Embed Size (px)
Citation preview
NASA CONTRACTOR REPORT NASA CR-159421
EMISSION MEASUREMENTS FOR A LEAN PREMIXED PROPANE/AIR SYSTEM
AT PRESSURES UP TO 30 ATMOSPHERES
t
By Gerald Roffe and K. S. Venkataramani
(NASA-CR-159421) EMISSION MEASORESENTS FOR H79-10165A LEAN PBSHIXED PROPANE/AIR SYSTEM ATPRESSURES OP TO 30 ATMOSPHERES Final Report(General Applied Science Labs., Inc.) 11 p OnclasHC A03/HF A01 CSCL 21B G3/25 33851
Prepared by
GENERAL APPLIED SCIENCE LABORATORIES, INC,
VJestbury, New York 11590
For Lewis Research CenterNAS3-20603
NATIONAL AERONAUTICS AND SPACE ADMINISTRATION, WASHINGTON D.C,
JUNE 1978
REPRODUCED BYNATIONAL TECHNICALINFORMATION SERVICE
U.S. DETRIMENT OF COMMERCESPRINGFIELD. VA. 22161
f
•fit?
1. Report No
CR- 1594214. T i t !» j.-.d S . , C T i t i »
oyster; a'_ Pressures up to Jv
2. Government Accession No.
Lea;-. Premised Propane/AirJ A u-nc spheres
7. Author (s )
Gerald R o f f e & K. S. Venkataramani
9. Performing Organ iza t i on Name and Addre«
General Applied Science Lab<Merrick and Stewart AvenuesWestbury , New York 11590
12. Sponsoring Agency Name and AddressAirbreathing Engines Divisi<NASA Lewis Research Center21000 Brookpark Road, Cleve
Dratories, Inc.
3n
Land, Ohio 44135
3
5.
6.
8.
10.
11.
13.
14.
Recipient's Catalog No.
Report DateAUGUST 1978
Per forming Organizat ion Code
Performing Organization Report No.
GASL TR 250
Work Unit No.
Contract or Grant No.
Type of Report and Period Covered
CONTRACTOR REPORTSponsoring Agency Code
15. Supplementary Notes
Final Report, Project Manager, Robert Duerr, Airbreathing Engines Division, NASALewis Research Center, Cleveland, Ohio 44135. Supplemental to "Experimental Studyof the Effect of Cycle Pressures on Lean Combustion Emissions." NASA CR-3032.
16. Abstract
A series of experiments were conducted in which the emissions of a lean premixedsystem of propane/air were measured in a flametube apparatus. Tests were conductedat inlet ter.peratures of 600K and 800K and pressures of 10 atm and 30 atm over arange of equivalence ratios. The data obtained were combined with previous datataken in the sane apparatus to correlate NO^ emissions with operating conditions.NOX emission index was found to be well represented by the expression:
In (- - 72.28 + 2.80/T~~ -38.02
where T is the adiabatic flame temperature (K) and T is the combustor residencetime (msec). The expression is independent of pressure over the range of 5 atmto 30 atm for inlet air temperatures of 727K and higher.
Sampling probe design was found to have a pronounced effect on measured CO levelsbut did not influence NOX measurements. The most effective probe tested was onewhich combined thermal and pressure quenching of the gas sample.
17. Key Words (Suggested by Author ( s ) )
Premixing Gas Turbine CombustorsCombustion Primary ZonesOxides of Nitrogen Gas Sampling Probes
19. St-r-jrity Clsisif . (of tms report)
UNCLASSIFIED
18. Distribution Statement
?0. Security Classif . !of this page)
UNCLASSIFIED1 i
" Fo: ssle by ihe N=!ior.3i T e c h r - c a ! ;n:o;ration Se/v ice. Soiingheld. Vifgima 22161
NASA-C-168 (Rev. 10-75)
TABLE OF CONTENTS
LIST OF FIGURES j j
INTRODUCTION . 1
APPARATUS 3
TEST RIG 3
FUEL SYSTEM AND FUEL PROPERTIES 7
TEST PROCEDURE 7
INSTRUMENTATION • 11
RESULTS 1.6
DISCUSSION OF RESULTS 23
SUMMARY OF RESULTS 31
DATA SUMMARY 32
LIST OF FIGURES
Page
FIG. 1 . COMBUSTION TEST RIG 4
FIG. 2. FUEL INJECTION/FORWARD INSTRUMENT SPOOL ASSEMBLY(VIEWED FROM UPSTREAM SIDE) 5
FIG. 3. WATER-COOLED PERFORATED PLATE FLAMEHOLDER 6
FIG. 4. .DETAILS OF COM8USTOR CONSTRUCTION 7
F I G . 5. PROPANE STORAGE AMD DELIVERY SYSTEM SCHEDULE 9
F I G . 6. THERMAL QUENCH PROBE 13
FIG . 7. PRESSURE REDUCTION SAMPLING PROBE WITH MODERATETHERMAL QUENCH 1*4 '
FIG. 8. PRESSURE/THERMAL-QUENCH PROBE 15
FIG. 9. COMPARISON OF EMISSIONS MEASUREMENTS USING THERMAL,PRESSURE AND PRESSURE/THERMAL QUENCH PROBEDESIGNS (T = 300K, p = 10 atm) 17
FIG. 10. EMISSION MEASUREMENTS 19
FIG. 11. EMISSION MEASUREMENTS 20
F I G . 12. EMISSION MEASUREMENTS 21
FIG. 13. EMISSION MEASUREMENTS 22
FIG. \k. CORRELATION OF NOX EMISSION INDEX FOR 2MSECRESIDENCE TIME WITH ADIABATIC FLAME TEMPERATURE 24(P=30 atm)
FIG. 15 COMPARISON OF PRESENT 30 ATM RESULTS WITH 20 ATMDATA OF REFERENCE (l) 25
F I G . - 1 6 CORRELATION OF NOX EMISSION INDEX FOR 2MSECRESIDENCE TIME WITH ADIABATIC FLAME TEMPERATURE(P=10 atm) 26
FIG. 17 CORRELATION OF NOX EMISSION INDEX FOR 2MSECRESIDENCE TIME WITH ADIABATIC FLAME TEMPERATURE(P=10 atm) 28
FIG. 18 COMPARISON OF PRESENT RESULTS WITH 5 ATM NO DATAOF REFERENCE (1) 29
INTRODUCTION
In a recently completed experimental program described in Reference (l),
the emissions of a lean premixed prevaporized propane-air combustion system
were measured over a matrix of inlet temperature and pressure. The matrix in-
cluded inlet temperatures of 600, 800 and 1000K and pressures of 5, 10, 20
and 30 atm. Matrix points at which data were obtained are indicated in Table I,
be low:
TABLE
><600K
300K
1000K
5 atm
-.';
•':
10 atm
-.';
.'.
20 atm
,•;
,•-
F
30 atm
U
*v
F
Matrix points marked with asterisks in Table I indicate conditions at which
data were obtained. Those points marked with the letter F indicate conditions
at which attempts to operate produced flashback. The letter U at the 30 atm/
600K operating point indicates that combustion at this condition was unstable.
w i t h the flame oscillating between the flarneholder and a station near the ex-
haust throat.
Table I indicates that relatively l i t t l e data was obtained at the 30 atm
operating condition. Since operation at pressures of at least 30 atm is an-
ticipated for future engines, there is considerable interest in obtaining ad-
d i t i o n a l data on LPP combustion at this condition. This report describes a
brief series of experiments in which the premixing combustion apparatus em-
ployed in Reference (1) was modified to produce stable combustion at 30 atm
Ref. 1. Roffe, G. and Venkatarantani, K. S., "Experimental Study of the Effectsof Cycle Pressure on Lean Combustion Emissions," NASA CR 3032, July1978.
pressure. Emissions data were taken at t h i s pressure for entrance temperatures
of 600K and 300K. Emissions neasurements were also made at a pressure of 10
atm, again for entrance temperatures of 600 and 800K, these latter tests
prompted by a desire to obtain a d d i t i o n a l data at a condition where consider-
able data scatter was reported in Reference (1).
-2-
APPARATUS
Test Rig - The combustion test rig is illustrated schematically in Fig-
ure (1). Heated dry air enters the apparatus through the bellmouth, passing
through an instrumentation spool where the entrance temperature and pi tot-
static pressure profiles are measured by an embedded rake. Fuel enters the
device through a heated plenum chamber which surrounds the instrumentation
spool and feeds fifty-two individual 1.6 mm diameter injection tubes. The
tubes extend 7 cm downstream from their entry point and inject fuel in the
streamwise direction to minimize the possibility of local flow separation.
The relatively long and thin injection tubes are supported at their midpoints
by a fine honeycomb structure 6 mm in streamwise extent representing a flow
blockage of 3%- The fuel injector assembly is shown in Figure (2).
The mixer tube is constructed of a heavy outer pressure wall and a thin
stainless steel liner. The two elements are separated by an internally vent-
ed air gap to m i n i m i z e heat loss. Four thermocouples are mounted 90° apart
2.5 cm from the downstream end of the mixer and placed so that their tips are
flush with the inner surface of the liner. The thermocouples serve as i n d i -
cators of autoignition in the mixer or flashback through the flameholder.
The flameholder assembly is illustrated in Figure (3)- The flameholder
is a water-cooled perforated plate, employing 21 holes 0.95 cm in diameter
to produce a porosity of 22%. At the 25 m/s reference velocity used in this
experiment, the flameholder produces a total pressure drop of 3%- It is pro-
vided with two wall surface thermocouples on the downstream surface and one
on the upstream surface and an integral hydrogen-air igniter which is used to
i n i t i a t e combustion. Flameholder depth, measured in the streamwise direction,
is 1.6 cm.
The conbustor assembly also employs a double wall design to protect the
heavy outer pressure wall. Here the air gap between the combustor liner and
•the outer carbon steel wall is kept cool by injecting a small amount of cold
air. In addition, an alumina tube is mounted inside the stainless steel liner
as illustrated in Figure (4) to provide an uncooled refractory combustor w a l l ,
m i n i m i z i n g convective and radiative losses from the gas.
-3-
CC
t-V)
CO
ou
ce
C9
iU£' : ' T^
-5-
Igni terMount
0.3R
Y///7/.
-0.8 Dia. (21 Places) Water Inlet
'1.6cm
-0.25
FIGURE 3. WATER-COOLED PERFORATED PLATE FLAMEHOLDER
SECTION A-A
-6-
ORIGINAL PAGE loOF POOR QUALITY
ooo:o
CDo
oto
cr:=3CJ
-7-
A dome-loaded pressure regulator is used to supply cold air to an an-
nular injection section just upstream of t'ne r;q exit orifice. By loading
the regulator to the pressure desired for the test, the appropriate amount
of cold air is added automatically to produce the correct total pressure in
the test rig. This method of pressure control offers the dual advantages of
automatic compensation for varying combustor exit temperature and thermal
protection for the choked exit orifice.
Fuel System and Fuel Properties - The fuel supply system is illustratedschema't ically in Figure.(5). Liquid propane is stored in a tank and pres-
surized with nitrogen. The l i q u i d is withdrawn from the lower section of the
supply tank, passing through a turbine flowmeter and pressure regulator before
entering a cavitating venturi which provides a constant fuel mass flow rate
independent of downstream pressure fluctuations. Fuel flow rate is controlled
during a test by adjusting the regulated pressure on the upstream side of the
cavitating venturi. The propane is heated to a temperature of 395K in a pres-
surized water bath and passed through a heated line to a metering venturi be-
fore being delivered to the injection plenum. An analysis of the commercial
grade propane used in these experiments is presented in Table II,
Test Procedure - In operation, the air flow through the rig is established
at the desired temperature and at a mass flow rate corresponding to a 25 m/sec
reference velocity at the test pressure and temperature. The rig pressure is
then brought up to the operating value by injection of an appropriate amount
of cold air at the exit orifice. The gas igniter is turned on, fuel flow
initiated and slowly increased u n t i l ignition is achieved. The rig equivalence
ratio is brought to the highest level desired during the particular test se-
quence, the gas igniter shut off and the rig operated to assure steady condi-
tions. Gas samples are then withdrawn at a combustor location corresponding
to a residence time of two milliseconds. The combustor location which corre-
sponds to this residence time is calculated from the 25 m/sec mixer tube ref-
erence velocity and the fuel/air ratio, assuming an instantaneous jump in temper-
ature from the entrance value to the adiabatic flame temperature. Once data
has been obtained at a given equivalence ratio, fuel flow rate is.lowered and
-3-
•oV
Q.a.3
V0)o
>re
Cre
•-9-
TABLE I I
ANALYSIS OF COMMERCIAL GRADE PROPANE USED IN TEST PROGRAM
Property
% Propane
% Butane
% Ethylene and Ethane
% Propylene
% Volatile Sulfur
Specific Gravity (air = 1.0)
Vapor Pressure, MPa
Hydrogen/Carbon Atom Ratio
-10-
the sampling probe moved to the new two millisecond position. The process is
repeated u n t i l the lean s t a b i l i t y l i m i t is reached, either extinguishing the
flame or producing CO levels beyond the range of the instrumentation.
The apparatus used here is essentially identical to that used for the
experiments reported in Reference (1). However, as that rig proved incapable
of holding a steady flame at the 30 atm/600K operating condition, a number of
modifications were made for the present tests. These consisted of increasing
the temperature of the steam bath in the propane vaporization heat exchanger
to 'tlOK and i n s t a l l i n g electric resistance heaters on the fuel distribution
plenum to preclude the p o s s i b i l i t y of liquid-phase fuel reaching the mixer
section; modifying the gas-phase venturi in the fuel line to better isolate
the upstream and downstream flow volumes; and increasing the mass flow of hy-
drogen and air to the maximum stable l i m i t in the gas igniter. A heated gas-
phase fuel storage reservoir was also inserted between the heat exchanger and
the gas-phase venturi. However,, i n i t i a l tests of the rig demonstrated that
w h i l e combustion s t a b i l i t y was unaffected by this reservoir, fuel system re-
sponse time was undesirably increased. The reservoir was removed from the
fuel system prior to final emissions testing.
Instrumentat ion - The design of the sampling probe constitutes an ex-
tremely important element of the experiment, particularly for the high pres-
sure conditions of interest here. It is necessary for chemical reactions in
the gas sample to be slowed to an acceptable level as soon after the sample
enters the probe as possible in order to avoid errors associated with residence
time. This problem is particularly severe in the case of carbon monoxide whose
oxidation reactions are pressure accelerated and further speeded by the high OH
concentrations present. The difficulty here is associated with the sample quench-
ing process. As the sample is cooled, the eq u i l i b r i u m concentration of CO de-
creases. If the sample quenching rate is too slow, CO levels w i l l drop during
the cooling process, attempting to remain in equilibrium.
Three sampling probe designs were evaluated during this test series: one
based entirely on thermal quenching and two based on a combination of pressure
and thermal quenching. The thermal quench probe design is illustrated in
-11-
Figure (6). The probe consists of a hollow V.27 cm diameter sheath terminating
in a hemispheric tip through which a 1 mm diameter sampling hole is d r i l l e d .
The gas sample flows through the tip into a 1.6 mm diameter tube which carries
it out of the probe. The 1.6 mm sample tube is Jacketed by an intermediate
tube through which cooling water enters the probe. The water flows past the
probe tip, returning through the annular gap between the intermediate tube and
the probe sheath. The cooling water is exhausted outside the combustion appara-
tus. The flow path of the cooling water is such that its temperature is as low
as possible where it contacts the sample delivery tube. In addition, the small
delivery tube diameter produces rapid cooling of the gas which it.carries.
Figure (7) illustrates the first pressure-quench sampling probe design
which employs only a moderate degree of thermal quenching. Here, the gas enter-
ing the probe through the 1.6 mm sampling port is expanded into a 6 mm diameter
dump tube w i t h i n which the pressure is maintained at 2 atm. Since the reactions
of interest here are all pressure-accelerated, the rapid drop in pressure pro-
duces an immediate reduction in reaction rates. Shortly after the expansion
station, a portion of the gas is withdrawn through a 1..6 mm sampling tube im-
mersed in the water cooling passage between the probe sheath and the dump tube.
In this design, the cooling water moves in on-ly one direction and is exhausted
into the combustor through holes in the sheath.
The third design, illustrated in Figure (8), employs a greater degree of
thermal quench along with the basic pressure-quench technique. Here, gas from
the 1 mm sampling port expands into a 3-2 mm dump tube w i t h i n which the pressure
is maintained at two atmospheres. A portion of this gas is withdrawn through a
1.6 mm quenching tube which extends through a water delivery tube to the probe
exit. The sample quenching tube protrudes to the centerline of the dump tube
to avoid capturing gas which may have been either recirculated or wa11-cata1yzed.
The water turns at the probe tip and returns to the probe exit as it cools the
outer sheath.
The gas sample analysis system and data reduction techniques are described
in detail in Reference (l). The entire system conforms to the requirements of
SAE ARP 1256.
-12-
ORIGINAL PAGE ISOF POOR QUALITY
Sheath (1.2? Dia)
Water Inlet Tube
0.1 Dia. .-'" * T" ' ' " ' " "—/*-water Out/»v__ :::;. ;__, _W~_J_L , ,_ -4--Water In
• •-— ••- - psi r--— . {
1.27 Dial -yv izi m._:/ Sample Tube
Probe Tip
(Dimensions in cm.)
FIGURE 6. THERMAL QUENCH PROBE
-13-
Sheath (1.27 Dia.) . . Sample Tube (.16 Oia.
T .
I
16 D i a .
Water Out
Sample Out
To Atmospheric Dump
Water In
~ Dump Tube (0.6 Dia.)
(Dimensions in cm.)
FIGURE 7 PRESSURE REDUCTION SAMPLING PROBE WITH
MODERATE THERMAL QUENCH
-1A-
ORIGINAL PAGE ISOK POOR QUALITY*
Sheath (1.27 Di
Sample Dump (0.32 Dia),
/ ///_///. / / / / / / /~
Sample Tube (0.16 Dia)
(Dimensions in cm.)
FIGURE 8. PRESSURE/THERMAL-QUENCH PROBE
-15-
RESULTS
The three sampling probes were f i r s t tested at inlet conditions of 10
atm and 800K using an uncooled perforated plate flameholder to determine the
most suitable design for higher pressure testing. Each probe was mounted at
a fixed position 30 cm downstream of the flameholder and gas 'samples withdrawn
as equivalence ratio was varied between 0.35 and 0-75- The test sequence was
repeated for each of the probes. The results of these tests are presented in
Figure (9).
Measured CO level is seen to be a strong function of probe design. The
pressure-reduction probe with moderate thermal quench delivers the lowest CO
concentration to the analyzers. The thermal-quench probe produces higher CO
levels, although its performance is s i m i l a r to that of the pressure-quench de-
sign at the higher equivalence ratios where heating load increases. The
pressure/thermal quench probe delivers by far the highest CO levels to the
analyzer, particularly at high equivalence ratio.
Measured NO level is seen to be independent of probe design, repeating re-
markably well for the three test sequences. Unburned hydrocarbon levels,
which are below the minimum accuracy level of the hydrocarbon analyzer u n t i l
just before the lean blowout l i m i t , do not provide a suitable medium of com-
parison although the data obtained using the three probes does appear con-
sistent.
From these results it would appear that although NO and UHC measure-X
ments are not strong functions of sampling probe design, CO measurements most
d e f i n i t e l y are. The large difference between the pressure/thermal and straight
thermal quenching design indicates the performance improvement which accrues
from providing a rapid pressure reduction. The poor performance of the pressure-
quench design employing only moderate thermal quenching indicates that con-
siderable care must be taken to assure that the pressure reduction step is followed
by effective sampling cooling. Based on the results of these probe evaluation
measurements, the pressure/thermal quench sampling probe was selected for use in
the remainder of the test program although even this design delivers CO levels
which were sornev;'nat lower than those for chemical e q u i l i b r i u m under conditions of
ORIGINAL PAGE IS
Ov POOR
L
r
IDcc<Uoctu
ID
00
0)
Dcn
(1) -g
0) C- Q U O )O -O 3\- o o1
D. I.O- I
JZ
0 ^ —C O ID0) C E3(1 )1 -01 3 0)
a)i. —3 <DU) Ein0)
u in0) 0)
a. I— o_
<aoa$<ul
toCO
QCQ.
oz<UJa:COCO
0£.Q-
a:UJ
I- E•u
(3 a.
— oCO i-
UJ .z: ^UJ Oo: o3 ooCO< iiUJ
CO
O •— coco -z.CO C3
s: coUJ UJ
au.O LU
CD2 Oo ceCO Q.
o; x< oa. -z.f. LUo •=>o o1
LUCC
CJ
U O | S S | U J 3
very high adiabatic flame temperature where one normally presumes chemical
e q u i l i b r i u m to exist.
The emissions measurements at 30 atm pressure at 600K and 800K entrance
temperature are presented in Figures (10) and (11). Measurements at 10 atm
pressure and 600K and 800K entrance temperature are presented in'Figures (12)
and (13)- All data is summarized in tabular form in the Appendix.
PAGE ISQUALITY/
I
lo
o
m
3CT
LU
oovD
II
CU1_
i_,(DiOD.E
H
E
(U
0
IIfl)1_3I/II/I
U0-
o(1)I/IE
(NJ
II
01E
J-
(Uoc<u
ool-zLU
LUD£)
t/1<LU1£
~r
0—COto^3:LU
d) d) .-— — inC C II
— — o;
O
oO
X3QNI NOISSIW3 -19-
(TJOH
uc0>
ooCO
II(U1_D
i.0)Q.
OJ
E4-1mo
Din
cui_Q.
o0)1/1e
CM
II
<LUZ
ento
4)UC0)
4-J 4-J "D4) 4) .-
_ _a:rsCJ
Oo
X3QNI N O I S S I W 3-20-
•-INAL PAGE IS;;i XJR QUALITY
oo
CO
O
r-
o
Oo -1°
Xo
-
0 0o —
1LT1
O
CO
O
r-o
tno
TOa:
uc
DCT
OO
0)
3
0)a.
e4J
(0
O
Dl/l(/>0)L.a.
zLUDC
inE
<u
0)uc0)
— — CC.
CO
o
vO
o
CO(/I
X3QNI N O I S S I W 3 -21-
O
oo
o
ino
CO
o o4-*03
80)
>
cr
LTV
O
Oo00
01>-
n.E<u
4-1<0
O
II
O)
u01
eO-J
II
0)uc
01 O) —— — I/IC C 4)
— — o:
00
LU
UJQ£
CO00
LUCC
O00
o
uno
'oo
X3QNI NOISSIW3 -22-
DISCUSSION OF RESULTS
Since MOX production is basically a post-flame reaction, one would
expect it to be a strong function of adiabatic flame temperature. Accord-
i n g l y , the (JO emission index measured at 30 atm has been replotted as a
function of adiabatic flame temperature with the results shown in Figure
(14). Here, the present data for entrance temperature of 600K and BOOK
are seen to collapse quite nicely to a single curve. The l i m i t e d 30 atm
data reported in Reference (1) is also shown in Figure (14) and.can be seen
to f a l l somewhat below the present data curve. Although the cause of this
discrepancy is not immediately obvious, the fact that Reference (l) reports
30 atm combustion to have become unstable at an inlet temperature of 600K
wh i l e no such i n s t a b i l i t y manifested itself in the present tests suggests
the p o s s i b i l i t y of. poorly stabilized combust ion in Reference (1) and a re-
s u l t i n g decrease in residence time at the adiabatic flame temperature.
Figure (15) compares the present 30 atm data curve with the 20 atm
data reported in Reference (l)- Here, one again sees good collapse of the
data when plotted as a function of adiabatic flame temperature and excellent
agreement between the 20 atm and 30 atm results, indicating no effect of
pressure at these operating conditions.
In Figure (16) the results of the 10 atmosphere tests are plotted as
a function of adiabatic flame temperature. In addition to tests at inlet
temperatures of 600K.and 800K, data is also presented in Figure (16) for an
inle t temperature of 727K. These latter data were acquired when the test rig
temperature controller malfunctioned, d e l i v e r i n g air to the rig at a tempera-
ture of 72?K rather than the 800K which had been desired. Since the sample probe
position during t h i s run was one which would have produced a residence time of
two milliseconds had the i n l e t temperature been 800K, it was necessary to apply
a small correction to the data so that it could be compared with the other
results on the basis of equal residence time. This was done by computing the
actual residence time at the probe station and m u l t i p l y i n g the measured emis-
-23-
"T
!
U
Q.
bi-
o—I o
o oo oo oo
I o-I O
D
<D1_D4-tro
Q.E
O 4>O ECM roCM —
LUo
a:
<_>UJ
CM
UJU Cc J cOJ I T
(0-QIB
O —O "^J
0)
ui 0)— ITO-a EE O>- i-
o. roo -o
.J I 1 I
XLU LJJo a:
z o:o LU— D.oo r;to uj
U. <_)o —•z. <O O3
< Q-I <LUa: zoc t-o —
ooCO
oo
01-' POOP.-2k-
OKIGINAL PAGE kUF POOR QUALITY
i-
oL.
a)
<u
c.5
l —*-*Dc.
**-
(0
5,
x^0ov^
o
ocQ>1-0)
14-
0o00
1 ;
I
0, o
i ^
'. O-1 1• ^
4)L-
4-J(D
OCM
O-
OJ
! O (DI csj —
rsl Li-
roJ3fO
1OOO
J I I I I M i l
OO00
<cO
o:o.
o LU</> o— :za: uj< tx-Q. LUS Ll-O LU<-> a:
LUa:o
oo
>opu| ON
r
h-
oo
ZLU.O
OLU
Oo
CM
06
ooCMCNI
ooo
<UQ.
<a
<DTO
OOCO
xS
C C£..— -=>
CO \-(/> <— az uuLU Q-
XLUO t-
o <
— t;
SiLU G3ec <cc —O QO <
vO
LUCC
C3
Lu
oo
(ianj-6>|/1 'ON-&) xapu| uoissiiug ON
o?
-26-
sion index by the ratio of two m i l l i s e c o n d s to t h i s tine. i h i i method of
correction assunes a li n e a r rate of NO. production •.-:! th t i m e and is based
on the measurements of emission index as a function of residence time docu-
mented in Reference (l). The maxinum correction applied was only k~, and
corresponds to the a d i a b a t i c flame temperature error near the lean s t a b i l i t y
l i m i t .
The 10 atmosphere data of Figure (16) shows a good collapse of the 727K
and SO'OK measurements which f a l l very close to the 30 atm data curve of Figure
(1*0. The 20 atm/600K data is conspicuous in that it does not collapse to the
higher pressure curve, rather running p a r a l l e l to it but displaced downward.
This same data is repeated in Figure (17), thi s time adding the 10 atmosphere
data obtained previously and reported in Reference (1). The data clearly ap-
proach the 30 atmosphere curve for inlet temperatures of 72/K and h.igher. The
previous data for i n l e t temperatures of 600K and 800K are lower than those ob-
tained in the current tests. It would appear that the a b i l i t y to anchor the
flame t i g h t l y at the flameholder decreases as the pressure and i n l e t temperature
decrease. The present series of experiments, employing a more powerful i g n i t e r
than was used in Reference (1), apparently produce a more positively s t a b i l i z e d
flame. In cases where the flame is not t i g h t l y anchored to the flarehc'der, the
assumption of an instantaneious temperature rise in the residence, time calcula-
tion is clearly i n v a l i d . Thus, the downward s h i f t in NO level may simply be a
reflection of a lower residence time at.the adiabatic flame temperature.
Figure (18) compares the present results with data reported in Reference
(1) for an i n l e t pressure of 5 atmospheres. Here again, the higher i n l e t tem-
perature data collapse to the 30 atmosphere data curve while the 600K i n l e t tem-
perature data falls below, this time to an even greater degree than occurred at
the higher pressure.
From an examination of the data summarized in Figures (1*0 through (18),
it would appear that NO emission index is p r i n c i p a l l y a function of adiabaticA
flame temperature and is not strongly influenced by either equivalence ratio
or i n l e t temperature except insofar as the combination of these two variables
. -27-
I I
TOOO-3- i.CM aj
c.0)
OOoCM
OOCO
oUJo
LLJDC
P "U- Q.
XUJ UJQ C£.
z a:O UJ— Q.
UJ UJ
x<c _i
U- CJo —z <O CO
< o
c —<-> 3
Oo
-28-
1s
UJoc
C£.
U-O
Xo
UJOC
00UJ
o00
a.oo
CO
cc13
-29-
determines the adiabatic flame temperature. Since data reported in Reference
(l) indicated that NOX levels were directly proportional to combustor residence
time, the results of this program can be summarized by the following expression:
VIn (--^) = - 72 .28 + 2.30/T
where ENQ is the NO emission index (g•N0,/kg-fuel), T is the ad iabat ic f lameX X £.
temperature (K) and T is the combustor residence time (msec). Over the range
of pressures from 5 to 30 atmospheres, there is no s i g n i f i c a n t observed departure
from this expression for i n l e t temperatures of 727K and higher.
-30-
SUMMARY OF RESULTS
1. NO emission index is p r i n c i p a l l y a function of adiabatic flame temperature
and cornbustor residence time. This function is well represented by the
expression
,n (^) = - 72.23 + 2.80/T-
where T is the adiabatic flame temperature (K) and i is the combustor resi-
dence t ime (msec).
2. Over the range of pressures from 5 to 30 atm, there is no significant
effect of pressure on NO emission index as a function of adiabatic flame
temperature for i n l e t temperatures above 727K.
3- The three sampling probe designs tested in this study had a pronounced
effect on measured CO levels but did not influence NO measurements. The •x
most effective probe design for the high pressure conditions prevailing in
this study was one which combined thermal and pressure-quenching of the gas
sample.
-31-
Q.O E
.- flj•a —< u.
o<_>
d)
iai
<u
IDl
0)
u_ '1CT1
X -*o ^z cn
LU • — •
t~~» cn CM cn CM oo r-«.-3" Cn *~~ VO \^ LA -3"
— O CM — O O O
cn-3"
0
E4-1
10
o cn —CM ^ .-vj
Ocno
O O O L Ar - - L A c N v ^ 3oo oo co co
O O O O O O L A Ovo fs ^o — r^ P^ un r^(N l tN l — — — O J C N I C M
CPA
o oCM cr\r-J —
\ PA r*»
CM — — —CO O
<sj oo
cr\ -3-o —
oCSI
— LA IA OO
CVJ — — —
CTl
o o o o • oo 0 0 ^ -
LA OO
O CM
-3-
CM
O
CM —
-3-O -3-
LArr>
CM —o~\ ro
-3"
I—
>•
sZJCO
<I —<
<h~<
GO
~Z.oGO
z:LU
<
s:o.
0 0-O >™*'
o^0 D_
Z Q.
^-
X Q.0 0_
O1
CMLA
Q
f~»
cn00
LAPA
0
00
LAO
-3"
PACM
OCMCM
PA
xD1
CMPA
P-.LA
0CMCM
PA
>_Q
r-
a\CM
PA-3-
000
ooLAO
cnCM
PA« —
0 0oo o
"
CO -3"
LA PAO CM
**o o— r-CM cn•—
0
r~
cncnC^J
cnPA
0
LA LA OLA v£> -3-
CM
OCM
LA LA LAr*» r^ cn
OOCM
— O
Orooo
LAr-v
r-.CM
^O ^D
O VO
O
vD -3"O
O-3-\O
O
oo
CM rjn— o
cn LAxO LA^O LA
-3-LA
o CMPA UALA LA
1COxOPA
CM
- r-
cn -3-VO LA
PA O\O ^-*^
-3"
-3-J-
LAr~.LA
CM\o oo^D ^^CM \O
-T
OO ^^O CMLA cn
r-PA CO
\o cn\D cn
vD ^«Dr*" \o
ov£>-3-PA
vO
1 —
—.cnCM
CM-3"r~-
CO
CM0CM
cncn
cncnLA
i*
°.
cnCMJ-
OLA
°.
OO
ooCO
oPA
-32-
o.u E
0)E
< LL.
LUIat
Ia>
01
O O O O LA Or-^ -T LA LA O PAo o cn co cc r-^CM CM r- — — —
co co oo CM cr>
O •— -J" N»O OCM CM — — rxj
-3" \D OO
CM CNI —
-T CTi
CM CSI
-3"
LA
oPA
CM
-3-
CM
O
LA O LA-3- PA -3-PA PA PACM CM CM
LA OO SOIA -3" OO
• 0O 0
LA O LA Oi — i — . i — cnPA PA CM CMCM CM CM CM
sO O OO -3-O -3" CM OO
cn soo —
0 • Oo o
oCM0CM
-3-
LA
0CM
O
0-3"cn
O
oo
P-.CM
O
0
cn
cn-3-
LALA
O
O
CO
LT\
00SO
O
OLA00
00
CM
-oDC
• —4-1
C
0o
cc<£2:
I/)
t-
o
<
•3.O
CO
ocoto
•s.LU
«
0!D
LL.
101
X ^0 ^z cn
LU - — •
f yQ-
O O_(_) ^-^
f ^O
0 0-•x. Q.iD ^—^
^~*
X. D-O Q-
P^ P** CO SO P^ P** - " -3" CA P1^ - " -3"C M L A P ~ C n P ~ - 3 - C M C n C n P A O O C O s D - T s O C M
P A P A - — o o o o o o o c n c n c n c n P A C Mw— *— f— ~-
L A O O O L A L A O O O O O L A L A O O L As o o o c n o o c n r - - CM — LA CM I — L A C M O O O OP^ P" -3" -CT LA CO f P^ CA -3" PA J" PA -3" CM CM
SO OO CM PAp A p ^ o s o c n c n — CM LA oo __ y£>
O L A C n O L A - 3 " — C O O O O J - - 3 - P ^ -O O C O — - 3 - s O O O C M C N — _ _ _
r*^ cn CM — so cnP A O O t ^ - 3 - - 3 ' - 3 - OO SO
P A P ^ L A P ^ - P A P ^ C M P A C M s O - 3 - O C M O O O < rp^ i PA «~ — cn •— • — cn i — • cn so so ' • «3*
CM -3" P^SO LA PA
r— ^ 1—
LA O LAP-~ p^ —- -3" -3"
vO — —CM - PA OO
LA CM LAPA J- P--
sO OO OO-3- — 0
— cn LAC N J p ^ r A c M C M C S I C M C M
-e- ^c_j
PA sO ^~ P^» sO ^~OO vO CM sO -3" •—LA LA IA -3" -3" -3"
P- LA LA LA SO LA OO- 3 - L A - 3 - L A P A S O O — L A — C n C O s O
o o o o o o o o o o o o o
oo00
oooo
PA £ oPA
-33-
o.UE
fI3 *—•.JDD ^r• -03T3 —
O O O O O L A O O L AC O O — C O O r<-\ r o O C OCM <NI — — O CT\ <Ti CO r~CM CM CNI CM CM •— '— '— ~-
\D LA —
rA OPA -3-
-3~ O
— O
cno LA
LA
I0)
O O O
\D -3"— CM
00-3-
OOCO
CO(NJ
— — O
-o•3>y£,.*UJ
Cn
c5
<5f<3;£$
<aH-*io
<£t-<to
55oCO00
—rLU
<
f ^
rv0 Q-t_j — -
^^O7"
(_j n^I Cu
^^
X Cu0 Cu
O O O OCA LA CM XOLA r r~- CM_ — CA Jf
rA O PACTv — CTv
O
— CM —
M3PA PA CM CO
PA rA — CMr^ r^ o Co
Of"\
COPA
vD
.0
o-3-
-3-LA
Oj-r
CM
\D
P—>O
1 —I —COCM
O-TOCM
vD1
O' —
LACM
vi5CM
LA O*~~ -~T•— \Q
CM —
PA OO^ CA
• •r- > \o— CM
-3- CACM CO
PA —
— —
CM CO
O O
LA LA LA . -3! -T
O O O O O
E4-1
fI3
o£
o ~<-> -
LU **v^
01
a;U-
00 0 1LU i ci03 LU -*o . ^~ce enCL. • — •
oLU . — .
O* ' Q)
_l Lu< 1s c^fy- J^ >/
LU O -s.ZC Z Ol|_ LU -
LUce
oo - — •00 2TLU LUa: o LUd. o —
o
LU Oa: 2:•=> ocuOO " Q-00 3 •UJa:Q.
_r —•z. x. a-C£ O Q.LU ^
pL4-0^ LU0 -9-31
—a.
0 0 -O .O >- (U
C *-• CT.1) — qj
CQ 1- o -1) O O
« o11 CD —
C£ > CO
-Qo1_0.
- a - r A j - < r o v r ) c n c n v O L A j - - 3 - f x i j - v o c n
O O O O O O O O O O O O O O O - J
-3-o o o o o o o o o o o o o o o
o
ex) P"""" CNI ^^ "— 'O*\ ^~ c*»j CNI o^ r*^ <o *-f^ r* - o*j o \fO -TJ" ^5 O t l- CO ^— LO ^^ ^D ^~ ^~ ^5 C5 O~l ^~
L A - = - r ~ - r ^ C n I - - ; - C V I < ^ C O v O L A P A < N | - -
LA O LA LA LA LA O O LA O LA LA LA O LA LA
0O O O O O O O O O O O O O O O
01
Is- LA LA O~l CNI
O f^- ^D PA - 3 " ' — f — c s l O O P A C O x O - 3 ' ' — P-~
• — i — — CM. — C M P A C s l — •— —
CO O~i CA CO -3" -3" CM T-- r ^ L A P A — C T \ ( M 1 — ^COCO ^^ I~~ PA -3" PA CO l~~ PA *~~- vO LA ^^ PA CM v^
o o o o o o o o o o o o o o o o
V,
E O
U '1 2 - 7« o <y 2 i -5 3
S § ! -g £ g £••c- c; c_ 5 -f] ^ ^
CTv CO
vC r~-
r~ oo
o o
O CTi
— o
o ovQ CO
PA -3"PA O
CM OOPA PA
— vD-3- LA
-3- PA
O —
PA PA
O O
-35-
T3
C._4-1
Co
~—
cc.<^2:z:to
<^1—Q
inLUCOOOiQ.
LU
O1
|
z:LU
h-
LU
toUJOiO_
£3•z.
LUCC— N
totoLUQi0-
„
1
*—2.
LU
1—
LuO
•z.O00
5Q-
0O
CO
dJ3
Lu
01
o \<-> 011 1 1 ^_^
Lu1Olv
<-J ^X 01
UJ ^^
^^^
ljD
^j
I
^ s/
ox ^•z. 011 1 1 ^_^
"z?Q_
O Q.tJ
.O*y*
C_J Q_T" Q_
Z3 *--
- — .
X Q-O Q-Z - —
2-LU
-s- iO
4) -0 >- 4)C 4-> Ol<l) — ID1- 0 .*4) O O
M- — O4) 4) • —
CC > CQ
4)-QO
OO — CO OO 1 — CM — O OO CM
— _ _ _ _ _ _ _ _ O L A
o o o o o o o o o o
<T\ ' — OO h^ OO • — -3" OO • — -3"O O r s / — L A v j S - 3 " f ~ ~ L n v D f M
O O ~ v - 3 ' - T O O O L A - 3 ' C M —' — ' — "~ — ' —
L A L A L A L A L A O C O - 3 - O O L AL A - 3 " O O O O r ~ . L A O O O O c M C M
r—
O O O O O O O O O O
OO CO OOo-j PO on -j* ^T csj r*N. r CM -~
o^» vD r*1^ . — ^o c\ ir\ i~r\ \o ooL T \ C M O — r ^ . C T \ C M C T l - T ' —CM CM -a- -3- oo CN '—
C N - — c o c N f s j o a ^ o o ^ ^ ^ ^Lf \ -3" L T v - 3 - CM L T \ O - 3 " L T \ r * ^\ \ r**» r**» r**» \o u*% LO - r oo
o o o o . o o o o o o
1/1*" ij 1
E 0
OO
o"
(D JZ 1-E U 4) 3U, C J3 CD4) 4) O —
cn cn— -3"
r- co- •
CM OO
MD LACO CO
0 0
0 O1 — . OOCM OO
-3- —• •
-3" rCM VO1 — 1—
r—
OO LA
, — , —i— >~-
-3" VOLA -3"OO OO
O 0
h- O1 Q-
I
5§
LALUCOOccC-
o1
2:o:
coLU
CL.
Q
LUo:
onCO
cnD.
<C£
o
o
cz
a.
o
DU.
O 1<_> CT.
LU -i
CT)
O3
U-0 1Z CT
LU -*
CD
W
LL.1cn
0 ^
UJ - — •'
0.0 0.(_) • — •
O
IE O.
x Sio a.
LU-e-a:
Re
fere
nc
eV
elo
ci
ty,
Blo
cka
ge
-Opa.
cr. CM o
co oo r-»
o o o
LO r-~ co**o co ro-3" ro —
O O O0 vO 0-3- CO ro
-3" p"" r*-*ro 0 0
O CM CM
o f» en-3" ro rsj
vD CM -3-— O COi — r~- VD
o o o
£ 0LO 1 —
•^ roi- — a)3 (TJ JZ i_l/> P O <U 3i . i- c J3 cn<D O (U O •-V_ _^ 3 ^- Li_0. 1— O' Cu ^
ro O CO CM O -3" O
LP» ro . — . — — LO vO
— CM CM CM CM CO CO
o o o o o o o
-3- CM vD LO CM Cn —CM -3- cn co r— .3- CM
O CO -3" CM CM — —
O O LO LO O LO OCsl -^f \O f*^ f*"^ f^ U^CM — — •—
r*. r^ oo co oo oo coo o r*** r**- r*** r*** r **
CM CM ro ro ro ro ro
— in r--.ro vD LT\ \£ r* - co Of^- vO \O CM O"\ CM OLA CT\ O LA -T CM CM
CM — —
xO O^ oO r**» ro - " C"\LT\ O LT\ r* . F"^ CT\ COsO vD LA *j~ -3- ro ro
o o o o o o o
-37-
