EFFECT ON NOX EMISSIONS OF AIR-PREHEATING

advertisement
lssN0i284740
BLrI-ETIN
FKKKSA8(l) : 50 , 1994
EFFECTOF AIR-PREHEATING
ON NOXEMISSIONSFROM
A CAS TURBINECOMBUSTOR
AiremanMustafa
(Dep!.ofPolymer& GasEnginerring, UTM)
ABSTRACT
A 76 mtn diatnetercan combwtor equippedwith a 45o curyed-bladetadial swirler was
used to stu.dythe efect of ligh inlet temperaturegas turbine combustion onNOa emission
characteristics. HiEh tnlet temperaturecombustionwas simulated usingpreheet air at
400 oK, 600 oK, 740 oK end 900 aK. Natwal eas was investigatedusing pqssage injecion
technique.It was denonstrated that increasingthe ihlet air temperaturesighificontly
widetuedthe weak extinction limit and iwroved the NOa emission ch,aracteisrtcs. An
optimurn Noxemission of less than 4-5 ppm at ISE7 orlgen, compaible with 99Va
fficiency was dertonstra.tedat 900 K i let temperature. It was also demonstrated thlrt
with high inlet temperature operstion leon wellmixed cornbustionrequired for an
adequate control of NO, emissions could be achieved. The major conchrsion reached is
that high inlet temperaturecotnbustioncould be a promising approachfor low NO,
emissionswith htgh combwtion efficiency qnd goodllame stability.
INTRODUCTION
As Malaysia moves towards becoming a developedcountry by rhe year 2020 more energy
intensive industries are expectedto emergeover the next 20 years. Consequendythere are
the ne€dsfor more efficient power generation.Gas tubines, in particular natural gas-fued
gasturbine systems,wi.ll be playing a key rcle of the text generarionof powerplanrsAs the demand for electdcity inqeases most powerplants using gas twbine systems will
opemte at full loads. The need to producemore electicity at peaking service requires the
turbine combustor to operate at higher combustionintensity and this can be achieved by
increasing the fuel-air mixture sfrength.In ofde! to inqease the air intake the combustor
has to operute at high compression ratio. Trend irl combustor operaring conditions
indicates a steady increase in inlet temperaturedue to increasing compression ratios (l).
At high combustior intensity harmful emission products in the exhaust gases consists
primarily of oxides of nitrogen (NO,). NO, consistsof 90-9570NO (nitrogen oxide) and
the balance NO2 (nitrogen dioxide). As a pollutan! NO" is thought ro be harmful in
severalways - as a contribuor to so-calledacid rain, as a destroyer of the ozone layet and
as a heat-fapping compound suspectedof increasing atmospheric temperatures-Ar low
elevations, NO also reacts with sunlight to create smog. As the rate of NO, formation is
strongly temperature-dependentany chalge in flame temperaturewill cenainly alfect the
levels of NO" emissions in the exhaust gases.The combustor idet tempemlrrc, ro some
extent, plays an important rcle in determiningthe flame temperatureand hence the
quantity of NO" emissionsin the exhaustgases. This paper aims to investigareand
discuss the influence of high inlet temperaturegas turbine combustion from rhe
aerodynamic
viewpointon NO* emissions
usinganultralow NOxgasturbinecombustor.
-50-
rssN01284740
BULETIN
FKKKSA8(l) ; 51 , 1994
OXIDES OF NITROGEN (NOX)
Oxidesof nitrogenwhich are formedduringcombustioninvolving air, commonlyreferred
to NO*, consistprimadly of Nitric oxide (NO) and lesseramounrsof Nitrogen dioxide
(NOz).
The lt{ecahnism of NOx Formation
Nitric Oxide (NO)
Niliic oxide (NO) cer be formed-b)dbEe-different
n]echanisms;
-Cornbustion-gene.ated
(i)
Thermal NO is formed by oxidation of atmosphericnitrogen in the posrfiame
gases(?)The well establishedmechanismof NO formatiqn in the combustion
plocesses was suggestedby Zeidovich(3).The main reaciion goveming rhe
formation of NO in fuel lean mixtures is the breakup of tire strong triple bond
holding N? molecule togedrenThis is accomplishedby the ftee oxygen from the
equilibrium dissociation of unbumt oxygen moleculeswhich iniriate rhe chain.
ii)
Prompt NO is prodlced by high speedreacrions ar the flame front in hydrocarbon
fuels(a)-Fenomore{s)showedthat NO was formed by enhancedreactionraresas a
result of interactions among the many intemediate hydl-ocarbonspecies. Claypole
and S1'red(6)
who invesiigatedthe influenceof levelsof swirl aerodynaq:Lics
on NOa
formation found that the major arcaof NOx formation wai at the reaction zone in tie
flame front. The moderate flame temperatureand the rapid formation of NO
ind.icate{i that NO was formed via the promt mechanism.
iii)
Fuel NO is produced by oxidation of nibogen contained in the fuel. The fuet
nitogen is converted to HCN in the flame zone (7) and depending on rhe degee of
nitrogen conve$ions the fuel NO can represent a considerable portion of the total
NO. The wolk of Appleton6)on vane swirled combustorat atmosphericpressue
sholgd &ar t!9 degee of fuevat mixhg was a major factor affecring the degreeof
conversion of fuel nihsen to NO.
Nitrogen Dioxide (NO2)
NO2 which is the major sourco of atmosphedcpollurant is formed by oxidation of NO in
the atmosphere.This is becauseat low temperatue N(}2 is more stable than NO. Normally
NO; emissions from combustion processesare dominarcd by NO. However evidencehas
shown that NO2 can be formed in the combustor where a lar:geamouot of excessair is
(9).
Preseot
Operating parametersaffecting NOx formation
The rcladve importance of the various gas turbine combustor operating parameters
affecting NO, formation has beenexamhed extensively in tie past few decades(2-9).The
results of these studies have shown that NOx formation is reaction-rateconfolled, a
function of flame temperature,residenceof the combustion gases,concentation of
nitrogen and oxygen present,and to a leseerdegree,the combustorpressure.Condition
ISSN0128{740
BULETINFKKKSA8(l) : 52 , 1994
f a v a o u r a b l e to N o xfo rma ti o n a rch ighflam etempelatule,]ongr esidencetime' hi gh
ther'le of reaction
increasesas
a"a f i*n oxygenavailabilityFlametemperalure
value
the
stoichiometric
"litt"*
zone fuel-airratio approaches
i".t"r..i ,"a it tti"p.l.-y
referencevelocities'
i-"tia"t"" time is aff;ted bi combustorprimary zone length€nd
the No"
air
also
encourages
fuel
and
the
mixing between
i.t"ifili*ifomration. "*.arnamic
EXPERIMENTAL APPARATUS AND TEST PROCEDURES
pressuretest g consistsof a 330nmloog 76mfi diaDeteruncooledcan
The atmosDheric
f""a blo*'er' eleitrical air preheater,otifice lrletering-device' l 5 m long
water
"lt
""-U"it*
ai"-"r"t approachpipe and meansa$Ple probe There is also a 152mm
ii"t-
windowon rhecoDbusrorcenterline. A
eiiiai-"rri".rr, pii" *nr,'in observarion
equipped
t"rt".^Oa f"y.t,'t'f the test rig is shownin Figure 1' The combustorwas also
with wall pressue tappings and thermocouples'
orifice
Combustion air and fuel were suppLiedby an air blower via a 42 mm-diameter
pressule
and
static
wall
temperature
rD"* una u tor"*","t, rcspectivet Inlet and oudet,
by un eie"ttonic nlicromanometerand were recorded in the data
i"i"
ihe preheatair was introducedin the Primaiy zonetbrougha swirler'
-t"it.i"J
.I.t
""qrititio;tytt"-.
ot^t t.ut"a ,o the required combustorinlet temperatureby an electrical heater' The
"*uit
*"s measuied using a chromel-alumelK thermocouplemounted 100mm
int-o1"o,p"tu,ot"
upstream$e flame stabilizer.
Experimental settings
Air and fuel were Partially Femixed in rhe vane passageof a 45o blade angle radial
loss_of 2.770) prior to entering the
t a swiil numbei of 0.54 and a pressure_
.*-i.ili<*i
fit t"sts were caried out a constantMach (M) of 0'03' M is defined as the
o;-u."
now velocity to the velociry of sound,as given by the equation below;
iutio o] ttr"
-o".
-"-
M=
( 1)
/R T
where U is the mean flow Yelocity in m/sec,R is the gasconstantfor air in JA(gK' T is the
inlet tempemture and Yis the specifrc heat ratio. The Mach number (hence the mear floF'
velocity), pressure loss (dP/P) and combusto! inlet tempemture(T) arc corelated by the
following equation;
=(+)(#t)'
s =r+-'r[$)(*l
(2)
whereAl is the uPsfeampipe flow are4 A2 is thecomblstorflow areaand CD is a flame
The pressuelossis defned as the ntio of presswedrop
saUitirci aisctratle
"oeffrcienl
acrossof the flame stabilizerto thecombustorinlet pressure'
rssN0128_0?40
BLI-ETINFKKKSA8(l) : 53 , 1994
EXPERI]\{ENTAL RESULTSAND DISCUSSION
i) Weak ExtinctionResults
Weat extinctionswere determinedat a constantMach numberand pressLtre
loss of 0.03
and 2.770,respectivelyand at a preheatinlet air temperarure
of either400K, 60CK,740K
or 900K by graduallyreducingthe fuel flow until the flame wa-sfinally exringuished.The
process \r,as observed direcdy from the control room thiough a 100rnn diameaer
observationwindow.The weal extinctionswerc alsoeasilyobsenedby a srddenincrease
il unburnthydrcarbon(JHC ) emissions.
The measuredweak extinctionresultsas a fu-4ctiotlof metredeo.uivalerceratio preseDted
in Figure 2. The testresultsshow that increasinginlet air temperature.from
400K ro 900K
reducedthe weak extinction equivalenceratio frorrO5 to S25, rhus widening rhe weai.
extinction region by 507a.This signihesa continualimprovementin the flarne stabiiiry
characteristics as th€ conlbustor inlet temperature increases.The reason for this
improvemenlmay be explainedby refen'ingto equation3 which corelatesthe rareof fuel
and air rdxing, e-laken as equivalent to the rate of energy dissipation per unit massto the
pressure
lossparzlmeter
andcombustor
inlettempemtlu_e(10)
cl,9lrl" 1
L
\P
(3)
,/
ehere C = constantRI.5,T = inlet temperature
andL = integralsca-le
of tur.bulence.
As the combustorpressureloss was kepl constant(and thereforecoDstantturbulence
intensity associatedwith pressureloss and Mach number) incleasing the iDlet temperature
ircreased the €cergy oithe aAjet entering the srabilising shearlayers. This resulted in at
increase in ihe shear layer temperatureand henceprcvided a better ignition source and a
rapid acceleration of buroing velocity of adjacent flame elements within the high
temperature stabilizing shear layers. As a consequencethe operation was extended to an
even leaner mixture.
The above discussion is further supportedby the developmenrof combustor wall
temperatweprcflles as shown in Figure 3(a-d). Comparisonof temperatureshows a
steadyincrease in t}e rate of llame prcpagation and the rrain region of beat release was
also much closer to the swirler outl€t asthe comttusior inlet tempera+ue increased. This
was mainly due to the reasonas discused earlier.
ii) NOy Emission Characteristics
In{luence of inlet air temperature
i) As a function of Mean EquivalenceRatio
The NO"/mean equivalenceratio cofelation is presenredin Figure 4. The minimum vales
of equivalenceratio plesented herc correspondedto the values just priot to Iean blowout.
The NOx emissionswere at maximum near the stoichiometric value and decreasedsteadily
as the equivalencemtio decreased.As mentionedearlier, the higher inlet temperature
operationwidenedthe stability maryin considerablyallowing operationat a much lower
value of equivalenceratio. Consequenrly,the minimum value of NO" was somewhat
-53-
ISSN01284740
BLTLETIN
FK(rS,{ 8(r): 54 , 1994
lower for rhe highercombustorinlet tempemturgThe lo\lest valueof NOx for naturalgas
This figure also showsthat
was 1.4ppm, This was recordedat the 900K inlet temperature.
equivalence
ratio
causedthe NO' level to
at
a
constant
increasingthe inlet temperature
increase.The reasonfor the higher NO' with higher combustorinlet lemperaturewas
(seesectionii).
roainly due to an increasein the flame temperature
ii) As a function of FlameTemperature
The effect of inlet air temperatureon flame temperatureand the sensitivity of NOr
emissionsto flame temper&tureare shownin Figure 5 and Figure 6, respectivcly.The
flame temperatureis prcferableto equivalenceratio as it takesinto accounta.nyvariation
in the combustioninefficiency.
As shown in Figure 5 increasingthe inlet ail temperatureat a constantequivalenceratio
increasedthe flame temperaturc.The combustorprimary zone temperaturerise, when
added to the high inlet temperatureflow tends 10 elevatethe flame temperalur€.An
increasein the NO* emissions with flame temperatiireas shown in flgure 6 was attributed
to an increase in the NO' formation rate becauseof increasing flame temperature.
However, due to the wider flame stability limit with higher inlet temperature,the
combustionprocesswas ablg to opemteat an even leaner mixture with a lo*er flame
temperatureand hence a significart reduction in NO* emissions.
Close examination of the flame temperatue^'Iox curve in Figure 6 shows'ihat for rhe
same flame temperature the NOr emissions decreasedas the inlet air tempelature
increased. The reason for the lower NO" with higher combustorinlet tempemffe was not
only due to the leaner operation but probably also due to the bett mixing of fuel and air.
The latter may be explained by rcfering to e4uation 3; if the pressureloss is fixed constant
( as in the caseof the presentexperiment)the rate of fuel-air mixing is stlongly dePendent
on rhe inler temperature.The increasedenergydissipationrate (hencathe mixing rate)
$,ith inlet temperaturereduced local fuel-rich zonesand hence thermal-Nox foming hot
spots.Thus, ii can be concludedthat low NOx gas turbine emissionsrequte the fuel
injected into the pdmary zone to bum at a uniforrnly lean value of equivalenceratio
Compadson of the axial teEperaturc profil€s for the same flame temnpemtule in Figure
3(a-d) iodicates that as the inlet temperatureincleased the main heat releaseregion moved
closer to the combustor head.Thus, from a purely flame temperatureviewpoint NO' was
most likely to form at the early stage of the combustion process if the combustor inlet
remperatureis suffrciently high. This may be associatedwith the combined effects of the
unmixedness promoted in the early stage(which give a high equivalence ratio) and the
higher inlet temperatue (which causesa rapid rise to local flame temperaturc).
iii) On Combustion Inefficiency
Low NOx systems are only viable if they are accomplishedwith low combustion
inefficiency. Il the preseattestresultsthe lowestvalue of NO, emissionswas foundjust
prior to flame blowout, but the combustion inefficiency was unacceptablyhigh. From the
practical gas turbine combustionpoint of view such an operation is not viable.
The combustion inefficiency as a function of the rig metred equivalencemtio is presented
in Figure 7. As the inlet temperatureincreasedthe inefficiency curves were shifted
towards a leanerregion and henceallowedthe minimum combustioninefficiency at an
-54-
ISSN0128{740
BL,'I-ETIN
FKKKSA8(l) : 55 , 1994
evenweakerequivalence
ratio.Such.a.move
wasmostlikelydueto rheenhanced
high
tenrpentureunburntandburntgasmi-\ingwithin$e recircuiation
zonewhichi.creased
thecombustion
intensityandhencetheefficiencyof thecombustion
process.
Figure8 showstheinfluenceof thecombustion
jnefficiencyon lheNOxcorrected
to 15Zo
oxygenand standard
day hunridiry.
As_shown
in rhisfigure,increisingthe inlet air
temperaturcnot only iDproved the efficiencyof the combustionprocess
but also
decreased
the ievel of correctedNOi. This showsthat high inlei iemperaiure
operatron
couldachievehighcombustion-efficiency
wiriourcomproriisingrf,eNb, emissi6ns.
fire
ulrm-lowNOx recordedat the900Kinlel_temperature
wasweliivirhrnrtreproposed
EpA
limir. Thus,highinlertemperarure
gasturbinjcombustion
coulaU"-ap.o#sing apo.oa.ti
.fo^rachievinglow NO* emissions
riirh no_associutea
peJo.,ou..!!"lufyr-na ro adverse
effecton-flame
The
highesr
valueof com6usdoa
inffA"n"y was toundjus!
,stability.
p.ior to flameblowou|
A largeanountof air presentat the leanestmixtue rnayLr,^_
decreasedthe mixing shearlayer tempenr!; una rn"."ly-iuip."ssed
the rate of
combustion
reactionasindicared
'
by theinireasedUHCernissio"ns.
Jr"isalso.rc4essary_ro
arcompanylow NO" systems
with an acceptablLlevel
of Cerbon
Monoxide(co) ernissions.
e-conetationof ae co *a wo,
o..e,tedto 15qa
axygenis presenledjn:Egure
9. This figureshowsa .neryi* "Jrion,
t.na "to Eigureg-ihe
minimumCO corrected
ro 157ooxygeniompatiblewirh 0.012,",
in"in"i"n"l", "*u. ipfrn.
Thus,miaimum emissionsof NOx
C-Ocbrrectedto t SZ,oxylen weresimulraneouslv
"nd efficiency.ff,"
achievedar a maximumcombustion
t"rt ,"rJfi, ufro ;;;;;;;;
minimumcombusrion
inefficiencyandCO emissions
o""urr"J ui it sameequivalence
ralio.
"
iv) On NO emission
As hasb€endiscussedearlierNitric Oxide(l.IO) emircd from a combusuon
sysremuslng
gas€ousfuel il mainly due to eitherpromptNO or thermal
frfO.-iig"r" ld showsrhE
variationofNO emissionswith flameiemperature.A compariion
oitfre NO level telow
1800K llame temperaturc
showsa signifriantreductionir rrr"Nb-"*"rions astie iniet
temperaturcmcreased.Below this flamercmperarure
NO is maintvformeavia thepiomfi
N0 mechanism(6).
The test resultsalso sug'gesrrfr"t rf*
uame temperature
high inlei tem?erature
"t ilnr""""Jiy
combustion,is
"roa'Jor'" promptNO. The
not;i-gnificamly
improvedcombustionefficiencvwlttl mlet tetolperature
may havereducedtheintermediate
producrs
in
the
iecirculation
;r"
;;
t;*
,h" fr6 enissions. Above
lltTqbo.
199-05
lame lerryemturenearly aI rhe.datafalt on a s".itf,t ii"i i"-o"uti.g tharrie rare
is indepeDdent
ofrhe inlettemperarure.
The NO elqrssrons
beyondLhis
:t"I9.11::"-"1
rrame
temperature,assuggested
by Zeldovich€)is formedvia tlre themal No m;hanis;
CONCLUSIONS
1.
2.
3.
Increasingthecombustorinlet tgm-peratwe
significaatlywidenedthe weakextinction
limit andhenceallowedal operationat aa
Io*". u"lo" oiequivalenceratio.
"r"n
Lean well-mixed gasturbinecombustionfor low NOx emissions
could be achieved
by high inlet tempirarur.e
operation.
Ulaa.low NOx emissions,wer.e.achieved
alongwirh high combustionefficiency
wirhouranadverse
effecton stabjlitymargin.
I
ISSN0128-0740
BULETINFKKSA 8(1): 56 , 1994
REFERENCES
1.
Aircraft Propulsion.NASA, SP259,1971.
2.
Bowman,C.T. : Kineticsof pollutantformation and destructionin combustion.
Prog.EnergyandConb. Sci..,Vol. 11,p9.33-35,1978
3.
Zeldovich,Y.B. : The oxidation of nitrogen in combustionexplosions.ACTA
Physicohirnic,Vol. XXI, No. 6, USSR,1976.
4.
Iverach, Kinov, N.Y. and Haynes,L.H. : The formation of i{itric Oxide in fuelrich flame. Prog. EnergyandComb.Sci..,Vol. 8, pg. 159, 1973.
5.
Fenimorc, C.P. : Formation of Nitric Oxide in premixed hydrocarbon flames.
13th Symposium(Int.) on Combustion.The Inst. of Combustion,Pittbugh, pg. 373380, 1971.
6.
Cla1pole,T.C. and SyredN. : The effect of swirl burner aerodynamics on NO1
formation. 13 th Symposium (Int.) on Combustion.The Inst. of Combustion,
Pittburgh,pg. 81-89,1981.
7.
Morely, E.L. : Formation and destruction of Hydrogen Cynide from
atnospheric anf fuel Nitrogen in rich atmospheric pressure flames. qgmbustion
andFlames, Vol. 27, pg. 189,1976.
8.
Appleton, J.P. and Heywood, J.B. : The effect of inperfect fuel-air mixing in a
burner on NO formation from Nitrogen in the air. 14 th Symposium(Int.) on
Combustion.The Inst. of Combustion,Pittburgh,pg. 777-786,1973.
9.
Lemansky, N.P and Sawyer,R.R. : NO and NO2 formation in a turbulence
hydrocarbon-air diffusion llames. 15 th Symposium(Int.) or Combustion.The
Inst of Combustion,Pittburyh,pg. 133-136,19'16-
10. Aadrews,G.E. ard Aboul, M.M. ; High irtensity burner with low NOx emissions.
British FlameDays, 1988.
tssN0128{J740
BULETINFKXKSA8(I)..57, 1994
Fis. (l) Scbcmarcdj?8d) ol$c 76'M coDbustonDg
1
900
r-i 8m
F
li 603
F
400
.
02
0.3
0.4
0.5
RATIO
EQT'IYALENCE
Ftgure (2), E Iecl ot tnlel te'npetalu|eon
efluvalence 'stlo for MftJBI
gas combusllon
'a:,:
lssN 0128-0740
B{,ILETIN
FKK(SA8(l) : 58 , 1994
:: s;I
E
;
ttll
+
tttl
:lttl
'rlll I
E
It
9-
3
a
;
.E
a
aaaaee?qo
J r!UnIY!3dt{3.1
I'aunr\.d3dt{!l'l1v^,r
1'lVIt
-
E
;.E
8
i.
ti
J '3UnM3dt{3r
:)
11V11
-58-
i 6I Y X S J I
._S*F
XO
ISI
-I-IYM
ISSN0]28,0740
BULETINFKKKSA8(1): 59 , 1994
100
4mK
600K
14tK
900K
60
;
20
0.3 0.4 0.J 0.6 0.7
EQtJIvALr^-CE RATIO
0.8
Fig. (4) NOx as E function of equlvalence
Etio at dittercnt lnlet Iempenlures
E
x
!
4@K
600K
l40K
a
9ooK
F
l
0.3
0.4
0.5
0.6
0..t
'
EQUIVALENCE
RATIO
Flg. (5) Hatue tempeftlure as a functlon ol
equlvalence ratlo at dlffeEnt tnlet
lemperaturcs
-59-
0.8
lssN 0128{740
BLI-ETINFKKKSA8(1):60 , 1994
.0 1
40
20
3A
TO
15%
OXYCDN'
CORRECTED
PPm
^{Ox
lnefficieqcy
gas
combustlon
Fig (8): Nafural
as a functlon of NOx corrected to
15o/"oxYgen
010
€
1000
i
*
r00
o
0102030
NOx co.rected to 15olooxygeh, ppm
Fig. (9) A coffelation of corrccted CO and
NOx at diffeient lnlet temperatures
-60-
40
I
rssN01284740
. BIILETINFKKXSA8(l): 51 , 1994
z
FLAMETEIUPERATURE,
K
Fig- (10) Nitrlc Oxlde (NO) as a tuncTion
ot ame temperctufe at dltferent
lnlet tempeatures
-61-
Download