AD-A151 464 AFWAL-TR-94-2104
INFLUENCE OF FUEL PROPERTIES ON GAS TURBINE COMBUSTION PERFORMANCE A.H.Lefebvre
Conihistion Laboratory Thennal Science and Propulsion Center School of Mechanical Engineenng Purdue University West Lafayette, Indiana 47907 I•ulary 1985
[vial Report for Period 3 January 1983 - 30 September 1984
Approved for public release, distribution unlimited.
BH1~ ilL UULPY E
AERO PROPULSION LABORATORY AIR FORCE WRIGHT AERONAUTICAL LABORATORIES AIR FORCE SYSTEMS COMMAND WRIGHT-PATTERSON AFB, OH 45433-6563
....-
.
". ....-.. "..... .
. •'
.-
*
l" .
'l+
•
. ". . .
,.•lip
. o ,,
. " .... -. .
.•
-". ". *". >•.
•
. ".. . . .-'J"-.. . .... -. .
,
•
,
,
08
03
85
065
.,;; .
,
.
•
;..
",
.",
-
',"-'•
'
"
-
:
"
•
when ý,overnment d"aw'ings, specifications,
ox other d•ita are used for any purpose
other than .n conneccion with a defirnitely related Covernment procursment operation the United State' Government there*y incurs no respolns.tbility nor any obligation whatsoever; and the fact that the government my have formulated, furnished, or in any way supplied the said drawings, specifications, or other data, is not to be regarded by implication or otherwise as in any ma&ier licensing the holder or any other person or corporateion, or conveying any rights or permission to manufacture use, or sell any ,atented invention that May in Any way be related thereto. This report has been rep'iewed by the Office of Pub4.c Affairs (ASD/PA) and as releasable to the National Technical Inforxition Service (NTYrS. At NTIS, it will be avatiable to the general public, including foreign nations. This technical report has been reviewed and is approved for publication.
CURTIS ll.REEVES Project Enqineer Fuels Branch Fuels and Lubrication Division FOR THE COMMANDFR
e
ARTHUR V. CHURCHILL, Chief Fuels Branch Fuels and Lubrication Division
4!
ROBERT 0. SHEROILL, Chief Fuels and Lubrication Division Aero Propulsion Laboratory
"If youz address has changedi, if you wish to be removed from our mailing list, or :5 che addressee 's no longer employed by your organization please notify AFWAL/POSF :W-PAFR, OH 4;433 to help us m'aintain a current mailing list". Copies of h report sLould not be returned unless return is required by security cons ier,-.tions, contractual obligations, or notice on a specific document.
UNCLASSIFI ED REPORT DOCCUMEMJATION PAGE
-~~
J 11 V -0 A1 141C'#4 4
MIAKPINIjS
II.
kIO eES IYIV '!LA %80ýI C,
UNC-LASSILED
~iCOW r 11,
OI10rng111UyVON/AVAILASILl1 4 OF REPORIT
C.LASSIb'ICATICJN AUTH-ORYY V~
2b
Approved for P'ublic Release; Distribiition Un)l~imited
*
5. MUNI T0RI NO ORGANIA TION AEIPOR1 NUMBER(S)
a PtolfORMING OPQANI%!ATI0N RIEPON11NUMI1E~145
N/A
AFWAL-TR-84-2104
t,&a NAML Of' PE.R
*.AQORVS!C
('iIl.
Iii. NAMA OF MQITO10RINCI QROANIZA'I(GN
It. OFFICE SYMBOL
OPIMING 0001ANIZATION
School of .Iecha,'ical l.nqineer*' ing,, Puirdu- Universlt, 1'
,iero Propulsion Laboratory (AFWAL/P6iS) Ar i.orce Wr ight Aeron~aui~cal Latir'atodies. w
(ifappicable)
N/
State and IIP Codri
7b.
NAME 'Lt' IlUN )INGISPONSORINC, 0140ANIZAII ( 14it/
Ski. OFFICE SYMBOL
AODDRESS WIs,I Stati, and /it'
W PROCUREMENT
AFWAL/POSF
NUMBE H
F33616-81-C-2O67
IRG~
¶0 SOUR~CE OF FUNDING NOS.
Cjd.0
Wright-Patterson Air Force Base OH 45433-6563 IIT TE(I
INSTRUMENT IDENTIFICATION
appoliebfle
A r usonLbrtr
Be
ZI1P Code)
Wright-Patterstn Air Forct' Base 014f 45433 -6563
West L.~fye~ttc .`K 47lO'i
Be
A001OR11ISS (itiy, Slatit and
ELEMENT
ChdissitteaoiIINFI UENCE OF FUEL
NO
62203F
u.Je irut~ivI
AJC
IS
NO,
NO
NO
3048
05
13
Vr . Mo., Vayp
15 PAGE LOUN'T
(A JI
L....
RP[~RT!FS' ON GAS TURBINE COMBUSTION PERFORAN 12. Pt RSONAL AUTHOVISl
13a. TYP& OF REPORT
13h. TIME COVERED
143
SUPPLEMENTARY NOTATION
I SUBJECT TERMS f12,ntinuu -1' -vers., 1___
f;OSATI ICODES ROU_
S LIItý
21 ;ý_J_ 04--19
*
January 1985
io3ffiepA41
SuavinAry Jchnical Rift I "0lM3 16
14 TJATF OF REPORT
AS r
YIAA. Can ifIAU,"
/Results
Fuels
Alternative Fuels Gas Turbine Comb. Vivae r.
ifnc."uo'v and idrntify by bicr
nkuN"-
Exhaust Emissions Lean Blow (Jul Fuel Atomization Comb. Efficiency
Ignition Liner Wall '1!mner1tur
itn-r',imav and iekntify by biuc* numberl
of an analytical and experimental program to determine the effects of broad Variations, in fuel properties on the performance, emiss;ions, and durability of several prominent turbojet engine combustion systems, including both tubo-annular and annular configurations, are presented. Measurements of mean drop size conducted at representdtive engine operating conditions are used to supplement the available experimental data on the effects of combustor design parameters, combustor operating conditions, and fuel type, on combustion efficiency, lean blowout limits, lean lightoff limits, liner wall temperatures, pattern factor, and pollutant emissions. The results of the study indicate that the fuel's phy.;cal properties that govern atomization quality and evaporation rates strongly affect combustion efficiency, weak extinction limits, and lean lightoff limits. The influence of fuel chemistry on these performance parameters is quite small. Analysis of the experimental data shows that fuel chemistry has a significant effec&7_20
UIST RIBUTION/AVAILAtilLIT Y OF ABSTRACT
ti4CLA'iIIE()/UNt.IMIri:
1
211NAME OF RESPONSIBLE
CURTIS MI. RILVEZDD FORM 1473, 83 APR1
SAIML AS ReT
iNrIVIOUAL
-J D71C USERS 0
21, A.STRIACT SECURITY CLASSIFICATION'
UNCLASSIFIED 22h. TELEPH4ONE NUMS. P i nci~ude llwq.Coe
(513) 255-3524
EDI rION OF I JAN 73 IS OBSOLE I E
22c
]TFIICE SYMUCI.
AFWAL/POSF UNLJjF I E D SECURITY CLASSIFICATION OF THiIS PAG E
N
-7.
-- :.:
-
~-.T-wr9C.
V-6
-
ItUNit ASS IFIiL M$CUA% 'f 4. AG04 4CAI Clh Of" THIN 1MAG
Block 18 (cont):
•" Pattern Factor S Sauter Mean Diamaeter
Block 19 (cont):
n flame emissivity, flaiue radiation, and liner wall temperature, but its influence on the of carbon monoxide, unburned hydrocarbons, and oxides of nitrogen, is small. Snvoke emissions are found to be strongly dependent on combustion pressure, primaryzone fujel/air ratio, and the mode of fuel injection (pressure atomization or airblast). Fuel chemistry, as indicated by hydrogen content, is also important 1 8nemis5ions
*
At the high power conditions where the durability of hot sectioiilomponents is of major concern, the influence of fuel type on pattern factor is shown to b negligibly small. Equdtlions are presented for the correlation and/or prediction o several key aspects
of combustion performance, including combitstior, efficiency, weak e tinction limits, lean combustor size, comlightoff limits, pattern factor, and exhaust emissions, in terms bustor geometry, engine operating conditions, fuel spray characte istics, and fuel type.
UNCLASSIFIED SECURITY CLASSIFICATION OF THIS PAO"
-
2.,*
,.
-.-
.
.-
-
.
..
.
S
-
1-
r-.
e-
.--- r--r-,-rn,
4--
.--
.'-
-i.--r ,
wr
- .
-
.
t
-
-
- .
r-w u*
v
•
.
V-
r-w
r7
w: --
,
17Tw
•a
FOREWORD submitted by the Combustion Laboratory
This final report is
of the Thermal Science and Propulsion Center, Purdue University.
cal Engineering,
conducted under Contract No.
School of Mechani-
The report documents work
F33615-81-C-2067
3 January 1983 to 30 September 1984.
during the period
Program sponsorship and
guidance were provided by the Fuels Branch of the Aero Propulsion Laboratory
Air Force Wright Aeronautical Laboratocies,
(APL),
Wright-Patterson Air Force Base,
Ohio.
The Air Force Technic 1
Monitor employed on this program was Mr Curtis M. Reeves.
ATecu•ston For
FTTS
",A&I
nTIr TAB
In . . ....
..
.
I.D ,t r.
. ~~~~~ .
..
. .
"I
A
.
.
.
..
.l
*
.
*
,p
.
-/o2
aa
rT
•
v=
IV
TABLE OF CONTENTS PAGE
3ECTION I.
INTRODUCTION
1
II.
FUEL ATOMIZATION
7
III.
.'OMBUSTION EFFICIENCY
25
IV.
1,,%AN BLOWOUT
39
V.
IGNITION
53
VI.
LINER WALL TEMPERATURE
63
1. 2. 3. 4. VII.
Internal External Internal External
64 65 65 66
Radiation Radiation Convection Convection
73
POLLUTANT EMISSIONS 1. 2.
Oxides of Nitrogen Carbon Monoxide
75 76
3.
Unburned Hydrocarbons
94
4.
Smoke
95
VIII.
PATTERN FACTOR
117
IX.
DISCUSSION AND SUI4ARY
125
1. 2. 3. 4. 5. 6. 7. 8. 9.
Combustion Efficiency Lean Blowout Lean Lightup Liner Wall Temperature NO Emissions COXEmissions Unburned Hydrocarbons Smoke Pattern Factor
X.
CONCLUSIONS
133
XI. XII.
REFERENCES LIST OF SYMBOLS
136 145
-V
".J
125 126 127 128 129 129 130 130 131
LIST OF ILLUSTRATIONS FIGURE
PAGE
1.
Distillation Characteristics of Test Fuels.
5
2.
Schematic Diagram of Spray Test Rig.
9
3.
Mean Drop Sizes Obtained for J79-17A Fuel Nozzle.
14
4.
Mean Drop Sizes Obtained for J79-17C Fuel Nozzle.
15
5.
Mean Drop Sizes Obtained for 7iul Fuel Nozzle.
16
6.
Mean Drop Sizes Obtained for TF39 Fuel Nozzle.
17
7.
Mean Drop Sizes Obtained for J85 Fuel Nozzle.
18
8.
Mean Drop Sizes Obtained for J85 Fuel Nozzle.
19
9.
Influence of Ambient Air Density and Atomizer Air/Fuel Ratio on SMD for F100 Fuel Nozzle.
20
10.
Influence of Atomizer Air/Fuel Ratio and Pressure Drop on SMD for F100 Fuel Nozzle.
21
II.
Variation of Effective Evaporation Constant with Normal Boiling Point at a Pressure of 100 kPa.
27
12.
Variation of Effective Evaporation Constant with Normal Boiling Point at a Pressure of 1000 kPa.
28
13.
Variation of Effective Evaporation Constant with Normal Boiling Point at a Pressure of 2000 kPa.
29
14.
Comparison of Measured and Predicted Values
31
of Combustion Efficiency for J79-17A Combustor. 15. 16.
Conmparison of Measured and Predicted Values of Combustion Efficienc:y for J79-17C Combustor. Comparison of Measured and Predicted Values
32 33
of Combustion Efficiency for F101 Combustor. 17.
Comparison of Measured and Predicted Values of Combustion Efficiency for TF41 Combustor.
34
18.
Comparison of Measured and Predicted Values of Combustion Efficiency for TF39 Combustor.
35
19.
Comparison of Measured and Predicted Values of Combustion Efficiency for J85 Combustor.
36
-
vi
-
20.
Compazison of Measured and Prtdictod Vatlues
37
of Combustion Efficiency for T133 Combustor. 21.
Comparison of Measured and Predicted Values of Combustion Efficiency for F100 Combustor.
38
22.
Comparison of Measured and Predicted Values of Lean
45
Blowout for J79-17A Combustor. 23.
Comparison of Measured and Predicted Values of Lean Blowout for J79-17A Combustor.
46
24.
Comparison of Measured and Predicted Values of Lean Blowout for J79-17C Combustor.
47
25.
Comparison of Measured and P•.zdicted Values of Lean Blowout for J79-17C Combustor.
48
;'6.
Comparison of Measured and Predicted Values of Lean Blowout for F101 Combustor.
49
27.
Comparison of Measured and Predicted Values of Lean Blowout for TF39 Combustor.
50
28.
Comparison of Measured and Predicted Values of Lean Blowout for J85 Combustor.
51
29.
Comparison of Measured and Predicted Values of Lean Blowout for F100.
52
30.
Comparison of Measured and Predicted Values of Lean Light Off for J79-17A Combustor.
55
"31.
Comparison of Measured and Predicted Values of Lean Light Off for J79-17A Combustor.
56
32.
Comparison of Measured and Predicted Values of Lean Light Off for J79-17C Combustor.
57
33.
Comparison of Measured and Predicted Values of Lean Light Off for 579-17C Combustor.
58
34.
Comparison of Measured and Predicted Values of Lean Light Off for F101 Combustor.
59
35.
Comparison of Measured and Predicted Values of Lean Light Off for TF39 Combustor.
60
36.
Comparison of Measured and Predicted Values of Lean "Light Off for J85 Combustor.
61
37.
Comparison of Measured and Predicted Values of Lean Light Off for F100 Combustor.
62
9
-
"-vii
0 ", . :': ... - A • '_, ....". .,..
-
'. . . .", .. ". .:. - .-- .:,.........:' :• .: :..2 .- :.{ .: :-- -.: " :
", , , :
.: - :. '"5 "'
"" " . - ",
38.
Comparison of Measured and Predicted Values on the
68
Effect of H2 Content on Liner Temperature for J"9-17A Combustor. 39.
Comparison of Meý.sured and Predicted Values on the Effect of H Content on Liner Temperature for J79-17C
69
Combustor. 40.
Comparison of Measured and Predicted Values on the Effect of H2 Content on Liner Temperature for Fl01
70
Combustor.
41.
Comparison of Measured and Predicted Values on the Effect of H2 Content on Liner Tenperature for TF41 Combustor.
71
42.
Comparison of Measured and Predicted Values of NO
77
..
Emissions
?or J79-17A Combustor.
43.
Comparison of Measured and Predicted Values of SOx Emissions for J79-17A Combustor.
78
44.
Comparison of Measured and Predicted Values of NOx Emissions for J79-17C Combustor.
79
45.
Comparison of Measured and Predicted Values of NO X "Emissions for J79-17C Combustor.
80
46.
Comparison of Measured and Predicted Values of NO
81
-
Emissions for F101 Combustor. 471.
Comparison of Measured and Predicted Values of NOx
82
Emissions for F101 Combustor. 48.
Comparison of Measured and Predicted Values of NO x Emissions for TF41 Combustor.
83
49.
Comparison of Measured and Predicted Values of NOx Emissions for T?41 Combustor.
84
50.
Comparison of Measured and Predicted Values of NOx Emissions for TF39 Combustor.
85
51.
Comparison of Measuxed and Predicted Values of NO Emissions for TF33 Combustor. x
86
52.
Comparison of Measured and Predicted Values of NO Emissions for F100 Combustor. I
87
53.
Comparison of Measured and Predicted Values of CO Emissions for J79-17A Combustor.
89
54.
Comparison of Measured and Predicted Values of CO
90
-
viii
-
Vil
.
.
- .
,
S.. .
.
*
*~Sf
Emissions for J79-17C Combustor. 55.
Comparison of Measured and Predicted Values of CO Emissions for F101 Combustoz.
91
56.
Comparison of Measured and Predicted Values of CO
92
Emissione
57.
for TF41 Combustor.
Comparison of Measured and Predicted Values of CO Emission for
93
F100 Combustor.
58.
Comparison of Measured and Predicted Values of Unburned Hydrocarbons Emissions for J79-17A Combustor.
96
59.
Comparison of Measured and Predicted Values of Unburned Hydrocarbons Emissions for J79-17C Combustor.
97
60.
Comparison of Measured and Predicted Values of Unburned Hydrocarbons Emissions for J79-17C Combustor.
98
61.
Comparison of Measured and Predicted Values of Unburned Hydrocarbons Emissions for F101 Combustor.
99
62.
Comparison of Measured and Predicted Values of Unburned Hydrocarbons Emissions for TF41 Combustor.
100
63.
Comparison of Measured and Predicted Values of
101
Unburned Hydrocarbons Emissions for TF41 Combustor.
*
64.
Graphs Illustratiag Influence of Hydrogen Content and Engine Operating Conditions on Soot Emissions for -779-17A Combustor.
106
6.5.
Graphs Illustrating Influence of Hydrogen Content and Engine Operating Conditions on Soot Emissions for .779-17C Combustor.
107
66.
Graphs Illustrating Influenc6 of Hydrogen Content and "Engine Operating Conditions on Soot Emissions for Fl01 Combustor.
108
67.
109
68.
Graphs Illustrating Influence of Hydrogen Content and Engine Operating Conditions on Soot Emissions for TF41 Combustor. Graphs Illustrating Influence of Hydrogen Content and
69.
Operating Conditions on Soot Emissions for "TF39 Combustor. Graphs Illustrating Influence of Hydrogen Content and
"Engine
Engine Operating J85 Combustor.
Conditions on Soot Emissions for
-
ix-
110
111
70.
Graphs
Illustrating
Influence of Hydrogen Content and
11i
Engine Operating Conditions on Soot Emissions for TF33 Combustor.
S71.
Graphs Illustrating Influence of Hydrogen Content and Engine Operating Conditions on Soot Emissions for
113
F100 Combustor. Comparison of Measured and Predicted Values of Pattern Factor for J79-17A Combustor.
120
73.
Comparison of Measured and Predicted Values of Pattern Factor for J79-17C Combustor.
121
74.
Comparison of Measured and Predicted Values of Pattern Factor for TF41 Combusi~or.
122
S72.
-x-
"LIST OF TABLES PAGE
TABLE 1.
Test Fuel Chemical and Physical Properties.
"2.
Values of A and B Employed in Equations (16) and (17).
3.
Values of C3 and C4 Employed in Equation
*0 "t.~
xi Z
4 43
(34).
i'4
SECTION I INTRODUCTION For the gas turbine and, heat engines,
in fact, for most other forms of
the most important fuel issues of today are t iose
of cost and availability.
The measures now being taken to ensure
future stipplies of fuels for gas turbines, forms of fuel conservation,
in addition to various
include the exploitation of alterna-
tive fuel sources and the acceptance of a broader specification for aviation fuels.
These developments highlight the need for
prediction techniques that will allow the impact of any change in fuel specification on hardware durability and combustion performance to be estimated accurately in the combustor design stage. "Unfortunately, the effect of a change in fuel properties is not constant for all combustors but varies between one combustor and another,
due to differences
ences in design.
in operating conditions and differ-
An additional complicating factor is
that the
various properties and characteristics of petroleum fuels are so closely interrelated that it
is virtually impossible to change
any one property without affecting many others. *.
However,
there
are several mitigating factors that help to ease the situation. One is that atomization quality is
influenced only by the physi-
cal properties of the fuel; namely, viscosity and surface tension, both of which are easily measured by standard laboratory techniques.
Evaporation rates are also closely linked to the
physical properties of the fuel, for example, vides a useful indication of fuel volatility.
ý-1-
fuel density pro-
Further
simplifications are possible because chemical reac-
tion rates vary only slightly among the various hydrocarbon fuels of
interest for the aircraft
gas turbine.
This is
these fuels exhibit only slight differences temperature, zone,
all
the fuels are largely pyzolyzed to methane, and hydrogen.
the reaction zone is
parent fuel.
Thus,
provided the discussion is
ignition performance,
efficiency,
Hence,
other 1-2
the gas composi-
substantially independent of the
anticipated range of aircraft fuels, in
adiabatic flame
and also because before entering the true reaction
carbon atom hydrocarbons, tion in
in
partly because
restricted
to the
any differences that occur
lean-blowout limits,
and combustion
will be caused mainly by differences in the physicaj
properties of the fuel insofar as they control the quality of atomization and the ensuing rate of evaporation. During the past decade,
the U.S.
Air Force,
Army,
along with NASA and the major engine manufacturers,
and Navy,
have ini-
tiated a number of programs to determine the effects of anticipated future fuels on existing engines. studies (1-6]
As a result of these
a substantial body of data has become available
that yields useful insights into fuel property effects on combustion performance. In addition to a considerable body of evidence on the effects of fuel property variations on the combustion performance and durability characteristics of the combustors investigated, references 1 thru 6 also contain detailed information on all the relevant chemical and physical properties of the fuels employed. -2-
Thene fuels were supplied by the U.S.
*
system evaluation. of the JP4,
Air Force for combustion
They include normal JP4 and JP8,
five blends of the JP8 and,
diesel fuel.
five blends
in some cases,
a No.
The blends were intended to achieve three oiffei'ent
levels of hydrogen content;
i.e.
12,
13,
and 14 percent by maas.
The key chemical and physic.al properties of these fuels are listed in Table 1.
Additional information on the distillation
characteristics of the test fuels is contained in Fig.
1.
"A major drawback to the data contained in references 1 thru 6 is that they include very little ,:haracteristics; drop size (SMD) ligation.
in particular,
information on fuel spray no measurements were made of mean
for any of the combustors employed in the inves-
In the absence of actual measured values of SMD,
pre-
v'ious analytical studies of those data [7-9] had to rely on values of SMD as calculated from standard equations for the mean drop sizes produced by pressure-swirl and airblast atomizers (10].
A main objective of the present investigation was to
remedy this deficiency by measuring the drop sizes produced by all the fuel nozzles employed in references 1 thru 6,
simulating
as far as possible the actual engine conditions of primary-zone gas density and fuel flow rate. While making these measurements, equipment problems prevented the acquisition of drop-size data *
for all the fuel nozzles of interest.
However,
:jufficient meas-
urements were made on several different types of fuel nozzles to provide the input needed to validate the analytically-derived ,quations
for the correlation and prediction of experimental data
-3--
3---
LLlw
1..
2
en,
0
L c
0%
cc
en
(JO
r" l
U
N 43
C4 N
m w'.
%0 '0i
V
C-1
')
=
al
0
.J
cJ Q,O+J
4)*~
C. 0
C) 0J
.4
W
+j.
0co+r
-4
0%
a--
) c
n
P
%
V- 9c 4
LA
0%
Ln vN
Ik)
Ln
NI
Ni
0
m
NN
t
(y)
5-
Pl
l
cIto o
4-)
V-4
4w
0-4'-
to.. 4A I--
V4.N04
LAI
.o
V
-4
O
L
C' 0)
4-' r- )
P" Go
0
fl
0
0-4-4
m
0b.
--
LC) 00
C
0%
C'
D
C
0)t a)
CO
v-I% ,
'-4
'.4 LO4
v-
-
O
CD C
0- >A N:) ci
N
r_.
0t 4)
0c N- 4
w
U 0'
0
00 CO
p-
0*
Ch
co 0k
WD
0
CD
C')
CD) -4)
m
LA
O%
r-0%
01
'
v4
CD
N
0 CD)
(.
0 N
0
C
CD
CD
'-
C.:
Ca)
N
.0
N= C
ND r_
0) N
0)
CD)
) CD
0_ V w'
0>?
i
n
CD-
N)
No
'
P
).
LL) j4..) etc
CC)C U-
kr-4cc
CO
004
co
00
a-
I0
CL
a.4.
v
"D
7
l-.)
0I17)
00
0d
0-
Ct
C
0-
07)
0-
v-.
CI
co
m
)(CO -3C
0) a-
17
LV)
0
0
-4CO
v.4-
O
d,
U)
r.)-4
t.
OD
Ch
CD
r4OY r-4 r-f-4
'-
c(N
CV)
mt
Ln
1.
r-.
CO0~ Ch-C 0
C')
m'
800
700JP 8 Blends
U;
No. 2Dise
w -Base JP 8
2
-Bass JP 4
S40^
JP 8 Blends 300 •-JP 4 Blends
2 DII
0
Figure l.
20
I
I
I
40 60 80 PERCENT RECOVERED
I
I00
Distillation Characteristics of Test Fuels (Ref.
-5-
2).
on all the key aspects of gas turbine combustion performance. The method used to measure spray characteristics results obtained are described ,*
the next section.
In subse-
quent sections the main combustor performance parameters are discussed in
turn,
reference 7. ,
in
and the
In
methods employed
following the style of presentation employed each case, in
a brief outline is
in
given of the
identifying the basic relationships between
the relevant fuel properties and each individual aspect of per-
"formance.
As liner wall temperatures and the emissions of oxides
of nitrogen are not materially affected by spray characteristics the findings 7,
in regard to these remain unchanged from reference
and are included herein for completeness
in a suitably reduced
form.
For each performance parameter,
the general approach has
been either to enhance existing correlations or to replace them with new correlations that are based on a firmer scientific footing.
It
is
hoped that the relationships developed in this pro-
gram will make a useful contribution to future combustor designs.
-6 ,S••. , . . , , . . , .
. . - . : . . . . . . •. . . . . .
. . . . , .. . . .. - - .
. - . ,- . - -
- -
.
.
I1
SECTION
FUEL ATOMIZATION The quality of the experimental data contained 1 thru 6 is
generally high,
in
references
Although the main liner dimensions
and a..rflow distribution are not always precisely defined,
it is
usually possible to deduce these parameters to an acceptable level of accuracy.
Reliable
area of fuel atomization. made to overcome this
information is
In
lacking only in
a previous study
the
[7] an attempt was
deficiency by calculating SMD values using
one of the following two expressions
(8]:
For airblast atomizers
SMD .
r
1 +
A
rnil~ i0.1 GoF 0.33 IPAUwDp
0.6
01r .
+ 0.068
[j
2 U •F2
0.5 j(l)
F'or pressure swirl atomizers SMD~~ .
0
0.25 p0.25 *0.75 1
P
vF
F
~-0.5 p
-0.25(2
P(2)
These equations take full account of variations erties (OF,
@F'
vF'
MF), air properties
geometry (D
and D Dh) p For example, Eq. (2),
However,
(pA and UA),
in fuel propand atomizer
they do have certain defects.
and all other published SMD equations for
pressure swirl nozzles,
are based almost entirely on measurenents
carried out in quiescent air at normal atmospheric pressure ind temperature.
For airblast atomizers the prefilmer diameter,
and the hydraulic mean diameter of the air discharge orifice, -- 7
--
D Dn,
in some cases,
.are often difficult to measure and, define.
difticult to
Usually they can only be established for any given atom-
izer by carrying out measurements of SMD at some convenient test condition.
After
(1) it
inserting these values into Eq.
can then
be used to predict values of mean dxop size at other operating conditions. The lack of measured SMD values in
references
rendered more serious by the fact that in tions,
Thus,
lean lightoff and
the mean drop size appears as SMI) squared.
the magnitude of any errors
effectively squared.
irs
many performance equa-
for example the equattons for predicting
lean blowout limits,
I thru
in
the estimation of SMD are
In a previous study (7] these errori were
minimized by replacing the absolute values of SMD in
thes.
equa-
tions with values expressed relative to the drop sizes ob.ained with the baseline fuel,
JP4.
This helped to compensate for the
lack of information on nozzle characteristic
dimensions,
but it
also diminished the practical utility
of the resulting equations.
In order to remedy this deficiency it
was decided at the outset
of the present study to measure the drop sizes produced by all the fuel nozzles described in references 1 thru 6.
The apparatus
employed for drop-size measurement is shown schematically in Fig. 2.
The main component is a cylindrical pressure vessel which is
mounted on a stand with its L20 cm long and 75 cm in
axis
in
diameter.
the vertical position.
It
The atomizer
is
under test
is
located centrally at the top of the cylinder and sprays downward into the vessel which
is
pressurized to the desired level using
-8-
0ii
. %•. %.'•."%*• •, %-,•.•Q• • %-•.-
•. •.•
.. •- .•.•-
.%............................................-....-.....-.....".......,.-........-
_-
•:
•. •
..
.
-' -
•,r
,r
w.r"Nw,
'•
'r
'-r9
q
-
r•r
. .•
-
- ..
Light
r•
-• --
Paid
Observation Window
Pressure
Lih..
at
------
,.Pressurizing
Figure 2.
Schematic Diagram of Spray Te~st Rig.
-9
-i
-i
-.
'I
¥• I; •
•
'L
.gaseous nitrogen that. is tapped from a large liquid nitrogen storage/evaporator of air
system.
The reason for using nitrogen instead
to avoid the risk of explosion at high pressures.
is
air the results
to those of
of nitrogen are very similar
"the phy;;ical properties
obtained with nitrogen are considered
tate into a collection tank at the bottom of the chamber,
i•s
for
valid
The droplets produced by atomization gravi-
systems using air.
whence the fuel is
As
from
The objective
returned to the storage tank.
to conserve fuel and to avoid potential pollution of the atmos-
phere
due to escaping fuel droplets. In addition to the nitrogen supplies for atomization and
tank pressurization two extra nitrogen lines are connected to the tank.
One line is
used to protect the windows from any contami-
nation by fuel drops or mist, while the other line is
connected
to a manifold located at the top of the tank which provides a
By
gentle downdraft of nitrogen through a large number of holes. kept to a
this mearis the problem of droplet recirculation is mi n imum.
Drop sizes were measured using the light-scattering techproposed by Dobbins,
nique first
,
is based on a direct meas-
It
later developed at Cranfield [121. light
and Glassman [11J and
Crocco,
a mono-
after
profile
intensity
.
,urement of the scattered
*
chromatic l..ght beam has passed through the spray.
The SMD is
obtained directly from measurement of intensity versus radius in
-
S".%'
-
Z.
10
.
this is
In practice,
the focal plane of the receiving lens.
-
.
.
.
- .
.. .
..-.
,
..-
.•.-
••
,
, . .
*•
air-mplished
by measuring the traverse distance (r)
optical
and a point on the profile
axis
sity
is
equal to one-tenth
scattered
profile.
*•
which the light
of the normalized
intensity
inten-
in the
The SMD of the spray can then be determined
using the relationship and Webb
at
between the
between r
and SM)
as derived by Roberts
(131.
When using the light-scattering
technique or,
in
fact,
eother optical technique for drop-size measurement,
it
is
any
impor-
tant not to attempt measurements of mean drop size too close to the nozzle.
This is because although all the drops leave the
nozzle with approximately the same velocity, the smaller drcps tend to lose momentum faster than the larger drops, resistance,
which leads to over-representation
in the sampling volume.
due to tir
of the fine Irops
Further away from the nozzle,
where al
thf drops are moving at roughly the same velocity as the downdraft of nitrogen,
the measurements indicate larger values of SMD
which are more representative of the actual spray. is
However,
it
equally important not to attempt to measure drop sizes too far
downstream of the nozzle as this could introduce errors due to fuel spray evaporation.
Calculations
indicate an ideal distance
of .15 cm for the conditions of the present experiments,
and this
is the value actually used. Due to the considerable time and effort that would be required to make detailed measurements of spray characteristics for all nozzles and all fuels,
it
measurements using one fuel only, -
ii-
was decided to conduct all and then to use these measured
values to estimate the corresponding mean drop sizes for all other fuels.
The fuel selected for detailed study was aviation
kerosine (Jet A),
a
-
which has the following physical properties.
a
0.02767 kg/s 2 ,
p - 784 kg/m 3
- 0.00129 kg/ms,
As fuel density has only a very slight effect on atomization quality, consideration need be given only to surface tension and viscosity.
(2)
For pressure atomizers Eq.
suggests that SD is
proportional to A0.25 , but some preliminary measurements of SkD carried out out on JP4 and DF2 fuels (representing the two extremes of viscosity) indicated a slightly lower viscosity dependence so that for pressure atomizers we have sMD co
For airblast atomizers,
0 .2 5
a0.
20
(3)
which are characterized by a
slightly higher dependence on surface tension and a lower dependence on viscosity [10],
is found that changes in SMD arising
it
ftom varietions in fuel type can be expressed tc a sufficient degree of accuracy oy" the relationship SMD - o Thus,
0
.3 5 if
for any given atomizer,
available for one fuel, then Eqs.
(4)
0.05
measured values of S14D are
(3) and (4)
allow mean drop
sizes to be calculated for any other fuel, provided of course its physical properties of surface tension and viscosity are known. For the fuels of interest to the present study, mean drop sizes for all operating conditions of fuel flow rate and ambient air -
P
01
12
-
density were obtained using the measured values of SMD for Jet A fuel,
in conjunction with one of the following two expressions. For pressure-swirl atomizers SMD .SMDl Jet A (aF/
*SMF
Jet A)
0.25
0.20 (•F/'Jet A)
(5)
Jet A 0.05
(6)
For airblast atomizers SMD F
SMD Jet A
A 0.35 (F
FJet
The SMD data obtained for the J79-17A, J85 and FI00 fuel nozzles, Figs. 4i
3 thru 10.
J79-17C,
using Jet A fuel,
F01, TF39,
are shown plotted in
Due to the difficulties encountered in the
procurement of an F101 fuel nozzle of the type employed in the it
F101 combustion program [2],
was decided to substitute a more
recent version for the atomization tests.
As these two types of
nozzles differ mainly in regard to fuel distribution characteristics rather than atomization quality,
it
is
believed that no
significant error was introduced by this substitution. equipment problems and time restraints,
Due to
no results were obtained
Thus for these nozzles,
for the TF41 and TF33 fuel nozzles.
SMD
values were calculated using Ea. '2,. The SMD 'nLa
contained in Figs.
3 thru 10 are presented
maLnly as plots of SMD versus fuel flow rate, miF' values of ambient air density,
PA'
for variou'-
but for the FI00 nozzle tie
SMD values are plotted against AFR in order to demonstrate tie effects of air/fuel ratio and liner pressure drop on mean drop size.
Not surprisingly,
Fig. -
10 indicates that atomization 13
-
*1,
100 J 79-17A Parker Hannifin Pressure Nozzle No. 1345-654010
S90
Fuel = Jet A TA=288 K
80
S80-
3 f:A' kg/rn
E70
1.22 3.66
.
60
8.54
50
40-
I0
Figure 3.
rh,
/
20
30
Mean Drop Sizes obtained for J79-17A Fuel Nozzle.
-
14
90 Z79 -17C 80
Parker HannIfin Hybrid Nozzle No. 670012 M2
70-
Fuel aJet A TA 3 0 0 K
kg/rn3
60-.A 4•-•
•
1.22
;i:::50,.'-
3.66
; .c
-.
6.10
40 08.54
300
20 "S10
,.
.II
10
0
"•F'
I
I
20
30
g/s
VF
"Figure 4.
Mean Diop Sizes obtained for J79-17C Fuel Nozzle.
-15 15
-S
-- - - - -
- -"
.
N.
"150
FIOI Parker Hannifin Pressure Noz:.re No. 18139 Fuel = Jet A TA= 2 8 8 K 100 -
'.3 *
0(
PA' kg/m
50 ---
3
6.10 1.22
0o
0
I0
20 F
Figure 5.
...
, ,I
I
30
g/s
Mean Drop Sizes obtained for F1OI Fuel Nozzle.
-16-
0
90 80(
Parker Harmif In Pressure Nozzle No. 468710 Fuel= Jet A TA= 2 8 8 K
70
[ 60 50
40
8,
30
201 0
Figure 6.
10
rhF, g/9
20
30
Mean Drop Sizes obtained for TF39 Fuel Nozzle.
-17
V.
150
.Doga (latin)
"'
q
~_g
reure
$IN H!740
zzle
I00"
508
50PA 101 k Pe 11 atm) TA=
288
K
1.225 kg/rm3
0
10
20 rh F
Figure 7.
g/s
Mean Drop Sizes obtained for J85 Fuel Nozzle.
F!
-
•
30
..
b*t,-
18-
150
-
D21ovan Proemure Nozzle 2 §/N U1740
100
50 PA kg/m3
1.225
3.67
6.12
10
0
20
30
tF, g/
Figure 8.
Mean Drop Sizes obtained for J85 Fuel Nozzle.
-
19
-
%" • ' '• "'• " %"• ' ." ' . ' ' .%° .•. . . ,N 'l 1 •• '.=% -•' o=,-,. ••",•rm I! ,• , - ' • •. '=' m
90
FIOO 80 (
Ex-Cell-0 Airblast Nozzle No. 21178 Fuel= Jet A
•~A
TA= 2 8 8 K
70o
PL/PA =2.5 %I
"60
E •50 2'A, kg/rn 3
(I,
.40-
1.22 3.66
0-
6.10
0
2
4
6
8
10
14
12
ATOMIZE R AFR
Figure 9.
Influence of Ambient Air Density and Atomizer Air/Fuel Ratio on SMD for F100 Fuel Nozzle.
-20-
4.,W
S
I
V.V
-V
r
"•
'
90-
"80
-F 100 Ex-Cell-O Airblast Nozzle No. 21178 Fuel= Jet A PA=I01 kPa (latin)
70-
~60-
500
L.5 2.5 "40-
3.0
30-
200
2
3
4
ATOMIZER
Figure 10.
5
6
7
AFR
Influence of Atomizer Air/Fuel Ratio and Pressure Drop on SMD for F300 Fuel Nozzle.
-
21
qiality
improves with increase
reults
obtained with a value of APL/PA of 2.5 percent were
selected for use in
in
this study,
liner pressure drop.
as this
represent the liner pressure drop in The variations curves drawn in dual-orifice quality level,
in
Figs.
is
The
considered to best
the dome region.
SMD with fuel flow rate exhibited by the 3,
atomizers.
5,
7 and 8 are characteristic of
,
Thus
it
improves with increase
is in
observed that atomization
fuel flow rate up to a certain
beyond which SMD values start
to rise
again.
The point of
minimum SMD coincides with the opening of the pressur..zing valve which admits fuel
into the main nozzle.
nozzle at relatively low pressure its poor.
With further
increase
in
atomizer
it
is
fuel flow the main fuel pressure
seen (Figs.
to improve.
in
Fig.
Inctease
4,
in AFR.
For the
9 and 10) that atomization
quality improves continuously with decrease i.e. with
fuel enters the
atomization quality is
increases and atomization quality starts ajrblast
As this
in
fuel flow rate,
For the hybrid nozzle,
as illustrated
the characteristic shapes of the SMD curves lie
some-
where between those of the pressure nozzle and the pure airblast atomizer, rate,
so that SMD remains sensibly independent of fuel flow
at, least over
the range of fuel flows tested.
The steep temperature
rise that accompanies combustion in
the primary zone causes a reduction offsets the
increase
in
in
density experienced by the air during its
passage through the compressor.
In consequence,
settings where atomization quality is
-
......
..•
.
.
.
.... ....
gas density that largely
22
at low powir
most limiting to combustion
-
.
*..
..
performance, sprayed
is
the density of the gas
into which the fuel
is
roughly the same as that of air at normal atmospheric
pressure and temperature.
For the results contained in Figs.
3
thru 10 the variation in ambient air density was obtained by changing air pressure while maintaining the air temperature constant at around 15 0 C.
Inspection of Figs.
3 thru 1.0 reveals that
atomization quality is generally improved by increases in ambient air density,
except for the F101 nozzle which exhibLts a slight
deterioration
in atomization quality with increase in pA'
With appropriate interpolations, Figs.
the results contained in
3 thru 10 can be used to establish formulae based on abso-
lute values of mean drop size for the prediction of combustion efficiency, factor,
lean blowout limits,
and pollutant emissions,
and smoke. discussed in
pattern
including unburned hydrocarbons
These various aspects of combustion performance are the following sections.
-
6V
lean lightoff limits,
23 -
III
SECTION
COMBUSTION EFF ICIENCY The separate effect•s
fuel-air mixing,
on combustion of fuel evaporation,
and chemical reaction rates,
described elsewhere
(7,9].
have been fully
For the aircraft gas turbine the main
factors affecting the level of combustion efficiency are evaporation rates and chemical reaction rates.
Mixing rates tend to be
limiting to performance only at operating conditions where the livel of combustion efficiency is deficiencies
*
so close to 100 percent that
in performance due to inadequate mixing are diffi-
cult to discern. Three separate ranges of operating conditions may be defined,
one in which combustion efficiency is governed solely by
reaction rates,
another in which combustion inefficiency
entirely to low evaporation rates,
is due
and a third region in which
the level of combustion efficiency depends on both reaction rates and evaporation rates.
For all three regions the combustion
efficiency is obtained as the product of the reaction :.ate efti"''
ciency,
71,
and the evaporation efficiency,
""•c"7 "c
7C ' c0
nc
(7)
X 7ce
c
"The second term on the right hand side of Eq.
O
i.e.
(7)
represents
the fraction of the fuel that is vaporized within the combu3tion zone. *
For
7ce
1,
c
77 Wc
and Eq.
reverts to the 0 param-
eter which denotes the fraction of fuel vapor that is
-
0,
(7)
25
-
converted
into combustion products by chemical reaction. From analysis of the available experimental data on combuswere
and 71
the following expressions for V
tion efficiency, derived (7].
•c
3
0.022 P•.
77exp
Vc exp
3
In Eq.
exp
-Do
V Vc
Co2
the temperature dependence
(8)
is
Xeff (9)
expressed in terms
the adiabatic flame temperature
of T , which is
in the combustion
assuming complete combustion of the fuel.
zone,
(8)
c
36 x 106 P 1
and re
(Tc/400)
f cmA cA
It
is
calculated
from the expression Tc
-
+ AT
T
c 3
(10)
c
where ATc is obtained from standard temperature rise charts for the fuel
in question, using appropriate values of P3 P T 3 and qc
(qov /f C). (8)
Equations
and
(9)
relate combustion efficiency to combustor operating conditions
combustor dimensions
(via Vc),
(via mA' P 3 and T c).
fuel nozzle characteristics
fuel type
(via D ) and
(via Xeff) 11, 12 and 13. at three levels of
Values of Xeff are shown plotted in Figs. These figures contain plots of Xe versus T
eff vbn 26 -
-
4•! d
.
-
• ,
• -
•% '. ''.' .
•'
. - . • m- . - . - - -
''-
-'
:•
•'
'-
. . . . , .
-
.
,
. , . , 0i . . . . . .
,'
.
- .,.'.
• " ...-
.
,
,
°
"
-
,
%
. " .
-
%
%
"
",
.
• %
.
'-
"
,
, "
,
', • '
" -,"''
. .
. "
".
."
m "
,
:
• •
'
'''•
.
:',
%" "
"•
1.4I
*P
1~00 kPa UDO
1.21.0-
50
0.68-200
1000 500
1.00.8E 0.6 -
10000 --- 000 zzz
2000
~0.4-_
_
_
_
_
S200
T :1200 K
0.2
1000
100 0
0.240.200.160.12-
10000o
*
2000
0.08-
500
0.4420
Figure 11.
440
460
480 500 Tbn , K
520
540
560
Variation of Effect4Ne Evaporation Constant with Normal Boiling Point at a Pressure of 100 kPa. -27-
I2.2
P 1000 kPa
UD0 ,-
10000 5000
1.4~2000
-
TOD= 2000
0.
1000 0
K<
10000
2000
0.6 1<
1000
500 200 100
0.4-
T -120 K00
D
0200
10000
0.10-
*
5000 ~2000 1000
-
0
0.2420
Figure 12.
6
440
460
480 500 Tbnt K T
520
54
560
Variation of Effective Evaporation Constant with Normal Boiling Point at a Pressure of 1000 kPa. -28-
UD., (M/s)(fk)
P-2000 kPO
2.
10000
~4..2.4
______________________
50
2.0
5000
0
1.6
2000 00
1.2
50000
.4
in
1.4 50C-
200 5000 __________________2000
0.6
0.2
__
_
_
_
__
_
_
_
_
100__
TrOD z 2_00K
0.22
0.18-
10000 2000 5000
500
20
0.02 4040
Figure 13.
60480 TK Tbn
500
50
540
560
Evaporation Constan of Effective point at a Pressureo variation 13oilifg with Normal 2000 kPa. -29-
-4
pressure,
namely 100,
ambient temperature,
1000 and 2000 kPa, namely 500,
and three levels of
1200 and 2000K.
For each value
of temperature several lines are drawn to represent different values of UD 0 , where U is the relative velocity between the fuel drop and the surrounding gas, *
and D
0
is the initial drop diame-
ter. From a practical standpoint the concept of xeff has considerable advantages since it
takes into account the reduced rate of
evaporation that occurs during the initial droplet heat-up period,
as well as the enhancement of fuel evaporation rates due
to the effects of forced convection the type shown in Figs.
(14].
Thus plots of Xeff of
11 thru 13 greatly simplify calculations
on rates of spray evaporation and drop lifetimes. The very satisfactory correlation of combustion efficiency data provided by Eq.
(7)
is demonstrated in Figs.
14 thru 21,
which include all the relevant data on combustion efficiency contained in references 1 thru 6.
-
7
30
-
4.
4.-..
__
2_-i
4
~
4'
J~{
100
0
Data from Table A-I (I]
98-
0 Test Point 2
96
6[ Test Point 10 All Fuels
A Test Point 30
0%
S
00 ... 94.-0I
92 J79-17A P 3 =256 kPa
g-90
88 8 -
T =421 K rhA= 1.53 kg/s
S86
84""-'86
88
90
92
94
96
98
100
Combustion Efficlency (predicted), %
Figure 14.
Comparison of Measured and Predicted Values of Combustion Efficiency for J79-17A Combuetor.
-
31
-
V. 0, .- ..•..,,% ,- ,,-. ,..
.
-.
..
-,•.,.,..-
•.,,,
-,,,
.-.-
..
-.
,.
.,
•. .... ?,
,.
..
.
.
.,.,.%
.,
.:.
•
•, •,,"-.
-. -,
'
I
Data from Table B-I (41 oTest Point 2
98 g
i4 &Test Point 3
All Fuels
2 96
00a 00
~92
94
3
w9
J79-IYC P=254 kPa
~88
T3 x421K -r86 L53 kg/s 884
86
88
90
92
94
96
98
100
Combustion Efficilency (predicted), %
Figure 15.
Comparison of Measured arnd Predicted Values of Comnbustion Efficiency for J79-17C Com~bustor.
-32
-
c
00
98
&lTest Point I Test Point 2
""0
0
A
N94All
00
A
Data from Table A-1[2] A
96
Fuels
9492-
P3 394 kPa
8EM
T 3 =466 K
.~
"o868
A78 .4 7 -9 .8 kg/s
o8
84886
8890
929496
98100
Combustion Efficiency (predicted), %
..
Figure 16.
Comparison of Measured and Predicted Values of Combustion Efficiency for FIOI Combustor.
-33-
100
::<
All Fuels
98
, 0
S96-
:•94-
0 0
00•.
00
~92-
TF 41
9010
T 3 '429 K rmA= 0 .9 6 2 kg/s
Be E
0
868 -88
90
92
94
96
100
98
Combustion efficiency (predicted), %
Figure 17.
Comparison of Measured and Predicted Values of Combustion Efficiency for TF41 Combustor.
>5 434
'.'
-
34
-
t ••.-
",,"', "pi"••'",,•' J,• •
I
,
". "
"*,l',,. ";,-•,h•
•',,,",. ",,..',
, "
,,"
•• •'•" '
'
•
..
•
• •.
.
.
.-.
.
,
•-.
..
,,-.*
.
..
100 •
Test Point I Table D-1[5]
,Test Point 2 Fuels: IB, 8B,19B,913B,319148115B1 ~98-
*960 94
,.
•:
0
50
o
I
92 f0'
929I94
96
TF :39 98
100
Combustion Efficiency (predicted), %
Figure 18.
Comparison of Measured and Predicted Values of Combustion Efficiency for TF39 Combustor.
-
35-
Data from Table C-1 (5]
10
,&Test Point I ge 98.
0 Test Point 2 VTest Point 3 E3 Test Point 4
V
Test Point 6 All Fuels
0
0o
0
A0
0 A6
S96-
0
-94-
J85 148-445 kKPa P T33 ==336-570
S9292-
rhAu 4 .4 -12.2 kg/s
90!
92
94
96
98
100
Combustion Efficiency (predicted), %
U
Figure 19.
Comparison of Measured and Predicted Values of Combustion Efficiency for J85 Combustor.
-36i,.
90 0 Natural Gas 0 Liquid Fuels
OF 0 33[
I80-
~70-
60-
05'0
0 P =207 kPa T a361 K .hA 2 0 kg7A
o
60
70
80
90
Combustion efficiency (predicted), %
Figure
20.
Comparison of Meabured and Predicted Values of Combustion Efficiency for TF33 Combustor.
-
37
-
*
100 st I98-
All Fuels
04
96-
0
0
0)
E 94 S9210,
90-
_
P 3 3=70-460 kPa T3=488 K hr-A 2. 8 kg/s
868484
86
90
88
92
94
96
98
100
Combustion Efficiency (predicted), %
*
-Figure
21.
Comparison of measured and Predicted Values of Combustion Efficiency for F100 Combustor.
-
38
-
SECTION IV LEAN BLOWOUT The problem of lean blowout has not loomed very large in the past,
due mainly to the widesptead use of pressure swirl atomiz-
ers.
The poor mixing characteristics of these atomizers allow
combustion to occur at mixture strengths that are well below the norhal, weak limit of flammability.
In fact,
of around 1000 air/fuel ratio (AFR),
lean blowout limits
based on overall combustor
values of air and fuel flow rates, used to be quite commonplace. In recent years the continuing trend toward improved primary-zone fuel-air mixing for the reduction of pollutant emissions and flame radiation has led to a narrowing of stability limits and to increasing concern over the attainment of satisfactory lean blowout performance. For homogeneous fuel-air mixtures,
flame blowout occurs when
the rate of heat liberation in the primary zone becomes insufficient to heat the incoming fresh mixture up to the required reaction temperature.
The lean blowout fuel/air ratio depends on the
inlet air velocity, pressure, the primary zone.
and temperature,
and on the size of
The relationship is of the form (15) x q
Equation
j'"rVip
exp (T 3 /b)
(11) may also be use& to predict the lean blowout
limits of combustion chambers supplied with heterogeneous
air mixtures,
fuel-
provided that the rate of fuel evaporation is -
39
-
•,•
• •
•
•
%', vw
-,r' w'rrrw
,rrrrrr rwr. -
.
- °
c
•
-
w=
'. ,
n'
. r-
r
r
y r
xr r r'rr'
.rr
.
••z.
. '.•
sufficiently hLgh to ensure that all the fuel is fully vaporized within the primary combustion zone. vaporize,
then clearly the "effective" fuel/air ratio will be
lower than the nominal value. that is
vaporized is
combined with Eq. blowout,
known,
However,
if
the fraction of fuel
or can be calculated,
it
can then be
(11) to yield the fuel/air ratio at lean
i.e.,
qLB
"LOx (heterogeneous)
where ff
the fuel does not fully
If
ffl
qU(homogeneous)
(12)
f
is the fraction of fuel that is vaporized within the
primary combustion zone.
*
From analysis of the factors governing the rate of evaporation of a fuel spray [14], it was found that D 2(13) 10 P Vf X fff =106 Pg pz eff/fpz "A 0 It
should be noted that Eq.
When this occurs it
(13)
allows ff to exceed unity.
simply means that ý_he time available for fuel
evaporation exceeds the time required,
so that the fuel is
vaporized within the recirculation zone. I
In these circumstances
should be assigned a value of unity. Substitution of qLBO(hom)
into Eq.
(12)
from Eq.
(11) and ff from Eq. (13)
leads to
9 pI qVo(VL+x)4 pz33
Mp_0m((+x) exp(T3/b)
(th)
-
I.
fully
40
-
1
eff LCV
(14)
,
term on the right hand side of Equation
The first
"* 14
is a function of combustor design.
The second term
represents the combustor operating conditions. *
The third term
embodies the relevant fuel-dependent properties, lower calorific value of the fuel.
including the
This property is
incluted
because lean blowout occurs at roughly the same temperature for all fuels,
so that fuels having a higher heat content are capable
of burning at lower mixture strengths (10]. Analysis of the experimental data for all engines indicates opti~num values for b, n,
and x of 300,
Inse-tion of these values into Eq.
P.3
q-Ko
where A'
is
(14)
0.3 and 0,
respectively.
gives
emp(T 3/300)1]15)
a constant whose value depends on the geometry and
mixing characteristics the value of A'
of the combustion zone.
at any convenient test
Having determined
condition,
then be used to predict the lean blowout fuel/air
Eq. ratio
(15)
may
at any
other operating condition. A difficulty that arises with Eq.
(15)
is that of assigning
appropriate values to Vpz, since for many combustors the primary-zone volume is not clearly defined. To surmount this problem it
into Eq.
was decided to substitute the pre-dilution zone,
Vct
instead of Vz. This may be justified on the pz grounds that V is easier to define and measure; also, values of
V
(15),
have already been used in the correlation of combustion -
41
-
""' .[.+.'++''"' .'', ' ' '•<,". ."" ' •.":• ++[ ++'% ...' ."-+L" ' ' ' """'- "+-
efficiency data.
as the ratio of primary-zone
Furthermore,
volume to pre-dilution volume tends to be fairly constant for using Vc instead of V
most conventional combustion chambers,
Wit
has the virtue of consistency without loss of accuracy. this modification Eq.
(15)
becomes 2
qL
mA[
Vc
-
1
1.3 expA(T /30DoefU0atT77j5Kg/kg(1 6)
J
The term (D at TF) 2/(D at 277.5)2 is
introduced into the above
equation to take into account the variation in drop size arising from a change in fuel temperature from the initial value, which is taken as 277.5K.
baseline
For lean blowout limits,
should be evaluated at an air temperature of 1400K, approximates the weak extinction temperature for all
X
since thLs fuels.
For each comfustor a value of A was chosen for insertion into Eq. data.
(16)
that would provide the best fit
to the experimental
These values of A are given in Table 2.
be advantageous values of A,
It
would clearly
if similar types of primary zones yielded similar
since this would facilitate the prediction of lean
blowout limits for new combustor designs.
Although the variation
in the values of A listed in Table 2 virtually prohibits such extrapolation,
it
should be borne in mind that these values
embody all the errors incurred in the estimates of combustion volume and the fraction of air involved in primary combustion, well as in the measurements of mean drop size. with f
the deviation is
reduced,
as
By combining A
as illustrated in Table 2.
pz -
-42
.
.. .
.
. .. . .
.
.. .
.
.
.
.
e -a .
..
~.
.
.
.
.
.
Table 2.
Values of A and B employed in equations (16) Engine
A
Afpz
B
Bfpz
J79-17A
0.95
0.22
0.477
0.109
J79-17C
0.70
0.22
0.544
0.'03
F101
0.54
0.22
0.700
0.287
TF39
0.60
0.18
0.360
0.108
J85
1.00
0.30
0.335
0.104
Floo
0.45
0.16
0.508
0.178
-
43
-
and (17).
As discussed
in reference 7,
the initial value of Af
culated for the P101 combustor was exceptionally high,
cal-
pz and this
was attriluted to an error in recording either the fuel flow rate or air flow rate when testing on a 540 segment of an annular combustor.
Dividing the reported values of qLBO contained in
reference 2 by (360/54)
not only gives more sensible value•i
but also reduces Af
to 0.22, which is
of
fully consistent
with the results obtained for the other combustors. The correlations of lean blowout limits provided by Eq. (16),
using appropriate values of A, are illustrated in Figs.
thiu 29 for the J79-17A, conbustors,
respectively.
J79-17C,
F101,
TF39,
J85,
and FIOO
The close agreement exhibited between
the predicted and the measured values of lean blowout fuel/air ratio is
22
clearly very satisfactory.
-
44
-
12 Fuel
Symbol0
2
0
9-
6 9
x [
8.
10 13
0
10-
xo1
7-
oxx 0
6-
xx
x
00 J 79-17 A
_5Cr
0
Z
P3 IOO kPa rhAz3.18 kg/s
3-
T3z=238- 300 K
2 2-I
TF
2 3 8- 3 0 0
8
9
1
2
3
4
5
6
7
qLB0 (predicted),
Figure 22.
Ub•,
Ji.
II
12
g/kg
Comparison of Measured and Predicted Values of Lean Blowout for J79-17A Combustor.
-
-:45
10
K
45-
Fuel
Symbol
4
3,
o0
0-
-
12- 5 3
13
4
E 2
x
7
0
0 0.
A
8
,,x "~~ ~~8
0
2 iA I J79-17
13X, ,TF2830
Cr7
5
6
7
8 ~
Figure 23.
1
A= 3 .18 kg/s T3 a238- 300 K 238- 300K
S44
4
P.3 zI00 kPa
AA
6-
S•10-
~
Z•T3"258-000
F
',•
3aO
9
0
11
12
13
14
(prodicted), g/ kq
Comparison. of Measured and Predicted Values of Lean Blowout for J79-17A Combustor.
-46-
15
Fuel
Symbol
IA
A
2A
x
I0A 9
13
x 8
x
0
8A
X
9A
0'
obr7-
o
IA
0
Do
5-
0
"05
0-
"
J79-17C
3-
P.= I01 k Pa 1 A=0. 3 18 kg/s
2
x"238- 278 K
TF 238 -278 KL I
Figure 24.
2
3
4 5 6 7 8 qLBO (predicted), g/kg
9
10
II
Comparison of Measured and Predicted Values of Lean Blowout for J79-17C Combustor.
-
47
-
x
II
a
Fuel "10 4A 5A
""9-
"
Symbol V 0
V
03
x E3
6A 7A
1
rn 0
IIA
S8-
0
hAo
172A
co 5-__ •i
o-'J
xx) x •- •0
-jx
JJ79-17C P3= 101 kPa rhA=0.3i8 kg/s
332
T3= 238- 278 K x 23 8
STF 0
4 3
5
q
figure 25.
•"-=
-•-,
.
"
% •.
w•
"".
`•'
7
8
910
K It
(predicted), g/kg
Comparison of Measured and Predicted Values of Lean Blowout for J79-17C Combustor.
-
•
6
278
TJ"
••
,:•
:•....
•-
48-
"•
">
"
••
'
•,•,:
••i
r•
,"
".'_
40
Ix0 3,3
*
5 6
A
0
8
0
o ,(
9
02x
12
0
A X x
U)
.20-
A
F 101 .0
P3 =100 kPa rhA =1.15 Iq/s T3 239-320 K
10-
TFu240-305 K
0
10
30
40
qLBO (predicted), g/kg
Figure 26.
Comparison of Measured and Predicted Values of Lean Blowout for FlOl Combustor.
-
49 -
~r
-W -Tr~
--
,-
7
r
a-
-r4-- WvWW
Wr
Tx -
it10
8-
60
Fuel
Symbol
IB 8B 9B
A 0 0
14B 15B
V x
E3 V
[E
5-
4TF 39 P5 102.6 k Pa
2 I.
T 3 =237-278 K 't A= 4.0 8 kg/s
0
1
2
3
4
5
6
7
8
9
10
qLBO (predicted), g/kg
Figure 27.
Comparison of Measured and Predicted Values of Lean Blowout for TF39 Combustor.
-
50
-
~~~ ,,+~
r+ +'
~
.
~.•,..r• ~ ~ ~ ~.,+• k
,l ..
'I~'l
k25
Fuel
Symbol
ICA 20
13C
13
15C
0
15-
00
.
00
*iue2.
op
soNfMasrdad
rdcedVle
-
5
A
0
10
Figure 28. r
20=44 2
(predicted), g/kg
~LBO
I
A=
comparison of Measured and Predicted values of Lean Blowout for 385 Combustor.
-
51
-
25-
S.
*•i
8-
/
:" •''
?'
7" •
•
0
"JP4S
BLENDI BLEND2 BLEND4
a 0
~JP4Shale
O0
7 A /
0 0'
I
4
040 -J
0
2-
P 3 =27-65 kPa
7
T =245-295 K "iA 0
1
2
3
4
5
6
0 2 2 - 0 .6 1 kg/s
7
8
9
"qBO (predicted), g/kg
Figure 29. A
Comparison of Measured and Predicted Values of Lean Blowout fox F1IO.
5-2-
4,$-Y.
KK
,2•~K'~~,
10
SECTION V I ON I T I ON It
is
increases
well-established that ignition in made easier by In pressure,
impeded by increases tion performance
is
through the way in vapor in
in
temperature,
and spark energy,
and is
velocity and turbulence intensity.
Igni-
also markedly affected by fuel properties which they influence the concentration of fuel
the spark region.
These influences arise wainly from
the effect of volatility on evaporation rates, and also from the effects of surface tension and viscosity on mean fuel drop size. The amount of energy required for ignition is
very much larger
than the values normally associated with gaseous fuels at "stoichiometric fuel/air ratio.
Much of this extra energy is
absorbed in the evaporation of fuel drops, the actual amount depending on the distribution of fuel throughout the primary zone and on the quality of atomization. Application of the theoretical concepts developed in references 16 and 17 to the ignition data contained in references 1 thru 6 leads to the following equation for lean lightoff fuel/air ratio.
•
*
*
qLLO " B ["' fDz~fmA P"
This equation is
O*
a
2 Do atTF Kg/kg(17) fT";JI%oaz/ at2'bK "[
exp (T 3 /300)
virtually identical to Eq.
higher pressure dependence; namely P "minor difference
is that X is ef f -53
(16)
except for a
instead of Pi.
Another
3 3 evaluated at the combustor inlet
-
:
air temperature,
T 3,
The correlation of lightup data obtained wtth Eq. illustrated in Figs. -'
TF39,
J85,
30 thru 37 for the 379-17A,
and F100 combustors,
respectively.
(17)
J79-17C,
is
F101,
The level of
agreement between predicted and experimental values is considered satisfactory, especially in view of the well-known lack of conignition data.
sistency that usually characterizes
and Bfpz for all combustors are listed in Table 2.
1,•5
-
54
-
Values of B
25 Fuel
Sy.bo.
I 2
0 x
5 4
0 V
5 6
0 A
20-
1
3
ox
TI
I0-
x
x
5-
:'1
x
A
(0
0"A 0
0
Cr
0
OO
J79-17A 0P 3 = 100 kPo =A3.1 8 kg/s
5-A
0
5
10
~~~~~~~~~~~~~~~~~~~~~....• , .... .•....,.....o ...-. .. "
238- 300 K
TF
23
15
8- 3
0 0
K
20
25
(predited), g/kg
qLL0
Figure 30.
T3
Comparison of measured and Predicted Values of Lean Light Off for J79-17A Combustor.
•... ...
...-
-
55
-
..-
... ,
.-..
.
.
..
..
.
- .....
.•,-
. .. ,.
.
25
2
Fuel
Symbol
7
V
8 9
0 0
10 12 13
A
0 I2 X
V
x
x
0
05 1f
0(0
V
0 00V
0 10 -J
J79-17A P3 z 100 kPa rA= 3.18 kg/s
5-
T3 = 238-300 K TF a 2 3 8 - 300 K
00
10
5
15
LL
Figure 31.
(predicted), g/kg
Comparison of Measured and Predicted Values of Lean Light Off for J79-17A Combustor.
56
-
"4.
20
"..
J
-
25
Fuel
Symbol
IA(R)
20
2A 3A 4A 5A 6A
0 x '7 0 0
S15-
'7v
0 0A
10-
0F
5-
__ %0/C
/
l0
%
IT ,~
O0
5
32.
b
- --
I
T3238a 278 K TF238-278 K
115
%
S~Figure
J79-17C P3 aI01 kPa hA a 3.1IS kg/$
20
q L LO (predicted), g/kg
P 3 =11
Comparison of Measured and Predicted V.,lues of Loan Light Off foJ J79-17C Combusto1.
- 57 -
25
fuel
20
_Symbol
7A 8A 9A
I0A IIA 12A 13A M
0 0 A x v' 01 ®
15-
""0
0 J 79-17C
P3 =101 kPo hA=3.18 kg/s T3 =238-278 K TF u 238-278 K
5
5
I0
q
Figure 33.
15
(predicted),
20
g/kg
Comparison of Measured and Predicted Values of Lean Light Off for J79-17C Combustor.
-58-
,....... .. ........ -... ...... . ..... "-"..."-. ............... . .
25
*
50
S
.05
,
A.
5
7 0
S6
OXc ,x
7
03
8
0
x
X x
.300
xV
20I 0*
P3 m0I
rha1.15,k/ T . 239-.320 K
59A
10-
kPo
TF z24 0- 3 05 K igtOCfr
ofLa
10 ~LLO
Figure 34.
20 q
30
40
(predicted), g/ kg
comparison of Measured and Predicted values of Lean Light Off ftor F10l Coiubustoc.
-59-
-'5-
0 F 101mbst
50
B,
*10-
Fuel
Symbol 0•0 0B 03
-~88
1
15-
9B
00
0
7 0
148
0
15B
x
SI3B
100
TF 39
5-
P3 -,02.6 kPa T =237-278 K rhA = 4 .08 kg/s
0 0
5
t0
IS
20
qLUO (predicted), g/kg
Figure 35.
Comparison of Measured and Predicted Value:s of Lean Light Off for TF39 Combustor.
-606
-,6
"4,L..
. • . . •,
.r .••_:'•'• t.•,• • ...., ,• '
• . •.• ,•,
• .•••••'.,
,,'"•5
:,'
,• tF • , ;'•'.'d " '
h,.r•
30 Fuel
IC 15C
symbol 0 A3
15C CD
00
P3 =101.5 kPa
ihAw 1.31 1,36 kg/s T• =223-272 K TF 2223-274 K
S(r,
10
20
qL LO (predlctd), g/kg
Figure 36.
Comparisont of Measured and Predicted Values of Lean Light Off for J85 Combustor.
-
61-
30
30 F100 CP
',20-
%00
10:-
8I3
Standard Day (T=288 K) 0 Cold Day (Ts 244 K)
0.4 8 kg/s rh sA
0
10
20
30
q LLO (predicted), g/kg
Figure 37.
Comparison of Measured and Predicted Values of Lean Light Off for FO00 Combustor.
-
L
%
.
"
"_".
_ _ _ _ ___ _ _
-'
,_"
62
-
__-
"
.
-.
"
"
.
SSECTION VI
LINER WALL TEMPERATURE For the purpose of analysis a liner may be regarded as a
container of hot flowing gases surrounded by a casing in which air is flowing between the container and the casing.
Broadly,
the liner is heated by radiation and convection from the hot gases inside it,
and is cooled by radiation to the outer casing
and by convection to the annulus air.
The relative proportions
of the radiation and convection components depend upon the *
geometry and operating conditions of the system.
Under equili-
brium conditions the liner temperature is such that the internal and external heat fluxes at any point are just equal.
Loss of
heat by conduction along the liner wall is comparatively small and usually may be neglected. rate of heat transfer
of heat transfer out.
Under steady-state conditions, the
into the wall must be balanced by the rate
under steady-state conditions R 1+
C1
R2
C2
(18)
The derivations of suitable equations for R1 , C1 , R2 and C2 are fully described in reference 10.
As these equations contain
no drop-size terms they are unaffected by the results of the present investigation.
Hence, the following discussion will be
confined to summarizing the key features of the calculation procedures for estimating liner wall temperature,
along with a com-
parison of measured and predicted values of Tw for various types of combustors. -
r"4
63
-
Internal Radiation
1.
This is the component of heat transfer that is most affected by a change in fuel type.
is given by (18]
It
Rh . 0.5 a (1 + 6w) ag T1. 5 (T2.
T 2 5)
(19)
= liner wall emissivity
ew
gas emissivity
g-
Tg9
gas temperature -
wall temperature
The 'bulk'
or mean gas temperature,
sum of the chamber entry temperature, rise u3combustion, ATcomb. s r
-
Stefan Boltzmann constant
where a
Tw
5
du e t o
c m u t o , A
T3 ,
Tg,
is
obtained as the
and the temperature
o b
Thus: Tg
-
T3 + 6 Tcomb
(20)
ATcomb may be read off standard temperature rise curves. appropriate value of fuel/air ratio is
The
the product of the local
fuel/air ratio and the local level of combustion efficiency. Most heat transfer calculations are carried out at high pressure conditions where it
is reasonable to assume a combustion effi-
ciency of 100 percent. For the luminous flames associated with the combustion of
heterogeneous fuel-air mixtures, the value of c Eq. (19)
is obtained as (18] -
64 -
for insertion in
g-
where q is
-
(21)
exp(-290 P 3 L (q 1b)05 T
the local fuel/air ratio and Ib
of the radiating gae.
is the
The luminosity factor,
L,
'beam length' is
an empirical
correction introduced to obtain reasonable agreement botween experimental data on gas radiation and predictions from Eq.
(2l).
Analysis of the experimental data contained in references I thru 6 led to the following expression for I1 (7] L - 336/(percent hydrogen) Substitution of this value of L into Eq.
2
(22) (21)
allows calcu-
lations of flame radiation to be carried out for all fuels over the entire range of test conditions. 2.
External Radiation The radiation heat transfer from the liner wall to the outer
casing,
R2 ,
cart be estimated only approximately due to lack of
accurate information on wall emissivitie6.
For this reason it
sufficient to use the cooling-air temperature,
T3 ,
the unknown temperature of the outer air casing. ation across a long annular space,
13
in place of Also,
for radi-
the geometric shape factor can
be assumed equal to unity, and the expression for net radiation flux then -ieduces to R
3.
2
0.4 o(T
w
-
42:) T4)
3
Internal Convection Of the four heat transfer processes which together determine
L -65-
_7'
this component is the most difficult to
the liner temperature, estimate accurately.
In the primary zone, the gases involveU are
at high temperature and undergoing rapid physical and chemical change.
introduced by the existence within
Further difficulty is
the primary zone of steep gradients of temperature, compositicn.
and
velocity,
Uncertainties regarding the airflow pattern, the
state of the boundary-layer development and the effective gas temperature make the choice of a realistic model almost aroitrary. In the absence of more exact data it
is reasonable to assume
tiiat some form of the classical heat-transfer
relation for
straight pipes will hold for conditions inside a liner,
using a
Reynolds number index consistent with established practice for conditions of extreme turbulence.
This leads to an expression of
the form (18]
. C1
4.
0.017
0.8 (Tg
DL
(24)
T.I
External Convection This is
4C
obtained as [18]
2
0.020
(5
an
~A an
The fluid properties are evaluated at the annulus air temperature, T3 .
In practice,
the cooling air temperature increases
-
A
66
-
°
.
but normally this amounts to no
during its passage downstream,
more than a few degrees and can reasonably be neglected. For equilibrium R1 + C 1
R2 + C2
(2.)
Solutior of this equation yields the wall temperature,
Tw.
'The value of Tw as determined by the method outlined above
represents the liner wall temperature that would be obtained in the absence of internal wall cooling.
As references 1 thru 6 do
not contain the detailed information needed to estimate film cooling effects on Tw,
it
was decided to calculate
'uncoojed'
wall temperatures for four combustors only, namely J79-17A, 17C,
F101 and TF41,
in order to ascertain if
J79-
the results obtained
reflected anticipated trends in regard to the effect of fuel hydrogen content on liner wall temperature. calculations are shown in Figs.
The results of these
38 thru 41 for all fuels as plots
,)f Tw versus hydrogen content. It may be noted in Figs.
38 thru 41 that the calculated
values of Tw are generally higher than the corresponding measured values due to neglect of internal wall cooling. power conditions,
Only at low
where the errors incurred through neglect of
internal wall cooling are partially balanced by the assumption of 100 percent combustion efficiency in the combustion zone,
do the
measured and calculated wall temperatures roughly coincide.
-
67 -
.... ................... ................. .. -w---
-
OW".M'Nw"
)DOP
M
-W
Mt
s
fm
wU
u
Dash---
CAp ------
'
14
160
120
"W,68-
*,.
-.
U-0 U'
900' a.'
U..
J________, ~~Calculated ~.o-~OEXPerirmt6Ital
0
-3
-B ý.w-
.
-.u.q.
-
Dash
0
00
0000
130
12 CUL$LCNENPEC.
1481
90 Boo
.
0 0. 70-
-
.--
-a
Caocuiated 01 ExPerIm~fltal
14
Ds W
S12000
Uj 1300
S1200 -J
S1100
*
cus
e
r. 1000 7i00
15 CONTENT, PERCENT
FUE 12 j~fl 40.
~d ~~d~ctdVa]USS onl the Effect
f bOaBt~ ompr
t CH 2 ~ conlen onl Ltfl of~re4
-70
t
TeMPerature for
-
l
CIbtO
140
CrCrui00 as
w -
Ak
71-o-
3:1000~~
-r
X1 900
z
.-
.
.hose factors are not considered too serious in a study that is
maLnly concerned with fuel type,
"force to all
fuels.
because they apply with equal
The fact that the measured and calculated
values of Tw follow the same trend, as evidenced by Figs. 41,
38 thru
tends to support the validity of using the luminosity factor
"concept as a convenient means for incorporating fuel hydrogen "content into the 'standard' "Eq. (21)
Thus
may be rewritten as e
g
N,72
equaticn for flame emissivity.
1
exp[
97440 -P (%H2 ) -2 (q 1)0.5 3
bg
T'.I
(26)
SECTION VII POLLUTANT EMISS IONS The pollutant emissions of most concern for the aircratt gae turbine are oxides of nitrogen (NOx), unburned hydrocarbons (UHC),
carbon monoxide (CO),
and smoke.
The concentration
Levels
of these pollutants can be related directly to the temperature, time,
and concentration histories of the combustor.
tories vary from one combustor to another and, combustor,
for any given
with changes in operating conditions.
pollutant formation is
These his-
The nature of
such that the concentrations of
irbon
monoxide and unburned hydrocarbons are highest at low-power conditions and diminish with increase in power.
In contrast, oxides
of nitrogen and smoke are fairly insignificant at low power settings and attain maximum values at the highest power condition. The basic causes of these pollutants and the various methods employed to dlIeviate them have been fully discussed elsewhere LlJ
. Most modeling of emission characteristics
with oxides of nitrogen,
has been concerned
but efforts have also been made to
predict the formation of other pollutant species.
To be success-
ful a model must accommodate the complex flow behavior
and
include a kinetic scheme of the important chemical reactions occurring dithin the combustor. combustion processes are, the prisent time,
The kinetics of some relevant
unfortunately.
not well understood
particularly for the production of carbon,
at cay-
bon monoxide and the hydrocarbon species that are intermediate -
73
-
in
the fuel oxidation process.
The primary requirement for a satisfactory emissions model for gas-turbine combustors is
that it
should represent an optimum
balance between accuracy of representation, economy of operation,
utility,
ease of use,
and capability for further improvement.
In
recent years, conlsiderable efforts have been directed toward the development of relatively complex mathematical emissions models that can be applied to gas turbines
[19-27].
The high cost and
complexity of the more sophisticated mathematical models have encouraged the development of semi-empirical models for NO CO emissions.
For example,
and x Hung's approach haa been used suc-
cessfully in predicting the influence on NOx emissions of water injection and wide variations in fuel type (26,27).
Other suc-
cessful semi-empirical models for nredicting emissions have been developed by Fletcher and Heywood [19,28] and by Hammond and Mellor [29,30]. Empirical models can also play an important role in the design and development of low emission combustors.
They may
serve to reduce the complex problems associated with emissions to forms which are more meaningful and tractable to the combustlion engineer who often requires only an insight and a quick estimate of the levels attainable with the design variables at his dispo-a1.
They also permit more accurate correlations of emissions
for any one specific combustor than can be achieved by the more general analytical models.
-74-
w4 A.- 4~I
4
1.
Oxides of Natro~en semi-empirical model for the prediction of poJ-
Lefebvre's lutanw. emissions
of mixing rates,
(9), based on considerations
chemical reaction rates, and combustor residence tame,
leads to
the following expression for NOx. NO
,,NOX
Vc exp (0.01 T
9 x 10-a P " W -..
) /l~g
(;d 7)
mA Tpz
Equation (27)
demonstrates that the only influence of frel
type on NOx formation is via the two temperatueao terms, The former is
Tsat*
T
and
as
calculated
Tpz -T 3 + AT pz where AT pz is the temperature rise due to combustion corresponding to the inlet temperature,
T3 , and the primary-zone fuel/air
Tat is the stoichiometric flame temperature corresponding
ratio.
to the inlet temperature,
Equation (27)
T3 .
suggests that, in
the combustion of heterogeneous fuel-air mixtures,
it is tne
stoichiometric flame temperature that determines the formation ot NO .
However,
for the residence time in the combustion zone, the appropriate tem-
which is also significant to NO
formation,
perature term is the bulk value,
Tpz,
as indicatod in the denomi-
(27)
is
nator of Eq. It
(27).
should be noted that Eq.
tional spray combustors only. combustors,
suitable for conven-
For lean premix/prevaporize
in which the maxiiaum attainable temperature is T
-75-
1pz
it may still be used, provided that T pZis substituted forT it
should also be noted that predictions of NOs based on Eq.
tend to be too high when the overall combustor air/fuel exceeds a value of around 100. fuel/air
ratio
This is
ratio
because with diminishing
the flame shrinks back toward the fuel
no longer occupies the entire combustion volume, V . "this is
(27)
nozzle and
However,
not considered a serious drawback since, in practice,
interest is normally focused on conditions of high fuel/air ratio, where NOx formation rates attain their highest values. The excellent correlation of data provided by Eq.
*'
illustrated in Figs.
42 thru 52.
(27)
is
These figures include all
combustors except the J85 for which the measured values are too low for satisfactory correlation. 2.
Carbon Monoxide For the prediction of CO emissions the relevant expression
9,
iBs
:;'•"CO
(9
-6
e(-D2 86 mA Tpz T exp( Vc
-05 x i6 0.5
0.00345"-p-•Tgz)Uz-'3
fPm
D0
o !
'pz
eff
As CO takes longer to form than NOx,
[PI F 3
1
g/X:g
(28)
115 3
the relevant tem-era-
ture is not the local peak value adjacent to the evaporating fue.. drops, but the average value throughout the primary zone, namely
*
T Also, because CO emissions are most important at low pre.,pz sure conditions, where evaporation rates are relatively slow, •.t
*i~i
is necessary to reduce the combustion volume, V*, by the volume "- 76
-
30 Fuel
Symbol 0
z
3
0
4
0
6
Xo
0
J79-17A P 3 -251-1374 kPa
P-0T
00
3
mA= I.I 20
10I *NOx
42 . Figure zEmissions
=43-78 7 .4
K
kg/$ I
30
(predicted), g/kg
Values and Predicted of Measured Comparison for I toof 6).NOx . (Fuels J79-17A CombustoJ rh 4
-
77
-
kg/ An1.7-
Fuel
,0
7
0
8 9 10
0
12
X
13
/
O x
a
0
I0 -dJ
79-IA P3 = 251- 1374 k Pa
T3 = 413 -787 K 4mA
0010
1.7 -7.4 kg/s
20
30
NOx (predicted), g/kg
Figure 43.
Comparison of Measured and Predicted Values of NO Emissions for J79-17A Combustor. (Fuels 7 to 13)W
-78-
30
.20-
S00
Fuel 21
Symbol
3 4 5 6
0 0
0 A•
O
0 0
7
x
-
(0
z 0
565k/7C mA=13 J79-1
P3 -250-410 kPa
Nx(peice),gk
T3 = 413-795 K rhmA = 1.5 -6.5 kg/s
Owe 0
Figure 44.
10 20 N0X (predicted), g/kg
Comparison of Measured and Predicted Values of NO (Fuels I to 6 )x, Emissions for J79-17C Com~bustor.
-79-
A
30
Fuel
"" 20
Symbol/
7
0
9
0
i1 12 13
0 x 0
0
/
So.
Z 1o-.
J__719-17C P 3 =250-1410 kPa T 3 = 41:5- 795 K
enA ,Pu"'"0
' 00
,Jf 0x
Figure 45.
I!
1.5 - 6.5 kg/s
10
20
NOx (predicted),
g/kg
"-80 "--.
30
Comparison of Measured and Predicted Values of NO Emissions for J79-17C Combustor.
,
I
-
(Fuelu 7
to 13
.
Symbol
Fuel
1
0 A
2
30-
3
-/
0/
4
0
5-
x
6
x
o
v
0
S20-
oo
/P
z
FIOI
IO-
/
P.= 386-1266 k Pa T3
464- 850 K
m A: 8.5- 20.5 kg/s
0/ 0
20
I0
NOx (predicted),
30 g/kg
Comparison of Measured and Predicted Values of NO X (Fuels 1 to 6). Emissions for FIO Combustor.
,ýigure 46.
-81-
I.. -
-
,I*~....*..~
-
*
*I
*~
Fuel
Symbol
*7
0
30-
8A 9 10 12 13
o 0
0
x
,D0O 0
A
,
x
0 00
E04
z
=820/
10I
F20 1 P3 386-1266 kPa
Ld
T3 =464-850 K mh =8.5 -20.5
kg/s
mA
N'x (predicted), g/kg
4*..
-.1+
Figure 47.
Comparison of Measured and Predicted Values of NO Emissions for F101 Combustor. (Fuels 7 to 13). x
-
-.7 ................................................................
82
-
-
'.* ..
I%'W'
-urw'.TUT'W
%
ý
Y¶1
Symbol
Fuel 1
0
02
30 5 6
OoX 0
'7a
x
~20-
TF 41 , 10
/2P
3
T 3 = 434-760 K
011,q x
rhA =0 .9 3 - 5.0 kg/s
/ 0/ O0
0
= 93-1870 kPa
10203 NOx (predicted), g/kg
Figure 48.
Values of NO comparison of Measured and Predicted x (Fuels I to E,).
Emissions for TF41 Combustor.
-
.4 4
83
-
Fuel 7
Symbol 0
8 9
S30-
A 00
10
0
A
11
0 -
a
0
S20-
10
/
TF 41
P3 =293- 1870 kPo 00
T 3 = 434-760K
rA 0.9 3 - 5. 0 kg/s 0I
0
20
30
NOx (predicted), g/kg
Figre 49.
Comparison of Measured and Predicted Values of NO (Fuels 7 to 12). X Emissions for TF41 Combustor.
84
Plotted Points Represent
.•.
Average Values for All Fuels"
•
30-
;:Ad h-0
Z"
TF 39 P3-=333- 1586 k Po
"10-
I0 -
T3= 452 - 760 K " 1.27- 4.62 kq /s
s
c-r
-
I'0
20
30
NOx (predicted), g/kg
Figure 50.
Comparison Emissions for of TF39 Measured and Predicted Values of NO Combustor.
85
I
20
/
All Fuels 160 0/
TF 33
0
0
0 z(:9_ 40
16
P3-200-1210 T3359- 662 kPo K
4-4
mA" 1.92 - 9.3 kg/s
•
••I
oO
Figure 51.
I
4
8 12 NOx (predicted), g/kg
16
Comparison of Measured and Predicted Values of NO Emissions for TF33 Combustor.
-
4H
I
66
-
40 All Fuels
30
o
*
88 0
0
0 z
FIO0 P 3 = 370-1520 kPO
10-
"T = 486-864 K
3
rhA= 2.8-73 kg/s
0
10
20
30
40
NOx (predicted), g/kg
Figure
52.
Comparison of Meteuted and Predicted Values of NO Emissions for FO00 Combustor. x
I.E
-87
IA
-
"occupied in fuel evaporation,
V
Tnis was evaluated
Ve - 0.55 x l0- 6 f
ý /P
m
as
&
off
pz
o
pz 'A
e
t(d
The correlations of experimental data achieved with Eq. (28)
illustrated in Figs. "FiOl,
53 thru 57 for the J79-17A,
FI00 combustoxs,
TMi, and
J79-17C, Sare
respectively.
It is perhaps worthy of note that although Eqs. (28) ,
nave no strong theoretical foundation,
t'ire.
primary-zone temperature. ture aind mean drop size.
tempera-
The effect of variations in overall
mbuitor fuel/air ratio is
• '(•
flow propor-
and operating conditions of inlet air pressure, and mass flow rate.
and
they do emoody the
main variables of combustor size, pressure loss, tions,
(ZI)
also included via its influence on Fuel type affects both flame tempera-
For NOx,
drop size is unimportant since
at the high pressure conditions where NOx emissions are moot. prominent,
the fraction of the total combustion volume employed
in fuel evaporation is so small that wide variations in fuel drop . ze have a negligible effect on NO . .)per.ir.Ion.
However,
at low pressure
where CO emissions attain their highest concentra-
tion5, a signiticant proportion of the primary-zone volume 13 needed to evaporate the fuel. Under these conditions, any factor influences futd evaporation rates, such as evaporation con-
that
stant,
or mean drop size, will have a direct effect on the volume
available for chemLcal reaction and, oft rO arid IJHC.
therefore,
on the emissions
Thus, for the correlation of CO data the effects
Suf fuel type cannot be ignored.
88
-
'-- "4,.
.
''
-
-
-
'
" -
-
'
"
'
',"
" .
-"
'
-
-
"
" •~~L'••l
"t,
-
.
",
"
.
".."•
"
.
"
=
"
•
'
,
.
.-
"
.
.
.
J 79-17A
All Fuels
I00-
"
50a' ~20
0
10-
C00 0 00
5-0
251-1374 kPa
0P 00
2
%%
T = 413-787 K
5
c8. -o
1.7 -7.4 kg/s
Coo (peice),g
2
5
K)
20
10t0
200
CO0 (pro ed•), g/kg
IA Figure 53.
"'• /
".'
Compariaon of measured and Predicted values for J79-17A Combustor. C'
1'
.- ,
1-
1-
1
L
- 89
-
-
H. I
- I
of
. -
CO Emi~sson6
200 J 79-17C All Fuels
100
b 0
•".1,
50
00
00
10
0
80
0
0
P3 = 250-1410 0
0
20 _0^
/
T3 =413-795 K
0
20o 0& 0h °A=
1.5-65 kg/s
% 0o&o 2
Figure 54.
kPo
5
50 20 10 CO (predicted), g/kg
100
200
Comparison of Measured and Predicted Values oif CO Emissions for J79-17C Combustor.
-
90 -
100All Fuels
50-
,8%
0
P-20-
*10
5
o
P3 =:386-1266 k Po
0o
2
T3 = 464-850 K
1 •
8.
pei52mA0
0 2
5
10
20
CO (predicted), g/kg
.5 kg /s 50
100
I,.
Figure 55.
Comparison of Measured and Predicted Values of CO Emissions for F101 Combustor.
-914 • ,
" ,,-,••c=,=,`.
•.•
,•
•.
, ,
.
• -
I00 41
.TF 400
All Fuels
50-
*daft '.°
110-
5~20 -
P3 293 -1870 k Pa 434-760 K
T•,
-
0.93 -5.0 kg/s
•,:r 2-o° 0
".0 0 ,y, -,'•.
..h..-. .-
-
. ,.
,-
12
.,
-• -
-
,
0
0
---
-,:;
5
,,
.
I0
. ,.,
,
,
,.
20
..
..
...
,.
.
.,.
50
..
.
. ..
100
..
..
.
CO (predicted), g/kg
Figure 56.
Comparison of Measured and Predicted Values of CO Emissions for TF41 Combustor-.
.92
,.,
%0 50
All Fuels
20 10145
0
00 c~0 1.0
P z370- 1520 kPa
0
T37486-864 K
0D 0
mA= 2 .8 -7 3 kg/s
mliii
II
Itui
-0
I0 .0.5
Fzgure 57.
2 5 10 CO (predicted), g/kg
20
50
Comparison of Measured and Predicted Values of CO Entission:i for F1O0 Combustor.
-
93
"3.
Unb~rneod Hydrocarbons
"Unburned hydrocarbons
incJtide fuel that emerges at the
combustor exit in the form of droplets or vapor,
as well as the
products of the thermal degradation of the parent fuel into species of lower molecular weight,
such as methane and acetylene.
They are normally associated with poor atomization, burning rates,
the chilling effects of film-cooling air,
combination of these. ally reduces
inadequate
An increase
in
or any
engine power setting usu-
the emission of unburned hydrocarbons,
partly
through improved fuel atomization but mainly through the effects of higher
inlet air pressure and temperature,
enhance chemical reaction rates in
which together
the primary combustion zone.
Analysis of the experimental data yields an equation of the form 11,764 mA Tpz exp(
fM
Vlp.L010-6
Ppz
I
-
0.00345 T z)
APL
2.
effJ
F3
.
This equation is very similar to Eq. of CO emissions,
L1
D
!_
(28)
for the prediction
except for a stronger dependence on liner pres-
sure drop and inlet air pressure.
This is
perhaps hardly
!iurprising, since the factors that control CO emissions also influence UHC emissions,
and in much the same manner.
Due to the well-known difficulties and uncertainties that are normally associated with the measurement of unburned hydroi;arbons,
close agreement between the predictions of Eq.
the actual measured values can hardly be expected.
(3U)
and
However,
-94-
"• " r
'
"
" ""
" :
"a."•
•
" " "
-" "" " - ""-
a. •.- ",
'"."..' -.7 *
.:
'•.'•.',::,•
,o.
, ••
•
'
although Figs.
58 thru 63,
which are drawn tor the J79-17A,
F101, and TF41 combustors,
17C,
exhibit more scatter than the
corresponding figures drawn for NOx and CO, achiuved is considered 4.
379-
the correlation
tairly satisfactory.
Smoke Exhaust smoke is caused by the production of finely-divided
soot particles in fuel-rich regions of the flame and may oe generated in any part of the combustion zone where mixing is inadequate.
With pressure atomizers,
the main soot-forming region
lies inside the fuel spray at the center of the combustor.
This
is the region in which the recirculating burned products move upstream toward the fuel spray, and where local pockets of fuel vapor are enveloped in oxygen-deficient gases at high temperature.
in these fuel-rich regions,
soot may be produced in
con-
aiderable quantities. Most of the soot produced in the primary zone is consumed in the high-temperature regions downstream. viewpoint
i combustor may be considered as two separate zones.
One is the primary zone, Lion,
Thus from a smoke
which governs the rate of soot forma-
and Lhe other is the intermediate zone (and,
temperature engines,
on modern nign
the dilution zone also) which determines the
rate of soot consumption.
The soot concentration actually
observed in the exhaust gases is an indication of the dominance of one zone over the other. Soot is not an equilibrium product of combustion except at
-
95
-
60 J79-17A All Fuels
50-
P 3 =251-1374 kPa T 3 =413-787 K r=
1.7- 7.4 kg/s
40
.
* *
0
0300E
00
000 00
20
0 0
0
00
%oo0
0
0
0 0
10 0 10
0 0
10
20
30
40
50
60
UHC (calculated), g/kg
Figure 58.
Comparison of Measured and Predicted Values of Unburned Hydrocarbons Emissions for J79-17A Combustor.
-
96
-
gt
•I
a.
:•
J79- 17C
,"
All Fuels
Dash and SLTO
00
0
00
S8 oc
0 0
Figure 59.
I 2 UHC (calculated), g/kg
3
Comparison of Measured and Predicted Values of Unburnec Hydrocarbons Emissions for J79-17C Combustor.
-974,-
.'
.. " - " .
" ,
" ." -
' " .
- } . e • .
, - .
' ',
- .'
'.
- ,',•,',
,.,L
. _ _ .,, •,.
,".-
••
J 79-17C 60-
All Fuels
50-
00
0 0° 000
5~0
so'
4
o
(9
0
0
00 S40-
0
300
==30-) o
00 0
M
0 20-
0
Idle and Cruise
I0-
0
0
10
20
30
40
50
60
UHC (calculated), g/kg
Fiqure 60.
Comparison of Measured and Predicted Values of Unburned Hydrocarbons Emissions for J79-17C Combustor.
-
Si
l
. •
. q •
. • • .
" "
" "b " " "
98
-
" " " "
" "
" " •
" "
' °
• *
" " = ° • • '
'
"
i
•
7
•-
-
-• ,• -..
•
-
,
w
*,
- %'cr.,
.
•
-
-....
-r
.o•
-;L r
-T
w•
-:
-.--
,-
,T
W'W
wr
*;
*y4 -
qI
r
40l P3 = 386-1266 kPa
All FuPis
T3 =464-850 K
030o
SA=8.5-2 0 5 kg/s
0
~20
0
0
0 0
Figure 61.
0
10
0
20 30 UHC (calculated), g/kg
40
Comparison of Measured and Predicted Values of Unburned Hydrocarbons Emissions for FlO1 Combustor.
99
-
V,W
3 Fuels
TF 4[All SLTO Dash and
i
'
2-
0
O000
2-
I
33
UHC (calculated), g/kg
Figure 62.
Comparison of measured and Predicted Values of Unburned
Hydrocarbons Emissions for TF41 Combustor.
-
100
-
00
6060
0
50-
00/ 0 0
~'40
S400
® SO0 0
S
200 10
TF 41 Idle and Cruise All Fuels
0
00
0I
20
30
40
50
60
UHC (calculated), g/kg
Figure 63.
Comparison of Measured and Predicted Values of Unburned Hydrocarbons Emissions for TF41 Combustor. 1-01-
z
L
Ik 1A
mixture strengths tar richer than those employed Thus,
zones of combustors.
it
in the primary its
impossible to predict
is
rate
of formation and final concentration from kinetic or thermo-
"dynamic data.
In
practice,
the rate of soot formation tends to
be governed more by fuel-spray characteristics and fuel-air mix-
"ing than by kinetics. Many specific mechanistic models for soot formation have been proposed.
Current thinking tends to favor the notion that
condensed ring aromatic hydrocarbons may produce soot via a different mechanism than do aliphatic hydrocarbons.
Aromatic hydro--
carbons can produce soot via two mechanisms:
condensation ot
(1)
the aromatic rings into a graphite-like structure,
or (2)
breakup
to small hydrocarbon fragments which then polymerize to form larger,
hydrogen-deficient molecules which eventually nucleate
and produce soot.
"t]. [31,32]
Based on their shock tube studies,
Graham et
concluded that the condensation route is much faster
than the fragmentation/polymerization route.
According to the
condensation-route model. aliphatics produce soot via the tfr;tqrrgentation/polymerization mechanism only.
As a result,
these
hydrocarbons do not form the quantities of soot produced by the aromatics.
Indeed,
during the fuel-rich combustion of a fuel
blend composed of aromatics and aliphatics, -
the aromatic hydro-
carbons would produce the major quantity of soot.
Combustion ot
,ie aliphatic port ions of the fuel would influence temperature and hydrocarbon fraqmc .t concentration,
but soot formatior via
trgmeritat. io0.'polymeri7ation would be minimal.
-
.,.,"%1?
..- ., ."r'Z
'?': ,
. . V...,
.
102
-
. ." " " -.
"
.. ..
. . . .
. .'" ' " "
' "" '
-
W-
MWr
.w
Experimental data obtained by Blazowaki (33]
using various
¢,
blends of iso-octane and toluene fuels were found to be con-
*
ssistent with this model.
..
study by Naegeli and Moses (34) suggest that the picture will oe
However,
the results of an experimental
more complicated for fuels with high concentrations of polycyclic aromatics. For gas turbine combustors the main controlling factors for soot formation and smoke have been determined experimentally as fuel properties, ratio,
combustion pressure and temperature,
fuel/air
atomization quality, and mode of fuel injection (10].
In order to analyze the smoke data contained in retererces 1
*
thru 6,
the first
numbers
(SN)
step must be to convert the quoted smoke
into soot concentrations
(Xc)
expressed in mg/kg.
This conversion was accomplished using the following different. factors for different levels of smoke number (35]. SN - 0 to 1
Xc -
SN - I to 5
log Xc -
0.136
SN -
5 to 10
log Xc -
0.06265 (SN)
-
0.769
SN -
10 to 20
log Xc -
0.03187 (SN)
-
U.4614
SN -
20 to 30
log Xc -
0.0301
SN > 30
0.1 (SN) (SN)
(SN)
log Xc - 0.02538 (SN)
-
1.136
-
0.42b
- 0.2845
The following equation was then used to convert engine
-
1. .
.
-.
-
.
'
-
103
-
exhaust soot concentrations
into corresponding
combustor exit
values.
XXc4 "-X
c8
q8 q4
it
is
For the purpose of analysis,
convenient to consider two
separate zones (a) a soot-forming zone, zone.
(31)
+ q4 Lq
and (b) a soot oxidation
The soot concentration measured at the combustor exit
represents the diffeerece in effectiveness between these two competing processes.
Unfortunately,
any attempt to derive suitable
expressions to represent rates of soot-formation and sootoxidation is
seriously hampered by lack of knowledge of the basic
mechanisms involved,
so that in practice there is
tive except to resort to an empirical approach.
little
alterna-
Useful guidance
is provided by the knowledge gained from past experience in attempting to alleviate the problems of smoke and soot formation in gas turbine combustors. erness and Macfarlane
Thus,
for example,
the work of Hold-
(36] has shown that soot formation
increases rapidly with increase in pressure, diminished by increase in AFR.
Moreover,
and is appreciably
sufficient is known to
indicate that soot oxidation proceeds most rapidly in regions of high temperature containing excess air.
These considerations,
conjunction with analysis of the experimental data,
lead to the
following expressions for the soot formation and soot oxidation processes.
Sf
H1.5 18 P2 q 2 m 3 2z fpz mA Tpz -
104
-
in
%2
2
.
z) (18
",3qpz exp (0.0011T fPZ
mA qsz
Now Xc - X
"
-
-
1.5 % H2 )
pz
Xo
Hence,
C
3
fpzmATpz
8x( ~ )
OZl
P qq
x [1 8
-
k3
H
8z
Application of this equation to the correlation of experimental data on soot concentrations yields results as illustrated
Figs.
64 thru 71.
in
The values of C3 and C4 associated with these
figures are listed in Table 3.
This table shows a large dispar-
ity between values of C3 for different combustors which is not surprising, zone,
since C3 relates to soot formation in the primary
and its numerical value will be very dependent on fuel
spray characteristics,
primary-zone fuel/air ratio, and primary-
zone mixing characteristics, combustor to another. dary zone where,
all of which vary widely between one
This is
in marked contrast to the secon-
in the hot gas stream entering this zone,
fuel is fully vaporized,
combustion is
almost complete,
the
and plug
flow of combustion products at fairly uniform conditions of temperature and composition is well established. secondary zone,
Thus,
for the
differences between different combustor types
should be appreciably less,
and this is
confirmed by the lack of
marked divergence between the experimentally-derived
-105-
values of C 4
70 J 79-17A Predicted Values
60--
A Take Off Cruise [3 Idle
0 S20
o,50-
.,
z"ý 40-0 40 "0
TakeOf
0
2I-
02010 I•1 0 II
::::::::::-:,:-::.%•
15
Craphs Illustrating Influence of Hydrogen Content and Engine Operating Conditions on Soot Emissions for J79-17A Combustor.
-
'p:.-
0l
14 13 HYDROGEN CONTENT, percent
12
Figure 64.
Idle
•:?. •:••,:.::-
,•
106
-
..-. :<,i:•:;....•::..•
.::::•:.•,,.•.:.•:
:••;
-
Predicted
Values
A Take Off 0 Dash
A -U
0 Ile
0
l.C
0A
(,A 0,nI0
DZ
Tk
00
Take 8f
00 0 12
Figure 65.
13 14 HYDROGEN CONTENT, percent
Graphs Illustrating Influence of Hydrogen Content and Engine Operating Conditions on Soot Emissions for J79-17C Combustor.
-
A'
'
% "'
'/
,
•
'
•,
•" ..
15
',.-
'•
,
- '
% ""% ,'\
107
.\ ,%%
-%",
-
- - , ,
.
.
.• . .
.
-,
-
.
.
,
.,
.
.
F 101 Predicted Values 1.4
A Take Off
1.2
V Dash 0 Cruise C3 Idle
V
V
1.0
AV
Aw N.
E .0.8-0 z
0
S
Dash
40
-0.6
0
0w0 Q4
.5.
80.2
*
01
Cruise
0 Idle 12
13
14
15
HYDROGEN CONTENT, percent
Figure 66.
Graphs
illustrating
Influence of Hydrocarbon Content and
Engine Operating Conditions of Soot Emissions for F1O0 "Combustor.
-
108
-
414 4TF41 --
Predicted Values tZ Take Off
12
Dash 0 Cruise "7
010-
Dash
0
z0 0
0)
Idle I
12
Figure 67.
13 14 HYDROGEN CONTENT, percent
15
Influence of Hydrogen Content and Graphs Illustrating Engine Operating Conditions on Soot Emissions for TF41 Combustor.
-
4
I
109
-
1.4 1.2
I
-
Predicted Values Ai Take Off 0 Cruise
~0.6-
w0 z
Idle
Cus
Q2
314 12 HYDROGEN CONTENT, percent
Figure 68.
Graphs illustrating Influence of Hydrogen Content arnd Engine Operating Conditions on Soot Emissions for TF39) Combustor.
q1
'
15
-no-X
44
Values
""-"Predicted
A Take Off V Dash 0 Cruise
Idle
8E
z
0
z
w
z 0
v HYRGEaONETprcn
10-
0
Enin
12
q
Figure 69.
Opeatin
Codtoso
ctE
14 13 HYDROGEN CONTENT, percent
0
s
OnforJ8
15
Graphs Illustrating Influence of Hydrogen Content and Engine Operating Conditions on scot Emissions for J85 Combustor. -
i1i
-
Values
S-PFPredicted
A Take Off
3-
V Cruise I Cruise 2 Idle
V0
00
2z
0
Take Off
*
CuiA
0
i-
z w z
0
8
Idle Ek 0
. *
0
12
Figure 70. O
Cruise I Cruise 2
14 HYDRGEN13 CONTENT, percent
15
Graphs illustrating influence of Hydrogen Content and E~ngine operating Conditions on Soot Emissions for TF33 Combustor.
-112-
F 100
3-
Predicted Values A Take Off V Dash
0 Cruise o Idle
00z
00 H0
*z 0
o
I
-Take
Off
0•O
Dash ak
0on
Cruise
Idle 0i
I
12
Figure 71.
II
13 14 HYDROGEN CONTENT, percent
15
Graphs illustrating Influence of Hydrogen Content and Engine Operating Conditions on Soot Emissions for FI00 Combustor.
-
113
-
A OW0
IN'
Table 3.
Values of C3 and C4 employed in equation (34).
I
C4
C3
Engine J79-17A
2.43
0.0046
J79-17C
0.045
0.0042
FlOl
0.017
0.0020
TF41
0.0785
0.0037
TF39
0.145
0.0044
J85
0.33
0.0038
TF33
1.0
0.0045
F100
0.0375
-
114
-
'0.0035
listed in Table 3. If
allowance is made for the difficulties involved in the
sampling and measurement of soot concentrations and the poor measurability of fuel aromatics content,
the level of agreement
between measured and predicted values of soot concentration. exhibited in Figs. although Eq. (34)
64 thru 71,
is
quite reasonable.
soot concentrations
it
However,
predicts quite well the influence of combustor
operating conditions on smoke output,
the, fuel,
as
and also demonstrates that
rise with decreases
in
hydrogen content of
also shows that the extent of this increase varies
from one combustor to another in a manner that cannot be predicted a priori.
Thus it
offers no guidance on the likely
smoke emissions to be anticipated from any new type of combustor. Only if
the values of C3 and C4 were sensibly constant for all
combustors would it
be reasonable to regard Eq.
(34)
as com-
pletely satisfactory for the prediction of smoke emissions. Another defect of Eq.
(34)
is
the absence of a term to
describe the degree of mixing of fuel and air prior to combustion.
This is
sions,
for example,
known to have a strong influence on smoke emisthe very large difference in smoke output
between the J79-17A and J79-17C combustors, large difference in their values of C3 , is
as reflected in the known to be due in
large measure to the steps taken to improve the premixing of fuel and air in the latter case.
Improvements in the prediction of
smoke emissions cannot be expected until more quantitative information on the influences of fuel-air preparation and fuel -
115
-
chamistry on soot formation becomes available. Eq.
In the meantime
can provide useful guidance on the effects of changes in
(34)
fuel type and combustor operating conditions on smoke output. For any given combustor,
all that is needed are a few smoke meas-
urements obtained with any fuel at any operating conditions,
ju3t
in order to establish values of C3 and C4 for insertion into Eq. Thi.* equation can then be used to estimate smoke levels
(27).
for other fuels and/or other operating conditions. It
is of
interest to note in Eq. (34)
represented by its hydrogen content only.
that the tuel is This is because hydro-
gen content was found to provide a slightly better correlation of experimental data than aromatics content.
Furthermore,
no con-
clusions could be drawn regarding the relative importance to soot formation and smoke of single-ring and multi-ring aromatics. This is because the data bnow that replacing single-ring aromatics by multi-ring aromatics increases the level of exhaust smoke in others.
In sowe cases and reduces it
-
a' ' . ." '" . " -
"-
"" ' ' " . "
. " " '
116
-
. ". ..
• • iII• , . - . ,"., " ,- • .
' •,' ' . L
SECTION VIII PATTERN FACTOR The attainment of a satisfactory and consistent distribution of temperature
in the combustor efflux gases is
one of the major Experimental
objectives of combustor design and development. investigations
into dilution-zone performance carried out on test
rigs and actual chambers have provided useful guidance, and error methods are still temperature-traverse
but trial
widely used in developing the
quality of individual combustor designs to a
satisfactory standard. The mixing processes in the dilution zone are affected in a complicated mariner by the dimensions, of the liner, liner holes, chamber,
the size,
geometry,
and pressure drop
shape and discharge coefficients of the
the airflow distribution to various zones of the
and the temnerature distribution of the hot gases enter-
ing the dilution zone.
The latter is
strongly influenced by fuel
spray characteristics such as drop size, penetration,
spray angle and spray
since these control the pattern of burning and hence
the distribution of temperature in the primary-zone gases. Several parameters have been proposed to describe the tem*
perature distribution in the combustor efflux,
the most widely
used being the "overall temperature distribution factor" which tends to highlight the maximum temperature found in the traverse and is,
therefore,
of special importance to the design and dura-
bility of nozzle guide vanes.
It
-
•'"J'• • •,'.' •" -• •-,. • % •,: ,,-- > 7•',.>'-
" '.
•.
• -
is
117
•
normally defined as
-
- , L ,'•.
•• • ," •'
"'
•""
•
•C . " " ."".-"40.
%
Pattern factor -
Tmax T4 _
(5) (
3
Of prime importance to pattern factor are liner length, which governs the time and distance that are available for mixing,
and the pressure loss factor of the liner which controls the
penetration and turbulence of the dilution jets.
At low pres-
sures, where evaporation rates are relatively slow, portion of the liner length is process,
a significant
occupied by the fuel evaporation
so that less length is available for mixing.
This may
be accounted for by reducing the liner length, LL, by an amount, Le,
in the following equation for pattern factor
Tmax _T 4- -i-exp
Lef
[
(36)
where Z - 0.07 for tubular liners and 0.05 for annular liners (10]. The evaporation length, Le,
is
obtained as the product of
evaporation time and the average gas velocity in the predliution zone.
me
In reference 9 it L
where p zone.
g It
e
-
is
shown that Le is given by
0.33 x 106
A
/pg o(7
AL
eff.
is the average gas density upstream of the dilution is calculated at a temperature T
which is obtained as g
Tg -T -3
+ AT
where AT
g
is the temperature rise due to combustion for a g fuel/air ratio of 0.6 q AL is the average cross-sectional *
- 118-
area of the liner.
It
is
estimated by dividing the volume of the
liner by its maximum length. height of the liner. as DL -
DL is the average diameter or
For a tubular liner it
is readily obtained
(4 AL/i)0.5
Substitution of L
e
from Eq.
T 4[APL Tmax " x T T4 - 3
_ zJl• '*f
(37)
into Eq.
J79-17C,
"15, and 19,
and TF41,
It
is
31- ,
namely the J79-
insertion of values for APLlqref of 14,
respectively into Eq.
lations of the experimental data, 74.
gives
D2 1-6 0.3 o 3 _K .77o mA 0 J LL Lg L L eff J
"For the three tubular combustors examined, 17A,
(36)
(38)
provides excellent corre-
as illustrated in Figs.
72 thru
of interest to note that the improvement in pattern
factor with increase in engine power,
as predicted by Eq.
(due to reduction in evaporation time), results contained in Figs.
is
(38),
fully borne out by the
72 thru 74.
The influence of fuel type on pattern factor is manifested throuqh the effects of mean drop size (via viscosity and surface tension) and effective evaporation constant •. evaporation time. *
(via TbnJ T on droplet.
Over the range of fuels examined,
of fuel type on pattern factor is
the effect
relatively small, at least at
high power conditions where the evaporation time is
always a
small fraction of the total combustor residence time, *
of fuel type.
However,
if
regardless
measurements of pattern factor are
conducted at low power conditions,
where the evaporation time
constitutes a significant proportion of the total residence time,
4
-119-
0.5 J 79-17A
0.4
'
&
~i
0.3
4
/
s
Cruise
Take Off
1- 0.2-
.Dash
a-0./ ,
00
I
0.o
II,
,
0.3
0.2
0.4
0.5
Pattern Factor (predicted)
Figure 72.
Comparison of Measured and Predicted Values of Pattern Factor for J79-17A Combustor.
S 1.2
"-":
-
120
-
%•'.•.
" , '',
S "
4'"
"•""
" " " -", '
;, '
; ''
": ''
-''.'•
-b: • '•
- '" ""' -. "" ''•
• ,- •••
J 79-17C/ •
0.1
Take Off and Dash 10/O
0.4
dl
cruise 0.4 "00"3
" Fato
fo79-17
// C Comusor
o/ 00
Figure 73.
0.!
0;2 0.3 0.4 Pattern Factor (predicted)
Comparison of Measured and Predicted Values Factor for J79-1.7C Combustor.
-
4z
05
121
-
of Pattern
IuIn
0.5 TF 41
*1
0.4 Take Off 00.5
S.0Cruise
-
0.2
a..
1 0' 0
Figure 74.
-
0.1
0.2 0.3 04 Pattern Factor (predicted)
Comparison of Measured and Predicted Values of Pattern Factor for TF41 Combustor.
-122-
6-
0.5
,
then a strong effect of fuel type on pattern factor should De expected. The practical utility
of Eq.
(38)
is that it
allows the pat-
tern factor at max power to be predicted from measurements of pattern factor carried out at reduced power, more convenient test conditions.
It
i.e.
at cheaper and
also demonstrates,
as stated
above, that at the highest combustion pressures where heat flux rates to nozzle guide vanes and turbine blades attain their maximum values,
the influence of fuel type on pattern factor is
negligibly small.
12
-
',..
,-
• .'- ".'l•'- :.".'-"."%".'-"","".''.• •.' ,, .. 'r•' '.'
123
-
i• .- •"• •-"
" ,
,'"•'.
-'.
'.'''%'
•'''•-..•
.'i
'.
."-
'-
•'."-,
.' ," .'V
SECTION IX DISCUSSION AND SUMMARY Analysis of the key processes occurring within gas turbine combustors,
along with examination of the experimental data con-
shows that although the impact of
. thru 6,
tained in references
fuel type on combustion performance and liner durability is
usu-
ally small in comparison with the effects of liner geometry and is
it
combustor operating conditions,
nevertheless of sufficient
magnitude to warrant serious consideration. parameters, is
For some performance
such as liner wall temperature and exhaust smoke,
found that fuel chemistry plays an important role.
it
For oth-
the effects of fuel type are manifested through the physical
ers,
properties that govern atomization quality and evaporation rates. In the following sections the effects of liner size, pressure drop,
combustor operating conditions,
liner
and fuel type on
various aspects of combustion performance are reviewed briefly in turn. 1.
Combustion Efficiency From analysis of the experimental data contained
references
1 thru 6 it
found that combustion efficiency
is
obtained as the product of the 8 efficiency, poration efficiency,
in
n
c
is
and the eva-
i.e.
nc
77
77
c
c
-
x 7c
125 -
(1)
e -0.022 P 1 exp()
hi ~~where nc
•
A
and
3
3
l-
7C
f -36
exp
Vc exp (T /400) f-
x 10
2 Tc
P3
vc
Xeff
0fDO fcmA
1
(9)
In common with other loading parameters for the correlation of combustion-efficiency data, Eqs. tion efficiency is air temperature,
(8)
and (9)
show that combus-
enhanced by increases in gas pressure,
and combustion volume.
Equation (9)
inlet
also demon-
strates the adverse effect of low fuel volatility on combustion efficiency, quality is
especially at operating conditions where atomization relatively poor.
practical experience,
This,
of course,
but the main attribute of Eq.
the direct quantitative relationships zation quality (via SMD), tion efficiency,
is well known from
it
(9)
lies in
provides between atomi-
fuel volatility (via Xeff)
and combus-
which allow the effect on combustion efficiency
of any change in fuel type or fuel nozzle characteristics to be readily estimated. 2.
Lean Blowout Weak extinction values of fuel/air ratio are obtained as
2r
D2 qL*Afm
DoI~ at1F
eff In this equation it
g/kg
(16)
is of interest to note that the depen-
dence of weak extinction limits on combustor volume and operating conditions is very similar to that for combustion efficiency. -
126
-
Also in common with combustion efficiency is
the slight effect of
whereas physical properties are impor-
fuel chemistry (via LCV),
and Xef-
tant due to their influence on D
The reasonable degree of similarity between the values Af pz sug-
listed in Table 2 for several different types of combustors, gests that prospects are good for predicting,
the lean blowout limits of future combustor
acceptable limits,
should also serve to encourage further experimental
It
designs.
within close and
and analytical efforts in this area. Lean Lightup
3.
The equation for lean lightup fuel/air ratio is
except for a
identical to that for lean blowout fuel/air ratio, We have
slightly stronger dependence on P3 "
fff~lfmA
pVr. .exp(T
qLLO=B
/300)
f
almost
2
12 FDatT at 277.5Kj g/kg
eff
(17)
The very satisfactory correlation of ignition data provided
(17)
by Eq.
demonstrates the important role played by the atomi-
zation process vapor
in
in
providing an adequate concentration of fuel
the spark region.
This equation also provides useful
quantitative relationships between fuel volatility (B or X operating conditions (P 3 ,
atomization quality (D ), and combustion volume (Vc). estimate the increase in
combustor volume and/or
atomization quality needed to recover the loss
can be used to
it
for example,
Thus,
T3, and mA),
in
improvement
in
altitude
relight capability caused by changing the fuel to one of lower -
I ,
''
, ' " •""'" ."-
", .
'- "•
'- ".7
".
- -,i .
-'
127 ,
'
-
, '
" ".
"'"- " " "
• "
"
' "
,2
." .. - -'.'-"-
volatility.
Despite the well-known inconsistencies ignition data, it
that tend to plague
the values of Bf
appreciable scatter.
listed in Table 2 do not exhibpz In fact, they are consistent. to within a
few percent for the four combustors featuring pressure atomizers; namely,
the J79-17A.
J79-17C, TF39,
and J85.
Lthe J79-17C nozzle is a hybrid type, •V
(Note that although
at lightup most of the fuel
issues from the primary which is a pressure swirl atomizer). These results may be regarded therefore as representing useful progre-s towards establishing accurate prediction formulae for lean lightoff limits. 4.
Liner Wall Temperature The most important factor governing liner wall temperature is
-
the combustor inlet temperature,
i'
T3 .
Inlet pressure
is also sig-
nificant due to its influence on the concentration of soot particles in the flame,
and hence on the magnitude of the luminous
radiation flux to the liner wall.
At max power conditions,
liner wall temperatures are of most concern,
where
evaporation rates
are so high that the physical properties of the fuel appear to have a negligible influence on Tw. quit.e small,
as shown in Figs.
Chemical effects are also
38 thru 41.
However,
even small
increases in maximum values of liner wall temperature can seriously curtail liner life. in this investigation,
Thus,
for the range of fuels covered
fuel type must be considered of signifi-
cance j:o liner durability.
"- 128 6 , •
. , . .
. .
. 4 " ,• 4
.
. 4
,
-
• , . . . , . .
'',
.. ' •', ' . . '
. , '. ' ' .
. . ,
'• '
,
.
In the calculation of liner wall temperatures,
the effect ot
fuel type can oe accommodated quite conveniently by introducing "the fuel hydrogen content into the existing equation for gas emissivity.
S
This approach leads to the following equation for
~g . S1
5.
[
-exp
1 b) u.S
97440 P 3 (%H 2 ) 2 (q
(26)
NOx Emissions
It
is
found that NOx emissions are very dependent on
combustor operating conditions,
and also on the size of the
combustion zone which governs the time available for NO tion.
formax The key factor controlling NOx is the stoichiometric flame
temperature which,
in turn,
tor inlet temperature.
is
almost solely dependent on combus-
As far as fuel type is
cal properties are of little
concerned,
physi-
consequence except at low power con--
ditions where NOx emissions are always quite small due to the correspondingly low values of Tst.
tle influence on NO
X
because it
Fuel chemistry also has lit-
affects only slightly the values
of bulk gas temperature and stoichiometric
flame temperature
in
the following equation for NO
x
5 9 x9x10 10--81.25 p Vc exp (0.01 Tt) NO x -g/kg
/c
( 27 )
mA Tpz 6.
CO Emissions These are correlated by the expression:
CO
86 m -
T
exp ( -0.00345 2 S1
-129-
0
1
-
T U P
g/kg
(28)
Combustor size and operating conditions also play a prominent role in determining the level of CO emissions. importance is
Special
attached to inlet temperature and primary-zone
fuel/air ratio,
due to their combined effect in resolving the
primary-zone temperature.
As in the case of NO
emissions,
the
influence of fuel chemistry is small and is manifested through sliqht variations However,
in Tpz with changes in lower calorific value.
since CO emissions attain their maximum values at low
power conditions,
where a significant proportion of the total
residence time in the combustion zone is tion process,
occupied by the evapora-
the influence of those physical properties which
affect. evaporation rates becomes important. 7.
Unburned Hydrocarbons It
is found that the factors which govern CO emissions
also influence UHC emissions,
and in much the same manner,
except
for a slightly higher dependence on inlet air pressure and liner wall pressure drop.
SUHC
fV
11,764 m A T z exp( -
8.
We have
0.55 x 10-6 fPlm
D effl
Pz
T Pz
2.5
Smoke Of all the parameters studied,
is
-0.00345
smoke emissions is the one that
most affected by changes in fuel type.
The physical proper-
ties of the fuel are important insofar as they influence the mean drop size in the spray and the penetration of the spray across the combustion zone.
Spray penetration is -
V,.
130
-
of considerable
importance from a smoke viewpoint because inadequate penetration leads to enhanced fuel enrichment of the soot-torming regions just downstream of the fuel injector.
Smoke emissions are also
strongly dependernt on engine operating conditions and primaryzone fuel/air ratio,
as indicated by the following equation for
exhaust soot concentration.
'q° 1C -ex(.0T
. 3P3q
;!.
Xc
fpzmATpz1
q
11-H J
Z
s
Although the correlations achieved, thru 71,
'g/kg(34)
show appreciable scatter,
as illustrated it
is
in Figs.
considered that Eq.
64 (34)
could prove very useful for predicting the effects of changes in operating conditions and fuel type on exhaust smoke levels. 9.
Pattern Factor This is
T -T max 4 T4-T3
described with good accuracy by the following equation AP
6' 11-1 mAD 0o L a_.~teff X
L
0p.33 x.10
L q~ref]U
-
where appropriate values of Z are 0.070 and 0.050 for tuboannular and annular combustors,
respectively.
The above equation
shows that the main parameters controlling pattern factor are the pressure drop across the liner wall and the liner L/D ratio.
It
also takes into account the influence of evaporation time in reducing the time available for mixing within the liner.
At the
high pressure conditions where pattern factoL is of most concern, the evaporation
time is
always quite short in -
.4 . . ,. .
. . ., . . . • - .
..
131
comparison to the
-
. . .. . . .. . , • , . - ,. . , ,•
, . ..
•
. • .
,•
total residence time of the combustor,
and so the dependence of
pattern factor on fuel type is fairly small. At lower power settings, to increase in D
the evaporation time increases due
and reduction in Xeff"
This produces a
deterioration in pattern factor as indicated by Eq. by Figs.
72 thru 74,
(38)
and alsu
which demonstrate for all engines that pat-
tern factor at idle is
distinctly worse than at take-off.
These
consideraticns highlight the importance of measuring pattern factor only at the correct combustor to engine operation at max power.
inlet conditions corresponding Tests carried out at lower
pressure levels give values that are overpessimistic.
Also,
they
show a dependence of pattern factor on fuel type which great ly exaggerates the dependence actually observed at high pressures.
-
132
-
SECTION X CONCLUS IONS 1.
The fuels'
physical properties that govern atomization quality
and evaporation rates strongly affect combustion efficiency, weak extinction limits, and lean lightoff limits.
The influence
of fuel chemistry on these performance parameters
is quite small
and stems from the effects of slight variations
in lower calorific
value on combustion temperature. 2.
For any given combustor and fuel type, Eq.
(7)
enables values of
combustion efficiency to be calculated a priori at any stipulated combustor operating conditions. 3.
The effects of changes in fuel type,
liner airflow distribution,
and engine operating conditions on lean blowout and lean lightup limits may be estimated with good accuracy from Eqs. (17), 4.
(16)
and
respectively.
Liner wall temperatures are controlled mainly by combustor operating conditions and combustor design, playing a minor role.
However,
ratio engines the combustor is
with fuel effects
since in modern high pressure called upon to perform
satisfactorily for long periods at extreme condit2ons on current fuels,
it
follows that any factor,
however secondary,
that creates a more adverse combustion environment,
will have
a large and disproportionate effect on combustion performance and liner durability.
Analysis of the experimental data, whic:h cover a range of -
133
-
shows that fuel chemistry,
fuel types from JP4 to DF2,
indicated by hydrogen content, flame emissivity, 5.
as
has a significant effect on
flame radiation,
and liner wall temperature.
The influence of fuel chemistry on the emissions of carbon monoxide,
unburned hydrocarbons,
quite small.
The fuels'
and oxides of nitrogen is
physical qualities affect the
exhaust gas concentrations of both carbon monoxide and unburned hydrocarbons at low power settings where fuel -evaporation "
raLes are relatively low.
However,
at the
high power conditions where the emissions of carbon
monoxide and unburned hydrocarbons become negligibly small,
the influence of physical properties on these
emissions [6.
is
negligibly small.
Smoke emissions are strongly dependent on combustion pressure, primary-zone fuel/air ratio,
and the mode of fuel injection
"idual-orifice or airblast).
Fuel chemistry,
by hydrogen content,
is also important.
as indicated
The data contained
in references 1.thru 6 do not support the notion that multi-ring aromatics exhibit stronger smoking tendencies than single-ring aromat ics. 7.
Fuel chemistry has no direct influence on pattern factor. However,
physical properties have an effect that is
•"appreciable in
at low power conditions but which diminishes
importance with increase in engine power,
becoming very
"-small at the highest power setting, where the durability of "- 134 V0
.
hot section components is 8.
a major concern.
Values of pattern factor measured at convenient
low
power conditions,
and used
may be inserted into Eq.
(38)
predict the pattern factor attainable at max. power Sto conditions.
1..
-
:35
-
. -r-w
-v-
-.a
•-
-U-J
•
U •,,
'r,
r-r. - m
r-. o
r--
r1
'n •
,rrW
W r
rU ,w•
-,,r1W,,•r fl1
1
U..•. ..
•L-.
•
-
-.
,•.
SECTION XI REFERENCES
L.
C. C.,
Gleason,
T.
L.
Evaluation of Fuel Character
Oller,
Effects on J79 Engine Combus-
M. W. Shayeson and D. W. Bahr,
tion System, AFAPL-TR-2Ulb, June 1979. 2.
Gleason, C. C.,
Evaluation of Fuel Character
T. L. Oiler,
Effects on FI01 Engine Combus-
M. W. Shayeson and D. W. Bahr,
tor System, AFAPL-TR-79-2U18, June 1979. 3.
Vogel,
R.
E.,
D. L.
and A.
J. Verdouw,
Fuel Character Effects on
Troth
Current High Pressure Ratio, Can-Type Turbine Combustion AFAPL-TR-79-207z,
Systems,
April 1980. 4.
Gleason,
C. C.,
T.
Evaluation of Fuel Character
L. Oller,
M. W. Shayeson and M. J.
Effects on J79 Smokeless Com-
Kenworthy,
bustor,
AFWAL-TR-8U-2U92,
November 1980. 5.
Oiler, T. L., M. J.
Kenworthy,
and D.
* 6.
C. C. Gleason,
Russel,
Fuel Mainburner/Turbine
Effects,
J. D. Cohen,
May 1982,
W. Bahr, P.
AFWAL-TR-i-21U0,
L.,
Fuel Mainburner/Turbine Effects,
AFWAL-TR-b1-208!,
Sept.. 1982.
-
136
-
7.
Lefebvre,
A. H.,
Fuel Effects on Gas Turbine Combustion, AFWAL-TR-83-ZUU4, 1984.
8.
Lefebvre,
A, 1H.,
Fuel Effects on Gas Turbine Combustion-Ignition,
ity,
Stalil-
and Combustion Efficien-
cy presented at 29th ASME International Gas Turbine Conference, 1984.
Lefebvre,
A.
H.,
June
To be publishea in
ASME J. 9.
Amsterdam,
Eng.
Power.
Fuel Effects on Gas Turbine Combustion-Liner Temperature, Pattern Factor and Pollutant Emissions, Aircraft, pp.
10.
Lefebvre,
A. H.,
AIAA Journal of Vol.
887-898,
21,
No.
II,
November 1984.
Gas Turbine Combustion, McGraw Hill, 1983.
11.
Dobbins, and I.
R.
A.,
L.
Crocco,
Glassman,
Measurement of Mean Particle Sizes of Sprays from Diffractively Scattered Light, J.,
Vol.
1886,
1, no.
8,
pp.
AIAA 1882-
1963.
-137P
.".T
" "....."
i
.
.C'.w
.
.'
...
12.
Lorenzetto, A.
G.
E.,
and
Measurements
H. Lefebvre,
a Plain Jet Airblast Atomizer, vol.
AIAA J.,
Roberts,
M. J.
J.
M. and
15.
no.
7, pp.
1977.
1006-1010, 13.
of Drop Size on
Measurement of Droplet Size
Webb,
for Wide Range Particle Distribution, AIAA .. , vol. no.
14.
Chin, A.
J.
S.
and
3,
pp.
583-585,
2,
1964.
Effective Values of Evapora-
H. Lefebvre,
tion Constant for Hydrocarbon Fuel Drops,
Proceedings of
the Twentieth Automotive Technology Development Contractors Coordination Meeting, pp. 15
Ballal, D. R. A.
325-331,
Weak Extinction Limils of Tur-
and
H. Lefebvre,
bulent Flowing Mixtuies, ASME. No.
16.
Lefebvre,
1983.
A. H.,
J. 3,
Eng.
pp.
Power,
343-348,
Vol.
t'
'rans. lul,
1979.
An Evaporation Model for Quenching Distance and Minimum Ignition Energy in Liquid Fuel Sprays,
Paper presented
at Fall Meeting ot Combustion Institute 1977. -
,-
"-."..
138
-
(Eastern Section),
17.
A.
General Model of SparK igni-
and
Ballal, D. R.
tion for Gaseous and Liquid
H. Lefebvre,
Fuel/Air Mixtures,
0
Eighteenth on
Symposium (International) Combustion, 18.
Heat-Transfer
A. H. and
Lefebvre,
pp.
1737-1746,
1981.
Processes in
Gas-Turbine Combustion Cham-
A. V. Herbert,
Proc.
bers, Vol.
174,
Inst. Mech.
No.
12,
pp.
Engrs.,
463-413,
"1960. 19.
R. S.
Fletcher, J.
A Model for Nitric Oxide Emis-
and
sions from Aircraft Gas Tur-
B. Heywood,
bine Engines,
AIAA Paper No.
71-123. 20.
Mosier,
S.
Development and Verification
A. .and
of an Analytical Model for
R. Roberts,
Predicting Emissions from Gas Turbine Engine Comoustors dur-
*
ing Low Power Operation, CP-125, 21.
Roberts,
R.,
R. Kollrack,
L.
Oxide Formation in a Gas Tur-
Teixeira
-
i•. ...°.° ...• •.• `.`•-•-•..`•
1973.
An Analytical Model for Nitric
D. Aceto,
D. P.
AGARI
139
-
.•`.•...-.•x.k-:".-•...'\..••.•.....••`•
\./..`.-...•• .•>•.•.•
.>.>•.
.>.``..•>.• .
Sand
M.
J.
oine Combustor,
Bonneli,
VoL.
10,
No.
t.),
AIAA JourRai,
pp.
b2u-S26,
1972.
22.
A Pollutant Emission Predic
R. J. and
Mador,
tion Model for Gas Turbine
R. Roberts,
Combustors, 74-1113, 23.
Edelman,
R.,
AIAA Paper No.
1974.
A Mathematical Model for the
and
Jet Engine Combustion
"C. Economos,
uollu-
tant Emissions AIAA Paper No. 71-74, 24.
Swithenbank, A. Turan,
1971.
Three-Dimensional,
J.,
Two Phase
Mathematical Modeling of Gas
and P. G.
Turbine Combustors,
Felton,
Gas Turbine
Combustor Design Problems, Hemosphere, 25.
pp.
249-J14,
198u.
Coalescence/Dispersion Modei-
Pratt, D. T.,
ling of Gas Turbine Combustors, Gas Turbine Combustor Design Problems,
Hemisphere,
pp.
315-33u,
1980. 26.
Hung,
W. S.
Accurate Method of Predicting
Y.,
"the Effect of Humidity on Inlected Water on NO x
|?
-140-
Emissions from Industrial Gas Turbines,
ASME Paper No. 1974.
74-WA/GT-6, 27.
Hung,
W. S.
An Experimentally Verified
Y.,
No
x
Emission Model for
Gas Turbine Combustors, Paper No. 28.
Fletcher,
R.
75-GT-71,
ASME
19/'/.
The Control of Oxides of
S.,
R. D. Siegel and
Nitrogen Emissions from Air-
E. K. Bastress,
craft Gas Turbine Engines, NREC 1162, Vols I,
FAA-RD-71-111,
I1, and III,
Northern Research and Engineering Corp., Cambridge, 29.
Hammond,
MA,
1971.
Analytical Predictions of
D. C. (Jr.)
Emissions from and within an
and A. M. Melior,
Allison J-33 Combustor,
Comnus-
tion Science and Technology, Vol. 30.
Mellor,
6,
5,pp.
No.
279-28u,
19/j.
Semi-Empirical Correlations
A. M.,
for Gas Turbine Emissions, Ignition, and Flame Stabiuization, Prog. Sci.,
-
141
-
Vol.
Energy Combust. 6,
pp.
347-3!j8,
19jL.
-31.
Graham, J.
B. Homer and
J.
L.
J.
The Formation and Coagulation
C.,
S.
of Soot Aerosols Generated by the Pyrolosis ot Aromatic
Rosenfeld,
Hydrocarbons,
Prou.
Roy.
344,
pp.
London A, Vol. 259-285, 32.
Graham, •.
B.
5. L.
S.
of Soot Aerosols, Tube Symposium,
Rosen'eld,
pp. 33.
Blazowski,
1978.
The Formation and Coagulatiun
C.,
Homer and J.
Soc.
621-631,
Shock
Int.
10th Proceeaings.
July 1975.
Dependence ot Soot Proauction
W. s.,
on Fuel Blend Characteristics and Combustion Conditions, Eng.
Power,
403-408, 34.
Naegell,
pp.
April 1980.
Structure on Soot Formation in Gas Turbine Engines,
ASME
Gas Turbine Conference,
New
Orle..jns,
36.
102,
Effect of Fuel Molecular
D. W. and
C. A. Moses;,
35.
Vol.
j.
Paper 80-GT-
Shaffernocker,
W. M.
Smoke Measurement Tecnniques,
and Stanforth,
C(. M.,
SAE Paper C80346,
H. arid
Suot Formation in Rich Kero-
Holderness, J.
F.
sine Flames at High Pressure,
3. Mactariarie,
-
0'
1968.
142
4.
Paper No.
18,
Atmospheric
Pollution by Aircraft Engiznes,
AGARD CP-125,
Advisory Group
for Aerospace Research and Development,
-143-
A"O
1973.
LIST OP SYMBOLS
S."
A,
B
in
constants
Eqs.
and (17),
(16)
m
respectively
2
Aan
combustor annulus area,
AL
liner cross-sectional area,
C1
heat flux from combustion gases to liner by W/m2
convection, C2
m2
heat flux from liner wall to annulus air by
W/m2
convection, C3?,C4
constants in Eq.
Dan
hydraulic mean diameter of combustor annulus,
Dh
hydraulic mean diameter of atomizer air duct at exit plane,
D L
(34) in
gm
liner diameter or height, m
D0initial
mean drop size of fuel spray,
um
D p
atomizer prefilmer diameter,
fc fp
fraction of total combustor air employed in comustyion fraction of total combustor air employed in primary-zone
m
combustion ff
fraction of fuel vaporized within combustion zone
k
thermal conductivity,
L
length,
L
or luminosity factor
length of combustion zone,
C,
.
-
V
..
.
m
Le
liner length employed in fuel evaporation,
LL
total liner length, m
LCV
lower calorific value of fuel, MJ/kg
ib
mean beam length of radiation path, m
m
mass flow rate,
,
.
,
•
•
-.
.
'-•.
.
"
.
•
m
kg/s -
'SI
J/ms K
"
'
145. r
•
'
.
,
'.-
--
•
.,
,
"
'-'
r
'"
•
.
Tw l•
*X,
'yr.
rr rW .'
rc,
vy•
r.,\
, C
ww
-.......
.... rrrvrrr .-
'
wY ,
'r
r, rrrwrwr
'-W
..- w r WY'
,r I'Y
reaction order
P
pressure,
AF
pressure differential,
(PL/P3
liner pressure drop av percentage of P 3
q
fuel/air ratio
'q
fuel/air
q
combustor overall fuel/air ratio
qref
reference dynamic head,
qL
fuel/air ratio at lean blowout,
g fuel/kg air
qLL.
fuel/air ratio at lean lightup,
g fuel/kg air
R
radiation heat flux from combustion gases to
rrr .,'r
'
,Wrrr.
kPa
ratio
liner wall,
in
kPa
combustion zone
kPa
W/m W/m4
R2
radiation heat flux from liner to casing,
SMD
Sauter mean diameter of fuel spray,
SN
smoke number
T
temperature,
Tbn
boiling temperature at normal atmospheric pressure,
AT
temperature rise, K
•U
velocity, m/s
VV
total combustion zone volume (-predilution zone volume),
C3
-
"'2
r,
n
.
...
rrw,,
gm
K
Ve
evaporation volume,
.Vpz
pr imary zone volume,
XC
soot concentration, mg/kg gas
Z
consLant in Eq.
C
emissivity
o
Stefan-Boltzmann constant
m m3
(38)
or surface tension, kg/s .146
(5.67 x 10-8 W/m
K),
K
m
nr, .
U
dynamic viscosity, kg/ms
V
kinematic viscosity,
keff 'tc
effective value of evaporation constant, combustion efficiency
Oc
.:ombustion efficiency based on chemical kinetics
m2 /a
mmz/e
9 7
combustion efficiency based on fuel evaporation
c
e
p
density,
3
kg/mr
Subscripts A
air
F
fuel
g
gas
ad
adiabatic value
st
stoichiometric value
c
combustion zone value
an
annulus value
pz
primary zone value
3•z
secondary zone value
max
maximum value
w
wall value
3
combustor inlet
4
combustor outlet value
8
engine discharge value liner value
L
value
-
4j
147
-
U.S.Government P inting Otfice, 1985
-
559-065/206 74
