is complete,
thereby averting combustion at high equivalence ratios which results
in high nitric oxide formation rates. Once a flame is stabilized in
the recirculation zone, propagation to the higher equivalence ratio
regions near the walls is prevented by the high velocity gradients
shown in Figure 2-9.
The computed flow pattern was utilized to calculate primary zone
fuel mixing rates and hot flow patterns. These data were then coupled
with the KVB analytical model that computes nitrogen compound concen-
trations and are discussed in the following paragraphs.
2.4.2 Chemical Kinetic Analysis
The chemical kinetic analysis of nitric oxide and nitrogen
dioxide formation was conducted by KVB engineering under subcontract
to AiResearch. The kinetic formation equations were expressed in a
form suitable for inclusion in a two-dimensional recirculating flow
AT-6097-R12
Page 2-29
-------
o
o
o-
o
o
o
o-
rH
H O.
o
o.
-------
analysis of a gas turbine combustion chamber primary zone with
nonuniform fuel-air ratio distribution. Chemical equilibrium may or
may not be attained at various locations in the flow field. The kinetic
analysis includes the following six reactions leading to nitric oxide
formation and two reactions for the oxidation of nitric oxide to nitro-
gen dioxide:
N2 + 0 j NO + N (1)
N + 02 + NO + 0 (2)
N + OH + NO + H (3)
N2 + OH j N20 + H (4)
N2 + 02 j N20 + 0 (5)
0 + N20 + 2ND (6)
NO + 0 + M j N02 + M (7)
NO + O2 + NO2 +0 (8)
For preliminary assessment of the initial combustor designs, a
simplified time step integration of the pertinent nitric oxide forma-
tion equations was performed by AiResearch. Nitric oxide formation
rates were computed at the design conditions of the Class A and Class
B combustors for a range of primary zone equivalence ratios assuming
mixture conditions constant at the adiabatic flame temperatures. At
0.6 equivalence ratio the formation rates were 14 ppm per millisecond
and 3.1 ppm per millisecond for the Class A and Class B combustors,
respectively. Characteristic primary zone residence times were com-
puted by dividing primary zone volume by the burned gas volumetric
AT-6097-R12
Page 2-31
-------
flow rate. These times were 3.0 and 3.6 milliseconds for the Class A
and Class B combustors, respectively. For the assumed conditions for-
mation rates are constant up to 10. milliseconds. Multiplying the for-
mation rates by the characteristic residence times results in primary
zone exit concentrations of 42 and 11 ppm for the two combustors.
These amounts reduce to 10 ppm and 5 ppm when adjusted by the dilution
zone mixing process. Conversion of the 0.4 gm/mile requirement for
oxides of nitrogen to a ppm basis for the two designs gives 8.2 ppm
and 15.8 ppm, respectively. The estimated formation rates are within
reasonable agreement with these requirements. This analysis is of
course highly simplified primarily in terms of neglecting mixture non-
uniformity. Analysis of data on existing combustors by this procedure
is planned to evaluate the merits of such a comparison.
Programming of the nitrogen oxide reaction mechanism was com-
pleted by KVB Engineering. A final report describing this effect is
included as Appendix II for your reference. This computation was
intended to be attached to the flow analysis program which solves the
flow pattern and hydrocarbon chemistry equations simultaneously.
After convergence, the nitrogen compound reaction equations were to be
solved. Introduction of the nitrogen oxide kinetic analysis after con-
vergence of the Gosman recirculating flow analysis was to provide
independent development of these two analytical models; therefore,
reduce computation time. For reference, the complete chemical mech-
ansim is shown in Table 2-8.
As a means of providing initial insight into the ability to solve
for NO, a simplified model was programmed to solve the NO formation for
the two-reaction Zeldovich mechanism consisting of the first two nitro-
gen chemistry reactions listed in Table 2-8. The combustor was con-
figured with a central methane fuel jet surrounded by a concentric air
annulus. Figure 2-10 shows the computed flow pattern and lines of
constant NO concentration. The peak level of 434 ppm is typical of
that expected at the primary zone exit prior to dilution air introduc-
tion.
AT-6097-R12
Page 2-32
-------
TABLE 2-8
FINITE DIFFERENCE FLOW PROGRAM
CHEMICAL KINETIC MECHANISM
Hydrocarbon Chemistry
CHX + ( + ) 02 + CO +
C0 + \ °2 "* C°2
Atomic Species
CO + OH + C02 + H
0 + OH + 0 + H
OH + OH H0 + 0
Nitrogen Chemistry
N + NO •<- N + O
N + 0 ^ NO + O
N + OH ^ NO + H
H + NO «- N + OH
A ^
0 + NO NO + NO
NO + 0 + M ? NO + M
N02 +0 + NO + 02
AT-6097-R12
Page 2-33
-------
CN —
K O
• *
II II
CM CM
z o
^••^^ ~
ui z °
o o
z z
CO
CM
O
CM
Z
-------
2.4.3 Flow Pattern and Chemical Kinetic Analysis
The following is a discussion of the integration of the oxides of
nitrogen computation procedure provided by KVB Engineering with the
AiResearch flow computation program. Several test cases were run to
check out the program on simple geometric configurations. The oxides
of nitrogen computations were found to converge at a rate comparable
to the fluid flow and mass fraction computations. The program was then
applied to the primary zone geometry for the Class B combustor for
which flow patterns without combustion had previously been computed.
A coarse 21-by-10 grid pattern was used for more economical computa-
tation during program checkout. This grid and the portion of the com-
bustor which it covers are shown in Figure 2-11. The computations are
performed in cylindrical coordinates with the axis of symmetry at the
axis of the cylindrical can combustor. Air enters through two inlets.
One, on the axis, represents the air-blast fuel insertion device
through which the fuel-air mixture enters the primary zone; the second
inlet, at the outer radius, represents the first cooling band.
The equations that are to be solved include vorticity (conservation
of total momentum), stream function (conservation of total mass), fuel
conservation parameter (conservation of all species and enthalpy), mass
fraction of unburned fuel, mass fraction of carbon dioxide, and mass
fractions of NO and N0?. Solution of these equations is dependent on
proper specification of boundary conditions, turbulent mixing rela-
tions, and reaction rates for interconversion of the various species.
The boundary conditions are specified for four types of bound-
aries: inlets, exits, solid walls, and the axis of symmetry. At the
inlets the vorticity is determined from the specified inlet velocity
profile. The stream function is determined by integrating the mass
flow rate across each inlet. If premixed conditions are being consid-
ered, the fuel conservation parameter and fuel mass fraction are set
AT-6097-R12
Page 2-35
-------
COOLING fc.
AIR
2'7
-•
IN.
— L-PIPE EXIT
1
I
i.e
•MM
f
52 IN.
PRIMARY
->.ZONE
EXIT
A-
t
PMBMNBD
WMimN
APPHOVKO
SCH
10/71
CLASS B COMBUSTOR
PRIMARY ZONE
COARSE GRID PATTERN
FIGURE
11
rwm PTMA-I
AT-6097-R12
Page 2-36
-------
equal to the mass fraction of fuel vapor in air at the inlet. Carbon
dioxide and nitrogen oxides are set to zero. It is assumed that the
inlet air is dry. Humid air can be considered in the program, but will
be neglected for initial predictions.
Boundry conditions at the exit are specified such that the stream
lines are parallel to the axis and mass fraction gradients are zero in
the axial direction. Vorticity is determined from the radial exit
velocity profile.
On solid walls the vorticity is computed by an implicit technique
from Reference (1), assuming a linear vorticity variation between the
wall and the first interior grid node. Stream function is constant
along a wall since the flow is parallel to it. Mass fraction gradi-
ents normal to the wall are zero for impermeable walls. Computation
assumes adiabatic conditions so that enthalpy is directly related to
fuel-air ratio. An alternate procedure is to specify the boundary
temperature and solve a conservation equation for enthalpy. The axis
of symmetry is a streamline and because of symmetry there are no mass
fraction gradients, so boundary conditions are identical to solid walls
except that the vorticity is zero.
Rates of turbulent mixing are specified in terms of an effective
turbulent eddy viscosity. For turbulent flow the rates of mixing for
all the conserved properties (mass, momentum, mass fractions, and
enthalpy) are of the order of unity. Accordingly, all the turbulent
Prandtl and Schmidt numbers are set to 1.0. The turbulent eddy vis-
cosity is computed from the Prandtl-Komologrov hypothesis relating
eddy viscosity to turbulent kinetic energy and length scale of turbu-
lence. The turbulent kinetic energy is assumed proportional to the
square of the inlet velocities, and length scale is assumed propor-
tional to the size of the inlets. Details of this computation are
given by Wolfshtein (Reference 2) and Gosman, et al (Reference 1) .
AT-6097-R12
Page 2-37
-------
The hydrocarbon-air kinetic interconversion of species is handled
by the following assumed two-step reaction mechanism:
CHx + (i + J) 02 CO + | H20
CO + 02 —* C02
Reaction rates are taken from Williams (Reference 3). Since there are
two reaction rates, the conservation equations for the mass fraction
of two species must be solved. The species CH and C09 are used in
X £*
the present program. The mass fractions of the remaining species plus
inert nitrogen are then directly related through stoic \iometry to these
two mass fractions.
Provision has been made in the procedure for fuel addition by
film vaporization from the wall. The formation of the fuel film is
computed by the method detailed in NACA Report 1087 (Reference 4).
Rates of vaporization are computed by the method of NASA Report
TR-R-67 (Reference 5}. Velocities and temperatures in the primary
zone are required inputs to the film formation and vaporization com-
putations. The computation is then iterated until temperatures ara
compatible with the rates of vaporization.
Initial computations for program checkout with the coarse grid
have been restricted to premixed fuel vapor and air entering the two
inlets. Computations with internal vaporization should be performed
after checkout of the integrated program is complete. However, this
phase was never completed. These two procedures for fuel introduction
provide a comparison of premixing and internal vaporization in terms
of the resultant emissions formation.
The computed flow pattern for conditions of the Class B combus-
tor full power design point is shown in Figure 2-12. The curves
AT-6097-R12
Page 2-38
-------
FRACTION OF TOTAL
INLET MASS FLOW
= 0.725
RECIRCULATION ZONE ENCLOSED BY
STREAMLINE =1.0
INLET VELOCITY = 300 FT/SEC
PRIMARY
ZONE
EXIT
PMIPAMB
WRITTEN
APPNOVCO
SCH
11-71
CLASS B COMBUSTOR
PRIMARY ZONE
HOT FLOW PATTERN
FIGURE
2-12
AT-6097-R12
Page 2-39
-------
F = FUEL CONSERVATION PARAMETER
= MASS FRACTION OF CONSERVED CARBON AND HYDROGEN IN
FUEL AND COMBUSTION PRODUCTS
0.04=
0.06 L-PIPE EXIT
PREMIXED VAPOR-AIR MIXTURE
AT BOTH INLETS
FNKPANID
WMITTKM
APPMOVCO
SCH
11-71
CLASS B COMBUSTOR
PRIMARY ZONE-
FUEL CONSERVATION PARAMETER
FIGURE
2-13
i
AT-6097-R12
Page 2-40
-------
40,000 =
31,425
17,140
712
17,140
2,850
000
EXIT
VALUE.S
LESS
THAN
0.1 ppmw
CONCENTRATION CVHV (PPM BY WEIGHT)
A x
PREMIXED VAPOR-AIR MIXTURE
AT BOTH INLETS
APFROVCD
SCH
11-71
CLASS B COMBUSTOR
PRIMARY ZONE
UNBURNED HYDROCARBONS
FIGURE
2-14
•MM FTMA-I
AT-6097-R12
Page 2-41
-------
represent lines of constant stream function so that a fixed amount of
mass flow exists between two lines. The hot flow pattern is quite
similar to the cold flow pattern presented in Figure 2-9, and is qual-
itatively consistent with patterns observed in the plexiglass model
flame visualization test and water model tests of the Shelldyne film-
vaporizing combustor.
The computation was performed with premixed mass fractions of
0.06 (0.95 equivalence ratio) at the L-pipe inlet and 0.04 (0.63
equivalence ratio) at the second inlet. Figure 2-13 shows the dis-
tribution of the fuel conservation parameter that expresses the rate
of mixing between the two inlet flows and the circulation zone. The
fuel conservation parameter is essentially an indication of the local
fuel-air ratio. It is the determination of the local fuel conserva-
tion parameter that is the key to balancing mixing and fuel introduc-
tion so that lean combustion is maintained for low NO emission. If
both inlets were set to the same value of initial premixed fuel mass
fraction, the fuel conservation parameter would be constant throughout
the field. With different values at the two inlets, local grid values
are intermediate between the inlet values. When internal vaporization
is introduced, local vaporization rates may be sufficiently high to
produce rich regions. It is the express purpose of this analytical
procedure to determine the combustor geometric changes necessary to
eliminate rich and near stoichiometric regions .
Figure 2-14 shows the mass fraction of unburned hydrocarbons.
For the premixed condition, hydrocarbon is converted to CO within a
very short distance from the inlets. Hydrocarbon emission at the pri-
mary zone exit is less than 0.1 ppm. Conversion of CO to C02 also
occurs rapidly with less than 1 ppm CO emission. Resultant combustion
efficiency is over 99.6 percent.
AT-6097-R12
Page 2-42
-------
The original Gosman-Spalding program assumed constant specific
heats of the gaseous species. Since temperature has a significant
effect on the reaction rates for both the hydrocarbon and nitrogen
reactions, the temperature calculation was improved by incorporating
variable specified heats through the use of sixth order equations
obtained from NASA SP-3001 (Reference 6). Figure 2-15 shows the
computed temperature distribution. For the calculation presented,
temperatures in the recirculation zone are of the order of 3500°F.
Peak temperatures occur at the exit of the air blast injection pipe
where the fuel-air ratio os near stoichoimetric.
After solutions are obtained for the flow pattern, species mass
fractions, and temperature, the conservation equations for formation
of NO and N0_ are solved. The procedure for solution is discussed in
the KVB Engineering Final Report (included as Appendix II of this
report). Figure 2-16 shows the distribution of lines of constant NO
mass fraction. Peak levels occur in the highest temperature region
near the air-blast injector inlet. Summation of the concentrations at
the exit weighted by the mass flow at each radius give a primary zone
exit NO level of 3 ppm by weight. This is well below the levels
required to meet the program goals. However, the calculation pre-
sented was performed for program checkout only when further work was
halted. Further computations should be conducted with refined finite
difference grid patterns, internal fuel vaporization, and at other
engine operating conditions for both the Class A and Class B combustors
Approximately 8 engineering hours and 6 computer hours would be
required to complete the analysis.
In addition to the foregoing analysis, a simplified procedure f^r
extrapolation of emissions from empirical data has been developed.
This procedure is being evaluated as a means of reducing the number of
test points required to assess the emission characteristics of a given
test combustor modification. This procedure has been integrated with
engine cycle mission analysis programs to allow rapid assessment of
emissions over the Federal Driving Cycle.
AT-6097-R12
Page 2-43
-------
1208 =
TEMPERATURE = 2850°R
PflKPAHID
wmrrm
AFPKOVCD
SCH
J.l-71
CLASS B COMBUSTOR
PRIMARY ZONE
TEMPERATURE DISTRIBUTION
FIGURE
2-15
,„, i --
roan »T»»A.«
AT-6097-R12
Page 2-44
-------
0. ~
NO CONCENTRATION
= 4.0 PPM BY WEIGHT
PREMIXED VAPOR-AIR MIXTURE
AT BOTH INLETS
INTEGRATED
EXIT LEVEL
=3.04 PPMW
fftftttD
WRITTEN
APPMOVCO
SCH
11-71
.
r*T AQQ n /^OMRTT^THTJ
PP TM fiPV 7nMP*
DISTRIBUTION OF NITRIC OXIDE
FIGURE
2-16
;
4
AT-6097-R12
Page 2-45
-------
2.5 Experimental Flow Visualization Model
A plexiglass model of a preliminary film vaporizing combustor
design was employed in flow visualization tests to verify internal flow
pattern computations by the Gosman method. The model was installed in
a three-dimensional water rig and the flow patterns were analyzed with
tracer particles. Figures 2-17 and 2-18 are photographs of the test
rig, Assembly No. SKP26283, as installed in the test facility.
In addition, high-speed motion pictures were taken of the flame
development during operation of the model in an ambient pressure air-
flow rig. The plexiglass model of a film-vaporizing combustor was
installed in the ambient airflow test rig, and air was drawn through
the model with a vacuum header. Motion pictures of the flame-
stabilization process within the combustor were taken at 4000 and 8000
frames per second during two-second operation tests. The motion pic-
ture of the film-vaporizing combustor flame stabilization process was
presented at the EPA in Ann Arbor, Michigan on August 24, 1971.
Minimal-volume intense recirculation and absence of combustion in the
high velocity fuel film along the combustor wall are illustrated in
the film.
The primary zone flow pattern observed during operation of the
model correlated very closely with the calculated streamlines from
the flow pattern analysis, thereby providing reasonable confirmation
of the analysis by the Gosman method. Further testing was conducted
on the three-dimensional water analog model to establish primary zone
flow patterns for comparison with the analytical model. Figures 2-19
and 2-20 are photographs of the trajectories of tracer particles used
to define the flow path inside the combustor. An overlay for Figure
2-19 has been made to indicate the direction of motion. Note the sim-
ilarity between the flow pattern of Figure 2-19 and the analytical
results shown in Figure 2-21.
AT-6097-R12
Page 2-46
-------
THREE-DIMENSIONAL WATER ANALOG RIG
SHOWING CLOSEUP OF INSTALLED COMBUSTOR MODEL
FIGURE 2-17
AT-6097-R12
Page 2-47
-------
THREE-DIMENSIONAL WATER ANALOG RIG
'vtv'
,«,,_,**•**•'- »AU,i.
*.-- «cifj#ig*W •V-'" •-" j
-,'-:
*'!
' ' V
.v .v -ii-,-. ^*^f.^' *
, i- i •»•,,;>,' v.
- f- »-«..' ^i; ' ^,
FIGURE 2-18
AT-6097-R12
Page 2-48
-------
FIGURE 2-19
AT-6097-R12
Page 2-49
-------
FLOW PATTERN RECORD COMBUSTOR OVERALL VIEW 3-D WATER ANALOG
LOW EMISSION COMBUSTOR STUDY
FIGURE 2-20
AT-6097-R12
Page 2-50
-------
O I—
O <
ii 7T
UJ
UJ
UJ
C*L
Q.
-------
The test was conducted at a Reynolds number of 1.42 x 10 based
on the combustor mean diameter. For Reynolds similarity with the
Class B combustor design condition, a Reynolds number of 4.02 x 10
would have been required. Once the Reynolds number is sufficiently
high to ensure turbulence, however, the flow pattern becomes indepen-
dent of Reynolds number, so the test condition was set up only to
establish turbulent flow (Re, > 10 ).
d
The photographs were obtained by introducing small polystyrene
spheres with a density approximately equal to that of water into the
rig inlet and allowing them to circulate continuously. The tracers
were illuminated as they passed through the combustor by projecting a
two-dimensional beam of light from a 1000-watt quartz lamp through the
combustor along the centerline of one set of orifices. Then with an
appropriate adjustment to the camera shutter speed a tracer direction
of travel was established by noting that the brightest illumination of
the particle occurs just as the shutter opens and then trails off in
the direction of motion of the tracer as the shutter closes.
Because of the two-dimensionality of the light beam and the
photograph, it is impossible to determine whether the tracers are mov-
ing into, out of, or completely within the flow pattern plane being
photographed. It is, therefore, impossible to record the effect of
swirl other than by actual visual observation of the model. In this
case, difficulty with plugging of the cooling air swirl passages with
tracer particles was encountered.
AT-6097-R12
Page 2-52
-------
3. COMBUSTOR TEST RIG AND INSTRUMENTATION
A full-scale test fixture with a removable test section, capable
of operating over the combustor operating range as determined from the
cycle analysis, was designed for use in the test program. A separate
component test section to simulate critical flow paths around the con-
bustion chamber was designed for both combustors. Each test section
was designed into a section of standard-diameter pipe flanged at both
ends to allow common inlet and exhaust plumbing.
A clean air supply that is capable of achieving combustor inlet
temperatures typical of recuperated engine cycle operation was pro-
vided for this program in three steps.
(a) From the start of the program to November 17, 1971, the
existing laboratory facility heater (a gas-fired heat
exchanger) was used to obtain a maximum temperature of 740°F.
(b) Added capability was achieved (up to 1000°F) by using the
discharge air from a cross-flow exhaust recuperator.
(c) Finally, by early November 1972, a new preheater (Model 1030)
was installed at the test facility. A combustor inlet tem-
perature of 1200°F was thought achievable, but, subsequent
test experience demonstrated only a 1030°F maximum temperature
was achievable at the test airflow rates.
3.1 Rig Design and Fabrication
•
It was decided to test the combustors in an 8-inch diameter pipe
to eliminate any effects due to non-uniform external flow conditions
and to ensure that combustor performance measured will be affected by
combustor design features only.
AT-6097-R12
Page 3-1
-------
Figures 3-1 and 3-2 show a section layout and a photograph of the
combustion high-pressure test rig respectively. Common inlet and
exhaust plumbing are used for both classes of combustors with a sepa-
rate test section for each. A pressure drop is taken across a screen
installed ahead of the test section to ensure uniform distribution of
the incoming airflow. The two test sections are fabricated from short
sections of standard diameter pipe flanged at both ends to facilitate
installation into the test cell. The combustor is mounted on four
half-inc., Lnreaded rods located at the inlet end of the test section
and slips into an exhaust collector at the discharge end. The mounting
rods can be positioned to accommodate various length combustors.
fill the gap between combustor delivery and high-pressure combustion
rig hardware delivery a simple low-pressure rig was designed and fabri-
cated in-house.
The low-pressure rig was assembled and installed in the test
facility as shown in Figure 3-3. A single-point probe capable of
traversing across the combustor exit in a single plane was used to
measure exhaust gas temperature and to pick up an emissions sample
simutaneously.
3.2 Instrumentation
The combustor rig test sections were instrumented to measure total
and static pressure and temperature at combustor inlet and discharge.
All temperatures were measured by ungrounded-junction, shielded, high-
recovery factor," thermocouples; combustor discharge-temperature-measuring
thermocouples were aspirated. Combustion chamber liner temperatures
were measured with temperature-indicating paint. A weight-'flow rate
system was used for fuel flow measurements, and standard orifice mea-
suring sections was used for airflow measurement.
AT-6097-R12
page 3-2
-------
>-
Q
<£>
5
o
u.
o
Ul
00
s
O
u
AT-6097-R12
Page 3-3
-------
LOW EMISSIONS COMBUSTOR STUDY, HIGH-PRESSURE TEST RIG
FIGURE 3-2
AT-6097-R12
Page 3-4
-------
LOW-PRESSURE (ATMOSPHERIC) TEST RIG
FIGURE 3-3
AT-6097-R12
Page 3-5
-------
3.2.1 Combustor Performance
Performance data were measured by the instrumentation described
below and recorded on data sheets shown in Appendix III.
(a) Pressure Measurements - Static pressure measurements were
taken with wall static taps, four each equally spaced at both
combustor inlet and discharge. Total pressures were mea-
sured with Kiel-type total pressure probes, one each cor-
responding to each static tap. The pressure readouts were
either vertical water or mercury manometers graduated in
0.1 in. increments or Wallace and Tiernan Bourdon-type gauges
graduated in 0.5-in. increments to 300 in. Hg. Advertised
accuracy of the gauge readouts is ±0.1 percent of full scale.
(b) Temperature Measurements - Inlet temperatures were measured
with shielded high-recovery factor iron-constantan thermo-
couples to 800°F and with similar chromel-alumel thermo-
couples to 1200°F. One thermocouple corresponding to each
total pressure pickup was installed. Discharge gas tempera-
tures up to 2000°F were measured with a separated chromel-
alumel thermocouples and with platinum/platinum-10 percent
rhodium thermocouples above 2000°F/ Eight equally spaced
two-point probes located at the area centers of equal areas
were used to determine the temperature distribution factor.
The temperatures were read out on Brown recorders: 0-1000°F
with ±2 deg accuracy for the I-C, 0-2400°F with ±5 deg
accuracy for the C-A, and 0-3000°F with ±2 deg accuracy for
the Pt/Pt-10 Rh. Capability to record temperature data
automatically with a digital acquisition system was also
provided.
AT-6097-R12
Page 3-6
-------
(c) Air and Fuel Flow Rates - Airflows were determined by
standard, regularly-calibrated orifice measuring sections
with flange taps. Two sections, an 8.0-in. and a 3.0-in.
are available as an integral part of the test facility
depending upon the required airflow and were used for test
conditions not requiring a heat exchanger. For inlet tem-
peratures exceeding 700°F, a heat exchanger was plumbed into
the system with a 6.0-in. measuring section between it and
the section.
Fuel flow was measured by a Cox Flowmeter. This instrument
determined fuel weight flow rate directly by measuring the
amount of time required to pass a specified weight of fuel.
The fuel weight was automatically measured on a balance
scale, with the known weight input by the operator.
(d) Metal Temperature Measurements - Combustor metal tempera-
tures were measured with the aid of temperature-indicating
paint.
(e) Emissions Sampling Probe - The gas sampling probe consisted
of three individual probes on a common support capable of
being traversed circumferentially and positioned axially up
to the combustor primary zone exit. The pickup points were
located at the area centers of equal areas such that a sep-
arate probe was required for each combustor class. The probe
was actuated by a rotary gear drive (Figure 3-4) supported
on the rig discharge elbow, and the positions were controlled
microswitches on the actuator mechanism. The probe position-
ing shaft was sealed with "O"-rings encased in a cooling
water manifold.
AT-6097-R12
Page 3-7
-------
g
EH
EH
O
§
I
n
w
CO
I
H
H
W
H
AT-6097-R12
Page 3-8
-------
The probe body construction consisted of three -separate
1/16-in. diam probes individually valved with remotely con-
trolled solenoids. The probes terminated in a common support
as shown in Figure 3-5. Cooling air was delivered to the
probe plenum through the outside passage formed by two con-
centric tubes which made up the rotating shaft and discharged
back through the center passage. The probe heads incorpo-
rated a convergent-divergent nozzle at the inlet to assist the
coolant flow in reducing the exhaust sample temperature by
increasing the sample velocity, thereby quenching any reac-
tions continuing at the combustor exit. Reaction quenching
assured that the composition of the sample did not change
significantly between the probe inlet and the analyzing equip-
ment. Use of air as the coolant ensured that the probe tem-
perature did not drop below the 300°F required to avoid con-
densation of the heavy hydrocarbons on the probe walls.
Later, to expedite testing, the emissions samples were taken
from the thermocouple aspirated flow (eight two-point probes)
as shown in Figure 3-6.
3.2.2 Emissions Analyzing Equipment
Continuous monitoring of pollutant levels was performed during
the test phase with the equipment described in this section. Manu-
facturer's data, including principles of operation and model specifi-
cations, are presented in Appendix II. The emissions-analyzing equip-
ment was installed in a truck (Figure 3-7) equipped with an environ-
mental control system. Figure 3-8 shows the instrumentation installed
in the truck. The following paragraphs briefly describes this equip-
ment.
(a) Carbon Dioxide and Carbon Monoxide Analysis - Carbon dioxide
and carbon monoxide concentrations were measured by the non-
dispersive infrared method. Concentrations were measured on
AT-6097-R12
Page 3-9
-------
-j
COOLANT FLOW ,IN
OOLANT FLOW OUT
COOLANT FLOW IN
EMISSIONS SAMPLING PROBE TIP
FIGURE 3-5
AT-6097-R12
Page 3-10
-------
z
o
5
D
CO
Z
OQ
O
o
I—
<
I
ro
o
H
AT-6097-R12
Page 3-11
-------
AT-6097-R12
Page 3-12
-------
oo
I
O
H
AT-6097-R12
Page 3-13
-------
a dry exhaust products basis since dessicants are used to
remove excess water vapor from the sample to minimize inter-
ference. Beckman Model 315A Nondispersive Infrared Analyzers
with the following specified ranges and accuracies were used:
Range: Carbon monoxide; 0-100 ppm, 0-500 ppm, and 0-2500
Carbon dioxide; 0-2 percent, 0-5 percent,
0-15 percent
Accuracy: 2 percent full scale for 100 ppm CO range
1 percent full scale for 500 and 2500 ppm CO range
1 percent full scale for 2 percent and 5 percent
C02 range
(b) Oxides of Nitrogen Analysis - Nitrogen oxide concentration
were measured by the homogeneous chemiluminescent method.
Water vapor was removed from the sample prior to entering
the analyzer so results were measured on the basis of dry
exhaust concentrations. A Thermo Electron Model IDA was
equipped with a thermal converter to break down NO2 to NO and
0-, thereby making it possible to monitor total oxides of
nitrogen, (N0_ + NO = NO ) instead of nitric oxide only.
£ JC
The instrument has the following advertised range and accur-
acy:
Range: 3-10,000 ppm
Accuracy: ±1 percent full scale (±1 ppm)
(c) Unburned Hydrocarbon Analysis - Unburned hydrocarbon concen-
trations were measured on a wet basis with a heated flame
ionization detector to minimize response errors due to
absorption-desorption of the heavy hydrocarbon molecules
between the probe and the analyzer. A Beckman Model 402
High-Temperature Total Hydrocarbon Analyzer was used. This
AT-6097-R12
Page 3-14
-------
instrument came equipped with an integral heated 10-ft long
sample line, thermostatically controlled at approximately
200°C. Advertised range and accuracy are:
Range: 0-5 ppm to 0-5000 ppm (as methane)
Accuracy: 1 percent full scale
AT-6097-R12
Page 3-15
-------
4. DATA REDUCTION METHODS AND PRESENTATION
The following analytical procedures were used to reduce the data
obtained during combustor performance testing. Data reduction was com-
puterized where practical.
4 . 1 Combustor Performance Data Redu -tioi
Conventional combustor performance parameters including press re
loss, temperature distribution, lean stability, efficiency, and me .al
temperature levels were measured at selected test conditions. Data
were reduced according to the following procedures .
(a) Pressure Loss
Combustor pressure drop was calculated as
(P - P ) /P
xtrp £ m ' / r m
in out in
where P- and P represent total pressures at the combustor inlet
in xout
and outlet, respectively. The pressure used in the denominator was a
circumf erentially averaged value from the individual probes while the
numerator value was obtained from a pressure gauge. The measured val-
ues of inlet total pressure were checked against values calculated
from a measured static pressure, airflow rate, and inlet temperature
with the aid of Mach tables.
(b) Temperature Distribution
Discharge gas temperature distribution was calculated as a Tempera-
ture Spread Factor (TSF) defined as
T - T
- max mean
mean inlet
AT-6097-R12
Page 4-1
-------
where T and T are the maximum measured and average measured
max mean
total temperatures at the combustor discharge, respectively, and T.
is the combustor inlet total temperature, all in degrees F.
(c) Ignition and Lean Stability
Ignition limits were determined as minimum fuel-air ratio required
to light as a function of combustor airflow. Combustor lean stability
was calculated as fuel-air ratio at lean blowout at the specified test
condition. The values of airflow and fuel flow were measured. Both
ignition and lean blowout were determined by monitoring combustor tem-
perature rise.
(d) Combustion Effficiency
Combustion efficiency was calculated as
(f/a)ideal
actual
where (f/a) . , , is the ideal fuel-air ratio required to obtain the
measured combustor temperature rise. This value is obtained from con-
stant pressure combustion charts. (f/a) . , is the actual fuel-air
ratio from measured air and fuel flow rates at the test condition.
The efficiency calculated from the above expression was checked
against an efficiency computed using the measured levels of carbon
monoxide and unburned hydrocarbons, assuming these constituents are
the only products of incomplete combustion.
(e) Combustor Metal Temperature
Metal temperatures were determined directly as isotherms on a
thermindex-painted combustor.
AT-6097-R12
Page 4-2
-------
AIRESEARCH MANUFACTURING COMPANY OF ARIZONA
A computer program available at AiResearch for engine emission
data reduction was modified to allow computerized data reduction of
combustion rig data. Both emission index and pollutant generation
rate in Ib/hr are calculated. The emissions concentrations are cor-
rected to concentrations in wet exhaust from a combustion process with
dry air. In addition, combustion efficiency is calculated from the
measured carbon monoxide and unburned hydrocarbon concentrations. A
typical computer printout is shown in Figure 4-1.
It should be noted that unburned hydrocarbon weights are calcu-
lated as methane, CK^. If it is necessary to convert the emission
indices to equivalent CH^gg or CgH^, the printed weight values should
be multiplied by 0.865 or 0.895, respectively.
In the reduction of the exhaust emission data, carbon monoxide,
carbon dioxide, and oxides of nitrogen volume concentrations will be
considered dry analysis data because of the use of a dessicant or con-
denser in the sampling train. Total hydrocarbons concentrations,
however, will include the water vapor initially in the air plus the
water vapor formed by the combustion process. For the purpose of
reducing the data on a volumetric basis, the concentration levels will
be corrected to percent (or parts-per-million) by volume of wet exhaust
gas from a combustion process with dry air. Accordingly, the wet
analysis data will be initially corrected to dry conditions as follows:
C II
S' = , (1)
1 - v - u
AT-6097-R12
Page 4-3
-------
o> o
no
O X I/I
cc z> •
< X X
CD t-i
Ol-
a u
a _i
in IA
•* e o
i
a.
-IX X
UI 13 X
3 X J
U. •
JO
UI _t
or u.
xr~ o
in o* IP o
«•* ft
!*- I
ui an
t- uiu. u
oc a.
•- z
in»- » _i
in in ui
«-
x a
a in x
cc t- _i
x <
o cc
« CO
o ^-
c
a i a
z u. «
in»- u.
o
U! •-«
o
or in
VO 0\
in o
•
in
o
•
in
m oo
(£ at
CT4
fvl O
(S
o » •
in
in
O
in
o
coino
ma>_«
»>*«CD ««•«-*
cc-cucc z
oa
O
1 UI
or
o x
o- x
rj
in
O
VO
O
O O CO
o « «
o k •-
O
n
%
0)
o
in in
« c
x •-
ui i/>
(!)
a —
u; a u
o
c «-H
C
^ H
e . • o
in -o «m
• 40"
m ri o> e
I
o doo
• • oo
o a oo
!o-o-
!o
i-*
2 ,- !
>-
a
•c *+ r- o* o
z
c
••4 « »< r-»-
ir in o " -
x
c
i (VI
o
X CC
a _j
i u
U1CE
: x
CC
1 O —I
• —• o
' < «z ui
I ' X ZO =>
, X O •-! U.
.I* on i-«_i
o o o" oor in i- z
m a. ui o
! , oo zee
: OO O
. H- O
1 UI _l C
I m *r~ ~* I3O N
mmco ozu.ori-1
• •» U.OT
Q O
, Z Z CVIO.
O «O «J
•-" Z >
I- «
1 < t- in in
:o(vi(Wz
c
H
*^ mmcoo
rvi o .-i m •» z t- o o o
uo za
UI ••*!&.
or x o
a: o z c
o cc ui
. cr
o
z
« =>Z i
CC U
ui in uj .uiz utuiuzz
- - i O —; 3 U UJ
u. o. u. x u. , , u. _^ u.zqu.»-<
u,a u. c •- : u. cc, ' u. ct uj U.CTOK —
i->-oxiu->ox UJD->-OX — >->-oxtni->-e>x ^inu.
ui cc ^uiwcc -^otuix xmuicc xouicc xincdZb.
D r: • i zxo s. a. t cccrz .CDICI- xxoui»-<
c c _; u- u. au:u. _i — u> ui u' ui zuiui IAZVIC
i- x ui
uj a D
or
• - u. I
ui ui c x
uj>-xu/oc_<_iocc_i_ioox — _i_ioa>->_i_ioco_i_iooX4Uiin
c!rcco.-j«-iOC'-_io;oo»--Ji oo^-lxoo^j 'COP-_IUJ> ~
•- u. x» u. a > > u.ccui>-> u.o>> u. in > > u. MVICC
x >- i- fu o or v cc _io • cc '«-• cc ui or _i g in x
CZZCI-ZWUII-TXVU,"^ «->->-ui>-ci>->-ui»-ci>->-ujo--ia«o
t^l^UlUXCCCCCCLX ECCO,XCXCCCCO,.T. CD CO a I v-t t
ctjuirx eo tr»-o isz ex1
aar«« xx««uixx«i-ii- xx««-(Uixx«»«r
— * — - - - — - - . aacoui
cr
cc
cr
z
a
»-
•-«
z
a
»-
•^
iz
1-1 '
«
»-
o
u
O
Z
AT-6097-R12
Page 4-4
-------
where:
S1 - volumetric concentration of pollutant species in dry exhaust
gas
S" - volumetric concentration of pollutant species in wet
exhaust gas containing water vapor in rig inlet air plus
water vapor formed by combustion process
v - mole fraction of water vapor in exhaust from initial
conditions of inlet air = h (M /M )
u - mole fraction of water vapor formed from combustion
process = R/2 (a1 + b1 + d1) = 2d'
h - Ib water vapor/lb dry air at rig inlet determined from
relative humidity of inlet air
M - molecular weight of exhaust products = 29.0 Ib/lb-mole
6
M - molecular weight of water vapor = 18.02 Ib/lb-mole
a1 - mole fraction of carbon dioxide in dry exhaust
b' - mole fraction of carbon monoxide in dry exhaust
d1 - mole fraction of total hydrocarbons including total
aldehydes (expressed as CH.) in dry exhaust. Use d" in
Equation (1)
R - hydrogen/carbon ratio of atoms in fuel
AT-6097-R12
Page 4-5
-------
The volume concentrations so determined and the dry analysis
carbon dioxide, carbon monoxide, and nitrogen oxide data will then be
corrected to fraction by volume of wet exhaust gas from combustion with
dry air by the following procedure:
S = S1 (1 - u) (2)
where
S - volumetric concentration of pollutant species in wet exhaust
gas containing only the water vapor formed by the combustion
process
An emission index expressed as pounds of pollutant per thousand
pounds of fuel consumed (equivalent to an index expressed as mg/gm fuel
consumed) will then be calculated according to the following equation:
PT (1 + f) S M
EI S N
where terms not previously defined are:
N = 3 if S is fractional
N = -3 if S is in parts/millions
El - emission index, Ib pollutant species S per 1000 Ib of fuel
O
consumed
f - fuel-air weight ratio determined from measured fuel and air
flow rates at the specified test condition, Ib fuel/lb dry
air
AT-6097-R12
Page 4-6
-------
Mg - molecular weight of pollutant species S, Ib/lb-mole
= 86.17 for CrE,,
D -L4
= 44.01 for C02
= 28.01 for CO
= 16.04 for CH.
= 30.01 for NO
= 46.01 for N02
Total oxides of nitrogen on an emission index basis will be computed
as N02 to comply with the 1976 Federal Standards.
The expression for the fuel-air ratio, as shown below, will be
used to cross-check the measured values:
MC + ™E . a' 4- b'
MA 1 + Sf + £ <* - 1) + d <|- 2) (4)
where terms not previously defined are:
M - molecular weight of air = 28.9 Ib/lb-mole
A
M - atomic weight of carbon = 12.01 Ib/lb-atom
M - atomic weight of hydrogen = 1.008 Ib/lb-atom
rl
Pollutant emission rates were also calculated to allow computeri-
zed integration over specific driving cycles to obtain an emissions
index in terms of pollutant weight per vehicle mile. The emission
rates were calculated by the following expression:
Ps - EIS x Wf (5)
AT-6097-R12
Page 4-7
-------
where
F - emission rate of pollutant species S, Ib/hr
Wf - measured test condition fuel flow rate, Ib/hr
The vehicle mass emission rate as a function of the driving cycle
was then computed from:
F
PS = 2.205 x 10~3 -§. (6)
where
P - vehicle mass emission rate, gm of pollutant species S per
D
vehicle mile
V - vehicle speed, mile/hr
4.3 Humidity Corrections to Emissions Results
It is known" that moisture in the air at combustion inlet can have
an significant effect on the emissions from a combustor. NOV emissions
A
are particularly affected. However, quantitative corrections for these
effects have not been established. Therefore, other than the small
volumetric corrections discussed in Section 4.2, Gaseous Emissions
Data Reduction, no attempt has been made to correct the emission test
results to any other combustor inlet condition than the dry air condi-
tion under which it was tested (0.0006 Ib water vapor/lb dry air).
AT-6097-R12
Page 4-8
-------
5. COMBUSTOR DEVELOPMENT AND EVALUATION
The experiment program has been divided into four test periods for
purposes of discussion of the test results. This division combines
groups of tests generally having common available test facilities and
program direction.
The first two test periods, May 12, 1971 through December 10, 1971
and December 11, 1971 through January 31, 1972, cover the preliminary
testing conducted prior to the contract hold period. The third test
period, February 1, 1972 through August 10, 1972 covers work conducted
by AiResearch during the contract hold period and up until a heater
failure occurred. The fourth test period covers the final calibration
test series conducted after the heater facility repair.
5.1 Test Period (5-12-71 to 12-10-71)
Preliminary tests were conducted on an atmospheric test rig on
four combustor configurations to compare combustion efficiency at high
loading with predicted levels. Atmospheric air was drawn through the
rig by the application of a vacuum to the combustor discharge. The
first test was conducted on a modified film-vaporizing combustor that
was initially designed for a high-velocity, high-heating-value-per-
unit volume fuel (Shelldyne-H) under a contract with the Air Force
Systems Command. The combustor was modified by the addition of a con-
ical surface on the baseplate with the apex of the cone located at the
plane of discharge of the primary pipe. The second configuration was
a baseline, unmodified configuration.
The combustors were tested at the Class A design point corrected
flow conditions to ensure a reasonable pressure drop during operation.
The pressure drop would have been too low at the Class B flow condi-
tions since the effective open area of the 'test combustors were approx-
imately three times that of the Class B design and two-thirds that of
AT-6097-R12
Page 5-1
-------
the Class A. Purpose of the test was to compare lean stability and
combustion efficiency of the film-vaporizer combustor with and with-
out the cone. The test results shpwed that performance of the combus-
tor with the cone was superior in both respects and also had a shorter
ignition delay. Specifically, lean limit blowout fuel-air ratio
decreased from 0.0045 to 0.003 and combustion efficiency increased
from 74 to 82 percent at the Class A design-point corrected flow con-
ditions.
Emissions data were taken solely for the purpose of calculating
combustion efficiency since the rig had insufficient instrumentation
for a thermodynamic efficiency determination based on the measured
temperature rig. The emission level was obtained by averaging the
value measured by a single-point probe as it traversed across the
combustor exit plane. The efficiency was then calculated from an
enthalpy balance using the actual measured combustion products.
The ideal enthalpy change per pound of fuel consumed is equal to
the heat of reaction or fuel lower heating value. The actual enthalpy
change is the lower heating value of the fuel minus the heating value
of each non-ideal combustion product formed. The efficiency is then:
(LHV) - ZW (HV)
where:
n = combustion efficiency, percent
(LHV)p = fuel lower heat value, Btu/lb
W = weight concentration of non-ideal combustion product,
Ib/lb fuel
AT-6097-R12
Page 5-2
-------
(HV) = heating value of non-ideal combustion product, Btu/lb
= 18,500 Btu/lb for JP-4
= 4345.2 Btu/lb for carbon monoxide
= 18,646 Btu/lb for CH.. »5 per Federal test requirements
The effect of oxides of nitrogen is negligible. Therefore:
18,500 - 4345.2 W_n - 18,646 Wu_
= 100 - 23.48 Wn_ - 100.8 WH_
L.U nt.
or on an emission index (El) basis in terms of pounds pollutant per
thousand pounds of fuel burned,
n = 100 - 0.02348 EInr. - 0.1008 EIU_
CU nv-
The emission index was calculated from the measured data using
the method detailed in Section 4.2 of this report.
Figure 5-1 shows the efficiency levels of the two test combustors
as a function of aerodynamic loading parameter superimposed on the
published efficiency map that was generated during the development
program. An increase in efficiency at both of the test conditions
established was observed for the conical-domed configuration. In
addition, the efficiency values were consistent with the earlier data.
Both the cone-domed and flat-domed versions of the Class B combus-
tor (SKP26312) were also tested on the vacuum rig. Figure 5-2 shows
the efficiency of the combustors compared with both the predicted
AT-6097-R12
Page 5-3
-------
P = INLET PRESSURE, ATMOSPHERES
T = INLET TEMPERATURE, °F
V = VOLUME PER CUBIC FOOT
W = AIRFLOW, POUNDS PER SECOND
FUEL-TO-AIR RATIO = 0.020
8
cm
w
H
u
H
fa
CO
ID
O
u
100
90
80
70
60
50
40
30
20
10
\
^
B
\
A
^
\
"X. G1
X
\
X
^
N
O
s^
k
k
^
^
•s
\^
^
i iii
A= T76
B= TFE731
C = SHELLDYNE-H COMBUSTOR
X-
n
^
c,
c,
Nc
v
WIT
STD
^
H CO
DOM
\
NE
E
"^
0.2
0.4
0.6
0.8
1.0
1.2
AERODYNAMIC LOADING PARAMETER, Q =
PERFORMANCE COMPARISON
W
VP
COMBUSTOR EFFICIENCY COMPARISON
FIGURE 5-1
AT-6097-R12
Page 5-4
-------
90
80
70
60
50
40
30
20
10
0
:i
t r
#
"4-
V
s
-
'"ff
- H-
-
+4
:£
- z -
-^H
- w "
_ n< Z_
~ * :r
I U |1
~ H i.
: U -4
"EP
- fa --
ULLLLLUi.LLUJ.i
MBUSTION I
1 1 1 ! 1 | 1 1 M 1 M | | 1
=1o ^
j
-r-; 1
" "^ — M
1
j
~~~ /•
_) —
; t~
:i
s
s,
\
-
-
-
N,
s
-
I!
S
I
r
1
1
-f
_E
_,..
M
J-
V
S
«
^
4-,
i
I
4.,
_
Jl
^
-
.
. +
t \
"•j~ ~
UJ
~1
I —
i
pU.,
L£
5£
; B
^
^
V
s»
S
V-
^
-T4
i>
El
FI
V
—
-
-1
\
\
-
-
'I
G
-
__
'n
u
V
\
s
s
4-
i
-
:
t
-
1'
:c
R
fs
h
,
\
s
S
^
T
V-
s
_L.
— .
:i
E
—
^
I
—
-
I
]
L-|
t-
-
-
\
S;
-
\
s.
I
sl
\
i i
-
s
-
S
1
— I —
I-
I
-I
r
-
h-
0.4-
:i>
5
K
-
:Y
2
1
-
I
H
I —
s
S
=
-
-
,1
^
•^
k
r
:\
i —
tL_
1
s
-
s
s
s
-
V
S
N
\
4-i
-
s
\
*>
$
S
^
VU-
r|r
V
s
-
-
r
"
^
>
• ^
Vl
-
^
"
::1
Q
*
v3
-
-
-4
^
k
s,
^
*
SJ
I
>
/
V
^
--
s
»-
— ,
^
i
j
„
\
"S
NJ
V
V
-
5i
•
—
-
-
-
V
!s
-
-:
i
V
-
—
-
X
C
_
s
-
-
s
-
s
C
i
-
-
;:
(.
V
<~
C
i —
•«
N
W
a
-0 ft
~|-j"H^~
re
;i
— 1
J
-4-
-4-
-1
--•-
\
i
P2
+-
—
-t-
u
-
rn •*-/ ^
0.8
—
~
=\
>
^a
SKP26312
SKP26312
SHELLDYN
^
i-
*
-
P =
V =
T.=
K =
444^
^
-
-
s
-
-
.1
.
E
..
J
W
S
Jl
ITH
TD
H C
-
-
CONE
DOME
OMBUS
r i
3ESIGN LINE
s.
V
-
1
_
^
-
s
-
s
-
1.0 -
s
-
-
-
-
-
-
-
-
-
.
T
-
.
J
J
0
-
-
P
P
R
,
.
-
|
'T~
"h
I
1
•~
-3
•-
LB/SEC
INLET PRESSURE ATM
COMBUSTOR VOLUME, FT
INLET TEMPERATURE, °
f/a > 0.02 (f/a)/0.
f/a < 0.012 0.012/(
--r-Li I UJrjd=j=H-H=HH - ++ \
L-
1
-
-
i
J
4
4
\\
-
-
-
—
--
-
-
-
=
3
R
020
f/a
4
f
--
--
--
_
-
I
)-
Page 5-5
-------
values and the curve from Figure 5-1. The data indicates that
development effort is required to improve the efficiency to the to the
desired level, but that the efficiency is not significantly worse than
the more highly developed Shelldyne configuration. The data did indi-
cate, however, that the flat-domed configuration was slightly more
efficient, but the low overall efficiency levels, in general, are
attributed to fuel maldistribution within the primary zone that was
observed during the testing.
Following the completion of this series of tests, the high-
pressure test rig was completed and installed at the combustion lab-
oratory. High-pressure combustion rig testing was therefore started
(November 17, 1971) on the Class B film vaporizing combustor, P/N
SKP26312. The test results are shown in Figure 5-3. Note in Figure
5-3 that the effect of pressure is the reverse of what would be
expected from simple kinetic theory. Figure 5-4 shows a typical chart
recording of emission readings.
These results are converted to grams/mile for the simulated
Federal Driving Cycle which reflects the variable geometry, free
power turbine engine performance as follows:
(a) Figure 5-5 gives data on the Class B engine assembled in
accordance with contract specified items and to provide
150 hp.
(b) Figure 5-6 summarizes data extracted from the optimization
study (68-04-0012) and provides the basis for selecting
5.5, 9.5, 18, and 27 hp conditions to represent the Class B
engine over the Federal Driving Cycle.
(c) Table 5-1 tabulates the Class B test points and gives the
constants used to convert the emission index (El) from com-
bustor test results to grams/mile for the Federal Driving
Cycle.
AT-6097-R12
Page 5-6
-------
CO
CO
U
I
CM
CO
•O
a.
^
CO
CO
J!Z U. xO
< o 8
r- _ OT
o
LU
I
in
Cvj
un
CM
O
O
X
UJ
(0
o
CO
s
in u.
3 I
a.
a. —
CM
O
O
O
01
D
a.
00
H
0 -">
\
0 D 0
ro
I
in
D
O
H
O
O
O
o
o
O CM O -00 «O
SIAI1V
AT-6097-R12
Page 5-7
-------
AT-6097-R12
Page 5-8
-------
LU
U
>-
u
O
z
CO
CO
CO
u
K
O
O
IU
LL
a
»•
a.
O
O
p
o
\
o o
§ § 9
?
ot oT
I
o
AT-6097-R12
Page 5-9
U>
SB
•LO- •«
CO-
CM'
.§
»-
g
ce
UJ
LU
V)
cc
O
_ .O
in
t/>
1
LU
in
i
§
g
H
-------
C_D
UJ
CL,
CO
m
en
CO
<
_i
o
CM
LU
CO
cc
V-
H
cc
CO
o
8
UJ
O
cc
<
o
u
CO
UJ
5
1
CM
O
UJ
in
s?
ui
u
CO
o.
a.
g
I
UJ
Q. Q.
X X
in in
in o'
CO
•z.
O
cc
UJ
_l
UJ
U
CO
UJ
co 3
UJ
UJ
O
UJ
UJ
00
i
a.
CO
111
5
0.
e Q
CM
I
UJ
3
a.
o
u.
_j
UJ
Q.
S2 es
CO
X.
UJ
_l
UJ
U
UJ
O
s
o
UJ
UI
85
UJ
o
<
oc
UJ
|
(3 P
H 2:
^ Q.
I o-
t q
.3 8
S g
i- „
•^ z~.
I i.
ui
UJ
D
csr
2
in
OC o
w o H
*- x fa
CM
cn
to
CO
°
3IAIIl'±IM30a3d
AT-6097-R12
Page 5-10
-------
a
CM
(N
I
in
W
CO
O
U
ffl
CO
CO
U
I
H
O
ft
EH
CO
n
Sets Combusto
I
H
00
cn
CN
0
o
0
0
0
CD
00
CN
O
O
O
CN
00
r-t
0
vo
VO
1-1
00
oo
r~-
o
^*
CD
H
0
0
O
,_,
CN
O
r-4
O\
H
ps
t»v
I
VO
CN
in
r~4
o
0
0
00
CN
O
f-
O
CN
0
CN
00
CN
vO
CO
o
o
o
0
,_,
m
o
VO
OO
CN
00
00
CTl
00
O
00
O
O
0
O
in
o
1-1
o
VO
oo
CN
00
H
CN
m
00
O
CN
VO
oo
CN CN
o
o
in
O
O
CN
CN
in
o
in
r-
r-
o
*
o
X
CO
CN
ro
CN
o
X
in
oo
oo
o
•
o
X
in
m
M
U
0)
co
XI
CN
•
O
oo
CN
CN
m
00
10
VO
m
in
U]
e
m
•
in
in
oo
CN
o
VO
o
m
tQ
C
O
•H
03
U)
AT-6097-R12
Page 5-11
-------
Note that the constants were derived as follows: At 5.5 hp, time
is 8.06 min out of a total of 22.709 min and 7.435 miles for the FDC.
With a flow of 9.0 pph, 1.209 Ib of fuel (JP-4) is burned. Then,
EI5>5 (lb/1000 Ib) x 1QQQ9 x y^| = 0.0738 x Elg 5
is the grams/mile contribution of the emission for the 5.5 hp segment
of the Federal Driving Cycle.
This procedure, using trend data from the Class A combustor allows
the data of Figure 5-3 to be converted to 5.0 grams/mile of NO-. The
CO and HC values were not converted because of the possible unreli-
ability of a value obtained from trend data of another combustor.
However, the values are believed to be below the 1976 Federal Standards
for carbon monoxide and hydrocarbons.
(T - T \
ffax—=— ), was not measured in these tests.
L2 ~ Ll I
The collection of data for the Class B combustor was terminated
because of damage to the combustor. Figure 5-7 shows a hole burned in
the downstream end of the combustor. In addition, a crack developed
in the exit elbow of the rig.
After repairing the rig, testing was resumed using the Class A
combustor. Use of only the Class A combustor for further testing
was justified for the following reasons:
(a) The Class A engine is preferred as a result of the automobile
engine optimization study (EPA Contract No. 68-04-0012) .
(b) It is generally more difficult to meet the 1976 Federal NOX
Standard with the Class A combustor, because of the higher
combustor inlet temperature of a regenerated engine.
AT-6097-R12
Page 5-12
-------
i,
""rS^W »it " t"> -m -.•".'*/»*
CLASS B COMBUSTION CHAMBER, P/N SKP26312 SHOWING
DISCHARGE LIP BURN-OUT
FIGURE 5-7
AT-6097-R12
Page 5-13
-------
5.2 Test Period (12-11-71 to 1-31-72)
5.2.1 Emissions Performance
Thirteen Class A combustor tests were conducted during this test
period, including various modifications of the following basic combus-
tor configurations:
(a) SKP26259 Vaporizer (7)
(b) PAP218770 Premix (2)
(c) SKP26489 Vaporizer (3)
(d) SKP26489 Pneumatic Impact (1)
Test results are presented graphically and discussed in the
following paragraphs. Test data were converted to a grams per mile
basis using the AiResearch 4-point test procedure (January 1972)
defined in Section 2.2 of this report.
(a) SKP26259 Vaporizer
Notable features of the baseline SKP26259 MQ (Cone Dome)
Test, Figure.5-8, were NO_ increasing at low fuel-air ratios,
and CO and C,Hn . decreasing at high fuel-air ratios with
D ±4
decreasing NO_. A further observation was NO2 emission index
considerably below the emission index for the Class B com-
bustor. This was attributed to the dilution ports being
closer to the primary zone (on an L/D basis) in the Class A
burner.
AT-6097-R12
Page 5-14
-------
<
u
CM
CN
o.
UJ LU UJ
_J J J
oc oc oc
o o o
Ifl
o o
x x
i85
UIV/13nd
d/d V %
AT-6097-R12
Page 5-15
-------
The first modification (NL) , Figure 5-9, was directed at
running a lean primary zone, high pressure drop and low pri-
mary zone residence time. The result was no change in NO ,
X
which was not the expected result.
The second modification (M2), Figure 5-10, removed the cone
and moved the primary pipe closer to the base plate. This
is a change in the direction of more conventional vaporizers.
The result was an improvement primarily due to reshaping the
NO,, versus fuel-air ratio curve.
The third modification (M_), Figure 5-11, moved the dilution
ports closer to the primary zone following a trend noticed
earlier. The result was a substantial improvement, and what
was not expected was an improvement in combustion efficiency
and temperature spread factor.
Modification 4, Figure 5-12, consisted of removing the pri-
mary pipe and injecting fuel and air into a plenum chamber
attached to the combustor baseplate. The mixture was
injected into the chamber through the radial air distributor
forming the vaporizer impact plate. The intent was to
improve the fuel-air mixing process to ensure a lean mixture
and to try to avoid bringing the fuel and air together from
opposite directions as in the L-pipe injection method, which
has a tendency to establish near-stoichiometric combustion
interfaces through diffusion. This modification exhibited
unsatisfactory lean stability, and a subsequent modification
intended to improve the stability did not result in suffi-
cient improvement to warrant further investigation.
AT-6097-R12
Page 5-16
-------
CO
CO
CM
O
CM
CO
CO
o
z
III
o
in
o
O
r*.
H S
• *•
O
* \
CD
1
CTl
I
in
w
Pi
D
03S/81 AACHdUIV
AT-6097-R12
Page 5-17
-------
<
CO
u
M
ON
m
Ill Ul Uf
AT-6097-R12
Page 5-18
-------
<
u
w
cs
o
CO
cc cc oc
o o o
co d d
£ x
ooz
zoo
aivnand
d/dV lN30U3d
AT-6097-R12
Page 5-19
-------
cc
UJ
N
E
o
a.
SK
J?
f D D
i
o o o
\
8X
ui
m
X
UJ
O
I
UJ
(N
3
6
p
O
«oN'l
AT-6097-R12
Page 5-20
-------
(b) PAP218770 Premix
This combustor (PAP218770) had originally been designed for
natural gas operation for an in-house research program on
combustion noise reduction. The first test, Figure 5-13,
was conducted on natural gas as a baseline evaluation since
prior testing on the vaporizer combustors had shown little
effect of fuel type on NO emission and because the combustor
X
had previously exhibited stability problems during operation
on liquid fuel. The preliminary data was very encouraging.
A modification (intended to reduce the premix chamber fuel-
air ratio) was then made to the combustor. Concurrently the
combustor orifice pattern was modified to reduce the open
area to that required by the Class A combustor in order to
test at the correct flow conditions. Test results, Figure
5-14, showed changes in NO and CO emission as a function of
X
fuel-air ratio that resulted in higher grams-per-mile values.
The change is attributed to the blockage of primary ports in
the combustor main chamber during the open area reduction
process. Apparently the ports had been effective in both
oxidizing primary-zone-generated CO and in partially freezing
the NO reactions.
(c) SKP26489 Vaporizer
The SKP26489 combustor was a vaporizer design derived from a
combustor developed for burning viscous fuel (Shelldyne H).
The first test on the secondary-pipe dome configuration (MQ),
Figure 5-15, resulted in a burned primary pipe and severe
distortion of the baseplate. Despite the failure, emissions
measurements indicated that HC and CO emissions were very
low (see Figure 5-15), corresponding to 92.5 percent
AT-6097-R12
Page 5-21
-------
O
K
K
00
333 a
cc.tc.tr.
u o o
at t o
n r+ q
«-' ^ d
oxoxx
z u u
O
O
"5
3
AT-6097-R12
Page 5-22
-------
I
in
P
O
H
oixva
a0N'i
AT-6097-R12
Page 5-23
-------
5
I
00
CN
o.
CO
UJ UJ UJ
d/dV%
«o
AT-6097-R12
Page 5-24
-------
2 1/2
efficiency at a loading parameter of Q = W/Vol P G ' = 0.80.
Eighty percent efficiency at Q = 0.80 had been used in sizing
the combustor. This indicated that a combustor of this type
3 3
with only 100 in. volume (40 percent of the existing 247 in.
volume) would still meet the HC and CO standards.
The combustor was rebuilt to the Ml configuration by adding
primary pipe cooling and a strengthened baseplate. Somehow,
the improved cooling with perhaps reduced airflow to the
primary zone resulted in an increase in NO as shown in Fig-
ure 7-16 .
At this time, a new high temperature air preheater (heat
exchanger adapted for rig use) was installed in the lab which
provided up to 1000°F combustor inlet air temperature capa-
bility as compared to 740°F. As a result, two measures were
taken to take advantage of this capability.
(1) The FDC emissions simulation test procedure was
changed from a multiple extrapolation involving
temperature, pressure, and fuel-air ratio
(AiResearch 4-point test procedure ME, - December
71) to an extrapolation on temperature only
(AiResearch 4-point test procedure TE, - January
72). As is described in Section 2.2, the later
test procedure involves setting the actual con-
ditions of airflow and inlet pressure and testing
at 500°F, 740°F, and 1000°F combustor inlet temper-
ature. Once the measured emissions data taken at
these points was reduced and plotted (El versus
T. ), the emission index, El, at the correct com-
in.
bustor inlet temperature corresponding to the test
points selected to represent the engine cycle, was
obtained by graphical extrapolation.
AT-6097-R12
Page 5-25
-------
VO
H
I
in
W
o
H
t,
aiv/ianj
AT-6097-R12
Page 5-26
-------
(2) For comparison, this same combustor (SKP26489-M1)
was tested using both the old and the new test
procedures. Figure 5-17 presents the new proce-
dure (TE) test results. The mass emissions levels
in grams per mile were determined for each and are
presented below:
SKP26489-M1
Test Method 1 (ME)
Test Method 2 (TE)
NO (as N0_)
X £*
6.18
4.19
GM/mi
CO
0.413
0.570
CHx (as CgH14)
0.0004
0.104
Note that the simpler and more reliable second
method yields lower NO mass emissions and 'far
X
greater unburned hydrocarbon levels. The slope
variations observed in the plots of unburned hydro-
carbons versus temperature illustrates that extra-
polation on temperature at constant fuel-air ratio
and inlet pressure, as in Test Method 1, may lead
to inaccurate results. Therefore, it is imperative
that all testing be conducted at the correct com-
bustor inlet conditions, if possible. Test results
obtained using the first test method (AiResearch 4-
point, December 1972) should therefore be examined
carefully with regard to interpretation.
(d) SKP26489 Pneumatic Impact
The test on this configuration was shortened because of fuel
dribbling problems associated with the injector that resulted
in high carbon monoxide and unburned hydrocarbon emissions.
Emission index data as a function of fuel-air ratio are
shown in Figure 5-18; the data were taken at 500°F combustor
inlet temperature.
AT-6097-R12
Page 5-27
-------
888
8 I ? *
AT-6097-R12
Page 5-28
-------
100
90
80
70
60
50
40
30
20
ui
to
3
xf
LU
Q
10.0
9.0
8.0
7.0
6.0
5.0
4.0
3.0
2.0
1.0
PNEUMATIC IMPACT
\
CO
HC
NOv
SKP26489-PI
V
\
FUEL
T|N - 500 °F
IN
a
18.8 PSIA
30.0 LB/MIN
0.000 0.0020 0.0040 0.0060 0.0080
CALCULATED FUEL/AIR RATIO
FIGURE 5-18
0.0100
0.0120
0.0140
AT-6097-R12
Page 5-29
-------
The injector was modified (SKP26489PIM1) to eliminate the
fuel dribbling problem but was not retested until later in
the test program.
5.2.2 Emission Pickup Probe Conversion
It was determined early in this test period that an unsatisfactory
correlation existed between emissions profiles measured with the rota-
ting emissions probe and temperatures measured during succeeding tests
with the same combustor and the rotating probe removed. This lack of
correlation was attributed to airflow distortion effects introduced by
blockage resulting from the probe presence at the downstream combustor
support discharge plane. For this reason a fixed probe, drawing its
sample from the discharge thermocouple aspiration air, was fabricated.
An average sample obtained from 8 circumferential pickup points at each
of two radii resulted. The new setup allowed both conventional perform-
ance and emissions performance data to be acquired simultaneously dur-
ing each test.
The new probe was checked out on a previously tested combustor to
determine if the 16-point pickup represented a true average sample.
Data were compared against the average emissions measured at equiva-
lent conditions with the rotating probe. It was concluded that the
new probe was satisfactory because of the higher values obtained in
this manner, as evidenced by the data tabulated below:
COMPARISON OF EMISSION INDICES MEASURED
BY TRAVERSING AND FIXED PROBES
Emission Index, El, lb/1000 Ib
Condition
1
2
NO
X
Rot
2.75
4.47
Fixed
2.75
4.67
CO
Rot
77.8
44.2
Fixed
78.5
46.5
HC (as C6H14)
Rot
32.2
6.5
Fixed
37.0
7.3
AT-6097-R12
Page 5-30
-------
5.2.3 Conventional Performance (Non-Emissions)
After the conversion to the fixed emissions probe was made, con-
ventional combustor performance parameters were also checked where
convenient. Among these were pressure drop, tempera'.ure spread, and
lean stability.
(a) Pressure Drop
The vaporizer combustors were initailly sized for 4-percent
loss of inlet total pressure at 150 hp I free-turbine engine
cycle) conditions. This value converts to 3.08 percent at
the reduced inlet temperature test ccndition. The following
pressure drops were measured as comrared with the 3.08-
percent value:
Measured
Combustor Pressure Drop, percent
SKP26259 MO 3.02
Ml 8.98
High AP/P mod
M2 3.30
M3 3.06
SKP26489 MO 3.60
Ml 3.37
PAP218770 Ml 3.22
The effect of combustor pressure drop on NO emission index
X
is illustrated quantitatively in Figure 5-19 for 5 combus-
tors. In general the L-pipe vaporizers have the same slope.
AP -°-176
El a (f-)
The slope of the line for the PAP218770 premix
AT-6097-R12
Page 5-31
-------
O
ui
a:
a.
a*
O
CO
D
fiQ
5
O
u
I
in
O
H
PH
—
an oooi/aT '
NoiiviAiaod ZON
AT-6097-R12
Page 5-32
-------
combustor is slightly steeper, and the exponent on percent
pressure drop is -0.327. No quantitative, theoretical, or
physical significance has been determined for these exponents,
although qualitatively the fact that NO decreases with
X
increasing pressure drop can be attributed to the lower resi-
dence time that occurs due to the higher combustor velocities
associated with higher pressure drop. Probably more impor-
tant, the higher pressure drop also promotes improved mixing
and hence more uniform primary zone fuel-air ratio which may
also reduce NO emission.
X
(b) Discharge Gas Temperature Spread
The combustor discharge temperature spread was also measured
by 16 aspirated C-A thermocouples, 8 at each of two radii in
the discharge factor (TSF) defined by
TSF =
T — T
max avg
T — T
avg inlet
where all temperatures are in °F. Typical results for the
various combustors tested are as follows:
Combustor
SKP26259 MO
M2
M3
SKP26489 MO
Ml
PAP218770 MO
Ml
SKP26489 PI
TSF
0.400
0.317
0.135
0.078
0.118
0.230
0.113
0.345
AT-6097-R12
Page 5-33
-------
(c) Lean Stability
Because of the low fuel-air ratios associated with recuper-
ated engine cycles, lean blowout was checked on several of
the combustors to determine whether they were suitable for
the particular application. Idle fuel-air ratio for the
AiResearch engine was to be 0.0025, so 0.002 was arbitrarily
set as an acceptable lower limit. Measured results for the
combustors tested are tabulated below.
Combustor
SKP26259 MO
Ml
M2
M3
M4
M5
SKP26489 MO
Ml
SKP26489 PI
Test Conditions
T.n, °F
750
500
500
750
990
500
Pin, psia
58.8
Same as MO
Same as MO
Same as MO
22.0
22.0
58.8
18.4
24.7
W , Ib/sec
a
2.18
0.64
0.64
2.18
0.47
0.76
Fuel-Air Ratio
At Lean Blowout
0.0032
0.0027
0.0015
0.0016
0.0058
0.0045
0.0019
0.0019
0.0019
No lean blowout tests were conducted on the PAP218770 premix
combustor because it was felt that the combustor would not
yield representative results on natural gas fuel. Also note
that the tests of the M4 and M5 modifications to the
SKP26259 vaporizer were terminated prematurely according to
the unsatisfactory lean blowout criterion.
On February 3, the program was placed on hold, and further
contract work suspended pending discussions and negotiations
with GAP/EPA.
AT-6097-R12
Page 5-34
-------
5.3 Test Period (February 1, 1972 to August 10, 1972)
5.3.1 Test and Analysis Activity
The test period covers a contract hold period (February 1, 1972
to May 15, 1972) in which experimental combustor evaluation sponsored
by AiResearch continued during the official contract suspension period,
The testing was conducted using the EPA-contract combustion rig per
agreement with the EPA and is reported in the sections that follow.
Contract activity was then reinitiated in May 1972.
During the period of AiResearch-sponsored testing, a series of
emissions tests were conducted on a baseline pressure-atomizing com-
bustor typical of the AiResearch Model GTCP85 Engine combustors. Sev-
eral other tests including vaporizing and premix combustion systems
were also conducted.
Subsequent to the reinitiation of the contract effort in accord-
ance with Amendment No. 3, four tests on the SKP26489 vaporizer con-
figuration, including one that simulated recuperator bypass air injec-
tion into the burner for a given engine configuration, and one test on
the pneumatic impact combustor, were conducted. These tests were per-
formed under off-design operating conditions as defined by the "Auto-
mobile Gas Turbine Optimization Study" under EPA/OAP Contract 68-04-
0012.
A new Federal Driving Cycle simulation technique was incorporated
in an existing mission analysis computer program which should closely
approximate the engine transient performance required for the vehicle
accelerations and decelerations of the FDC. On the basis of this sim-
ulation technique and with an assumed part-load operating schedule for
an engine, vehicle mass emissions (grams/mile) were predicted from the
combustor emission indexes.
AT-6097-R12
Page 5-35
-------
Table 2-6 showed the new test conditions as determined from the
selected engine cycle performance that resulted from the simulated
Federal Driving Cycle portion of this study. Additional details on
this AiResearch 5-point simulation procedure are presented in Section 2.
5.3.2 Test Results
Emissions test results are presented graphically on an emission
index (El) basis defined as pound of pollutant per 1000 Ib of fuel
burned. For comparison purposes, the average El corresponding to the
FDC standards were determined by the following relationship:
gm _ °'454 Df
mT
(MPG)
where
D_ = fuel density, Ib/gal
(MPG) = vehicle fuel economy, mi/gal
(El) = emission index, lb/1000 Ib fuel or gm/kg fuel
The average allowable El values, based on 14 miles/gal and a fuel
density of 6.25, are as follows:
1976 Federal Emission Standards
Emission Specie
NO (as NO_)
X &.
CO
HC (as CH1>85)
HC (as CH4)
gm/mi
0.40
3.40
0.41
0.474
El, lb/1000 Ib fuel
1.98
16.80
2.03
2.34
The allowable El values were superimposed on the test results which
follow.
AT-6097-R12
Page 5-36
-------
(a) Baseline Atomizer Tests (P/N 899930)
Baseline atomizer test results are presented in Figure 5-20.
When compared to vaporizer data at similar test conditions,
Figure 5-21, the atomizer exhibits slightly higher NO emis-
X
sion with a greater fuel-air ratio dependence at the higher
inlet temperatures and substantially increased CO and
unburned hydrocarbon (CH ) emissions. These results are
X
attributed to the fact that the test conditions were not
typical of those for which the combustion system was initi-
ally designed, and consequently, the primary side of the
dual-orifice atomizer being used was oversized for the appli-
cation involved. Corresponding grams-per-mile levels calcu-
lated by the procedure defined in AT-6097-R9 (Page 6) for
both the atomizer and vaporizer combustors are tabulated
below:
Combustor
P/N 899930 (Atomizer)
SKP26489 (Vaporizer)
1976 Federal Standards
Gm/Mi
NO (as NO,,)
X ^
4.18
3.36
0.40
CO CH2
5.65
1.98
3.4
(0.41
< (as CH4
1.31
0.04
0.475
as CH1>8
)
[- /
(b) Premix Combustor Tests (PAP 218770)
A modification to the PAP218770 premix combustor tested
earlier in the program was made to improve its performance
on liquid fuel. A successful liquid fuel test was then con-
ducted on JP-4 injected into the combustor prechamber through
a pressure atomizer. Poor efficiency attributed to unsatis-
factory atomization of the fuel in the prechamber resulted
in excessively high CO and unburned hydrocarbon emission.
AT-6097-R12
Page 5-37
-------
LINE REFLECTS 1976 FEDERAL STANDARD AT 14 MI/GAL FUEL CONSUMPTION
100.0
w
oo
o
o
o
CQ
X
w
Q
a
§
H
CO
2
w
400 600 800 1000 1200
COMBUSTOR INLET TEMPERATURE, TJN» °F
1400
ATOMIZER COMBUSTOR P/N 899930
(FROM GTCP85-118 ENGINEI
FIGURE 5-20
AT-6097-R12
Page 5-38
-------
LINE REFLECTS 1976 FEDERAL STANDARD AT 14 MI/GAL FUEL CONSUMPTION
10.
w
D
fn
CQ
o
o
o
03
i4
X
W
o
H
a
1000
1200
1400
1600
1800
0.1
2000.
COMBUSTOR INLET TEMPERATURE, T_.T, °R
IN
VAPORIZER COMBUSTOR - SKP26489
FIGURE 5-21
AT-6097-R12
Page 5-39
-------
NO emission was correspondingly very low-estimated at 0.84
X
gm/mi over the Federal Driving Cycle. NO emission index as
JC
a function of combustion efficiency based on the test data
is shown in Figure 5-22. The allowable Federal Standards
require a combustor efficiency of 99.4%. Increasing NO with
X
increasing combustion efficiency is attributed to the higher
flame temperature associated with increased heat release.
(c) Vaporizer Testing
A series of tests were conducted on a recuperated AiResearch
Engine Model GTPR36-61 to determine the effect of recuperator
bypass air (expressed as percent of total engine flow) on
NO emission levels. The major tests associated with this
X
program consisted of the following:
(1) An atomizer combustor baseline
(2) A vaporizer combustor baseline (PAP226608)
(3) 1.5-percent bypass air to the vaporizer combustor
primary pipe or L-pipe (PAP226608)
(4) 3.0-percent bypass air, half to L-pipe and half to
baseplate (PAP226608)
(5) 3.0-percent bypass air plus early quench (PAP226608)
Test results are shown in Figure 5-23. With respect to the
atomizer baseline, the vaporizer showed a 28-percent increase
in NO at 900°F. With respect to the vaporizer baseline, the
X
bypass air and early quench modifications resulted in the NO
X
reductions shown in Table 5-2. It should be noted here that
pressure losses in the test setup limited the maximum bypass
flow available to 3 percent of the total engine flow. The
engine performance was unaffected by the recuperator bypass
setup, however.
AT-6097-R12
Page 5-40
-------
u
55
w
H
u
H
1*4
O
H
EH
O
U
CM
-------
30
rg
O
Q
W
E-
D
O
U
^
W
E
OQ
O
o
o
OQ
X
W
Q
2
M
O
Cfl
en
M
S
w
O
2
20
10
I
2
3
4
5
6A
6B
7
8
9
O
D
0
0
ATOMIZER, STD. CAN
VAPORIZER, HOT L-PIPE AIR
VAPORIZER, COLD L-PIPE AIR
VAPORIZER, COLD L-PIPE AND DOME AIR
VAPORIZER, AS NO. 4, L-PIPE INSERT REMOVED
VAPORIZER, NO. 5, W/3 COOLING PIPES TO L-PIPE
VAPORIZER 6A WITH LOW N2 FUEL
VAPORIZER 6A WITH 3 PRIMARY ZONE COOLING
TUBES ADDED
VAPORIZER 7 WITH BELLMOUTH ENTRY TO 3 PRIMARY
ZONE COOLING TUBES
VAPORIZER 8 WITH (3) 5/8 IN. BELLMOUTH ENTRY
TRANSVERSE PRIMARY HOLES VIA TUBES AND
SECONDARY HOLES COVERED 1/3
1
600
700 800 900 1000 1100
MEASURED COMBUSTOR INLET TEMPERATURE, TIN, °F
DEVELOPMENT PROGRESS ON THE
L-PIPE VAPORIZER COMBUSTOR PAP226608
FIGURE 5-23
1200
AT-6097-R12
Page 5-42
-------
TABLE 5-2
PERCENT N0x REDUCTION OBTAINED ON GTPR36-61
ENGINE TESTS WITH BYPASS AIR AND EARLY QUENCH
PAP226608 Combustor
Configuration
1.5% bypass
3.0% bypass
3jO % bypass + early quench
Percent NO Reduction at
Jt
T. = 900°F
in
10.5
26
51
1000°P
4
22
51
(d) L-Pipe Patternation Tests
Patternation tests were conducted on a vaporizer primary -pipe
to determine what modifications would improve the fuel dis-
tribution from the primary or L-pipe.
The configuration which gave the most uniform distribution
of fuel is detailed in Figure 5-24.
The test setup and a comparison of test results from eight
circumferential positions for two L-pipes and the improved
L-pipe are presented in Figure 5-25 and 5-26, respectively.
Table 5-3 presents a quantitative comparison of the resultant
fuel rates per unit time at the eight positions.
TABLE 5-3
Configuration
Max-Min
£ lReading-Avg |
JT
L-Pipe No. 1
1.579
0.484
L-Pipe No. 2
1.819
0.484
Modified L-Pipe
0.64
0.170
AT-6097-R12
Page 5-43
-------
o
un
JL
0.030
0.50 O.D.
0.030-
U—0.60
0.30
1.20
IMPROVED PRIMARY PIPE
FIGURE 5-24
AT-6097-R12
Page 5-44
-------
D
METERED
AIR
PRIMARY PIPE
COLLECTOR
\ y*-"*TUBES
FUEL
COLLECTOR
ATMOSPHERIC FUEL PATTERNATION
TEST RIG
FIGURE 5-25
AT-6097-R12
Page 5-45
-------
PRIMARY PIPE INLET
IMPROVED PRIMARY PIPE
PERFECT
DISTRIBUTION
UNMODIFIED L-PIPE £2
UNMODIFIED L-PIPE #1
FUEL PATTERNATION TESTS.
ON L-TYPE PRIMARY PIPE
WITH DOUBLE WEIRS
FIGURE 5-26
AT-6097-R12
Page 5-46
-------
(e) SKP26489 Vaporizer Tests
Four tests were conducted on the SKP26489 vaporizer combustor.
The first test consisted of a baseline run of the original
configuration with an additional radial cooling skirt on the
combustion baseplate. The test was conducted per the con-
ditions shown on Table 5-1. Figure 5-27 shows the basic
dimensions to the SKP26489-M2 configuration tested, and Fig-
ure 5-28 shows the test results. Representative shaft horse-
power (shp) loads shown on this and the following curves are
based on N112V engine cycle performance demonstrated over the
simulated Federal Driving Cycle (shown in Table 5-1) as
determined from the "Automobile Gas Turbine Optimization
Study" (EPA/OAP Contract 68-04-0012) . N112V denotes a 1975
technology single-shaft engine cycle with recuperation and
variable inlet guide vanes, the resultant emissions reflect
an increase in NO , gm/mi, which is attributed to the test
4\
and calculation procedure change discussed in Section 5.3.3(a)
As a result of thermal distress noted on the combustor base-
plate, particularly on the ends of the secondary pipes, a
baseplate redesign, without the secondary pipes, was con-
structed for subsequent test. Figure 5-29 is a detail sketch
of the redesign and Figure 5-30 shows a photograph of the new
design. Test results for this configuration, Figure 5-31,
showed a slight increase in carbon monoxide and unburned
hydrocarbons attributed to a slight drop in combustion effi-
ciency and essentially no change in NO emission.
AT-6097-R12
Page 5-47
-------
r-
CM
I
in
g
o
o
H
CM
a
AT-6097-R12
Page 5-48
-------
LINE REFLECTS 1976 FEDERAL STANDARD AT 14 MI/GAL FUEL CONSUMPTION
100.0
w
£10.
m
o
o
o
CQ
.J
H
1.0
0.10
.6.77 GM/MI NOV AS N0
•F/A = — SHP
0.0086-80.5
0.0076_53.5
0.0058 29.5
0.0050 IDLE AND
8.5
800
1000 1200
COMBUSTOR INLET TEMPERATURE, T
2200
IN'
VAPORIZER COMBUSTOR - SKP26489-M2
FIGURE 5-28
AT-6097-R12
Page 5-49
-------
(Tl
fM
I
in
W
U
AT-6097-R12
Page 5-50
-------
VAPORIZER COMBUSTOR BASEPLATE REDESIGN
P-45410
(USED IN SKP26489SD)
FIGURE 5-30
AT-6097-R12
Page 5-51
-------
•LINE REFLECTS 1976 FEDERAL STANDARD AT 14 MI/GAL FUEL CONSUMPTION
100.0
SYMBOL F/A SHP
0.0042 IDLE
0.0046 8.5
0.0058 29.5
0.0076 53.5
0.0086 80.5
HC AS CH
0.1
800
1000 1200
1400 1600
1800 2000 2200
COMBUSTOR INLET TEMPERATURE, T , °R
VAPORIZER COMBUSTOR - SKP26489SD
FIGURE 5-31
AT-6097-R12
Page 5-52
-------
The combustor was then fitted with a new primary pipe
equipped with an internal weir to improve fuel distribution.
Test results, Figure 5-32, were essentially unchanged from
the baseline test of the new baseplate design. An overtem-
perature failure of the primary pipe occurred during the
test, as shown in Figure 5-33, and since the test data indi-
cated that no benefit was obtained from the new L-pipe
design, the combustor was repaired with a conventional
L-pipe with no internal devices.
(f) Recuperator Bypass Simulation
To more fully explore the GTPR36-61 Engine recuperator bypass
test results, a recuperator bypass test was simulated in the
combustion rig by delivering cooled air to the L-pipe. The
test condition was the 29.5 hp point from Table 2-6. The
effects of both bypass air (percent of total engine flow)
and bypass air temperature were determined.
Test results are presented in Figure 5-34. NO emission was
X
reduced 93 percent with 10-percent bypass flow (300°F) and a
combustor inlet temperature of 1060°F (1520°R). A comparison
of emissions obtained at 10-percent bypass flow with
emissions that would be obtained with 300°F combustor inlet
temperature is presented below.
COMPARISON OF EMISSION INDEX VALUES
NO
X
(as NO2)
CHx
(as CH4)
CO
Emissions at
10-percent bypass; 2.02 3.02 7.01
1060°F combustor inlet
Emissions at
300°F combustor 4.0 6.0 60.0
inlet
AT-6097-R12
Page 5-53
-------
-LINE REFLECTS 1976 FEDERAL STANDARD AT 14 MI/GAL FUEL CONSUMPTION
100.0
w
£10.
m
o
o
o
03
X
w
Q
£3
H
w
0.10
0.0042 IDLE
0.0046 8.5
0.0058 29.5
0.0076 53.5
0.0086 80.5
800 1000 1200 1400 1600 1800
COMBUSTOR INLET TEMPERATURE, T
IN'
2000
°R
2200
L-PIPE VAPORIZER WITH WEIR
COMBUSTOR - SKP26489 NL
FIGURE 5-32
AT-6097-R12
Page 5-54
-------
PRIMARY PIPE OVERTEMPERATURE FAILURE
P-45425-1
FIGURE 5-33
AT-6097-R12
Page 5-55
-------
TEST CONDITIONS
= 1060°F
PIN = 31 IN. Hg G
TDISCH = 145°°F
TBYPASS - 3°°°F
HC AS CH
BYPASS FLOW
RATE BASED ON
AVAILABLE AP
5 10 15 20
PERCENT RECUPERATOR BYPASS FLOW
EFFECT OF RECUPERATOR BYPASS AIR
ON EMISSIONS OF SKP26489-SD COMBUSTOR
FIGURE 5-34
AT-6097-R12
Page 5-56
-------
The introduction of 10% bypass air at 300°P into the primary
zone achieved lower NO levels than were obtained with all
X
the combustor inlet air at 300°F. There are several reasons
that could account for the lower NO levels of the bypass
X
test compared to NO levels with all combustor inlet air at
A
the same temperature. The most important of these are primary
zone equivalence ratio and primary zone turbulence level.
The introduction of 10% bypass air into the L-pipe while the
combustor pressure drop is essentially constant adds 45% extra
air to the primary zone. This reduces the primary zone fuel-
air ratio from a calculated 0.042 to 0.029 (PZ equivalence
ratio from 0.62 to 0.43). With a combustor inlet temperature
of approximately 1000°F, this change in primary zone fuel-air
ratio would reduce the equilibrium combustion temperature
from 3350°R to 2750°R and the corresponding equilibrium con-
centration of NO by a factor of two . Therefore, the change
in primary zone equivalence ratio is sufficient to account for
most, if not all, of the measured 2/1 reduction in NO from
X
an ET value of 4.0 to 2.0. The availability of increased
quantities of excess oxygen with the leaner primary zone
would also account for the lower measured quantity of CO
emission.
An increase in general turbulence level would be expected as
the result of the high airflow rate and high pressure drop in
the L-pipe at 10% bypass. This higher turbulence level could
improve the mixture homogenity and would improve primary
zone mixing and combustion temperature uniformity in cases
where substantial variations were present during no-bypass
operation. A subsequent test (reported in Figure 5-50) indi-
cated that the turbulence effect on NO formation rate was a
n
relatively minor one in this case.
(1) "A Combustion system for a Vehicular Regenerative Gas Turbine
Featuring Low Air Pollutant Emissions", SAE Paper 670936, (1967)
by Cornelius, Stivender and Sullivan.
AT-6097-R12
Page 5-57
-------
Figure 5-35 shows the effect on NO emission index of
X.
varying the temperature of the bypass air at 5 percent by-
pass flow from its nominal value of 300°F. At 200°F, the
NO reduction was 2.7 percent compared to 300°F bypass air
X
temperature. At 400°F, the NO increased 5.5 percent.
AT-6097-R12
Page 5-58
-------
12
11
O
53
W
<
Q
W
O in
u iu
CQ
O
O
O
CO
X
w
n
§
H
H
S:
W
X
COMBUSTOR SKP26489SO
TEST CONDITIONS
TIN = 1060°F
P = 31 IN. Hg G
F7K = 0.0063
SHP =29.5
CONSTANT BYPASS AIR FRACTION
100
200 300 400
BYPASS AIR TEMPERATURE, °F
500
600
NO
EFFECT OF BYPASS AIR TEMPERATURE ON
El AT 5 PERCENT RECUPERATOR BYPASS FLOW
FIGURE 5-35
AT-6097-R12
Page 5-59
-------
By comparison, the results shown above reflect a larger than
expected decrease in NO emission index than for the same configura-
X
tion under the same conditions without recuperator bypass flow [based
on data extrapolated to 300°F (760°R) combustor inlet temperature
(shown on Figure 5-31)]. Possible reasons include increased turbulence
and mixing because of high pressure drop, and higher primary zone
equivalence ratios.
The test results also verified that an engine performance penalty
must be paid in the form of increased fuel flow for control of NO
X
emission by the recuperator bypass method. For example, an increase
in fuel-air ratio from 0.0063 to 0.0072 was required to maintain a
constant combustor discharge temperature of 1450°F at 10-percent by-
pass flow. The increased fuel-air ratio did have the added benefit of
delaying the drop-off in combustion efficiency associated with cooling
and leaning the combustor primary zone as illustrated in Figure 5-36.
These results were further investigated in the cycle studies reported
in Section 5.3.3(b).
In order to determine the optimum amount of bypass air, the El
values with bypass from Figure 5-34 were converted to grams-per-mile and
plotted versus percent bypass, as shown in Figure 5-37. This conversion
takes into account the increased fuel flow required when using bypass,
and assumes that the 29.5 hp condition is representative of the emis-
sions over the complete range of operation. For the fixed bypass air
temperature case, the influence of nonbypassed air from the recupera-
tor into the burner was assumed to be negligible, although an air
temperature difference of 260°F was observed between the fixed bypass
air temperature test condition and the 29.5 hp shaft load operating
condition. It can be seen from the plot that the optimum bypass air
percentage is approximately 9 to 10 percent. This could be achieved
with a 3/4-in. primary pipe instead of 1/2 in. The final calibration
will be done at 10-percent bypass, based on the above, to verify the
results obtained and to determine if bypass is as effective at other
horsepower points.
AT-6097-R12
Page 5-60
-------
100
o-
99
98
<£—-4--.
F/A ADJUSTED FOR CONSTANT
DISCHARGE TEMPERATURE
§
w
Qu
97
w
H
u
H
PM
H 96
cn
§
2
8 95
F/A HELD CONSTANT
= 0.0063
\
94
93
5 10
PERCENT RECUPERATOR BYPASS FLOW
EFFECT OF BYPASS AIR ON
SKP26489-SD COMBUSTOR EFFICIENCY
FIGURE 5-36
•8
AT-6097-R12
Page 5-61
-------
2.0
1976 FEDERAL EMISSIONS STANDARDS
N0x (AS N02) 0.40 GM/MI
CO 3.40
HC (AS CH1 85) 0.41
HC. (AS CH4) 0.474
456789
PERCENT RECUPERATOR BYPASS FLOW
EFFECT OF BYPASS FLOW ON EMISSIONS
FROM SKP26489 COMBUSTOR ON
A GRAMS/MILE BASIS
FIGURE 5-37
10
AT-6097-R12
Page 5-62
-------
(g) Pneumatic Impact Testing
The fuel injector of the pneumatic impact combustor configuration
was modified by the addition of a flow diverting cone welded to the
impact plate to prevent coalescing of the fuel droplets in the center
of the plate as had been observed during earlier bench testing of the
atomizer. The fuel delivery pipe was then modified for radial fuel
injection to minimize eccentricity problems between the apex of the
flow-directing cone and the delivery tube centerline. Additionally a
shroud was added around the radial spray ports of the fuel tube to
keep the fuel from impinging on the atomizer venturi walls. These
modifications are illustrated in Figure 5-38.
The modified injector was then bench tested at atomospheric condi-
tions. Test results indicated that the fuel dribbling noted on the
unmodified injector had been eliminated.
The new configuration was then tested in the rig. Test data,
Figure 5-39, showed that carbon monoxide and unburned hydrocarbon emis-
sions were substantially reduced, but the NO level was high, esti-
mated at nearly 9 ,gm/mi. In addition stability problems were
encountered at the low power test conditions.
An examination of the data showed that the injector passed too
much air resulting in unacceptable fuel-air ratios near the idle con-
dition. However, the unburned hydrocarbons and CO emissions are suffi-
ciently low enough that the injector could be resized to give lower
NO without exceeding the CH and CO limits. This work, however, was
x x
not done.
AT-6097-R12
Page 5-63
-------
FLOW DIRECTOR
MODIFIED PNEUMATIC IMPACT
FUEL INJECTOR
FIGURE 5-38
AT-6097-R12
Page 5-64
-------
•LINE REFLECTS 1976 FEDERAL STANDARD AT 14 MI/GAL FUEL CONSUMPTION
100.0
ffl
o
o
o
H
ffl
X
§
2
M
§
H
W
w
H
S
a
10.0
NO (computed aS NO-)
-------
5.3.3 Analytical Effort
*a) Federal Driving Cycle Simulation for Optimized Engine Cycle"
The following modification to the test procedure was made to
incorporate new test conditions based on the optimized engine cycle.*
The recommended cycle was a recuperated single-shaft engine with vari-
able inlet guide vanes (labeled NII2V).
The Federal Driving Cycle (FDC) was simulated by a mission anal-
ysis computer program*with each route segment represented by a speed
change phase followed by a sustained speed phase to achieve the correct
segment average speed and end speed of the automobile. Then the com-
plete mission (FDC) was surveyed to obtain the total time spent within
each horsepower range. All horsepower levels were covered, using 1-hp
intervals to 31 hp, and 3-hp intervals from 30 hp to 121 hp. For the
Federal Driving Cycle, the NII2V Engine does not operate at more than
91 hp at any time.
From the mission analysis program output, a set of test condi-
tions can be chosen that satisfactorily represents the ranges in fuel
flow, pressure, and temperature over which the engine combustion system
must operate. A 5-point representation was chosen as shown in Table 2-6
the end points of the horsepower ranges that are simulated by the 5
test points are shown in Table 5-4. It can be seen from the table that
the points were chosen so as to minimize the range of operating vari-
ables associated with each point in order to ensure maximum accuracy
in the conversion to grams-per-mile.
*Refer to "Automobile Gas Turbine Optimization Study," Final Report
(AT-6100-R7), Contract 68-04-0012.
AT-6097-R12
Page 5-66
-------
TABLE 5-4
TABLE OF 5-POINT TEST EVALUATION
Test
Points
HP Airflow, Temperature, Pressure, Fuel Flow, _ , ,,. .
Ib/sec Tr °R Plf psia Ib/hr Fuel/Air
Min 1.3
1 Av 1.3
Max|Min 6.5
2 Av 8.5
-Max|Min 18.5
3 Av 29.5
Max IMin 38.5
4 Av 53.5
Max|Min 62.5
5 Av 80.5
Max 121.0
0.331
0.331
0.392
0.412
0.506
0.615
0.705
0.850
0.935
1.080
1.380
1960
1960
1935
1915
1850
1780
1722
1660
1632
1580
1485
20.5
20.5
21.9
22.5
25.8
29.3
32.8
38.2
42.0
49.4
65.0
3.64
3.64
5.30
5.85
9.30
13.40
17.00
23.50
27.00
35.00
54.00
0.00305
0.00305
0.00373
0.00394
0.00500
0.00605
0.00705
0.00768
0.00835
0.00900
0.01050
The use of the 5-point evaluation above accounts for all steady-
state conditions, including a detailed integration of horsepower vs
time during engine accelerations and decelerations. It does not
account for exhaust emissions present during cold and hot engine starts
nor the variation in emissions associated with engine transient opera-
tion.
The effect of the new driving cycle simulation is to increase the
predicted emission levels in grams-per-mile. This is illustrated by
values calculated according to the OAP-suggested procedure compared
AT-6097-R12
Page 5-67
-------
with values from the two AiResearch procedures (original 4-point
simulation versus revised 5-point simulation). Calculations for the
SKP26489 vaporizer yield the following values:
FDC Simulation NOx (as NO2) Percent
gm/mi
AiR 5-pt 6.38 137
AiR 4-pt 5.46 117
GAP 4.67 100
Engine Cycle-Recuperator Bypass Study
Off-design cycle studies were conducted on the NII2V Engine to
determine the effect of recuperator bypass on engine performance. Per-
formance data were generated for bypass percentages of 0, 5, and 10
percent of total engine flow at each of the 5-power points in the driv-
ing cycle simulation. The increase in recuperator effectiveness as a
result of decreased throughflow was included. Small secondary effects
resulting from pressure drop changes were assumed to be negligible.
Results of the cycle study are presented in Table 5-5; the data show that
the effect of recuperator bypass decreases as engine power increases
because of the increased combustor temperature rise at higher power
conditions.
The fuel flow rate (W_) data shown in Table 5-5 has been plotted
against output power for 0-, 5-, and 10-percent recuperator bypass
flows (Figure 5-40). Fuel consumption penalties are small at the higher
power levels. The penalties can be significant at low power levels,
AT-6097-R12
Page 5-68
-------
TABLE 5-5
RECUPERATED SINGLE-SHAFT, VIGV AUTO ENGINE
SIMULATED OFF-DESIGN PERFORMANCE AT SEA LEVEL,
85° 1700°F T4. FOR RECUPERATOR
BYPASS FLOWS OF: 0%, 5%, 10%
Engine design point match at T4 = 1900°F
Sea level, standard day, AP/p)_ = 4%
hi
HP
AP/PT,
' Burner
T ° R
comp disch
T ,_ °R
rec out
0%
5%
10%
*T, . °R
burn-in
0%
5%
10%
eregen
0%
5%
10%
W , Ib/hr
0%
5%
10%
f/a
0%
5%
10%
1.3
0.0282
641.3
1968.8
1997.6
2012.9
1968.8
1934.6
1885.0
0.9582
0.9790
0.9900
3.4
4.0
4.9
0.0031
0.0037
0.0044
8.5
0.0333
650.7
1916.4
1942.0
1964.6
1916.4
1881.9
1841.9
0.9508
0.9700
0.9870
5.8
6.61
7.56
0.0040
0.0046
0.0052
29.5
0.0431
709.7
1783.4
1807.0
1829.1
1783.4
1756.7
1723.0
0.9227
0.9430
0.9620
14.0
15.0
16.2
0.0061
0.0065
0.0070
53.5
0.0433
773.8
1676.4
1686.9
1716.9
1676.4
1653.0
1627.0
0.8996
0.9200
0.9400
23.4
24.5
25.7
0.0077
0.0081
0.0085
80.5
0.0420
832.5
1591.2
1609.5
1627.7
1591.2
1572.2
1551.3
0.8778
0.8990
0.9200
34.5
35.6
36.8
0.0091
0.0094
0.0097
*T
includes mixing of
burn-in
recuperator bypass flow.
AT-6097-R12
Page 5-69
-------
depending on the amount of bypass required to achieve the desired NO
H
emission reduction at these lower power levels. This can be seen from
the plot of percent fuel consumption penalty vs output power in Figure
5-40. If a 9-percent bypass flow is needed-at low power levels, the fuel
consumption penalty would range from about 6 percent at 80 hp to 39
percent at idle. While the results shown in Figure 5-31 would indicate
that bypass flows of 9 percent might be desirable/ several factors
could make the optimum bypass flow lower than that value. The test
represented in Figure 5-40 was run at a fixed power condition and with
the bypass temperature fixed at 300°F. If the bypass air temperature
is assumed to be at compressor discharge temperature, the range of
bypass temperature expected is from 180°F to 370°F. Because of the
low temperature of the bypass air at low power conditions, the amount
of bypass flow may be reduced for a desired NO reduction. Figure
X
5-41 shows a comparison between 5 percent bypass flow and 9 percent
bypass flow with respect to fuel consumption penalty.
(c) Data Reduction Program
During the contract hold period, a computer program available at
AiResearch for engine emission data reduction was modified to allow
computerized data reduction of rig data. Both emission index and
pollutant generation rate in Ib/hr are calculated. The emissions con-
centrations are corrected to concentrations in wet exhaust from a com-
bustion process with dry air. In addition, combustion efficiency is
calculated from the measured carbon monoxide and unburned hydrocarbon
concentrations. A typical computer printout is shown in Figure 5-42.
It should be noted that unburned hydrocarbons are calculated as
methane, CH.. If it is necessary to convert the emission indices to
equivalent CI^ g5 or CgH^, the printed values should be multiplied by
0.865 or 0.895, respectively.
AT-6097-R12
Page 5-70
-------
10 PERCENT
BYPASS AIR
5 PERCENT BYPASS AIR
0 PERCENT BYPASS AIR
CONDITIONS
SEALEVEL,85°FDAY
T4=1700°F
40 60 80 100
OUTPUT POWER, Po. HP
FIGURE 5**40
FUEL FLOW RATE VS OUTPUT POWER FOR
0%, 5%, 10% RECUPERATOR BYPASS AIR
120 140
AT-6097-R12
Page 5-71
-------
UJ
o
oc
Ul
a.
UJ
O.
z
o
I
i
I
UJ
40
35
30
20
15
10
^ 9 PERCENT BYPASS AIR
CONDITIONS
SEALEVEL,85°FDAY
T4=1700°F
5 PERCENT
BYPASS AIR
20 40 60 80 100 120 140
OUTPUT POWER, Po,HP
FIGURE 5-41
EFFFCT OF RECUPERATOR BYPASS AIR
ON PART LOAD FUEL CONSUMPTION
AT-6097-R12
Page 5-72
-------
MO
CL UJ
O. _l
v) tn
o
c o
o o
-t Z Z
Ul 3 X
3 I _l
Ul _>
ir li-
ar
ZIM
rr- o
3 I • •
v>o> in o
i<- i
ui • a
r-UlU. U.
vo
in O
in
o
in
o
m
ro oo
o
o
o
r- «•>•>)•*
ooo>
0 » •
oc o.
VII- 9 _l
inin ui
rH UJ O I/I
Zr-
U1C
in
T If
o a> • •
EC * o o
•r-e M
UIIM in
via a>
ui * M
It in
rH It R U
x a
a in x
IE »- _/
z «
o a
u
• o
O r-
(Ti
ro
CTi
o m
o in
o o
CN H
O
in
o
r-l
o
VO
VO
o
o
VO
o
r-l
CM
o
rH
o
vo
o
•cue
o e V
o • •
ooo
o • •
• IT
o> •* * at «-« tnmmiMO* «or*<
fun-* • • «(vj« •
ooinrvj
PI
I
i- '
i(Vl
O
; O ** X i
Vim • •>- !»- .
' •: s? ,52
(WC\J —
• • • O r* »« (
• «
oui
"> o
9 ui
a
o> z
0- z
(r)fr| (V
«f> !•)
O
0, rZ
UJ
BUI
u; a u
z i x
«->»- i
a
Z r- _l
u. < u-
O 2
(0 U
I) ^j
c a
or c u.
• « c
02 at
_) U.
a ui
u t-> o
z •-
c tuu,
r- « Ul
< o c
U.U1CJ
» No«
ui •« iv
ru M o •
r- CT- ru»
• • • in
U1
« o »-t in n m •»
ntnin
r>n«
(VI
o»r-o «x
*- rj« in in
rv/rxi > •
(Vl (V. *
n
orvi
j o —
in a o ni o-
•••-iino
o'oiu e
• »o o
iviNin e
un
in in « o
u
OOO* GOT Vlr-Z
rant • «oozri
i*> o. ui o
a z or*
i , o oza
. oo o
r— O
1 UJ J O
*» Ul Z Ul
. •* no N
nn
: < »- in tn
orvjiv«a<<<
. • • O m r- Cj
r- O O O
o o z a
ui * u.
a z o
(u at o z c
o o or UJ
Z<\J-i(h«UU.Or-
• • r-l in Z «*
ui ui c x
T Z •
33-1
I • »C • •
•» 00
rjc
u
in
u u.'
ZJ S 3
u. a u.
u. a u. c
• o x i •- >- o x
: x in u) o:
• • Z T C
a i o i. a
•«VI
c
x
e
(V
' • • Z Ul 3
in o i- » o
-• •— «« (V _J
U.
>-O
JQ:
CE U O-
X — r- >- O X
v. ru UJ or .x
>« r- I— M
C Z Z O r-
•-.u, uiox
coo e
a ac »i-i
z ui uio ui
o a c. -I*
a
o
x: > > u.c>> u. tr
o a: > tr -JQ • tr
ZWUlr-X>->UJr- r- > > UJ
CEIXO.X ffaraxcxttao.
z eo IT »-o
ZZ •I1UJXZ «r-lr- XI .
z a a. a)ujza_o-!ruj O O
Z Uf UJ O Z Z
w 3O UJ
u. z q u. >->
' Z • O w c O
UJ U. IT Or- r-.
oi->-ox < tn u.
ouitr xcncezu.
ct9O««z ouj
r- OC iC O UJ •-
z uj uj t« z in C'
ZZO »l/>UJrH»H
U,330»r-l_JZr-
O_l_IOOZ«JUll/l
c o -~ _iuj > r
V» > U. rHt/l CC
Ul IT —I P Ul I
n> a>a x
rlUJ Z Z ••-•
uito a. a CDUI
a.
«
u
IT
Z
3
O!
r-
z
or
I-
»-l
|Z
,
i-. 1
Mtr rr a x « ujz o
>c o • • • •
O Z Z • •-) •
a a. m ui
-jo. a J 3 in ,
4 . UI I
r- I- i
£ I ?!
AT-6097-R12
Page 5-73
-------
5.4 Test Period (8-10-72 to 11-15-72)
5.4.1 Test Results
Tests were conducted during the test period on several vaporizer
combustor configurations and on the pneumatic impact injector combus-
tor. The testing included operation with simulated recuperator bypass
operation as well as at the zero bypass flow condition. The combustion
rig mouified for bypass simulation is shown in Figure 5-43.
(a) Bypass Simulation Plus Transverse Primary Jets - Fuel build-
up on the outlet leg of the primary pipe at low fuel flows
was observed during primary pipe fuel distribution tests.
The combustor that had undergone the recuperator bypass simu-
lation test was therefore modified (SKP26489M SD) by the
addition of three primary ports arranged as in Figure 5-44
to inject air at the primary pipe outlet leg to eliminate
fuel cohesion at the pipe outlet at low power conditions and
to reduce the local fuel-air ratio at the primary pipe dis-
charge plane. This test was suggested by trends noted dur-
ing testing on an earlier vaporizer configuration and sub-
stantiated by subsequent testing with recuperator bypass in
an AlResearch GTPR36-61 Engine.
The GTPR36-61 engine tests had indicated that NO emission
X
reductions could be attained by increasing the amount of
air being injected through the primary ports. This trend
held true up to the point where an additional 30 percent
primary air was being introduced; beyond that point the com-
bustor stability deteriorated. From these data it was
decided that transverse primary ports injecting an additional
25 percent primary air would be satisfactory for the combus-
tion rig recuperator bypass simulation test, and the combus-
tor was so modified. The dilution ports were blocked off an
equivalent amount to maintain the same overall pressure drop.
AT-6097-R12
Page 5-74
-------
25
O
H
H
co
CO
co
<
CM
X
O
Pn
1
O
H
§
H
E-<
w
P
O
u
ro
O
H
AT-6097-R12
Page 5-75
-------
O
h-
CO
O
u
O
CO
li
~§i
is|
i£a
Sfcz1
M
ii
o£
or-
UJ
U)
UJ
cc
D
UJ
UJ
o
2
UJ
a
z
o
UJ
>
cc
UJ
UJ
o
cc
a.
AT-6097-R12
Page 5-76
-------
The test was conducted at pressure and flow conditions cor-
responding to the 29.5 hp power condition in order to compar
results with the original bypass test without primary jets.
Test results are compared in Figures 5-45, 5-46, and 5-47.
The data showed similar NO reduction potential to the
earlier test, but the drop-off in combustion efficiency asso-
ciated with bypass operation occurred at 10 percent rather
than 15 percent bypass flow. This is because the primary
jet configuration provides a leaner primary zone. Since the
primary zone with transverse jets is already more lean than
the nonjet combustor and the fact that bypass operation
reduces the primary zone equivalence ratio and average tem-
perature, less bypass flow may be added in a primary-port
combustor before the stability limit of the combustor is
reached.
(b) Bypass to Primary Pipe and Dome - Since the bypass test with
primary jets verified that 10 percent bypass still gave the
maximum NO reduction, it was concluded that without a change
in the engine cycle, it was impractical to attempt to take
advantage of the additional reduction potential demonstrated
at lower bypass percentages. Instead, a modification was
made to the combustor to attempt to introduce the bypass air
into the combustor at lower pressure drop. To this end an
additional bypass line was added to the system that delivered
bypass air to a plenum attached to the combustor baseplate.
A schematic of the dual-bypass combustor is shown in Figure
5-48 and a photograph of the dual-bypass combustion rig is
shown in Figure 5-49. The bypass flow split was controlled
by the area ratio between the L-pipe and the center base-
plate rosette discharge annulus. This flow split was approxd
mately 58 percent through the dome, and 42 percent through
the L-pipe.
AT-6097-R12
Page 5-77
-------
CO
LU
u
ex.
<
CO
O
H
Q
ui
at 0001/81 'xaami NOISSIWH
AT-6097-R12
Page 5-78
-------
100
99
98
97
w
H
U
H
w
§
ID
CQ
96
95
94
93
CONSTANT f/a
» .0063
NOTE: 1056 BYPASS HAS FUEL FLOW
OF f/a AT CONSTANT T
CURVE f/a » .0076
DISCHARGE
5 10
BYPASS FLOW, PERCENT
DISCHARGE
15
COMBUSTION EFFICIENCY COMPARISON BETWEEN VAPORIZOR COMBUSTORS
MODIFIED FOR BYPASS ONLY AND BYPASS PLUS TRANSVERSE PRIMARY AIR JETS
FIGURE 5-46
AT-6097-R12
Page 5-79
-------
11
10
w
M
O
o
O
w
13
O
H
§
SKP27489-SD
(5% BYPASS)
SKP26489-M3 (SD)
BYPASS PLUS
PRIMARY HOLES)
100
200 300 400
BYPASS AIR TEMPERATURE
500
600
NOX EMISSIONS COMPARISON BETWEEN VAPORIZER
COMBUSTORS WITH 5% BYPASS AND 5% BYPASS PLUS PRIMARY HOLES
FIGURE 5-47
AT-6097-R12
Page 5-80
-------
Q.
LU
U
z
o
a.
>
CD
D
a
-Iffl <
01
00
QX
01 i=
OC
LU
§ 9
ca— en
S< §
O w . I/O
E -i
0<
-------
O
H
25
O
H
EH
O
u
CM
(Ti
•^
I
in
O
H
AT-&097-R12
Page 5-82
-------
Test results for this configuration, again at the 29.5 hp
flow and pressure conditions and maximum facility inlet
temperature, are shown in Figure 5-50. The data essentially
repeated that obtained with the original configuration with
bypass through the primary pipe only, except that the deteri-
oration in combustion efficiency at 15 percent bypass was
more severe with the dual-bypass system (88.5 percent com-
pared with 93.0 percent).
(c) L-Pipe Vaporizer With Staged Dome, SKP26489SD - Final
Calibration - Because the dual bypass test indicated that by
injecting all bypass air through the primary pipe at high
pressure drop was not causing artificially low NO readings
X
and because there appeared to be no advantage from an emis-
sions standpoint in retaining the dual bypass system, it was
decided to conduct the final test at 10 percent bypass
through the primary pipe only. The primary objective of the
test was to evaluate the effect of the recuperator bypass
NO control concept at test conditions other than the 29.5
X
hp condition and to generate sufficient emissions data to
predict a grams-per-mile value for the Federal Driving
Cycle (FDC).
The test procedure consisted of testing at the pressure
levels defined for the 5-point FDC simulation defined in
Section 2.2. The flows were corrected up by a constant fac-
tor to maintain 3 percent combustor pressure drop, and the
fuel flows were increased by a similar amount. Ten percent
bypass flow was delivered through the primary pipe at a tem-
perature corresponding to compressor discharge pressure at
each of the 5 test conditions. The fuel flow was again
increased to account for the performance penalty incurred as
a result of the recuperator bypass operation; resultant fuel
flow increase as a function of bypass percentage is shown in
Figure 5-51.
AT-6097-R12
Page 5-83
-------
D
UJ
tt
CO
CL
>-
CO
CO
o
o
LLJ
s
S
§
an oooi/an 'XIQNI Noissiwa
AT-6097-R12
Page 5-84
-------
VAPORIZER COMBUSTOR SKP26489-SD
50
40
E-i
K
W
8
W
H 3°
cn
20
W
10
4 6
BYPASS FLOW, PERCENT
8
SHAFT
HORSEPOWER
i.3 (IDLE)
8.5
29.5
53.5
80.5
10
EFFECT OF RECUPERATOR BYPASS ON
FUEL CONSUMPTION AT PART-LOAD OPERATING CONDITIONS
FIGURE 5-51
AT-6097-R12
Page 5-85
-------
It was intended that emissions data would be taken at
constant airflow, fuel flow, bypass flow, and combustor inlet
pressure at each of three inlet temperatures up to the
capability of the facility in order to establish a means by
which data could be extrapolated to the inlet temperatures
defined by the cycle. While at the maximum combustor inlet
temperature conditions, the facility preheater tube bundle
developed an internal leak that precipitated a failure in
the test rig combustor as a result of diverting air from
the rig. The loss of air to the rig at constant fuel flow
was accompanied by a step increase in the discharge tem-
perature at which time the facility was shut down, but by
that time the combustor was already destroyed.
Subsequent disassembly of the rig and visual inspection
of the combustor indicated that failures had occurred in
the baseplate and in the combustor liner at the first
cooling skirt. Photographs of these failures are shown in
Figures 5-52 and 5-53.
In order to complete the final test, both the preheater
and the combustor required repair. A period of two months
elapsed while repairs were being completed. The final test-
ing was subsequently reinitiated in October, 1972.
(d) L-Pipe Vaporizer With Staged Dome, SKP26489SD - Final
Calibration - This configuration is shown in Figure 5-54.
Results of the continued final test are presented in
Figures 5-54, 5-55, 5-56, and 5-57. The zero bypass data
are shown in Figure 5-58 for reference. It can be seen
from the data that a substantial reduction in NO emission
ji
is available with recuperator bypass with significant
changes to CO and unburned hydrocarbon emissions.
AT-6097-R12
Page 5-86
-------
SKP26489-SD COMBUSTOR BASEPLATE FAILURE
FIGURE 5»52
AT-6097-R12
Page 5-87
-------
AT-6097-R12
Page 5-88
-------
1000
B
CQ
O
O
O
CQ
X
W
H
§
H
H
s
w
O
CJ
100
10
400
600
800
1000
1200 1400 1600
COMBUSTOR INLET TEMPERATURE, TT._, °F
IN
CO EMISSION RESULTS FROM COMBUSTOR SKP26489-SD
WITH 10% BYPASS AIR THROUGH PRIMARY PIPE ONLY
FIGURE 5-54
AT-6097-R12
Page 5-89
-------
100.0
10.0
OQ
o
o
o
H
CQ
X
W
Q
2
H
§
H
W
CO
H
S
O
O
1.0
0.10
^^
ot
^
^S
53.5
^^^^^^"
80.5
29.5
SHP
^
"^ -^
\
\
\
\
\
\
^j^~ ~
•^v
\
-\^ ^
\
\
\
\
\
•^
00 600 800 1000 1200 1400 1600 1£
1U.U
1.0
0.10
100
COMBUSTOR INLET TEMPERATURE, TTla, °F
IN
CO EMISSION RESULTS FROM COMBUSTOR SKP26489-SD
WITH 10 PERCENT BYPASS AIR THROUGH PRIMARY PIPE ONLY
FIGURE 5-55
AT-6097-R12
Page 5-90
-------
400.0
100.0
w
m
o
o
o
m
x
w
§
H
CO
CO
H
s
w
u
High f/a
(Stab.
Problem)
10.0
1.0
0.10
0.01
400 600 800 1000 1200 1400
COMBUSTOR INLET TEMPERATURE, T:
AT-6097-R12
Page 5-91
1600
w
CM
H
CM
O K
E"1 ^
CO S
§i I
OL
_ . *^
S &<
o
o w
s§
§ 2
a
E-i
CO
H
<
o ><
H PQ
CO E-i
H ^
S W
S W
O 0J
H
P
CO
en
oo
in
i
m
O
H
-------
100.0
D
fa
pa
o
o
o
rH
CQ
X
H
Q
2
2
O
H
10.0
o
2
en
X
o
2
OPEN SYMBOLS
SOLID SYMBOLS
INITIAL DATA
CHECK POINTS
(BAD STAB)
HIGH f/a
0.10
400
600 800 1000 1200 1400
COMBUSTOR INLET TEMPERATURE ,TTXT,
AT-6097-R12
Page 5-92
1800
in
i
in
O
H
fa
-------
•LINE REFLECTS 1976 FEDERAL STANDARD AT 14 MI/GAL FUEL CONSUMPTION
100.0
SYMBOL F/A
0.0042 IDLE
0.0046 8.5
0.0056 29.5
0.0076 53.5
0.0086 80.5
HC AS CH
800
1000
2200
COMBUSTOR INLET TEMPERATURE, T
IN'
VAPORIZER COMBUSTOR - SKP26489SD
(ZERO BYPASS)
FIGURE 5-58
AT-6097-R12
Page 5-93
-------
(e) SKP26489SD Vaporizer - GAP 6-Point FDC Simulation - After
completion of the initial final calibration testing of the
L-pipe vaporizer combustor with staged dome, represented by
the initial data shown on Figure 5-57, a test was conducted
using the OAP 6-point simulation procedures. Some oper-
ating conditions of the AiResearch 5-point test and the OAP
6-point test were then repeated by additional runs to deter-
mine if the emissions performance had changed because of dis-
tortion of the dome by the combustion. The check run results
for the AiResearch 5-point testing are superimposed on Figure
5-57.- Significant emissions differences were measured during
these check runs, and it was decided to repeat the OAP test
with a newly fabricated dome of identical design.
The OAP test results and calculations are tabulated in
Figure 5-59 for the five test points at 10 percent bypass
with the distorted dome. The emission index values are given
for each test condition. Fuel flows and fuel-air ratios
were adjusted during the test to account for the estimated
fuel consumption penalty, as in all previous testing.
All emissions were aboVe the 1976 standards. However, it can be
seen that 98 percent of the HC and 94 percent of the CO comes from
points 2 and 3. It was thought that dome distortion in combination
with a high bypass ratio may have been excessively leaning gut the
primary .zone during these low heat-release points.
The dome was then replaced with a new one of identical design.
The test was repeated and extended to include variable bypass opera-
tion (6 percent, 8 percent, and 10 percent) at OAP test points 2 and
3. Figures 5-60 and 5-61 show the variable bypass results for points
2 and 3, respectively. These figures, along with the El values from
the other test points, were then used to predict the emission at
several assumed variable-bypass schedules. The results were as follows
AT-6097-R12
Page 5-94
-------
GAP 6-POINT SIMULATED FEDERAL DRIVING CYCLE MASS EMISSIONS
WITH 10-PERCENT RECUPERATOR BYPASS AIR
(DISTORTED DOME)
POINT
1
2
3
4
5
6
El X
0.96
0.394
1.14
1.84
4.25
13.28
(At 0% bypass, the gm/mi was 4.67
CO
POINT
1
2
3
4
5
6
HC
POINT
1
2
3
4
5
6
*Fc = Wf at 10%
El X
<1.0
58.4
8.5
3.84
<0.1
12.72
El X
<0.1
7.1
3.58
0.24
<0.01
0.29
BP/Wf at 0% BP
KFc*
0.045
0.580
0.447
0.478
0.447
0.035
or a 77% reduction
KFc
0.045
0.580
0.447
0.478
0.447
0.035
KFc
0.045
0.580
0.447
0.478
0.447
0.035
FIGURE 5-59
AT-6097-R12
Page 5-95
EIxKxFc
0.043
0.228
0.453
0.880
68%
1.900
0.465
3.969 v 7
EIxKxFc/7.5
(0.0058)
(0.0305)
(0.0679)
f (0.117)
1 (.0.253)
(0.061)
.5 = 0.53 gm/mi
from the original burner NO emission)
EIxKxFc
<0.045
33.87
94%
3.80
1.84
0.045
0.445
40.05 T 7
EIxKxFc
<0.0045
4.118
98%
1.600
0.115
<0.0045
0.010
5.852 * 7
EIxKxFc/7.5
(<0.006)
f (4.514)
(. (0.507)
(0.245)
(<0.006)
(0.059)
.5 = 5.3 gm/mi
EIxKxFc/7.5
(<0.0006)
f (0.549)
1(0.213)
(0.015)
(<0.0006)
(0.0013)
.5 = 0.78 gm/mi
-------
35 -T-
DOME COMBqSTOR(DISTORTED)
DOME
0 -»-
789
PERCENT BYPASS
EFFECT OF BYPASS AT OAP CONDITION 2
FIGURE 5-60
AT-6097-R12
Page 5-96
11
-------
7 -r
4 - -
CQ
O
O
O
i-(
>
3 --
2 — —
1 - -
0 •*-
14 i
t
35.7
HC
7 8
PERCENT BYPASS
EFFECT OF BYPASS AT OAP TEST CONDITION 3
FIGURE 5-61
AT-6097-R12
Page 5-97
-------
TABLE 5-6
EFFECT OF VARIABLE BYPASS
Schedule
No.
1
1
2
3
4
Configuration
Distorted'
dome
New dome
New dome
New dome
New dome
Assumed BY Variations
OAP Test
Point
All
All
1,3,4,5,6
2
1,3,4,5,6
2
1,4,5,6
2,3
Percent
Bypass
10
10
10
9 1/2+
10
7
10
8
Emissions, gm/mi
N0x HC CO
0.53 0.78 5.3
0.66 0.12 1.90
0.66 0.09 1.82
0.78 0.09 0.9C
0.80 0.06 1.19
5.4.2 Discussion of Test Results
The various tests conducted on Class A vaporizer combustors, in
conjunction with the recuperator bypass concept, established several
trends that warrant further investigation. These findings are discus-
sed below in the order in which the tests were conducted.
(a) Bypass Simulation Plus Transverse Primary Jets - Figure 5-45
presents additional NO reduction over ,\nd above that avail-
A
able from bypass alone and appears to be attainable at low
bypass percentages, with primary air jets. Specifically, the
data show that the NO emission reduction in percent, changes
X
from 46 to 66 at 3 percent bypass and from 68 to 76 at 5
percent bypass with the addition of primary jets at the 29.5
hp test condition flow and inlet pressure and maximum avail-
able test temperature. These data suggest that this com-
bined method of NO control night be attractive in an auto-
Jv
motive application with a schedule of combustor inlet
AT--6097-RI2
Page 5-98
-------
temperatures optimized with respect to the reduction of NO
X
emissions over the range of engine operating conditions.
The bypass flow percentage required would then be optimized
to the lowest value to minimize the fuel consumption penalty
associated with bypass operation thereby improving overall
engine performance.
Note that the benefit derived from the primary jets decreases
with increasing bypass percentage. This is attributed to
the effects of the bypass operation and the primary jets
simultaneously acting to reduce primary zone equivalence
ratio (hence, combustion efficiency). At the lower bypass
percentages the primary pipe internal fuel-air ratio is above
the rich extinction level and the air injected through the
primary jets is required to complete combustion. At 10 per-
cent bypass, however, the internal fuel-air ratio is slightly
below stoichiometric such that combustion can be completed
prior to the point of influence of the primary jets. The
bypass ratio at which the primary pipe internal fuel-air
ratio decreases to stoichiometric is the apparent point at
which there is no further effect of transverse primary jets.
For example, the internal fuel-air ratios at 3, 5, and 10
percent bypass are approximately 0.21, 0.13, and 0.065 at the
29.5 hp test conditions.
(b) Bypass to Primary Pipe and Combustor Dome - As was illus-
trated in Figure 5-46 the dual bypass arrangement appears to
have potential for additional NO emission reduction at
bypass percentage in excess of 10 percent. This is in contra-
diction with the test results for the single bypass system in
which a NO emission increase was noted in going from 10 to
X
15 percent bypass. Further testing is required to establish
the validity of these trends and to determine the ultimate
benefit available from recuperator bypass.
AT-6097-R12
Page 5-99
-------
In the event that such a program is undertaken, methods would
have to be explored to delay the drop-off in combustion effi-
ciency associated with high bypass percentages. In partic-
ular it can be seen from Figure 5-46 that carbon monoxide and
unburned hydrocarbon emissions at 15 percent bypass for the
dual delivery system are much greater than for the primary
pipe bypass only. If cycle studies demonstrate that bypass
percentages greater than 10 percent are feasible from an
engine performance standpoint, then a dual delivery system
with combustion efficiency control is an attractive candi-
date NO control system. Efficiency drop-off point might be
Jv
predictable from primary zone aerodynamic loading consider-
ations .
(c) L-Pipe Vaporizer With Staged Dome (SKP264895D) - Final
Calibration - The data from the final calibration test with
SKP26489SD were converted to grams-per-mile by the method
described in Section 2.2 and compared to the zero bypass
values of the best vaporizer, SKP26489M2.
This comparison is the most conservative comparison possible.
The emissions at 10 percent bypass for a vaporizer combustor
which had high zero-bypass emissions (SKP26489 SD) are com-
pared with the zero-bypass emissions of the vaporizer com-
bustor that had the lowest zero-bypass emissions. The
values are tabulated in Tables 5-7 and 5-8 below.
AT-6097-R12
Page 5-100
-------
TABLE 5-7
MASS EMISSIONS PREDICTION OVER THE SIMULATED FEDERAL DRIVING CYCLE
(for a vaporizer combustor with and without recuperator bypass)
GRAMS /MILE
EMISSION
CO
HC (as CH4)
NO (as N0_)
X £
ZERO BYPASS
(SKP26489M2)
0.13
0.012
6.38
10 PERCENT BYPASS
(SKP26489 SD)
0.071
0.005
1.73
GOALS
3.4
0.47
0.4
REDUCTION
(°/o)
45
58
73
The emission index to grams-per-mile conversion constants at
each power level as tabulated in Section 2.2 were multiplied
by the fuel consumption penalty factors resulting from bypass
operation as determined from engine cycle calculations. The
penalty factor is the ratio of the fuel flow required at the
specified power output with 10 percent bypass to the fuel
flow required with no bypass. The fuel flow increase for
bypass operation is plotted versus bypass percentage in
Figure 5-47 for lines of constant engine power. Tabulated
below is a comparison of the zero bypass conversion constants
with those corrected for 10 percent bypass operation.
TABLE 5-8
EFFECT OF 10 PERCENT RECUPERATOR BYPASS ON El CONVERSION
CONSTANTS AND REQUIRED FUEL FLOW
POWER
SHP
1.3
8.5
29.5
53.5
80.5
CONSTANT, KQ
(NO BYPASS)
0.0103
0.0881
0.0565
0.0246
0.0038
FUEL FLOW
PENALTY FACTOR, Fc
1.441
1.303
1.158
1.099
1.067
CONSTANT, KIQ
(10 PERCENT BYPASS)
0.0148
0.1144
0.0650
0.270
0.0041
AT-6097-R12
Page 5-101
-------
Bypass operation at the low power conditions results in a
larger fuel flow penalty because the combustor inlet temper-
ature is effectively reduced and is reflected as a greater
percentage of the combustor overall temperature rise at the
lower fuel-air ratios. The fuel flow increase, is, there-
fore, a greater percentage of the zero bypass fuel flow in
order to bring the combustor discharge temperature back up
to its initial value.
While the absolute value of the NO emission in grams-per-
X
mile is still in excess of the 1976 standard, the magnitude
of the reduction achieved with the recuperator bypass con-
cept is highly significant. A 73 percent reduction in NO
X
emission was realized with 10 percent bypass over the best
configuration of the SKP26489 vaporizer combustor without
bypass, namely the M2 configuration with secondary pipes.
The magnitude of the reduction is greater than 75 percent
compared with the zero bypass value of the staged-dome con-
figuration currently in use. Furthermore, realize that the
SKP26489 vaporizer combustor with no bypass generated the
greatest amount of NO of all the baseline combustors tested.
x^
The potential to meet the 1976 standards still exists with
other combustor configurations, particularly if the variable
engine geometry of the automotive engine cycle is exercised
to optimize the combustor inlet temperature from an emissions
standpoint.
Figure 5-62 shows the NO emission index plots for the best
X.
SKP26489 zero bypass configuration (M2) and the 10 percent
bypass staged-dome configuration (SD) superimposed. Note
that the reduction available with bypass decreases with
increasing power level. The significance to this is that
the engine/vehicle spends greater than 50 percent of its
operating time at power levels below 10 HP over the Federal
Driving Cycle (FDC). A comparison between the emission
AT-6097-R12
Page 5-102
-------
w
<
CM
100. 0
OQ
o
o
o
PQ
X
w
H
§
H
H
2
w
CN
O
13
10.0
(BAD STAB. )
HIGH f/a
0.10
400 600 800 1000 1200 1400 1600
COMBUSTOR INLET TEMPERATURE, TIN> °F
o
13
H
o
u
u
(Tl
00
(N
Q CM
| H
5 £
tn
H
u
CQ
!3 5
s I
en y
2 w
a *
w
x -1
o K
2 EH
H
I
CTl
00
•^
VO
AT-6097-R12
Page 5-103
-------
indices extrapolated to the combustor inlet temperatures
specified by off-design engine performance analysis and the
percent reductions achieved with 10-percent bypass are shown
in Table 5-9 as follows:
TABLE 5-9
NO REDUCTION ACCOMPLISHED WITH 10 PERCENT RECUPERATOR BYPASS
JC
POWER ,
SHP
1.3
8.5
29.5
53.5
80.5
EMISSION INDEX
ZERO BYPASS
(SKP26489 M2)
41.6
38.0
31.3
29.4
31.0
EMISSION INDEX
10 PERCENT BYPASS
(SKP26489 SD)
7.0
9.4
7.0
9.4
16.2
PERCENT
REDUCTION
83.2
75.2
77.6
68.0
47.7
A significant reduction potential exists at the low power
levels typical of vehicle operation over the Federal Driving
Cycle. The data suggests that a variable bypass arrangement
might be effective in maximizing the NO reduction as a func-
X
tion of part load power level. This possibility could be
explored by conducting a bypass evaluation test at each of
the five conditions chosen to simulate the FDC to determine
whether the optimum bypass percentage varies as a function
of power output.
The reduction in bypass effectivity with increasing power
level is believed to be related to the primary zone equiva-
lence ratio. Since the bypass air is delivered with the
fuel through the primary pipe, the system tends to function
similar to a premix combustor in that the primary zone
equivalence ratio is strongly influenced by the equivalence
ratio in the primary pipe (fuel plus bypass air).
AT-6097-R12
Page 5-104
-------
At low power levels, 10-percent bypass air is sufficient to
reduce the primary pipe fuel-air ratio well below stoichio-
metric with a corresponding decrease in primary zone tempera-
ture. For instance, at the 1.3 and 8.5 hp test points, the
equivalence ratio in the primary pipe is 0.45 and 0.58,
respectively. Corresponding equilibrium flame temperatures
would be below 2700°F resulting in low NO formation rates
X
even if no further dilution of the primary pipe mixture is
assumed to occur prior to combustion. Dilution would fur-
ther reduce this temperature.
At high power levels (above 40 hp), the primary pipe fuel-
air ratio with 10-percent bypass is slightly greater than
stoichiometric yielding higher primary zone temperatures and
corresponding higher NO emissions. At the 53.5 and 80.5 hp
X
test points the equivalence ratio in the primary pipe is
1.13 and 1.32 respectively. Corresponding equilibrium flame
temperatures would be in the range of 3700°F to 3800°F pro-
vided the amount of additional primary air, over that sup-
plied in the primary pipe, is not greater than about 20 per-
cent for the 53.5 hp test point and not greater than about
50 percent for the 80.5 hp test point. These high primary
zone temperatures yield high NO formation rates.
X
A variable bypass system would offer the potential of main-
taining the beneficial effect of a lean primary premix
system throughout the load range. Previous attempts to con-
trol primary zone equivalence ratio without premixing or
burner variable geometry have been unsuccessful, except for
bypass operation. Such results are typical of combustors
that operate with diffusion flames.
AT-6097-R12
Page 5-105
-------
A further indication that variable bypass or premix with
bypass might be effective as a NO control technique is
X
illustrated by a comparison of the effect of overall fuel-
air ratio on NO emission for the zero bypass configuration
X
as shown on Figure 5-63 for the maximum inlet temperature test
conditions. The data show that at the higher fuel-air ratios
there is a strong dependence of NO emission on fuel-air
X
ratio. The smaller difference in NO emission index between
X
i--rc and ten percent bypass at the high fuel-air ratios
implies that the bypass percentage could be optimized at the
higher power levels to obtain additional reduction.
(d) L-Pipe Vaporizer - OAP 6-Point FDC Simulation - A conclusion
of the test was that the distorted dome had caused excessively
lean operation of the primary zone with resultant high values
of HC and CO.
The results further illustrate that variable bypass can be
used to achieve desireable compromizes in total Federal
Driving Cycle emissions. NO and CO emissions would be
2C
further reduced by operation of several of the points at
bypass ratios greater than ten percent. This is illustrated
by Figure 5-61.
The desirable CO and NO combination from assumed Schedule 3
X
from Table 5-6 (7 percent bypass at point 2, 10 percent bypass
at all other points) should be noted. It should be possible
to achieve the 76 FDC emissions standards by; 1.) reducing
the combustor volume to increase the loading, thus achieving
lower NO and higher CO and HC, 2.) additional optimization
X
on the combustor fuel delivery systems and 3.) combined
optimization of the engine cycle and the combustor character-
istics.
AT-6097-R12
Page 5-106
-------
30
w
p
En
CQ
O
O
o 20
X
w
@ 15
§
H
H 10
CN
O
CO
0
ZERO
BYPASS
10 PERCENT BYPASS
0.004 0.005 0.006 0.007 0.008 0.009 0.010
FUEL-AIR RATIO
EFFECT OF FUEL-AIR RATIO ON NO EMISSION FROM
COMBUSTOR SKP26489-SD WITH 10 PERCENT BYPASS AND ZERO BYPASS
FIGURE 5-63
AT-6097-R12
Page 5-107
-------
5.5 Effect of Inlet Temperature on NO Over FDC
J\.
The computer simulation of an automobile performing over the
Federal Driving Cycle (FDC) involves a number of assumptions which
affect the predicted emissions. Obviously, the type, size, and con-
figuration of the vehicle's engine have important influences. The type
of transmission selected has a strong influence. Another of the most
important influences is the part-load operational schedule that is
selected for the engine. Once the transmission is defined in a fixed-
geometry gas turbine engine, the part-load operational schedule is
fixed. However, in variable-geometry turbine engines, the designer has
*
the freedom to select, within limits, the part-load operational sched-
ule which optimizes the engine in a desired manner. This optimiza-
tion could be for best fuel economy, lowest NO emissions or various,
A
tradeoffs of these or other desirable performance qualities.
The part-load operational schedule used in the AiResearch mission
simulation program was optimized for good part-load fuel economy as
determined for the NII2V engine studied in the Automobile Gas Turbine
Optimization Study. Good part-load fuel economy was achieved by main-
taining high combustor discharge temperatures at reduced power levels
by the use of variable inlet guide vanes. This selected schedule has
high combustor inlet temperatures at reduced power levels as shown
in Schedule A on Figure 5-64.
NO formation is exponentially dependent on local combustion tem*-
perature. The fuel-air ratio, heat loss rate to the combustor walls,
combustor inlet air temperature, bypass rate and bypass air temperature
all have an important influence on the NO formation and emission
<&
rates.
Automobile Gas Turbine Study, EPA Contract 68-04-0012, Final Report
(AT-6100-R7), July 14, 1972.
Page 5-108
-------
The emissions shown in Figures 5-55, 5-56, and 5-57 may be used
to illustrate the effect of selected combusted inlet temperature on
the simulated FDC emissions at constant bypass late (10 percent) and
bypass air temperature (300°E;.
Figure 5-64 presents, for example, two cco-rjustor inlet temperature
(CIT) schedules, A and D as a function of output power.
«
Schedule A reflects the AiResearch 5-point FDC simulation
procedure selected for best fuel economy. (Defined in
Section 2.2^
o Schedule D corresponds to arbitrarily defined schedule at
lower combustor inlet temperatures .
Table 5-10 presents the computation of simulated NO , CO, and HC
•X,
emissions over the Federal Driving Cycle for the two CIT schedules.
The predicted NO emissions for these schedules are 1.73 and 0.85
X
grams/mile for Schedules A and D, respectively. Thus, a NO reduction
X
of 50 percent is predicted for changing the CIT schedule from the
schedule used in the AiResearch 5-point simulation procedure to
Schedule D. Concurrently, the CO increases from 0.044 to 0.65 grams/
mile and the HC increases from 0.004 to 0.25 grams/mile. Even at the
increased values of the lower CIT schedule D, the CO is only 19 percent
of the 1976 Federal Standard and the HC is only 6 percent of the 1976
standard. These margins would permit further reduction in NO while
yL
continuing to meet the emission standards £'.>' •"•"> ~"d HC . The fuel
economy reduction was estimated at 6 ^ar^?,,c , •-- h 2'uC and has been
accounted for in the results. This %:,-;'. ; . £ -; ^:;< "i.'-s pre-
dicted advantage. However, thi? n "' "'T-"tr?.n-=>c r.na r>rj+-_o«t- i_a i
significant FDC NO^ reduction? j.; r.ht : ^hifu ?i of i. .e
-------
C/J
w
j
g
w
ffl
u
w
PJ
P
w
EH
EH
W
o
EH
O
u
I
in
H
Pn
annxvHadwai
AT-6097-R12
Page 5-110
-------
O
rH
«
y '
rH
M
O
rH
*:
o •
-H
K>
O
rH
M
X
O •
i-H
W
UrH
< U
8rH
O H
z w
K10
Corrected
For 10%
Bypass and
Low CIT
-a
(U Of
•P O 01
0 U i-H 01
•-HO OJ
« Ll H CU
1-1 O >i
O fe «
u
o
1-1 01
(!) 01
O N IS
•P >1
nl m
Combustor
Inlet
Temperature
°F
4J
3 M
•M S .C
3 0
O 04
i
§
u
1-1
a
CM
EH
in
%
%
T SIMULATION
H
O
CM
ID
<
w
J
3
a
H
en
r-i to ro to •o'
O O O CN O
O O O O O
o o o o o
o o o o o
CN -^i* o ro in
in H CN »» o
O H O (N O
O 0 O O O
O O 0 O O
•si* — 1 ro to ^r
r- to oj r-4 in
o en ^r CN o
o o o o o
in in in in
o o o o\ a\
o o o o o
O 0 O 0 0
e f ^
m CN
•stf f-H O CTl rH
O O O O O
o ** m o H
in oo to oo ro
r-l
in o o o o
o o o o o
CM
CN m O\ oo <*
rH CN tO CN O
O H O O O
o o o o o
00 -ej* O O rH
^ •*!< in r^ ^j*
rH rH tO CM O
0 rH 0 O O
O O 0 O 0
m rH in to oo
O 00 to ^ CO
rH 00 in CN 0
o o o o o
o o o o • o
o o o o o
o to r~ f- i-H
CN H O CJ> O
rH rH rH rH
n in in in in
rH oo o-i en o
CN in co
in
CN
O
O
in
to
o
in
CO
o
§
H
§
w
s!
o
EH
o
rH
I
M
AT-6097-R12
Page 5-111
-------
combustor inlet temperature schedule in combination with bypass rats.c.
Since reduced CITs and the use of recuperator bypass, both have detri-
mental effects on fuel economy, a complete part-load analysis wltr var-
ious combinations of bypass ratio and CIT is necessary to arrive at cha
optimum trade off between minimum NO emissions and reduced fuel
-rC
economy.
AT-6097-R12
Page 5-112
-------
5.6 Development Test Summary
Table 5-11 presents a Summary of Test Results for the combustor
development program.
During development testing, an examination was made of various
methods of measured NO emissions data presentation. Preliminary
X
indications suggest a suitable NO correlating parameter in the form
X
of an aerodynamic loading parameter
W
Q =
_n „ , T/540
P Vol. e '
where W = combustor inlet airflow, Ib/sec
P = combustor inlet pressure, atmosphere
T = combustor inlet temperature, °R
Vol. = combustor volume, ft
n = exponent
The following Figure 5-65 is an example of a machine plot of
NO emission index as a function of the combustor loading parameter
X
based on data obtained from the vaporizing combustor, SKP26489. It
is significant to note that as combustor aerodynamic loading is
increased, by a corresponding decrease in combustor volume, lower NO
X
emissions should result for this combustor configuration. However,
additional study should be conducted on this and other NO correlating
X
parameters for a number of combustor configurations prior to making
any generalizations about combustor emissions performance.
AT-6097-R12
Page 5-113
-------
w
3
D
W
W
nS
H
1 H
in W
W
rj P4
S
« fa
<<0
T
IMMARY
tf
c
u
gfi
*J TJ
Q i-t O
32
8
§ a3
•H \0
33" S
gVf
fc.
§ 8
O
as hj "
H t **
^ £ C
i§sl
e a
o j o^1
3 oxas
n z v>
H ni
I
c
o
1
u
4)
a
j?
Combustor
P/N and Mode
M
ca
0
m
«g
? ? =i
ATM preaaure i
ATM preaaure i
Higb pxeaaure
•Claaa B condi
•SI at 740*F
s i
a i
S
$ d
s
< «5
! o
S 2
*
« 09
*
d •»
* «
r- co
GTP30/Shelldyne H/Cone
GTP30/Shelldyne H/STD
GTP30/Shelldyne H^Coue
Cone dome
SO) $ 01
N H H
M H M M
000 0
& 1? & «
SKP26312
SKP26312
SKP26312 MQ
SKP26259 MQ
0 £
r- r»
g $
&
a
CO
d
CO
o
2
* *
O in
en 0
d ^
* *
Cone dome/blocXed oorts
Flat dome (full ports)
S S
W H
0 0
a, &
> >
SKP26259 ML
SKP2&259 M2
h
o
u
n
5
a
a
d
(N
«
*
3
«
i
s
I
c
>,
i
u
p.
X
•a
4J
a
P4
S
h
a
>
en
M
a
en
1
g
CM
O
4J
M
S
1
^
i
r
3
tn
S
d
*
•*
P remix chamber
?
o
X
o
r>
CO
1
'u
e
H
2
s
I
s
CD
O
O
P remix chamber
o
X
o
£
1
O
i
a
s
i
CO
s
n
*
r-
CD
d
in
1
n
o
Er
I
o<
i
PAP218770 MI
S
a
CO
o
!
in
2
4
s
Secondary pipe
O
N
L<
a
s°
O\
CQ
^
kD
CN
I
n
^
B
H
O
3
S
N
*
d
«
s
a
1
0)
S.
Du
&
•D
S
S
8.
| SKP26489 1^
* . all
R * g * a » ft » t
* « H^S^^^^S-^ r88 ggs=
0*J 4J*>4l(jflio((O4i4)oai|52fl'>' •A'o
.. .. KJ-u'fJ,'2."Hi"'^'H'.l^'.l* • « * -S -H B *« 'a -* S
C u WM&eec&BHe 4JWW c C w k
5 55 SSEEESSSEft,
H ***« S26SS"PBS332g
S wo S 1 A>83
d ui N OH ^ d H
onfj 2* GOT.*
(M CQi-iM
« if *
Oco ™ ""I t£ ft i* it •£*j],ifti-3P4>4)a> OH 41 o o o d o
H VO--$VH TJCU c 4J-a'0'O'0"o
«3 1 (D.IICO J5Hfl| 0 S
C CO £)'~> *«<<«» -HQJ ( 0)04)414)41
Su Qi^j^'H H nj H •Hp.o.nniOio
H f-iOOOtOWOlBIBWH 4J fl) H S E 4J 4J 4J 4->
en o U3:uu<<<<'<<<O4iaiat4l4)lU4i4>4l4l 4)0)4)41 41 £L Q) OJ 41 a;
> £ a.>&>>>>>>>>>> >^>> >£>>>>
„ ^,
§S issssssssSsSssSssH&s^sS
u] i/i coinda,cuaiai0i
0) U S «C H
U ^ C . g
s s ° s <
•D «* -i
5 S . ° S
multiple extrapolation, (A) or (B)
temperature extrapolation, (A) or 1
nt, temperature extrapolation, Clas
hp only, temperature extrapolation,
oint, temperature extrapolation, Cla
4? J 8. " 0
C C 01
4> (1) r^
SO U d
. * >1 >, ft
3 U 41 O O 3
(i r- r- IB SB
C O O 1 [
3 a.
CD
AT-6097-R12
PAGE 5-114
-------
03-
O LOW NOX t
* LOU NOX 2
+ LOW NOX 3
X LOW NOX 4
» LOW NOX 5
* 10 NOX 1
X LO NOX 2
Z tO NOX 3
V LO NOX 4
X LO NOX 5
X LO NOX 1
X LO NOX 2
1 LO NOX 3
• LO NOX 4
' LO NOX S
1 LO NOX 1
1 LO NOX 2
6 LO NOX 3
1 LO NOX 4
« LO NOX S
SKP26489TI2
SKP26489I12
SKP26489M2
SKP26489TIZ
3KP26489f12
SKP26 48930
SKP26489SO
5KP26489SO
3KP26489SO
3«P26489SO
SKP26489P!
3KP2S489PI
8KP26489PI
SKPZ6489PI
SKP26469PI
SKP26489NL
SKP26489NL
3KP26489NL
SKP26489NL
5KP26489NL
'l «r
0~* 2
3 456
COMBUSTOR
7 B 3
LORDING
"10-
(W/P«
2 3
•N/V/EXPIT/540))
T 5 6 7 8 9 110'
EPO-ORP LO-NOX COMBUSTOR PROORflM
FIGURE 5-65
AT-6097-R12
Page 5-115
-------
-------
.CONCLUSIONS AND RECOMMENDATIONS
6.1
An analytical model was formulated and programmed for a digital
ccrripnter The model predicts the formation of the trace species
•:iiti;c .-.rt.a^ !? : or>,,:,l "':Iov, fi^.ld in ,-s gas turbine combustor.
Ccmt. arisen of the nitrogen oxide concentrations with one test case
i:iccrpcr;-.t::Lr^r racra rigorous calculations using a one-dimensional flow
model showed good agreement '.see Appendix I, Pages 33 through 35).
'Li "he experimental program, a variety of combustcr configura-
tions were rig tested, including vaporizers,, atomizers, radiant, pneu-
niar.ic impact and premix combustors. Both an atomizer combustor and a
rapcriser ccmbustor with various recuperator bypass ratios were tested
-.,.1 an engine. Significant test, conclusions have been identified and
1-i.seussed under Secr.icn 5 Ccro^ustor Development and Evaluation.
Conclusions reacnad regarding the relative merits of each system
tes _ec .ii.e presented be. Low
«s Tii-3 conventional gas turbine pressure-atomizing combustor
would require some modification to be applicable to a low
emission automotive engine. This is because of the problem
•:•::' off -design combustion efficiency as well as relatively
v.icr;>. NO enn ssious . Such modification would probably include
?c
type of ritr a,«.qistRd atomization.
?. film-vaporizing combustion system, by itself, offers sig-
niCiearit improv?menr. ever a pressure atomizing system with
re-s^eci; to combustion efficiency, and its fuel delivery sys-
tart i<3 more adaptable to NO emission reduction by inlet
3C
aturo ana fuel-air ratio control .
AT-6097-R12
Page 6-1.
-------
« The premix combustion system shows potential for reduced
NO emission. This is due to control over the primary»zone
fuel-air ratio inherent with this design, as opposed to
pure diffusion flame combustors of the atomizing or vapor-
izing types. Flashback is a potential hazard with this
system, however, particularly at the high inlet temperature
characteristic of recuperated engine cycles.
• Insufficient data were generated for the pneumatic impact
fuel injection system from which to draw legitimate conclu-
sions. The baseline NO emission level for this configura-
X
tion was generally higher than the other systems. By virtue
of its design, however, the pneumatic impact atomizer is an
attractive system on which to apply the recuperator bypass
technique because by injecting all primary air and all
bypass air through the injector venturi the designer can at
once control both primary zone inlet temperature and fuel-
air ratio.
The experimental program was able to achieve high combustion effi-
ciencies, significant reductions in NO emissions while maintaining
X
acceptable low HC and CO emissions, and low temperature spread factors
in practical combustor designs.
The concept of recuperator bypass to the combustor was evaluated
both in an engine and on the combustion test rig. NO reductions of
up to a 97-percent reduction were measured at individual test condi-
tions. The predicted NO reductions over simulated Federal Driving
X
Cycles were from 72 to 77 percent depending on the simulation method
used.
When a high (10 percent) and fixed bypass ratio was used in a
vaporizer combustor at all operating points selected to simulate the
Federal Driving Cycle, the low power conditions with low combustor
AT-6097-R12
Page 6-2
-------
inlet temperatures yield excessive HC and CO emissions. This is
believed to be caused by excessive lean fuel-air ratio and by excessive
cooling or quenching in the primary combustion zone. Two methods of
alleviating this condition could be used.
(a) Using a lower, but fixed, bypass ratio for all operating
conditions
(b) Using a bypass ratio which is varied as a function of engine
load (that is, reduced bypass-ratio at the low power condi-
tions)
Based on the test data taken during the program, it is reasonable
to postulate that at any operating condition, a bypass ratio exists
where major NO reductions are possible and where CO and HC emission
X
increases are acceptable.
The predicted values of emissions from a gas turbine powered
automobile when driven over the Federal Driving Cycle were found to be
as much as an order-of-magnitude difference (for HC and CO) depending
on whether a particular pollutant was predicted using the EPA/OAP
simulation method or the AiResearch-developed 5-point simulation
method. (NO predictions were different by a factor of 3 to 1, with
x
the AiResearch 5-point method predicting the higher values.)
The predicted values for the Federal Driving Cycle emissions are
also very sensitive to the assumed part-load operating schedule for
the hypothetical gas turbine engine assumed. A part-load operating
schedule which is optimized on minimum fuel consumption, as was done
in the cycle analysis performed for this study, will generally result
in low values of HC and CO emissions. Because of the strong dependence
of NO formation on burner inlet temperature, this optimization on
X
fuel"consumption will generally result in adverse rates of NO forma-
X
tion. The variable geometry gas turbine, such as provided by VIGVs,
AT-6097-R12
Page 6-3
-------
provides the ability to select the burner outlet temperature (thus,
also burner inlet temperature) at part load so as to optimize on the
desired low NO emissions.
X
There is considerable potential for optimizing the bypass tech-
nique in combination with other NO reduction techniques and for simvj.-
X
taneously optimizing the schedule of burner inlet temperature as &
function of load at low power levels so as to minimize the formation
and emission of NO . The potential exists for meeting the required
X
total emission goals over the FDC with less fuel consumption than the
1976 spark ignition engine and with a fixed burner geometry in a gas
turbine engine,
Summary of Conclusions
(a) Recuperator bypass, correctly applied; is an effective NO
jn,
control technique
(1) An 82 percent reduction in NO using 10 percent bypass
j?C
was demonstrated on a vaporizer combustor (SKP26489 SD)
during AiResearch 5-point FDC simulation. The calcu-
lated fuel consumption penalty for 10 percent bypass
compared to zero-bypass over the FDC was 18 percent.
(2) A 73-percent reduction using 10-percent bypass was demon-
strated during AiResearch 5-point FDC simulation. This
result was determined from two vaporizer combustor con-
figurations; a vaporizer that exhibited the best zero-
bypass emissions compared to a vaporizer that exhibited
the highest zero-bypass emissions.
AT-6097-R12
Page 6-4
-------
(b) Recuperator bypass as a NO control technique is applicable
Ji
to a variety of combustor concepts including the following:
(1) Premix (2) Air-Blast Atomizer
(3) Vaporizer (4) Pneumatic Impact
(c) An engine cycle can be optimized to provide the best balance
of the emission constituents and fuel economy
(d) Variable recuperator bypass is a simple and convenient
alternative to variable combustor geometry, with the
required control system being simpler and with the
potential of having:
(1) Lower cost
(2) Higher reliability
(3) Better maintainability
<4) Higher fuel consumption
6.2 Recommendations
On the basis of the potential demonstrated by the test results
obtained, it is recommended that additional development activity be
conducted in the following areas.
(a) Combustor Optimization:
(1) Exploratory testing on the combustor to develop the
optimum baseline configuration varying the following
parameters:
AT-6097-R12
Page 6-5
-------
• Volume
• Primary Zone F/A (Lean)
• Fuel delivery
• Mixing
(2) Baseline configuration optimization for each of the
fo1.lowing combustor types:
• Vaporizer • Pneumatic Impact
• Premix • Air-Blast Atomizer
The relative potential of candidate combustor types for
achieving low NO emissions is relative to achieving a
X
highly homogeneous mixture in the combustion zone that
can be controlled at an optimum lean equivalence ratio.
This objective can be realized by striving to achieve
the following combustion system characteristics:
« Premix the fuel and air before entry into the com-
bustion zone
• Enr.ire * homogeneous fuel-air mixture by providing a
suitable fuel presentation system (vaporizer or
atomizer)
© Provide a high degree of air turbulence throughout
the mixture to ensure a uniform fuel dispersal
through the mixture
AT-6097-R12
Page 6-6
-------
The following preferred list of candidate combustor
types are listed relative to their predicted NO reduc-
tion potential based on their ability to match these
design considerations:
• Premix, preferably prevaporized as well
• Vaporizer
• Air blast atomizer
• Pneumatic impact
(3) Optimize the bypass air delivery system on the most
promising baseline combustor type varying the quantity
and distribution of bypass flow.
(4) Develop complete emissions performance maps of
selected combustor type.
(b) Match selected combustor design with Optimized Engine Cycle
to arrive at optimized emissions at minimum fuel consumption
penalty for FDC operating condition. The following order of
analysis is recommended:
(1) Optimize bypass percentage at all load conditions
for fixed bypass
(2) Optimize bypass percentage at all load conditions
for variable bypass
(3) Optimize combustor inlet temperature schedule at
part load with VIGV for minimum NO emission
X
(4) Combine 1 and 3 or 2 and 3.
AT-6097-R12
Page 6-7
-------
6.2.1 Recommendations for Future Programs Include the Following;
• All experimental data should be obtained at the correct
combustor inlet conditions to avoid uncertainties associ-
ated with extrapolated results. AiResearch has recently
ordered an indirect-fired heater capable of simulating
recuperated engine operation with nonvitiated air up to
1700°F at the heater discharge flange. Projected avail-
ability of this heater is May 1, 1973.
• Additional effort should be expended to improve the
simulation procedure for the Federal Driving Cycle.
Specifically, a more adequate representation of the power
levels associated with acceleration transients and of
ignition emissions is required.
• A special study on emissions correlation should also be
conducted to determine a suitable NO correlating parameter.
X
Preliminary indications at AiResearch suggest that the aero-
dynamic loading parameter used for combustion efficiency
correlations may be suitable with additional development.
AT-6097-R12
Page 6-8
-------
APPENDIX I
TWO-DIMENSIONAL MATHEMATICAL MODEL
OP NITRIC OXIDE AND NITROGEN DIOXIDE
FORMATION
By
C.A. Bodeen, J.G. Setter, and V. Quan
K V B Engineering, Inc.
Prepared for
AlReeearch Manufacturing Company of Arizona
October 15, 1971
-------
TABLE OF CONTENTS
Paee No.
SUMMARY ill
INTRODUCTION 1
1. NITRIC OXIDE AND NITROGEN DIOXIDE
FORMATION HATES 3
2. MODIFICATIONS MADE TO THE GOSMAN-SPAEDING
PROGRAM 7
3. SUGGESTIONS FOR FUTURE IMPROVEMENTS 16
APPENDIX A. TEST CASES 18
1. SMAC.l IS
2. OHZ-BXKE3SIQNAL KINETICS COMPARISON 33
ii
-------
TWO-DIMBNSIQKA1 MATHEMATICAL MODEL 0?
NITRIC OXIDE AND KI2H03ZN DIOXIDE FORMATION
The work described in this report was completed
in support of the low non-emission combustor study for
automobile engines being carried out by the AiResearch
Manufacturing Company of Arizona for the Environmental
Protection Agency.
The task was to provide a mathematical model to
predict the formation rates of oxides of nitrogen, given
AiResearch1s two-dimensional analysis of the combustor flow
field including local heat release rates. The nitrogen, oxides
model was to be based on chemical kinetics, and was required
to "be valid under any or all of the following conditions:
a. Both nitric oxide (NO) and nitrogen dioxide (NO^)
being forned.
b. Either fuel-rich or fuel-lean combustion.
c. Local zones in which the nitrogen oxide reactions
may be either near to or far from equilibrium.
d. Local zones in which the major heat-releasing
reactions may be either near to or far from
equilibrium.
e. Significant amounts of the nitrogen oxides being
formed in regions of recirculating flows, where
the average streamlines form closed loops.
f.'Large spatial variations in temperature and con-
centrations so that different chemical reac-ions
ray predominate at various locations or under
certain operating conditions.
The node! which was chosen to meet these requirements
is described in Section 1 of this report. Section 2 describes
how the programming of "Che model for digital computer was
AT-6097-R12
•Appendix I
Paae I
-------
accomplished. In Section 3, some suggestions for further
refinements to the model, in possible future efforts, are
outlined. Finally, results of test cases run using the
present version of the program are given in the Appendix.
In its present form the model, when coupled with the
AiResearch flow-field analysis, appears to be the most
advanced method available for analytical study of pollutant
formation in combustion systems^
AT-6097-R12
Appendix I
Page 2
-------
1. NITRIC OXIDE AND NITROGEN" DIOXIDE FORMATION RATES
Expressions describing the formations of nitric oxide,
NO, and nitrogen dioxide, N02, in combustion products of hydro-
carbon-air mixtures are given below. These expressions can be'
used as source terms in the NO and N02 two-dimensional mass
conservation equations. The mass conservation equations,
which contain the convection and diffusion terms in addition
to the source term, can txien be integrated simultaneously to
obtain the NO and N02 concentrations in a given flow field by
using the same computer program (Ref. l) which has been modi-
fied by AiResearch to do the combustor flow field analysis.
The reactions of importance for NO and N0« formation
are considered to be:
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
Reactions (1) to (6) account for NO production, while (7) and
(8) account for N02 production. Among reactions (1) to (6),
which have been considered in Reference 2, (1) and (2) are the
dominant ones when NO is far below equilibrium and are usually
referred to as the Zeldovich mechanism. Under fuel-rich con-
ditions, (3) may be significant. For lean mixtures under
relatively low temperatures (4) to (6) may dominate over (1)
to (3); although NO is nearly frozen at low temperatures.
Since production of N02 is- significant only under fuel-lean
conditions, it is believed that (7) and (8) are the major
reactions for N02«
AT-6097-R12
Appendix I
Page 3
-------
Ihe rate equations for NO, N, NgO, and NOp can "be
written as
(NO) * klb(N2)(0) + k2f(N)(02) + k3f(N)(OH)
) + k?b(N02)(M) + kQf(N02)(0)
- k2b(NO)(0) - k3b(NO)(H)
- 2k6b(NO)(NO) - k?f(NO)(0)(M) - kQb(NO)(02) (9)
(N) = klb(N2)(0) -i- k2b(NO)(0) + k3b(NO)(H)
- fclf(N)(NO) - k2f (N)(0£) - k3f(N)(OH) (10)
(N*0) = k4b(N2)(OH) + k5b(N2)(02) + k6b(NO)(NO)
- k4f(H)(N20) - k5f(0)(H20) - k6f(0)(N20) (11)
- k7f(!TO)(0)(M) + kQ
- k?b(N02)(M) - kQf(N02)(0) (12)
where (X) indicates concentration of X in moles per unit volume;
the dot denotes creation rate per unit time; and k.f and k.,
(which are functions of temperature) are the forward and backward
reaction rate constants, respectively, of reaction i. Any mole-
cule in the system can act as the third body, M, so that the con-
centration (M) = C°/M , where f is the density and M is the mole-
cular weight of the gas.
It is assumed that (N) and (N20) are at steady state;
i.e., the net (N) produced in reactions (1), (2), and (3) and
the net (NgO) produced in reactions (4), (5), and (6) are zero.
With (N) = 0 and (NgO) = 0, Equations (10 ) and (11 ) yield
klb(N2)(0) +k2b(NO)(0) +k,b(NO)(H)
" klf (NO) -i- k2£ C0) + k C
AT-6097-R12
Appendix I
Page 4
-------
+ kgf CO;
In addition to the rate constants kif and k.. where
i a 1 to 8, one needs (N2), (02), (0), (OH), and (H) to solve
for (NO) and (N02). For hydrocarbon-air combustion, it is
assumed that (N2),(02), (HgO), (CO), and (COg) are known. The
following reactions are assumed to be infinitely fast in adjusting
the concentrations (0), (OH); and (H):
CO + OH ===== C02 + H (15)
0 + OH ===== 02 + H (16)
OH + OH ===== H20 + 0 (17)
Note that this is not the same as assuming equilibrium values
for all the species involved in these reactions. The non-equili-
brium values for (H20), (CO), and (C02), computed in the combus-
tion analysis, are used to get the concentrations of the much
more reactive species 0, OH, and H by assuming that a quasi-
equilibrium is established for ^eaotions (15) through (17).
Thus, for example, although (OH) may be present in such a small'
concentration that a large percentage change of (CO) takes a
comparatively long time, H and OH are highly reactive and the
ratio of their concentrations is assumed to be rapidly adjusted
according to the ratio of C02 to CO so that the relationship
(C02)(H)
= Keq
is satisfied. The equilibrium constant K is a function of
temperature only. Reactions (15), (16), and (17) yield
(co)(o2)
AT-6097-R12
Appendix I
Page 5
-------
(OH) = [tL7-1(H20)(0> ] * (19)
- V (00)(OH)
- Ki5 (6o2;
where K, are the equilibrium'constants for reaction i.
Equations (13), (14), (18), (19), and (20) allow one
to evaluate the source terms for (NO) and (N02) which are
given by equations (9) and (12), respectively. To convert
moles per unit volume to mass fraction for species X, one
merely multiplies (X) by the molecular weight of species X
and divides by-the local fluid density. The terms (NO) and
(NOg), multiplied by the molecular weights of NO and N02,
respectively, constitute the source terms in units of mass of NO
and N02 per unit volume per unit time to be used in the dif-
ferential equations describing conservations of mass of NO and
respectively.
AT-6097-R12
Appendix I
Page 6
-------
2. MODIFICATIONS MADE TO THE GOSMAN-SPALDING PROGRAM
A. General Technique
Nitric oxide (NO) and nitrogen dioxide (N02) are
treated as trace species being formed kinetically in a
flowfield in which temperature, density, and the concentra-
tions of Np, 02, H20, CO, and C02 are specified. These
concentrations need not be equilibrium values, but it is
assumed that the formation of NO and N0» does not signifi-
cantly affect the concentrations of the other species. The
iterative solution for the trace species concentrations is
performed separately from that for the fluid mechanics and
hydrocarbon kinetics.
The N0-N02 source terms as used in the pro.gram are
derived using the following definitions:
Gl = kib(N2)(0)
G3 = k2f (02) + k3
G4 = k4b(N2)(OH)
G5 = k4f (H) + k5f (0) (21)
Gg = kgf (0)
G? = k?b(M) + kQf(0)
GQ = k?f (O).(M) + kQb(02)
In terms of the G's, equations (9) and (12) become
(HO) = Gx + G3(N) + 2Gg(N20) + G?(ir02)
- klf(N)(NO) - G2(NO) - 2 k6b(NO)2-GQ(NO) (22)
) = G8(ITO) - G?(N02) (23)
AT-6097-R12
Appendix I
Page 7
-------
Equations (13) and (14) become
G2(NO)
UK))
(N20) =
G. +
,(24)
(25)
Substitution of equations (24) and (25) into (22) yields
(NO) * 2
G1G5 '-
J4"6"~6b"5v
Gt + G.
- GQ(NO) + G?(N02) (26)
Unnecessary repeated evaluation of terms in equation
(26) can be avoided by defining
2G1G2
Z3 = G3
Z. = 2k
lf
(27)
z
^ -
Equations (26) and (23) can now be written as
(28)
AT-6097-R12
Appendix I
Page 8
-------
(N02) = Z8(NO) - Z?(N02) (29)
B. Subroutine NOXCON
The subroutine NOXCON contains all reaction rate data
and performs the function of evaluating the Z's of equations (27)
Data for the reactions directly involved in the produc-
tion.: of_NQ and N02 include: (l) logarithms of eleven equili-
brium constants as functions of temperature and (2) lists of
activation energies/?., frequency factors A.,, and temperature
exponents N. for calculation of eight reaction rates according
to the equation
kfd * Ad T"N;J exp ("A/RT) (30)
These data are given in'Table ..1. The eight kinetic reactions
are equations (l) through (8) and the required equilibrium
constants are for these plus equations (15) through (l?).
The logarithms (base 10) of the equilibrium constants
are stored for several values of temperature (°R). Except for
E«. the equilibrium constants are dimensionless; K». has units
' / \ '
of (moles per unit volume). A list of values of K~* was taken
In units of cm /g-mole, and the logarithms of these values
are stored in the data statement. Before these values can be
used, they must be converted to values of logarithms of
which has units of ft'/lb-mole used in the program. The
relationship is
cm 1000 g-mole
g-mole 2.20462 l
0160185 £7* (31)
AT-6097-R12
Appendix I
Page 9
-------
H
%
$
&»
8
O
&ES
pu
erf
w
s
p
3
CO
55
0
M
CO
CO
Oj
PN
w
EH
|
CO
§
o
w
EH
«aj
P5
•»
55
O
M
g-l
£j
««!
S
^
^*
O
g
pa
H
o
f
o
Vl
.
£
1-»
CM
rH
rH
^-
CM
O
to
O
rH
X
CM
CO
•
to
CM
O
^
CM
^5
11
0
CM
J25
JL
6
in
0
|H
CM
rH
rH
•«»-
CM
0
to
O
rH
X
co
in
•**•
0
^.
o
55
11
O
CM
55
+
O
*
O
M
VO
•o
p-
00
H
O
in
O
•H
X
in
O
rH
S
CM
o
1 1
\\
o
o
55
^
p- «o •«• in
H^j ^™i
rH rH
P- 1 1 1
0
O
VO
O
O », 1 1
rH
O
rH
X
O I 1 1
rH
CM K O
O + W +
+ CM + O
O O CM CM
55 0 0 W
11 11 11 11
0 W «
+ 0 W 0
CM + O +
O O + K
55 0 O O
CO Cn O rH
rH rH
Q
C
O
jjj
0
0)
h
0)
o
(4
O
Vl
Q>
•H
ttf)
0}
0}
•a
Q
0
rl
Vl
tJ
0)
4>
fl}
O
iH
0
Q)
0
0)
C
O
•H
•P
O
Ctf
Q)
m
CQ
Q)
•P
^1
O
Vl
'^p
03
-------
Subroutine NOXCON adds log1Q .0160185 to the tabulated values
of log K and overatores the new values.
The J of equation (3O) are stored in the DATA State-
ments in kcal/g-mole, and they are used in conjunction with the
universal gas constant R. NOXCON continues data conditioning
by overstoring and redefining^}' as
(32)
where R is giv«n by
R » .001987 „
-001104 ** (33)
then has units of °R.
The units of the A^Cwith the exception of A-) as stored
in the DATA statements are cm5 °K ;)/(g-mole sec). For consis-
tency of units, we convert:these to ft^ °R j/(lb-mole sec),
overstoring the original values according to
. ft' . 1000 g-mole
(12*2 54)^cm3 2.20462 Ib-mole
or
Ai m JL x 0.0160185 x (|) * (34)
Again, equation (7) is the exception to the rule. The
stored units of A_ are (cur/g-mole), so. the conversion factor
.0160185 is applied twice.
AT-6097-R12
Appendix I
Page 11
-------
Reaction 7 requires (M), the molecular concentration
of the jgas as a whole. This quantity is given by
(M) * |p (35)
and numerically becomes
f*rt f« tN Ib-mole °R
(M) * (p j-J) 1545 ft lb
The forward reaction rate constants are next evaluated
according TO
kfj s Aj T"N;I exp (-/^j/T) (37)
where it will be remembered that/3. contains the factor 1/R.
Backward rates-are required for reactions 1 through 8,
and these are calculated according to
k
Subroutine NOXCON next converts the species mass
fractions (produced by the main iteration of the Garrett-
modified Gosman-Spalding program) into molar concentrations
in Ib-mole/ft (^ised in the NO-NOg source terms). Then the
functions S (Equation (21)) and Z (Equation (27)) are evaluated.
AT-6097-R12
Appendix I
Page 12
-------
C. Subroutine SORCCK
The source term subroutine SORCC1 was modified to
Include the terms for NO and NO-, which are assumed to be
called "Mass Fraction 5" and -"Mass Fraction 6" within the
structure of Garrett's program.
The Gosman, et al text (Ref . 1) describes a technique
in using source terms which serves to avoid divergence or to
improve convergence in, many cases. The source term, d^ ,
which appears in the algebraic statement of the equation is
represented within the program as (in cylindrical coordinates)
= -SOURCE + ZQ ' 0 (39)
where
r is the local radius;
is the rate of production of 0, in this case NO or N02,
Ib /(ft3sec);
ID
is the mass fraction of NO or N02;
SOURCE and ZQ * 0 are an artificial separation of terms
such that SOURCE is not a function of
In the case of N0« the basic concept in the separation
of d^ into the two terms, -SOURCE + ZQ *0 , is well illustrated:
N00
2
: SOURCE = rMjjQ ZQ(NO) (40)
*
where MJJQ is the molecular weight of NOg, m»g is the mass
fraction of N02,and P is the local density.
In the case of NO the task is not quite so simple
since the NO rate function is nonlinear in (NO). After
AT-6097-R12
Appendix I
Page 13
-------
discarding one attempt which sometimes exhibited instability,
the following definitions were found to be satisfactory:
: SOURCE = rMjjQ fa + Z7(N02) + z^
(42)
NO Zn * J& l\ZK+ g J4/WJ (NO)2+Z8 (NO))
(43)
where M^0 is the molecular weight of NO and m^c is the mass
fraction of NO.
D. Subroutine SOLVCK
The subroutine SOLVC1 was modified to avoid certain
unnecessary functions during the NO-NO, iterations. Purtherncre
SOLVCK rearranges and restores the solution matrix "A" and
a number of other lists to facilitate the trace species iteration
using the same program log-'c as the main iteration.
E. Subroutine BLOCKK
The input subroutine BLOCK! was modified to read a
new card_which controls the behavior of the program with the
three variables IOLD, INEW, INOX according to the following
schedule:
IOLD Absolute value of IOLD is the logical unit
number of a magnetic tape (or other peripheral
device) which contains the solution to a pre-
viously worked problem.
IOLD = 0 implies "no such tape exists"
IOLD<0 implies "read in the tape, but do not
iterate - just generate plots".
AT-6097-R12
Appendix I
Page 14
-------
INEW Logical unit of tape on which solution is
to be saved.
INEW = 0 implies "do not save solution."
INOX Absolute value of INOX is the number of trace
species equations to be solved: 1—>• NO only
2 —-»• NO & N02
0—>• neither
INOX
-------
3. SUGGESTIONS FOR FUTURE IMPROVEMENTS
Two improvements to the Gosman computer program are
suggested for future efforts. One is the computation of
the effective chemical species production terms. Since the
formation rate of nitric oxide is an exponential function
of temperature, the rate computed using the average temperature
in an element of volume as currently done is always lower
than the effective rate for which the temperature variation
in the element of volume is taken into account. The error
can be substantial, since the temperature variation within
the element is often of the order of hundreds of degrees.
A method of calculating the effective rate has already been
developed by KVB Engineering. The method consists of inte-
grating analytically the chemical species source terms over
the element of volume -considering the variation of temperature
in both of the coordinate directions. It is recommended
that this method be implemented in the computer program in
a form suitable for gas turbine analyses. The effective
rate computation is not only useful for nitric oxide evaluation,
but it can be applied to more accurate computation of fuel-
oxidizer combustion as well.
Another recommended modification to the current com-
puter program is the inclusion of heat transfer. Although
gas turbine walls are nearly adiabatic, there is great spatial
variation of temperature in the fluid. Radiation can reduce
the peak temperature to the extent that nitric oxide produc-
tion is substantially reduced. Therefore, heat transfer by
radiation and convection should be incorporated into the
computer program.
If calculations made with the present version of the
program indicate that reactions 3 through 8 and the backward
directions of reactions 1 and 2 are of little importance, a
simpler algebraic expression for the nitric oxide source term
can be utilized and the nitrogen dioxide solution can be omitted.
AT-6097-R12
Appendix I
Page 16
-------
The calculation of the improved effective source term mentioned
above is also considerably simplified. These changes are quite
easily made and could be added as an option in the present pro-
gram. It is recommended that this possibility be investigated
because a substantial saving in computer time, as well as im-
proved accuracy in the source term, may be possible.
REFERENCES
1. Gosman, A. D. , et al., Heat and Mass Transfer in
Recirculating Flows. Academic Press, London, 1969.
2. Lavoie, G. A., et al, "Experimental and Theoretical
Study of Nitric Oxide Formation in Internal Combustion
Engines," Combustion Science and Technology, Vol. 1.
PP 313-326, 1970.
3. Baulch, D.I. , et al., "Critical Evaluation of Rate Data
for Homogeneous Gas-Phase Reactions of Interest in
High-Temperature Systems," Dept. of Physical Chemistry,
The University, -Leeds, England, Vol. 4, p. 1.
4. ibid., Vol. 4, p. 11.
5. Roberts, R. , et al., "An Analytical Model for Nitric
Oxide Formation in a Gas Turbi&e Combustion Chamber,"
presented at AIAA Seventh Propulsion Joint Specialist
Conference, Salt Lake City, 15 June 1971. '
6. Baulch, et al., op. cit., Vol. 5, p. 15.
7. ibid,, Vol. 5, p. 1.
8. ibid., Vol. 3, p. 14.
9. ibid.fVol. 4, p. 40.
10. ibid., Vol. 4, p. 44.
11. ibid., Vol. 3, p. 1.
12. Schofield, K. , "An Evaluation of Kinetic Rate Data for
Reactions of Neutrals of Atmospheric Interest," Planet.
Space Sci. 1£, 1967, p. 654.
13. Baulch, et al. , op. cit., Vol. 1, p. 1.
14. ibid., Vol. 3, p. 14.
15. ibid. , Vol. 2. , p. 20.
16^ Kliegel, J.R. , Frey, H. M.. One-Dimensional Reacting _Gas
Konequilibrii'ca ^Perfonnance Prc.graa, TEW Systems Group,
Redondo .Beach, CA.1967.
AT-6097-R12
Appendix I
Paqe 17
-------
APPENDIX A
TEST CASES
1. SMAC.l
Garrett supplied KVB with a sample test case for the
program. A small premixed methane-air burner was modeled
under the name "SMAC .1". The program input for this case
ie given in Table 2 and the results of the basic Gosman
solution are shown graphically in Figures 1 through 5.
Solutions for NO and NOg mass fractions are shown
in Figures b and 7.
The NO-NOp iteration system was also tested on several
simple modifications of the basic problem results. Figures 8
and 9 give NO and N0_ mass fractions for the basic case with
500°R added to the temperature field. Figures 10 and 11 give
similar results for the basic case with all densities multiplied
by 10. Finally, Figures 12 and 13 present NO and N02 solu-
tions for the combination of the 500^41 temperature increase
and the 10-fold density increase.
The NO-NOp iteration scheme converged to 1% relative
residual in 59 or fewer iterations for all four test cases.
AT-6097-R12
Appendix I
Page 18
-------
TABLE 2
INPUT FOR SMAC .1 TEST CASE
SMAC . 1
TURBULENT
NON-UNIFORM DENSITY _ .
~CONSTA«T PRESSURE "~~~~
UEP VA.UAOLES ARE:
VOKTICITY - A(IrJrNW)
STKEAM FUNCTION - A'('I~»j7NFT
MASS FRACTION NO 1
MASS FRACTION NO 2
E^GEOMETRY: -"-—
**"•' 8 ROWS 8 COLUMNS
XMXN IMAX
1
2
3
!£;.•"•' 6
7
8
1
1
1
1
1
1
1
8 • •
8
8 • i
"•"•'."!." ;'~" 8 " ".'^'f';-: V":-:_. /-~^^fB:£-?: "K ^l-???*"? ^.^"j.-/"'--1™'"".
fl /^>-^'-~.;/;----i^^\--i-V.^-V/^-7--/:v:..TV:--' .'•-••• \'': 't :'; ': " *'.
8 . .
8 ' '
JBr JCrlABf 1C
3' '3' 8 1" 8 --.r.-:- , ^ ^
RADAr . RADAl/.-V-i'-i RADBr"
A'J?00.- —l^rJ5?0 •....'-•'1.000 J
"~" i?A» " " DA"1» " DB»
.0000 .0000 .0000
.-"• RADCf'•:-.''-:.~' RADN
'1.000 •;•-1.000
DN"
.9845-01
.9845-01
SJp'H'-'v' X2AX1S -•' XlC'ONV •••-:" X2CONV^Jk-:^:.v<:^-;-,4;^^;r. ^; :, -i:;i^:.._^-
^..-; •:•'" 1.000 .2813. .1142. rx-> r:;r .-;.:; .^^ " -r-- ' v:^'"-»" -, .^
f.- 2 OR j xim X2(j> v. j.. <;.• j;:.-^;^ >-..... ;;^;^: . ..-.;- • •:,- •-'.-.
1
2
3
te:-."jJ-iI-'-"= 4 " '. .-
^••r" 5
?:rr,.-..- 6
7
8
PHYSICAL
r • v PR i
i ROKEF
' ' . ZMUREF
CP R£F
GC
L HP(FU)
.0000
.1^*06-01
• .2813-01
,4219-01 -,.r
.5626-01 - /
.70i2-Ci
.8439-01
.9845-01
DATA:
1233.
.7500-01
.2000-04
.240U
32.20
17.167"-"
. 17856.__
.0000
.4167-02
.8336-02
,.1459-01 ^S^>:OK^^K:^^^^-^^-¥-^--- "^
-.2284-01 -..•;,•" — ••---•ii'iii'.-J^-T*-- • • -;.:,; .• :' .
.2855-01 "•.""''. -: ^.:^V;-V.^x:*/-!'*i:i-.-'?.l-.v -•;••.':
.3426-01
.3997-01
- v SEC -: .0000 . . - "•••-•• '••.-:• ••••-•••
. P REF 2116. '
T REF . 3000.
ZJC 778.0
"""HC " -17415. " ' ' " '••"""•:', ;.' ' .'
HS(OX) H41.00 • .
AT-6097-R12
Appendix I
Page 19
-------
K
O
U.
o
o
a.
o
o
2
8
O
OL
i
o
o
lu
u.
Ul
a:
u
a
s:
D
in
CT
in
ep
in
it
in
a-
o
+
•-»
c\»
in
«SJ
•
o
It
r-
a-
n
•
K)
II II
IO CO
in in
o o
* +
vO O
CM K>
o* in
•-! K)
• •
rt o^
I I
II II
CM r-
\n m
o o
* +•
in
-------
cr
u.
o
i-
o
i
(A
4.
O
o
1-4
UJ
c.
u:
Q
LJ
O
O.
l/l
UJ
in
o
u
h-
O
u.
:u
IX
to
o:
UJ
:a
2"
3
AT-6097-R12
Appendix I
Page 21
f.'::
i
i ••
i
[«
O I
•H •
+» i
O i
O
H
-P
CO
03 -
cc
o
o
•H
CQ
rH
*
O
CO
CM
0)
-------
u.
u_
o
o
_l
a
i
i
i
i
_
•
i
!•
|
1
•
•
•
3
j
•«
J
M
n
j
5
»
%
>
ft
3
t
>
J
L
(M
c-
1
O>
(M
«O
o
,
*
u ;
in !
• .
cv.
?
a- 1
10 '
S '•
•
O
CO
CO
AT-6097-R12
Appendix I
Page 22
-------
UJ
. -J
o
LJ
O
OL
-
z
^
in
IL:
o
I
I-
o
U
UJ
o t-
a. o
U.
uJ
U
'1^
5».
3
•
o*
it
in
o
i
o*
tn
CO
CVJ
o
I
o
S'
(Vi
a-
II II
IO 03
CM =>
O •-«
I I
10 o
If) O
co sn-
CM O
II II
a o
i i ,
r- o
CM \o !
f O •
-1 O
• •
(M -4
II II .
-« -O
• '• • •;• • • • OS • • •
1 ^»-«n
o i
ca i
ra
w
oi i
o '
O
a-
in
10
AT-6097-R12
Appendix I
Page 23
-------
:
i
i- =J:
'T H»
f \ti.
< y
» -*.
D
U
'= i
Q-
u
a «
Z I-
IU I/)
»- r
o ,
fe r-!-
Uj .
o P.
a. o
7> c:
-J UJ
< u.
> UJ
i C£
< a:
>*• u
V/i O
€ 5
<*> ^
; L
«o
o
+
01
o
in
CO
*
K)
ii
in
10
to
IO
CC
(O
•
10
II
*•
10!
o
10
• _
10.
II II
«O 00
IO iO
o o
+ +
-i in
»o o*
CJ C»
II II
ecr-
IO CO
•3 O
•»• +
10 T
O '\»
' , J
II II
•H O
•-•'' " .'.i
••
10 ••••
AT-6097-R12
Appendix I
Page 24
,,'M
jjiX;,,
If) 'V:
•p
0)
• a*
CD
CQ
O
o
•H
(0
ftf
o
CO
0) I
t>0
•H
fe
-------
0
cc
l/l
UJ •
O'
UJ .
o.
0_
u
I X
t- UJ
0 t'J
in
o
IO
«n
*
«n
ii
in
in
o
I
O
o»
r-
oi
II
o
I
I
1! U
K) fj
O
\D O
o> o
LO O
r- o
it it
wr-
O -3
I I
10 :n
-•^o
J*- :\J
03
CQ ;
ca •
CQ
O
« i
vo
0)
M
3
60
.O II II
m .
AT-6097-R12
Appendix I
Page 25
-------
CO
o
10
J...
l/>
%
a
w;
o !'
cl;
< i
>
t- •
T*
<. i
o
o
-5
/_!
Q.
'jj
_J
<
~>
I
U '
u. .•
•J •
a:'
a:
u .
II
in
co
o
in
M
CO
O
I :
O
II II
o o
i i
.O -0
r» .-r
' a- •-»
il II
-4-0
.
;
: t
i. .
t'
!
' 'i
'..'•'.
y"i' il
• : * . i
j ' .. : .: ' r/j ;
i. '• ^ .'• '.••• , '
• ''/I ::i. ••
' \ ' •• .' i
\y' • • •'' •
.X^/i^n ,r> >n .r« »n \n \O
;o:
! 3
: [1
n
AT-6097-R12
Appendix I
Page 26
-------
o
I
if
in
UJ
CO :
U) I
o,
< !'
^ i
o
UJ
h-
H- .
O
_J
c.
UJ
I
vt
-*
6
o
_;
0.
.J
•t
•?
UJ
U. :
vt
G:
tj .
a,
II •
m
in
o
-------
•
•
o
.'._
t-4
W '
\fl
UJ '
£'
U. irt
* 6
o
b '
o >-
5! o
_l tu
< u. .
if uJ
« cc
o
I
o
iT
CO
M
II
in
CO
o
I
,
to
II
ii it
n to
II II
-< -O
^ f*
AT-6097-R12
Appendix I
Page 28
-------
y,-1
ij
<
>
a!
UJ .
i-
H-
O
a..
1/1
3
U..
O
III
, O I-
a! o,
tu
^ j;
-1 a!
i <• U.
> LJ
rr
K)
o
I
m
on
in
o
u
ID
O
CO
«•
\o
II
sr
o
I
vO
II II
IO O
o
I
MO
-------
in
o
I
O
w)
f-
«-4
U? i
02 "
I/)
UJ
a :
. LU ..
o
t*...
i
Ul
o
u. .
^ iL1
O P
j! o
•'u
O '.i
U UJ
1C U.
II
m
o
rt
-<
(7* O
•>3 O
M O
'/) O
II II
O O
I I
•^
-------
O
I
LL
O
tu
_J •
.a. o ;
H-
ia
in
n
in
10
o
i
o
II .
0
7.
^^
u; .
ia
I/) ''
U '
o •
_J
<
>
Q.
UJ
P
o •
_l
Q- .
LJ
D •
< '
>
1
H
**.
< '
a-
*
>-
IO '
o
1
CJ
o
^
&
ro
II II
* »
SO
o
1
too
O 0
oj o
O o
« o
u n
»O K)
o o
I I
j- o*
lO U.
»« »-»
10 ro
; in
c
o
O
03
CO
O
O
UA
-f
i
O
II II
-4 O
AT-6097-R12
Appendix I
Page 31
-------
in
o
I
I
o
u
O J-
o. o
13 i
<•£ U.
I OC
»-
./)
It
in
o
t
r-
ro
u
in
o
i
o-
vO
Kl,
II U
•O sj
O
I
O O
53 O
in o
r^- o
• •
•1 O
u u
«r tn
o o
I I
-u
•
-r
ro
o
•H
-P
O
-------
2. ONE-DIMENSIONAL KINETICS COMPARISON
The one-dimensional chemical kinetics program (ODX)
of Reference 16 was run for 26 reactions involving 13
species as given in Table 3. The pressure was constant
at 10 atmospheres and the initial temperature was 4300°R.
Temperatures, densities, and mass fractions of CO,
COg.I^OjOg, and N_ from the kinetics run were preloaded into
a slightly modified version of the Gosman program in which
one-dimensional flow was simulated. Since ODK assumes
inviscid flow, with no diffusion or heat transfer, similar
assumptions were introduced into the Gosman comparison'case
by setting the viscosity to an extremely small value and
the Prandtl and Schmidt numbers to unity.
The results shown in Table 4 and Figure 14
show very good agreement between ODK and the Gosman
program. The quasi-equilibrium and steady state assumptions
concerning the species 0, OH, H, N, and NpO are not exactly
valid. The worst cas3 is that of NpO, which does not really
react fast enough to be in steady state on the time scale
of this- example. Still, the assumptions seem to be adequate
for the calculation of NO and
AT-6097-R12
Appendix I
Page 33
-------
TABLE 3
REACTIONS AND INITIAL COMPOSITION FOR ODX-GOSMAN COMPARISON
. . .^^-.. ------ : ---- 7- -SPECIES TABLE- INPUT- ODK START CONDITIONS-
Mass Fractions
...... I N2 ,7501
2 02 .2259
2 N .2441-7
--... -NO • — ~*.-2600-3 — -r-.
" 5 0 .1590-2
'<•.•&. H .2863-5
:,. 7 _,.,.. OH -.1566-2
8 N02 .1171-5
9 N20 .9951-5
?.. iQ. -~ ,- H20 ^-^rS 380-2- -^-,—
, n co •• .3062-3
12 • C02 ./ . .1483-1
13 H2 •- • .3442-5
.......
•i REACTION
NO
'*-•• 02
N2
H2
^^-H +
*-'- N +
.•- o +
I- CO
Ei.O
w *
n»- N *
*•-: N +
H *
- 0 +
0 +
u02
- 'CO
(j +
OH
<-0
M *
0
Ofi
Oh
+ 0
= 0
= i4
= h
OH
0 =
N2
+ 0
Tbk
NO
Ok
OH
CARDS
= N02'
i- 0
+ N'
•*• H'
=- HiO»
NO'
= N20»
= C02»
REAX
= N2 +
= NO +
= ,NO •«•
ti»r tr?^rt~-'.
• A=2
A=3
A=7
-, A=3
A=l
A=l
A=6*
0'
0'
H»
t;cO =N
^^
•05E15
N=l.
N=l.
r
00
60
PK:Ih -™—
N=0.
•i '
' . . --'
t
r B=118.
"i ' -" '?"•
B=-l
7'
' 3=226-0'
N=-oO' B=92.60'
— N=l •
N=L
N=l.
N=O.OO
•10E13
•43E9'
•22E13
•9bE13
•82E13
•58E13
•OE13'
.6E11'
•3E13'
.75E12
00
50
00
t
t
-
r
f
f
f
—
»
r,--8=0
» B=0
» B=0
B=2.50
N=0.
N=-l
i4=0.
(•4=0.
N=0.
N=0.
N=0 •
.. N=0.
N=U.
N=0.
• 04
.00
• 00
r
•
f
* t
t
t
f
f
>
» .
f
•*
r
r
RX29
3=0.
0=6.
8=0.
B=10
3=24
8=24
8=0.
3=1.
8=0.
.870'
RX 21
RX 22
RX 23
RX .25
RX 26
RX 23
•.
RX 7»LEEDS 4.1
'JOHNSTON (1958) -
»APP£LTON(19&6)
'3ROKAW (.1970) >
'PREHN (1967) -_-«
»WRAY (1963)
'BORTN£R(1967)
' BAULCH(1968)
334'
250'
0'
.77'
.1'
.1'
600'
OtiO'-
»
' B=.780»
A=5.60£ll' M=0.00» B=1.0U'
A=l.
» A=k .
A=a-
A=L/.
A=7
A=l.
A=L
44E14'
96i£13.'
19E13'
7bL12'
•20E14
93E11'
90E13'
»
N=0.00
N.-O.OO
N=U.OO
N=0.00
N=O.QO
N=0.00
M=0.00
f
f
t
t
f
t
t
B=16
8=9.
U=ii-
B=0.
9=1.
8=3.
B=54
.60'
flU'
lt>'
7W»
93'
4U'
.15'
RX 1'LEEDS 4.11
- RX 2» ROBERTS -
RX 3'ROBERTS
RX 4»R03ERTS
- RX 5 ROBERTS -
RX 6 ROBERTS
RX 8 LEEDS 5.1
— RX 9 Ll.l
RX 10 L3.14= '
RX 11 L2.201-
RX31'GAULCH(68)
RX3U» BELLES (70)
RX39 BRAUHS 70
UX40'OAULCll(oa)
KX41'UAULCH(o6)
RX4d'SCHOFELD(67)
RX 50'AVRA,^ENK(65)
RXi>l'BAULCH(68>
LAST KE.AX
ThI,'
-------
m
to
§
o
CO
o
o
co
o
M
«
rH~
I
H
CO
3
W
§
§
§
CO
H
23
o
o
Fractions
co
CO
S
>O
o
rH
X
O
CVJ
&
cn
0
rH
X
Jg
0
H
X
tc
*a-
O
rH
X
w
o
T
0
rH
X
o
in
O
iH
X
CM
0
fe
to
O
rH
X
O
S3
Program
o
CQ
ja
•H
m o
Cn H
Cn rH
rH 00
^r c-
-sr C—
CM rH
(0 O
VO tO
CO "3-
CM OJ
VO rH
VO -sr
in -«r
r-l rH
o cr\
<7\ CVJ
in -
rH
CM CO
VO
-------
ODK
•Go smaii
12-
•1"
0
2468
m, t ,-,'-,- -, > Species
Time (milliseconds; ""^1
NO
N02
0
OH
H
N
10
Scale
IO"6
ID'4
lO-4
io-7
io-9
10"
Figure 14. Comparison of One-Dimensional Kinetics
and Gosman NO Calculations.
AT-6097-R12
Appendix I
Pacre 36
-------
APPENDIX II
EMISSIONS ANALYZERS
PRINCIPLES OF OPERATION
A. ANALYZER INSTRUMENTATION
This section reviews the gas analyzers and their principles of
operation as used to conduct emissions measurements at AiResearch
facilities.
A.I Heated Hydrocarbon Analyzer (Beckman Model 402)
The hydrocarbon analyzer, Figure A-l, is designed to measure the
total hydrocarbon contents of exhaust emissions from gasoline, diesel,
gas turbine, and jet engines. The analysis is based on flame ioniza-
tion, a highly sensitive detection method.
The instrument consists of:
(1) Heated, temperature-controlled sample line.
(2) Analyzer unit, incorporating a flame-ionization detector and
associated sample-handling system, with critical sample-
handling components contained within a temperature-
controlled oven.
(3) Electronics unit, containing an electrometer amplifier and
associated circuitry, readout meter, and recorder output
provisions. The electronics unit is attached directly to
the analyzer unit as shown in Figure A-l.
Sample from the source is drawn into the analyzer through the
sample line. To prevent the loss of higher-molecular-weight hydro-
carbons, the sample is maintained at an elevated temperature during
its passage through the sample line and the interior of the analyzer.
Temperature setpoint for the sample line and the analyzer oven will be
approximately 350°F for the tests carried out according to the plan.
AT-6097-R12
Appendix LI
Page 1
-------
Access Door for Burnei
Sample Pump, and Gas
Selector Valve
Access Door for
Sample Filter
* NOTE: Electronics Unit may be detached from Analyzei
Unit, for remote mounting at maximum distance of
75 feet.
ELECTRONICS UNIT *
A. ANALYZER UNIT
HYDROCARBON ANALYZER (BECKMAN MODEL 402)
(COURTESY BECKMAN INSTRUMENTS, INC.)
FIGURE A-l
AT-6097-R12
Appendix II
Page 2
-------
The hydrocarbon sensor is in a burner where a regulated flow of
sample gas passes through a flame sustained by regulated flows of a
fuel gas and air. The flame formed when fuel gas (hydrogen diluted
with an inert gas) burns in air contains an almost negligible number
of ions. Introduction of traces of hydrocarbons into such a flame,
however, produces a large amount of ionization. Within the flame,
the hydrocarbon components of the sample stream undergo ionization,
producing electrons and positive ions. Polarized electrodes collect
these ions, causing current to flow through measuring circuitry
located in the electronics unit. The ionization current is propor-
tional to the rate at which carbon atoms enter the burner and is
therefore a measure of the concentration of hydrocarbons in the
original gas sample.
The flow diagrams for the analyzer and burner are shown in
Figure A~2 A stainless-steel bellows-type positive displacement pump
of four cu ft/min capacity draws the sample into the analyzer and
through a glass fiber filter that removes particu.Iate matter. The
sample is then supplied to the burner under positive pressure. An
internal sample-bypass arrangement provides high-velocity sample flow
through the analyzer, thus minimizing system response time. A front-
panel flowmeter indicates bypass flow. Since the ionization level is
related to the flow rate of sample through the flame, the flowmeter
must be set to read a sample flow identical to the calibration gas
flow. The analyzer has rear-panel inlet ports for connection of
suitably pressurized zero and span standard gases. Flow of each
standard gas is controlled by a corresponding front-panel needle
valve. A front-panel three-way valve permits selection of either the
actual sample or the desired standard gas (i.e., "zero gas" or "span
gas" referenced in Figure A-2) used in calibrating the instrument.
The oven, which uses air-bath heating, is maintained at the
selected temperature by a solid-state temperature controller utiliz-
ing a thermistor sensor. A vertical partition divides the oven into
AT-6097-R12
Appendix II
Page 3
-------
TEMPERATURE-CONTROLLEOI
SAMPLE LINE
FLO*METER
1 vMrnf
FUEL INLET |^
IGNITOR DISCHARGE POINTS
THERMISTOR SENSOR FOR
BURNER-FLAMEOUT/FUEL-
SHUTOFF CONTROL CIRCUIT
SHIELD ASSEMBLY
ANODE TERMINAL
COLLECTOR ASSEMBLY
ELECTRODES
HOUSING
JET
AIR INLET
BASE
•SAMPLE INLET
FLOW DIAGRAM
SPECIFICATIONS
Analysis Temperature
Line Voltage
Ambient Operational
Temperature
Ambient Operational
Humidity
Polentiometric Output
Sensitivity
Ranges .
Response .
Electronic Stability ...
Repeatability
Temperature Controlled
Piobe
.Adjustable from 800' F to 400* F
.107-127 VAC 50/60 Hz.
1000 watts max.
,.32'FtollO'F
. 95* R.H.
.. 10 mV. 100 mV. IV
.5 ppm Id 10'. lull scale as CH.
With H,/N, or H,/He Fuel
..XI. X5, X10. X50. X100. X500,
X1000. XSOOO with continuous
electronic span adjustment
. Less than 1 second lor 90*.; ol
linal reading (with CH« from ana-
lyzer input without sample prose)
, . £1% lull scale/24 hrs. with less
than 10* ambient temperature
change
.±1% lull scale lor successive
samples
.. 10 IL length, teflon surface in con-
tact with sample (proportional
temperature controlled and ad-
|ustabl« from 200'F to 400 - F)
BURNER DIAGRAM
HEATED HYDROCARBON ANALYZER SPECIFICATIONS AND DIAGRAMS
(FIGURES COURTESY OF BECKMAN INSTRUMENTS, INC.)
FIGUREA-2
AT-6097-R12
Appendix II
Page 4
-------
two compartments with separate doors. The left-hand compartment con-
tains the burner, the sample pump, and the three-way gas-selector
valve. The right-hand compartment contains only the sample filter and
permits access to the filter without disturbing temperature equilib-
rium of the other elements.
Within the analyzer, the fuel gas is routed to the burner through
a solenoid valve controlled by the burner-flameout/fuel-shutoff cir-
cuitry. A thermistor sensor continuously monitors the status of the
burner flame. In event of flameout, the valve closes to stop the flow
of fuel gas; simultaneously, a front-panel indicator illuminates to
alert the operator.
The electronics unit has front-panel controls for range selec-
tion and adjustment of zero and span. Readout is on a front-panel
meter calibrated linearly from 0 to 100. In addition, a selectable
output of 10 mv, 100 mv, or 1 volt is available to drive a voltage-
type recorder.
A.2 Chemiluminescent/NO - NO Analyzer
X
The chemiluminescent analyzer is packaged as four separate units:
(1) control unit, (2) an analyzer unit, (3) a reaction chamber mechan-
ical vacuum pump, and (4) a converter for the thermal conversion of
NO2 to NO. For these tests, a Thermo Electron Corporation analyzer
model 10A was used, which was operated in the NO and NO modes.
X
A typical arrangement of the Model 10A chemiluminescent analyzer
is shown in Figure A-3. Other equipment needed for use with the ana-
lyzer are the NO and NO~ standard gases, an oxygen source for the
ozone generator, and an accumulator and suitable sample bypass pump to
provide two to two and one-half cu ft/hr sample flow.
The control unit contains the switch for selection of sensitivity
from seven available full-scale ranges (10, 25, 100, 250, 1000, 2500,
and 10,000 ppm) and potentiometers which provide for instrument
calibration.
AT-6097-R12
Appendix II
Page 5
-------
ANALYZER UNIT
tange Sensor Switch
PPM meter
Calibrate (Gain)
djustment
Photomultiplie r
Dark Current
• (Background)
Suppression .
NO -to-NO
onverter Power
Main AC Power
CONTROL UNIT
MODEL 10 CHEMILUMINESCENT ANALYZER
(CONTROL AND ANALYZER UNITS)
FIGURE A-3
AT-6097-R12
Appendix II
Page 6
-------
The analyzer unit contains tlve reaction chamber, the photo-
multiplier tube, the ozonator, the ozonator power supply, the oxygen
and gas sample lines, capillaries, and pressure regulators.
Figure A~4 presents a schematic drawing of the entire chemilumi-
nescent instrument with the portion inside the dashed rectangle repre-
senting the analyzer unit. The heart of the analyzer is the cylin-
drical reaction chamber where sample gas containing NO molecules mixes
with 03 molecules from the ozonator. Electronically excited N02 mole-
cules are created that emit light (chemiliminescence) as the orbital
electron's decay to their ground states.
The chemiluminescence is monitored through an optional filter by
a high sensitivity photomiltiplier positioned at one end of the
reactor. The filter-photomultiplier combination responds to light in
a narrow wavelength band unique to the desired electron decay. Sample
flow is controlled so that the output from the photomultiplier is
linearly proportional to the NO concentration.
Oxygen, 0-, enters the analyzer unit, passing through a pressure
regulator that is used to regulate the flow rate, and enters the
ozonator. A fraction of the O2 is converted to O,, and the mixture
passes through an orificing glass capillary to the reaction chamber.
Sample gas enters the instrument, passing through another glass
capillary, and is bled off to the reaction chamber. That portion of
the entering sample not diverted to the reaction chamber passes
through a front panel flowmeter adjusted to two standard cubic feet
per hour and a regulator to the instrument exhaust system. A bypass
pump is used to pull the sample through the instrument. That portion
of the entering gas sample diverted toward the reaction chamber is
directed to the rear of the analyzer unit, where the sample gas will
AT-6097-R12
Appendix II
Page 7
-------
w
H
H
U
AT-6097-R12
Appendix II
Page 8
-------
enter the converter if the instrument is operating in the NO mode
(i.e., NO - NO_ mixture mode). As explained below, use of the con-
verter is unnecessary if the instrument is operating in the NO mode.
The basic chemiluminescent analyzer is only sensitive to NO mole-
cules, as opposed to N0_ molecules, since 03 does not react with NO_
to create chemiluminescence. Therefore, to measure NO (N0_ + NO),
X ^
the N0_ must first be converted to NO. The conversion is accomplished
by passing the sample gas through the converter, a thermally-insulated
resistance-heated stainless steel coil at 1292°F. With the applica-
tion of heat, NO^ molecules in the sample gas are reduced to NO mole-
cules. Two three-way valves located on the front of the converter
direct the sample gas either through the converter to measure NO or
X
past the converter to measure NO.
A mechanical vacuum pump is supplied to evacuate the analyzer
reaction chamber to pressures in the 12 torr range. A metal bellows
hose connects to a molecular sieve installed above the mechanical
pump. The purpose of this sieve is to absorb O_ in order to prevent
breakdown of the pump oil.
The gas sample is pulled through the instrument by a small dyna-
pump, after which it is exhausted to ambient. This pump improves the
overall system response by moving the flow in the main sample line by
about 20 liters per minute, while each instrument in the analyzer
group removes flow from the main sample line at a much lower rate.
AT-6097-R12
Appendix II
Page 9
-------
A.4 Non-Dispersive Infrared (NDIR) Analyzers for CO - CO-
To measure the differential absorption of infrared energy, this
instrument employs a double-beam optical system contained in the
analyzer section. A simplified functional diagram and instrument
specifications are shown in Figure A-5
Two infrared sources are used, one for the sample energy-beam,
the other for reference energy-beam. The beams are blocked simul-
taneously ten times per second by the chopper, a two-sigmented blade
rotating at five revolutions per second. In the unblocked condition,
A of Figure A-5 each beam passes through the associated cell and into
the detector.
The sample cell is a flow-through tube that receives a continuous
stream of sample. The reference cell is a sealed tube filled with a
reference gas. This gas is selected for negligible absorption of
infrared energy of those wavelengths absorbed by the sample component
of interest.
The detector consists of two sealed compartments separated by a
flexible metal diaphragm. Each compartment has an infrared-
transmitting window, to permit entry of the corresponding energy-beam.
Both chambers are filled, to the same sub-atmospheric pressure, with
the vapor of the component of interest. Therefore, each chamber will
absorb infrared energy from its source and will respond.
The response of the two detector chambers differs, since in
operation the presence of the infrared-absorbing component of interest
in the sample streams leaves less energy available for the correspond-
ing detector chamber. There is, thus, a difference in energy levels
between the sample (containing the component of interest) and the
reference (non-absorbing) sides of the system. This energy difference
results in the following sequence of events.
AT-6097-R12
Appendix II
Page 10
-------
SPECIFICATIONS
MAXIMUM ZERO DRIFT;
* 1% of full scale per 8 hours.
MAXIMUM SPAN DRIFT:
£ 1% of full scale per 24 hours.
SENSITIVITY:
0.556 of full scale.
ACCURACY:
ee~l won Abwrbing I. • • I Infrared-Absorbing
V. i Molecule* I fft\ Molecules
AMPLIFIER RESPONSE SPEED:
90% response in O.S second.
AMBIENT TEMPERATURE RANGE:
IR315. IR315L, -20' to +1ZOT,
OUTPUT (Options Availaole) :
Current output— 0 to 5 ma into SOO ohms maximum,
Voltage output-Adjustable to match any potentiomrtric
recorder having a span of between 1 and 100 my.
VOLTAGE AND FREQUENCY
(Options Available) :
116 * 15 volts. 60 ±0.5 Hz (cps) : or
115 ±15 volts, 50 ±0.5 Hz (cps).
MAXIMUM POWER CONSUMPTION
IR315. 530 watu ; IR315L, 640 watts.
MAXIMUM SEPARATION OF AMPLIFIER
CONTROL SECTION AND
ANALYZER SECTION
500 feet.
SHIPPING WEIGHT:
IR315S. 115 pounds ; IR315L, 170 pounds.
NET WEIGHT :
IR31SS. 85 pounds : IR315L. 120 pounds.
CATALOG NUMBERS:
95700 Model IR31SS Infrared Analyzer (Short Path)
95701 Model IR315L Infr^r-d Analyicr (Long Path)
Sample
Source
Sample In
••• •- •-*:•$ NOT BLOCKED BY
"K'POtP PASS IHCOU
.'.-ILLS AND INroOLT£CTOR
Samp.. Out
Stationary
MM* Button
Oscillator
Unit
To Amplifier/Control Section
Chopper Motor —j I
Reference
Source
Chopper
Sample
—Source
•c'.l*' — Sample In
8 BExMS BLOCKED
•;•' CHOPftR DO NOT
=»F/,On DETECTOR
Sample Out
NON-D IS PE ^ < 1 \-'£ INFR.2- -RED
ANALYZER SECT TON SPSCIFIC'ATI
AND FUtfC.OiON.'VL DIAGRAI-j
FIGURE A-5
B
ur-tnti
-------
(1) Radiant energy absorption: In the sample cell through which
the infrared radiation passes on the way to the sample
chamber of the detector, part of the original energy of the
sample infrared beam is absorbed by the component of inter-
est present in the sample. In the reference cell, however„
absorption of infrared energy from the reference beam is
negligible, and the energy of this reference beam is highere
(2) Temperature effect: Inside the detector, each beam heats
the gas in the corresponding chamber, because of absorption
of infrared energy by the component of interest. The gas in
the reference chamber is heated to a higher temperature,
however, since the energy available from the reference beam
is higher.
(3) Pressure effect: The higher gas temperature in the refer-
ence chamber raises the pressure of this compartment above
that of the sample chamber.
(4) Mechanical energy effect: The higher gas pressure in the
reference chamber distends the diaphragm toward the sample
chamber. The energy difference between the two chambers in
thus expended in flexing the diaphragm.
(5) Capacitance effect: The diaphragm and an adjacent station-
ary metal button (see Figure A-5) constitute a two-plate
variable capacitor. Distention of the diaphragm away from
the button decreases the capacitance.
When the chopper blocks the beams, as in B of Figure A-5, pres-
sures in the two chambers equalize, and the diaphragm returns to the
undistended condition. As the chopper alternately blocks and unblocks
the beams, therefore, the diaphragm pulses, thus changing detector
AT-6097-R12
Appendix II
Page 12
-------
capacitance cyclically. The detector signal is passed through the
electronic circuitry, where it is treated and sent to a meter and
recorder.
The meter reading is a function of the concentration of the
component of interest in the sample stream. When the instrument is
put into operation, it is adjusted so that a reading of zero or any
desired arbitrary reading corresponds to a concentration of zero per-
cent of the component of interest, while a fullscale reading corre-
sponds to the highest concentration in the operating range covered.
Each instrument is provided with a calibration curve for converting
meter readings to concentrations.
AT-6097-R12
Appendix II
Page 13
-------
-------
H
H
H
w
EH
W
W
W
cn
X <
H EH
O <
& Q
W
cn m
& A
ft, ft
<
en
-*
^
(
hi
U
ft
K
§
H
M
z
3
g
R
o
s
J
as
\
o«
1
•-j
ai
u
S
u
g
»-»
H
x^
-
cc
o
»•
M
c
0
.
<
£
o
!2
^
4
fs
§
t:
^
>1
^
1
1
*
1
x.
-4
X
•••
S
1-
ik.
is*
N
^
I'
^
K
c
'•N
»
C
^
-A
.X
*-\.
u
*-*
<
60
X
Z
(6
u
a
^
«Q
V)
*
^
?-
o
^
r-i
0
v
•
\
>
Q
v-
C^>
-v
m
(X,
re
v
i/
u.
re
Cu
H
U
M
W
C
^
«
*
^
V
"*
r
^
X.
^
V°
^
•
^
•V.
>
V
5
t'.
IT
C\J
U-
o
rv
F
u
P
u
H
H
U
r;
H
u
c
*T-
'
•
S
1
p-
u*
o
=
Q
O
v
Cx
ft;
^
^
>
~~^
,
•N
~S
r\
^
|
|
[:
o-
c
fL
U
E
g
t
I
i
\
f
1
r
i
\
\
t
t
t
t
1
vc
en
w
X
c_
t
t-
o
^
y
B
f
1
I
\
4
i
A
n
1
1
1
1
A
F\
V
—
ro
INCHES
FICE DIAMETER
S
2
X
to
t-
V
(V
I
u
1
~o
^
0
cs
\
*
r^
Q
X
S
X
V
m
CM
u
X
5C
V
tn
8
^
ft
(V
X
Cu
rt
g
o>
it-
^
0-
g
t
v
^>
"N.
^
1
V]
s
0]
i>
O.
^
S?
§
c
^
\
V. \
CD
t>
V)
u
ft
(X
(J
P
f-
tn
K
(r
H
o
£
r-4
^
H
e
s
^
e
a
0
s:
r\
v.
X
\
^(
2
^
^
*
>
i
^
X.
1
\s
\
Cv
e
v
u
r
C
H
in
c.
C/l
S
c
H
t/1
|
r
£
5
s
-
2
C
C
^
V")
ij
3
*
^
^
V
«i
§
,\
K
o
^
tbt
N
vc
^
en
K
fr
(L.
O
H
E
t-l
C
H
v
ft
U
C
*~*
ir
r
^
r
e
V)
V
u
(
H
o
H
u
t/1
M
c
a1
c
fr-
i
r
6
\
:
^
r
i
a
\
•.
:
1
(0
a
V
1
r
I
f?
C
^
t;
V
A
c-
^
\
IT
(VI
IT
*•
U.
o
gj
u
H
a
H
c
h-
t
i
i
AJ
/•
r
p
K
r-
0
r-
^
f
(<
a'
c-
r.
i.
\
TV
U>
*•
O
V
)
;
rx
^
i
^
V
^
_^
J
5
1
r
^
\
c
t-H
w
p-
u
j
s
IS
^
^
-
?
[NJ
X
s:
i^>
r
X
^
^
c
if>
s
c
c
u
E
S
U
u
l(
N,
s,
^
in ^
\
Xl r
-I' 'x
?7
^
^
u
^
*~
**>
< t-
1 V)
^
e
O V
X u
a: o.
H t-
bJ u
o **
al u
w u
e g
2 5
^
V.
r; '
^
£
>^
^
g
0
c>\
^' 1
Q
S
\
£:'•
1
\
\
-1 «
u< u<
0 O
dl
H
C
g'
•^
^
^/f
^.'
<•
S'
^
?lf
v
v^'
,c '
S'
5'
V)
^
^
T
3C :
ft
UJ f
U f
U^F
gi
1
«
l'
^
A
M.
l^
\ »
\i
f^
\ ^
^(D
i s
J U
& z
rFICE DIAMETER
S g
2
J6031P
0
LU
§
C
u
;
«.
5
MtlTUOC CO
(
.j
0
f
^J
>
-^
P
m
X
=»•
X
3
a
s
•-t
g
^
c*
(A
. ^
^
5 tn
0 '
•3
^5
^
;l
a w
s s
• a
5 I .
5 S '.
0 S £
o
H
§>
. 3 F-
.wui ;
/I ci W C
auH s
^e^r
3'g^fe
^•jj
S"1""
-gut
3£yfc
S^hr
2. IF COLUMN0 ARE SKIPPED,
3. DECIMAL 'NT MUST APPEA
4. FOR'BELLi-lOOTH: AIR OR
9 3
s i
R s
~N
a.
DIAMETE
gs
« a
P?
LLMOUTH
\L RECE
w n
n «
c-S
t> i
!gg:
ig«
s,., s
ORIFIC
DAI
Appendix III
Page 1
-------
£^
N
B e
3S
e is
<
a
as
o
O, u
ait:
H _ H
SS
2S
n! <
O4 £.
<"*
f-t S
38
gy
a s
Appendix III
Page 2
-------
FUEL SPECIFICATION
OIL & FOEL ANALB3S // .-f / /
MATERIALS ENGINEERING
Requestor
Copies To PtttKEXSTfet
A. Dept. 93" 20 Customer^
Bfenufgcturer
Date ^dT. SL?, *9?a.
Engine Serial No._
Operating Hours
Sample Origin
Charge No.
Date Required
ttlLt>
/»7!L
B. P. Distillation
I9P/.TZ. °r
&
Piilllii'o'"
Or«Tltr
llOO°F
O Flash Point
DT.A.N.
QOther
A <%>-
Vr
Vap. Press.
JViseosity
cs* 9
77
100
••••mM
210
O
AT-6097-R12
Appendix IV
Page 1
-------
-------
:\PPENDTX V
."lyAFLf: EMISSIONS RECOK
tt
lain
o
~,~~Ll "'a ," " " '
AT-6097-R12
Appendix "7
Page 1.
-------
-------
APPENDIX VI
COMBUSTOR PHOTOS
PHOTOGRAPHS
VI-1 PNEUMATIC IMPACT INJECTOR
VI-2 SKP26259M, VAPORIZER COMBUSTOR-
(CONE DOME, BLOCKED PORTS)
VI-3 FLAME TUBE, SKP26489, VAPORIZER
COMBUSTOR
VI-4 DOME, SECONDARY PIPES (SKP26489M1)
AT-6097-R12
Appendix VI
Page 1
-------
AT-6097-R12
Appendix VI
Page 2
-------
AT-6097-R12
Appendix VI
Page 3
-------
AT-6097-R12
Appendix VI
Page 4
-------
G. P. O. 1973 - 747-788 / 328, REGION NO. 4
-------
-------
-------
-------
|