the reheat concept: the proven pathway to ultralow emissions

1

cGttle

�wcGcatlwhlbsGfn

cA�ocwms

pi

p

J

Downlo

Felix Güthe1

e-mail: [email protected]

Jaan Hellat

Peter Flohr

Alstom,Brown-Boveri-Strasse 7,

CH-5400 Baden, Switzerland

The Reheat Concept: The ProvenPathway to Ultralow Emissionsand High Efficiency and FlexibilityReheat combustion has been proven now in over 80 units to be a robust and highlyflexible gas turbine concept for power generation. This paper covers three key topics toexplain the intrinsic advantage of reheat combustion to achieve ultralow emission levels.First, the fundamental kinetic and thermodynamic emission advantage of reheat combus-tion is discussed, analyzing in detail the emission levels of the first and second combustorstages, optimal firing temperatures for minimal emission levels, as well as benchmarkingagainst single-stage combustion concepts. Second, the generic operational and fuel flex-ibility of the reheat system is emphasized, which is based on the presence of two funda-mentally different flame stabilization mechanisms, namely, flame propagation in the firstcombustor stage and autoignition in the second combustor stage. This is shown usingsimple reasoning on generic kinetic models. Finally, the present fleet status is reported byhighlighting the latest combustor hardware upgrade and its emission performance.�DOI: 10.1115/1.2836613�

IntroductionIn the mid 1990s, Alstom introduced two similar sequential

ombustion gas turbines: the GT24 for the 60 Hz market and theT26 for the 50 Hz market. Since its first launching in 1995 �1�,

he advanced class GT24/GT26 engines have demonstrated thathis technology platform does offer superior operating flexibility,ow emissions, and high part-load efficiency with world-class lev-ls of reliability and availability.

The main technology differentiator of Alstom’s GT24/GT26Fig. 1� gas turbines is the sequential combustion principle, whichas already introduced in 1948 into the market as a way of in-

reasing efficiency at low turbine inlet temperature levels. TheT24/GT26 combustion system is based on a well-proven Alstom

ombustion concept using the environmental �EV� burner in annnular combustor followed by the SEV �sequential environmen-al� burner in the second combustion stage �see Fig. 1�. This dryow NOx EV burner has a long operating history and is used in thehole Alstom gas turbine range. Sequential combustion, “the re-eat principle for gas turbines,” had already been applied to ear-ier �Brown Boveri� engines but using two side-mounted silo com-ustors. Integrating the concept of dry low NOx EV burner andequential combustion into a one shaft engine resulted in theT24/GT26—a machine with a high power density and a small

ootprint. With a pressure ratio of �30:1, the compressor deliversearly double the pressure ratio of a conventional compressor.

The compressed air is heated in a first combustion chamber �EVombustor� by adding about 50% of the total fuel �at base load�.fter this, the combustion gas expands through the high-pressure

HP� turbine, which lowers the pressure by approximately a factorf 2. The remaining fuel is added together with some additionalooling air in the second combustion chamber �SEV combustor�,here the combustion gas is heated a second time to the maxi-um turbine inlet temperature and finally expanded in the four-

tage low-pressure �LP� turbine.Relative to a conventional nonreheat cycle, the same specific

ower output is achieved at lower turbine inlet temperature. Thiss illustrated in Fig. 2.

1Corresponding author.Manuscript received May 30, 2007; final manuscript received October 8, 2007;

ublished online December 24, 2008. Review conducted by Dilip R. Ballal.

ournal of Engineering for Gas Turbines and PowerCopyright © 20

aded 09 Jan 2009 to 91.199.43.41. Redistribution subject to ASME

2 Basic Reheat Features for Low EmissionsThe low emission levels, which can be achieved with a reheat

system, are the combined effect of three key mechanisms: First, areheat combustor makes a more efficient use of the oxygen byburning twice in the lean premix mode. Second, there exists achemical advantage of reheat combustion, which can be exploited.Third, the flame stabilization by autoignition leads to increasedflexibility, which allows operating in low emission mode at a wideload range by avoiding high peak flame temperatures Tflame, whichlead to exponential increase in NOx, for both combustors. Weexplain these effects in detail below.

2.1 Impact of Lower O2 Level. The reheat concept alsomakes very efficient use of the O2 from the air by operating thesecond stage at approximately 15% O2 at the inlet. Overall thisleads to very high power densities. The O2 content in the exhaustis in the order of �10%, yielding a beneficial normalization factor�for 15% O2� of �21–15% � / �21–10%-� �0.55 for the absoluteNOx, as shown in Fig. 3 on the right hand axis. The correctionfactor decreases with higher temperature �� consumed fuel �O2�and is lower for the reheat engines due to its increased O2consumption.

This leads to varying absolute levels for a fixed “15%” guaran-tee value depending on reheat or nonreheat combustion and Tflameand allows higher absolute emission values after two combustionchambers. This correction has been introduced to make a faircomparison of power plants based on their emissions relative toconsumed air or fuel. Emissions quantified in the widely used

Fig. 1 GT24/GT26 sequential combustion system

MARCH 2009, Vol. 131 / 021503-109 by ASME

license or copyright; see http://www.asme.org/terms/Terms_Use.cfm

ef

rrbhb

tflpmcd

sftaoe

0

Downlo

mission index defined as �atmospherically produced� g�NO2� /kguel are reduced accordingly.

2.2 NOx Characteristics of Reheat Combustion. A secondeason for the low NOx emission is the combustion in the SEV ateduced O2 levels and increased H2O contents from the first com-ustor ��5% �. This allows the SEV combustor to operate atigher firing temperatures and produce less NOx than an EV com-ustor would produce at the same temperature.

Results of CHEMKIN �2� simulations using the GRI-MECH �3� forhe different flame types are shown in Fig. 3, where the laminarame speed module PREMIX has been used for the EV �4� and thelug flow reactor �PFR� module for SEV. A detailed networkodel based on PREMIX and PFR has been developed for an ac-

urate emission prediction for realistic engine conditions, but itsetails are out of the scope of this paper.

A simple CHEMKIN calculation shows that NOx produced in theecond combustor is essentially unaffected by NOx levels arisingrom the EV combustor as long as they are of moderate magni-ude. The SEV calculation seems simply to be offset by themount of EV NOx given at the start. Therefore, a chemical modelf the engine treats both combustors as independent NOx produc-rs. Although the GRI is assumed to describe the SEV combustion

Fig. 2 Principle of the reheat cyc

Fig. 3 NOx emissions on log scalecombustor „EV…, the SEV combustor,factors calibrating to 15% O2 depend o
to the right hand scale.
21503-2 / Vol. 131, MARCH 2009


sufficiently, further efforts are taken to investigate possible effectsof NOx and water as well as reduced O2 content in ongoingprojects in connection with shock tube studies.

Shown in Fig. 3 are the NOx contributions versus Tflame of asingle combustor �EV� at approximately 30 bars and of a SEV-type combustor at approximately 15 bars, as well as a combina-tion of the two with fixed Tflame for the first combustor �EV� forthe idealized situation of perfect mixing without heat losses. NOx

emissions are given on an absolute scale normalized to the valuefor a single combustor �EV� at Tref. This allows the SEV contri-bution �at reduced O2� to be shown on the same plot. The curvesare analytical fits to CHEMKIN II results of the GRI3.0. The reheatengine results in lower emissions at the same Tflame. The calibra-tion factor for 15% O2 refers to the right hand scale. The NOx

emissions of the full reheat engine are taken as the sum of theabsolute emissions of the two combustors assuming no interactionof NOx from the first combustor on the second.

The reduced NOx production of the reheat combustor is amongother possible factors due to the decreased concentration of Oradicals, which is influenced by the reduced O2 and the increased

ompared to a standard GT cycle

rmalized, noncalibrated… for a singled a reheat combustion system. Theflame and are shown in green and refer

le c

„noan

n T

Transactions of the ASME


Hre

t

F„

airi

J

Downlo

2O content. The O radicals are an essential part of the mainoute to produce NOx by oxidizing the otherwise stable N2 mol-cule according to the kinetic rate law:

d�NO�dt

= k�T,p��N2��O� �1�

�NO� = k�T,p��N2�t�O� �2�

Since k and the N2 concentration do not vary much for givenemperatures and pressures, the NOx formation depends mainly on

ig. 4 O-radical concentration profiles for EV and SEV flamestaken from PREMIX and PFR calculations… for similar Tflame. Therbitrary time scale extends over 4 ms. The shaded area in the

ntegral �t†O‡dt determines the most important NOx formationoutes and is approximately three times higher in the EV thann the SEV.

Fig. 5 Schematic diagram explaining the physicflame „right… for natural gas. The intermediates inupper graphs. The shaded areas indicate the mo
axes were given in reverence to the typical mixing t
ournal of Engineering for Gas Turbines and Power


the time t and the concentration O radicals. The time profiles ofthis species are plotted in Fig. 4, revealing the integral to be ap-proximately three times smaller, which is the combination of shortresidence times and low O concentration due to H2O and O2contents of the vitiated air as explained above. For such laminarcases, the prompt NOx formation due to the O radicals is esti-mated to be approximately three to six times smaller at givenTflame considering O-concentration profiles. To invoke the ofteninvoked reburn chemistry to explain the low NOx emissionsseems not to be necessary and is believed to be unlikely since theheat release is occurring entirely in fuel lean conditions.

Another factor reducing SEV-NOx emissions is the signifi-cantly reduced residence time in the SEV compared to the EV.Some additional benefit might arise from the fact that the fastSEV gas velocities of part of the rich combustion products in theSEV result in faster quenching of this hot pockets and subse-quently less time to produce NOx. It can be summarized that theNOx production in the SEV starts at a significantly lower value.

2.3 Reheat Combustion With Autoignition. The reheat con-cept allows a very flexible operation of the engine, enabling us tofind an optimized operating range for the given conditions andfuel compositions. The reason for that is described in the follow-ing section.

While the EV inlet temperature Tinlet�EV is determined by thehigh-pressure compressor, in the SEV the inlet temperatureTinlet�SEV is governed by the outlet temperature of the HP turbine.The SEV flame temperature, Tflame�SEV, determines the inlet tem-perature of the LP turbine. Accordingly, the HP-turbine inlet tem-perature �and, along with it, Tinlet�SEV� is controlled by the EVflame temperature, Tflame�EV.

The reactor inlet temperature Tinlet in the chemical model isderived from mixing cold fuel and hot combustion air before thereaction, yielding an effective Tmix as Tinlet. The shorter residencetime of the products in the combustor is accounted for in themodel.

nd chemistry of the EV flame „left… and the SEVating preflame zones „CH2O… are displayed in therelevant regions of reactivity. The different time

s adicst

imes.

MARCH 2009, Vol. 131 / 021503-3


fsbtmnai

Fn

SuntaNs

aeiottw�c

rpceN�ilTofl

sflfladTtthab

tpznflvwlprdaf

0

Downlo

The higher Tinlet�SEV results in a much higher reactivity of theresh gases, a different mechanism being important for the de-cription of the flame. This higher reactivity is partly compensatedy higher burner velocity leading to shorter characteristic mixingimes �mix�SEV �� ignition time �igntion�1–2 ms� available forixing for the SEV combustor, leaving some residual unmixed-

ess in the flame region compared to the EV flame, where a char-cteristic mixing time �mix�EV ��5–10 ms� is given by the travel-ng time from fuel injection to flame front.

It should be pointed out that Fig. 3 indicates ideal conditions.or a real system, effects of local unmixedness and heat losseseed additional consideration.

This high reactivity at high combustor inlet temperatures in theEV has the consequence that flame stabilization occurs alwaysnder premixed conditions by autoignition, and no piloting iseeded to stabilize the SEV flame. The Tflame of the two combus-ors are ideally optimized primarily with respect to performancend turbine lifetime, but also at a given load for equal amounts ofOx produced, deriving an absolute minimum in engine emis-

ions.

2.4 Optimal Firing of Combustor for Minimal Emissionsnd Maximum Flexibility. The reheat concept enables us to op-rate the two combustors at different temperatures with only littlempact on the overall power output. This increases the flexibilityf the combustion system, allowing us to burn more or less reac-ive fuels in the same engine by adjusting the relative load be-ween the two combustors and to mitigate ranges where operationould become difficult. This can be utilized for different fuels

C2+ or even H2 containing fuels�, exhaust gas recirculation, orhanges in ambient conditions.

In general, the engines are adjusted so that the first combustoruns at low firing temperatures, keeping the emissions as low asossible for conventional premix combustion, while the secondombustor is operated at higher temperatures, producing similarmissions �on absolute scale� due to the reasons mentioned above.ote that while the EV runs more or less at constant conditions

see Fig. 10�, keeping low NOx even at very low loads, the load-ng is achieved by varying the fuel input in the SEV starting at asow as �25% GT load �as well as increasing the mass flow of air�.he high Tinlet�SEV allows operation without piloting for the entireperation range, entirely relying on autoignition for the SEVame.The conceptual difference between EV and SEV flames is pre-

ented in Fig. 5. The flame has to be stabilized between limits forashback �FB� and lean blowoff �LBO� determined by the laminarame speed SL for the EV flame, which depends on conditionsround the adiabatic Tflame and less on Tinlet. The SEV flame isominated by the chemistry occurring in the preflame zone atinlet. This includes the buildup of a radical pool up to a critical

hreshold. The flame position is here determined by the ignitionime. To compare the two different combustion regimes, the timeas been scaled to zero for the region of maximum heat releasend normalized to the characteristic mixing times as definedefore.

For the inlet conditions of the EV, autoignition times are muchoo long to play a role even for the most reactive fuels. Flameropagation is the leading mechanism for a stationary reactionone stabilization, as can be calculated with PREMIX for the lami-ar case �laminar flame speed SL�. The description of the turbulentame propagation is derived from that concept. As for every con-entional premix burner, the EV flame stabilization happensithin the boundaries of flame blowoff at a low flame temperature

imit and flame FB at a high flame temperature limit. These flameropagation limits are sketched on the left side of Fig. 5. Theeactivity is controlled by transport across the flame front andetermined by the kinetic rates at high temperatures close to thediabatic one. Intermediates are abundant only close to the flame
ront. The reactivity �Fig. 6� depends on the kinetics of the fuel
21503-4 / Vol. 131, MARCH 2009


and its transport properties at high temperatures close to the finalTflame�EV and only slightly on pressure and Tinlet�EV. For the EV,the relative reactivity RREV is defined with reference to one op-erating point as the ratio of laminar flame speeds in

RREV = SL/SL−ref �3�This flame stabilization is fundamentally different for the SEV

flame, which is controlled by autoignition �calculated with PFR�arising from reactive intermediates forming and accumulating upto a critical threshold before the flame front. The chemistry atlower temperatures ��Tinlet�SEV� is responsible for the reactivity.This means that in the engine the SEV flame is controlled andgoverned by the Tflame�EV, while Tflame�SEV has only little influenceon the flame position. The relative reactivity RRSEV is defined inEq. �4� from the ignition times as

RRSEV = �ig�ref/�ig �4�These reactivity definitions enable comparing combustor specificdefinitions on one scale �Figs. 6–8�.

Plotting RR over Tflame ��1 /�� for lean conditions� in Fig. 6leads to the contrary effects: For this laminar approximations, theEV flame becomes more reactive with Tflame, but the SEV flamebecomes less reactive because the cooling effect of adding morecold fuel to the heated air from the first combustor controls thereactivity while the reactivity �toward ignition� is not increased by

Fig. 6 EV and SEV reactivities as defined by Eqs. „3… and „4…for varied fuel contents and constant Tinlet „compare with Fig.8…. The shaded area defines an operational range.

a higher concentration of fuel. In the very lean region, an increase



otottlomt

tapbph

tmEcr

tc

Ffhb

J

Downlo

f fuel molecules can increase RRSEV slightly. Taken together,hese effects counterbalance to give a reactivity maximum and,verall, very small dependence on Tflame. The small variation forhe SEV reactivity with fuel is the cause for the high flexibility ofhe reheat concept. The reactivity increases with fuel only for veryean conditions. The chemical acceleration of rich mixtures levelsff at intermediate levels, and the reduction of Tmix �due to theixing with cold fuel� outruns this effect, slowing down the reac-

ivity toward rich mixtures.The EV reactivity RREV varies much more with fuel; therefore,

he EV flame position has to be controlled aerodynamically byttaching it to a flame anchoring point, in this case a stagnationoint in the flow field �vortex breakdown� of the swirled EVurner. Effects of autoignition in the EV combustor or flameropagation in the SEV are assumed to be minor and are neglectedere.

The operating range of a burner is limited by a minimum reac-ivity ��SL for the EV�, causing a LBO of the flame and a maxi-

um acceptable reactivity preventing the flame to flash back. TheV is governed by the aerodynamic stabilization, allowing thisombustor to hold the flame at the same positions over a wideange of reactivity.

The reheat combustion system has the particular property thathe flame stabilization in both premix combustors, EV and SEV, is

ig. 7 Probability of reactivity derived from a distribution ofuel for EV and SEV combustors at fixed inlet conditions. Theistograms show the probability for a given reactivity, while theubbles are calculated reactivities for a given fuel-mass ratio.

ontrolled by a single physical quantity—the flame temperature of



the first EV combustor. This way, the high flexibility of the enginecan be maintained by being able to operate the EV over a range ofTflame, therefore enabling us to find an optimum HP-turbine outlettemperature for SEV operation. This optimization can be used forthe lowest emissions as well as to handle a variety of fuels andinlet conditions.

The SEV flame position is mainly controlled by its chemicalreactivity, which is mainly influenced by Tinlet�SEV and the fuelcomposition. The different control mechanisms of flame positionsare utilized in the ALSTOM gas turbine design by attaching theEV flame aerodynamically and by controlling the SEV flamethrough Tinlet�SEV. The fact that the SEV flame is relatively insen-sitive to Tflame leads to a high operational range: SEV load can beincreased without increasing FB risk �emissions and pulsationsstill have to be considered, of course�.

This leads to an interesting effect for not perfectly mixedflames. The reactivity probabilities resulting from a given fuel-airunmixedness are plotted in Fig. 7. The same variance in the upperand lower graph lead to a wider distribution of flame speeds forthe EV and a very narrow distribution for the SEV �at given HP-turbine outlet temperature�. Here the different mechanisms be-

Fig. 8 EV reactivity for varied Tflame�EV „upper graph… for twodifferent fuels on logarithmic scale. In the lower graph, the re-activity of the SEV for varied inlet conditions „Tinlet�SEV… is pre-sented, referenced to a standard inlet temperature Tin�ref. A re-activity increase due to the fuel is compensated by reducedTinlet�SEV. Note that the RRSEV is plotted versus TflameSEV in Fig. 6.

come clear once more by spreading out over a wide range of

MARCH 2009, Vol. 131 / 021503-5


rS

3

ris

fltstrt

ltipTtp

cahrtptf

rsEa

ttsSias

GbwEnbfhst

opo

0

Downlo

eactivities for the EV flame but only a very narrow range for theEV-flame.

Generic Operational and Fuel Flexibility

3.1 Fuel Flexibility for Highly Reactive Fuels. While theeactivity for the EV is controlled by Tflame�EV, the SEV reactivitys mostly controlled by the inlet conditions and Tinlet�SEV. This hasome interesting conclusions for the engine operation.

1. The SEV can be varied in Tflame�SEV without changing itsflame position much.

2. An increased reactivity due to a more reactive fuel can beeasily compensated by reducing the Tflame�EV and, therefore,Tinlet�SEV.

This is also illustrated in Fig. 8, which demonstrates the fuelexibility achievable by varying Tflame�EV. The increase in reac-

ivity �vertical arrow� due to a more reactive fuel can be compen-ated by reducing Tflame�EV �horizontal arrow�. The reduction ofhe EV fuel leads to a Tinlet�SEV reduction, which simultaneouslyeduces the SEV reactivity. In the SEV, a reactivity increase due tohe fuel is compensated by reduced Tinlet�SEV.

The described insensitivity of flame reactivity to Tflame�SEV al-ows the inlet temperature of the LP turbine to be constant �inerms of protection for the combustor�. Therefore, only little lossn overall power and efficiency due to derating the EV since theower output is largely recovered by adding more fuel to the SEV.his is in contrast to the nonreheat engines, where a derating due

o increased fuel reactivity leads to decreased turbine inlet tem-erature resulting in severe loss of overall power and efficiency.

This key feature can be used to operate the engine �5� such thathanges in fuel composition do not impact emission performancend reliability significantly. For example, if for given Tflame higherydrocarbons �“C2+ ” � or even H2 or CO are added to the fuel,eactivity will increase �i.e., flame speed will increase and ignitionime will decrease� and, as a result, the EV combustor will likelyroduce higher NOx. Likewise, the SEV combustor may encoun-er FB risk due to shorter ignition times if no adjustment isoreseen.

For highly reactive fuels, the fuel input in the EV combustor iseduced, which leads to lower SEV burner inlet temperature, totay below the burner protection limit. This is feasible because theV burner lean extinction limit is also reduced for increased re-ctivity in the fuel.

However, a simple adjustment by decreasing the fuel input inhe EV combustor and simultaneously increasing the fuel input inhe SEV combustor �to keep the total fuel flow constant� willtabilize combustion performance: EV emissions remain low, andEV flame stabilization is back at its design point. Due to the

nherently low emissions of the SEV combustor, the emission pen-lty will be negligible if the engine is operated in such a way aschematically shown in Fig. 9.

3.2 Dual Fuel Capability. The primary fuel of the GT24/T26 fleet is gaseous fuel, while fuel oil is often needed as aackup. The sequential combustion system can also be operatedith fuel oil, engines in this case being delivered with dual fuelV and SEV lances instead of the normal gas-only lances. As isormal practice, demineralized water is used for NOx control, thiseing combined with the fuel oil in a mixing block before beinged to the lance as an emulsion. The GT24/GT26 engines alsoave the ability to switch between the two fuels while remainingynchronized to the grid and so providing the operator with arbi-ration capabilities and full backup security.

3.3 Part-Load Flexibility. The sequential combustion systemffers the intrinsic advantage that premix combustion can be ap-lied across the entire load range because the EV combustor is
perated at a constant flame temperature through the entire load
21503-6 / Vol. 131, MARCH 2009


range, while the �premix� SEV combustor is used to vary loadsfrom approximately 12–100%. This is schematically shown inFig. 10.

The so-called “low load operation concept” utilizes the possi-bility of shutting down the sequential SEV combustor at low partloads and is therefore a unique feature of the GT24/GT26 tech-nology. This concept allows the plant to be operated in a com-bined cycle mode at a very low combined cycle load �about 15–25%� with the EV combustor operating in the lean premix mode,giving low emission levels as well as a homogenous HP-turbineinlet temperature distribution while keeping the water-steam cycleup and running. The result is a concept that has a number ofadvantages to the operator because it avoids start-stop cycles andthe related cyclic thermal stress to the top and bottom cycle equip-ment. It enables the operation at low emission levels close to baseload levels and reduces cumulative emissions compared to park-ing a plant at a higher partial load. Also, it assures a homogenousturbine inlet temperature distribution, which is more difficult toachieve for nonreheat engines due to burner piloting at part load.

4 Engine Experience

4.1 Combustor Hardware Status. The continuous drive �6�toward higher performance and lower emissions motivated an up-grade of the EV combustor. The combustor upgrade development,based on the so-called “staged EV burner” is described in detail ina recent paper �7�. Its main feature is the replacement of the cen-tral pilot injector by a premix injector. This leads to a burnersystem with two fuel stages, which can be operated both in thepremix mode to vary the fuel profile. In the premix mode, the

Fig. 9 GT26/GT24 operation concept as a function of fuel re-activity, here given by the content of higher hydrocarbon con-tent „C2+ … in the fuel

Fig. 10 Schematic of the operating concept of the GT24/GT26
with load variation done through the SEV combustor fuel flow


pt�lfo

tsltobeE

tSSeIitc

Fo

J

Downlo

rofile is optimized for minimal emission performance, while inhe startup mode, the staging is optimized for maximal stabilitysee Fig. 11�. As a result, this burner has much reduced emissionevels in part-load operation, as well as in the base load where theuel staging between the two stages can always be chosen to beptimal to the base load adjustment point of the engine.

4.2 Emission Characteristics. The emission performance ofhe staged EV combustor against the standard EV combustor ishown in Fig. 12. This graph shows two effects: First, emissionevels are reduced and thus confirm the expected improvement ofhe staged EV burner. Second, the emission gradient is changedver a wide load range �in this plot between the measured pointsetween 65% and 100% loads�. This change in gradient is inter-sting because it is a direct consequence of the optimization of theV combustor.The emission levels of the EV combustor are now so low that

he total emissions of the engine start to become dominated by theEV combustor. As was shown in Fig. 10, the fuel input in theEV is increased with load and, as a result, now visible in thengine emission performance even if absolute values are very low.ndeed, a more detailed analysis of the combustion system, whichncludes effects of heat losses and burner imperfections, showshat the relative emission contribution of the SEV combustor in-reases.

Fig. 11 Staged EV-burner principle. Din Ref. †7‡.

ig. 12 Emission levels achieved with the staged EV burnern logarithmic scale



5 SummaryIn this paper, the key advantages of the reheat system as imple-

mented in the GT24/GT26 engines have been highlighted, namely,its inherently low emission levels as well as its operational flex-ibility. The contribution of the two combustors to overall emissionlevels has been discussed in detail and highlights how the latestEV combustor upgrade has improved the overall engine emissionperformance. Also, it has been explained how the proven opera-tional flexibility of these engines are inherently linked to thecomplementary combustion kinetics in the two combustion sys-tems.

The achievement of ultralow emission levels while keepingmaximal operational flexibility �both in load variation and fuelcomposition changes� and reliability will remain the key driver forALSTOM’s combustor development. The sequential combustionsystem is suited very well for this challenge, both today and forthe future.

AcknowledgmentThe support of Reaction Design in using the CHEMKIN code is

gratefully acknowledged.

References�1� Joos, F., Brunner, P., Schulte-Wenning, B., Syed, K., and Eroglu, A., 1996,

“Development of the Sequential Combustion System for the ABB GT24/26Gas Turbine Family,” ASME Paper No. 1996-GT-315.

�2� Kee, R. J., Rupley, F. M., Miller, J. A., Coltrin, M. E., Grcar, J. F., Meeks, E.,Moffat, H. K., Lutz, A. E., Dixon-Lewis, G., Smooke, M. D., Warnatz, J.,Evans, G. H., Larson, R. S., Mitchell, R. E., Petzold, L. R., Reynolds, W. C.,Caracotsios, M., Stewart, W. E., and Glarborg, P., 1999, CHEMKIN Collection,Reaction Design, Inc., San Diego, CA.

�3� GRI-MECH �http://www.me.berkeley.edu/grimech�.�4� Biagioli, F., and Güthe, F., 2007, “Effect of Pressure and Fuel-Air Unmixed-

ness on NOx Emissions From Industrial Gas Turbine Burners,” Combust.Flame, 151, pp. 274–288.

�5� Riccius, O., Smith, R., Güthe, F., and Flohr, P., 2005, “The GT24/GT26 Com-bustion Technology and High Hydrocarbon �C2+ � Fuels,” ASME Paper No.GT2005-68799.

�6� Philipson, S., Ladwig, M., Lindvall, K., and Schmidli, J., 2006, “A FurtherUprate for Alstom’s Sequential Combustion GT26 Gas Turbine,” Power GenEurope 2006, Cologne, Germany.

�7� Zajadatz, M., Lachner, R., Bernero, S., Motz, C., and Flohr, P., 2007, “Devel-opment and Design of ALSTOM’s Stage Fuel Gas Injection EV Burner forNO Reduction,” ASME Paper No. GT2007-27775.

ils of its development are described
eta
x

MARCH 2009, Vol. 131 / 021503-7


the reheat concept: the proven pathway to ultralow emissions

Documents