GER 4211 - Gas Turbine Emissions and Control Turbine Emissions and Control.pdfWorldwide interest in gas turbine emissions and the enactment of Federal and State regulations in the

GE Power Systems

Gas Turbine Emissions and Control

Roointon PavriGerald D. MooreGE Energy ServicesAtlanta, GA

GER-4211

g

Contents

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1Emissions Characteristics of Conventional Combustion Systems . . . . . . . . . . . . . . . . . . . . . 1

Nitrogen Oxides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2Carbon Monoxide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3Unburned Hydrocarbons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5Sulfur Oxides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6Particulates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7Smoke . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

Dry Emissions Estimates at Base Load . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8Dry Emissions Estimates at Part Load . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

Simple-Cycle Turbines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8Exhaust Heat Recovery Turbines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

Other NOx Influences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13Emission Reduction Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

Nitrogen Oxides Abatement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16Lean Head End (LHE) Combustion Liners. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17Water/Steam Injection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18Carbon Monoxide Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22Unburned Hydrocarbons Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24Particulate and Smoke Reduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

Water/Steam Injection Hardware. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25Minimum NOx Levels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27Maintenance Effects. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29Performance Effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29Summary. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30List of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31List of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32


GE Power Systems ■ GER-4211 ■ (03/01) i


GE Power Systems ■ GER-4211 ■ (03/01) ii

IntroductionWorldwide interest in gas turbine emissions andthe enactment of Federal and State regulationsin the United States have resulted in numerousrequests for information on gas turbine exhaustemission estimates and the effect of exhaustemission control methods on gas turbine per-formance. This paper provides nominal esti-mates of existing gas turbine exhaust emissionsas well as emissions estimates for numerous gasturbine modifications and uprates. (For site-specific emissions values, customers should con-tact GE.) Additionally, the effects of emissioncontrol methods are provided for gas turbinecycle performance and recommended turbineinspection intervals. Emission control methodsvary with both internal turbine and externalexhaust system emission control. Only the inter-nal gas turbine emission control methods —lean head end liners and water/steam injection— will be covered in this paper.

In the early 1970s when emission controls wereoriginally introduced, the primary regulatedgas turbine emission was NOx. For the relative-ly low levels of NOx reduction required in the1970s, it was found that injection of water orsteam into the combustion zone would producethe desired NOx level reduction with minimaldetrimental impact to the gas turbine cycle per-formance or parts lives. Additionally, at thelower NOx reductions the other exhaust emis-sions generally were not adversely affected.Therefore GE has supplied NOx water andsteam injection systems for this applicationsince 1973.

With the greater NOx reduction requirementsimposed during the 1980s, further reductionsin NOx by increased water or steam injectionbegan to cause detrimental effects to the gasturbine cycle performance, parts lives andinspection criteria. Also, other exhaust emis-

sions began to rise to measurable levels of con-cern. Based on these factors, alternative meth-ods of emission controls have been developed:

■ Internal gas turbine

—Multiple nozzle quiet combustors introduced in 1988

—Dry Low NOx combustors introduced in 1990

■ External

—Exhaust catalysts

This paper will summarize the current estimat-ed emissions for existing gas turbines and theeffects of available emission control techniques(liner design and water/steam injection) on gasturbine emissions, cycle performance, andmaintenance inspection intervals. The latesttechnology includes Dry Low NOx and catalyticcombustion. These topics are covered in otherGERs.

Emissions Characteristics ofConventional Combustion SystemsTypical exhaust emissions from a stationary gasturbine are listed in Table 1. There are two dis-tinct categories. The major species (CO2, N2,H2O, and O2) are present in percent concen-trations. The minor species (or pollutants)such as CO, UHC, NOx, SOx, and particulatesare present in parts per million concentrations.In general, given the fuel composition andmachine operating conditions, the majorspecies compositions can be calculated. Theminor species, with the exception of total sulfuroxides, cannot. Characterization of the pollu-tants requires careful measurement and semi-theoretical analysis.

The pollutants shown in Table 1 are a functionof gas turbine operating conditions and fuelcomposition. In the following sections, eachpollutant will be considered as a function of


GE Power Systems ■ GER-4211 ■ (03/01) 1

operating conditions under the broad divisionsof gaseous and liquid fuels.

Nitrogen Oxides

Nitrogen oxides (NOx = NO + NO2) must bedivided into two classes according to theirmechanism of formation. Nitrogen oxidesformed from the oxidation of the free nitrogenin the combustion air or fuel are called “ther-mal NOx.” They are mainly a function of thestoichiometric adiabatic flame temperature ofthe fuel, which is the temperature reached byburning a theoretically correct mixture of fueland air in an insulated vessel.

The following is the relationship between com-bustor operating conditions and thermal NOxproduction:

■ NOx increases strongly with fuel-to-airratio or with firing temperature

■ NOx increases exponentially withcombustor inlet air temperature

■ NOx increases with the square root ofthe combustor inlet pressure

■ NOx increases with increasingresidence time in the flame zone

■ NOx decreases exponentially withincreasing water or steam injection orincreasing specific humidity

Emissions which are due to oxidation of organ-ically bound nitrogen in the fuel—fuel-boundnitrogen (FBN)—are called “organic NOx.”Only a few parts per million of the available freenitrogen (almost all from air) are oxidized toform nitrogen oxide, but the oxidation of FBNto NOx is very efficient. For conventional GEcombustion systems, the efficiency of conver-sion of FBN into nitrogen oxide is 100% at lowFBN contents. At higher levels of FBN, the con-version efficiency decreases.

Organic NOx formation is less well understoodthan thermal NOx formation. It is important tonote that the reduction of flame temperatures



Major Species Typical Concentration Source(% Volume)

Nitrogen (N2) 66 - 72 Inlet Air

Oxygen (O2) 12 - 18 Inlet Air

Carbon Dioxide (CO2) 1 - 5 Oxidation of Fuel Carbon

Water Vapor (H2O) 1 - 5 Oxidation of Fuel Hydrogen

Minor Species Typical Concentration SourcePollutants (PPMV)

Nitric Oxide (NO) 20 - 220 Oxidation of Atmosphere Nitrogen

Nitrogen Dioxide (NO2) 2 - 20 Oxidation of Fuel-Bound Organic Nitrogen

Carbon Monoxide (CO) 5 - 330 Incomplete Oxidation of Fuel Carbon

Sulfur Dioxide (SO2) Trace - 100 Oxidation of Fuel-Bound Organic Sulfur

Sulfur Trioxide (SO3) Trace - 4 Oxidation of Fuel-Bound Organic Sulfur

Unburned Hydrocarbons (UHC) 5 - 300 Incomplete Oxidation of Fuel or Intermediates

Particulate Matter Smoke Trace - 25 Inlet Ingestion, Fuel Ash, Hot-Gas-Path

Attrition, Incomplete Oxidation of Fuel or

Intermediates

Table 1. Gas turbine exhaust emissions burning conventional fuels

to abate thermal NOx has little effect on organ-ic NOx. For liquid fuels, water and steam injec-tion actually increases organic NOx yields.Organic NOx formation is also affected by tur-bine firing temperature. The contribution oforganic NOx is important only for fuels thatcontain significant amounts of FBN such ascrude or residual oils. Emissions from thesefuels are handled on a case-by-case basis.

Gaseous fuels are generally classified accordingto their volumetric heating value. This value isuseful in computing flow rates needed for agiven heat input, as well as sizing fuel nozzles,combustion chambers, and the like. However,the stoichiometric adiabatic flame temperatureis a more important parameter for characteriz-ing NOx emission. Table 2 shows relative ther-mal NOx production for the same combustorburning different types of fuel. This table showsthe NOx relative to the methane NOx based onadiabatic stoichiometric flame temperature.The gas turbine is controlled to approximateconstant firing temperature and the products ofcombustion for different fuels affect the report-ed NOx correction factors. Therefore, Table 2also shows columns for relative NOx values cal-culated for different fuels for the same combus-tor and constant firing temperature relative tothe NOx for methane.

Typical NOx performance of the MS7001EA,MS6001B, MS5001P, and MS5001R gas turbines

burning natural gas fuel and No. 2 distillate isshown in Figures 1–4 respectively as a function offiring temperature. The levels of emissions forNo. 2 distillate oil are a very nearly constantfraction of those for natural gas over the oper-ating range of turbine inlet temperatures. Forany given model of GE heavy-duty gas turbine,NOx correlates very well with firing tempera-ture.

Low-Btu gases generally have flame tempera-tures below 3500°F/1927°C and correspond-ingly lower thermal NOx production. However,depending upon the fuel-gas clean-up train,these gases may contain significant quantities ofammonia. This ammonia acts as FBN and willbe oxidized to NOx in a conventional diffusioncombustion system. NOx control measures suchas water injection or steam injection will havelittle or no effect on these organic NOxemissions.

Carbon Monoxide Carbon monoxide (CO) emissions from a con-ventional GE gas turbine combustion system areless than 10 ppmvd (parts per million by vol-ume dry) at all but very low loads for steady-state operation. During ignition and accelera-tion, there may be transient emission levelshigher than those presented here. Because ofthe very short loading sequence of gas turbines,these levels make a negligible contribution tothe integrated emissions. Figure 5 shows typical



NOx (ppmvd/ppmvw-Methane) NOx (ppmvd/ppmvw-Methane) @Fuel Stoichiometric 1765°F/963°C – 2020°F/1104°C 15% O2, 1765°F/963°C – 2020°F/1104°C

Flame Temp. Firing Time Firing Time

Methane 1.000 1.000/1.000 1.000/1.000

Propane 1.300 1.555/1.606 1.569/1.632

Butane 1.280 1.608/1.661 1.621/1.686

Hydrogen 2.067 3.966/4.029 5.237/5.299

Carbon Monoxide 2.067 3.835/3.928 4.128/0.529

Methanol 0.417-0.617 0.489/0.501 0.516/0.529

No. 2 Oil 1.667 1.567/1.647 1.524/1.614

Table 2. Relative thermal NOx emissions

CO emissions from a MS7001EA, plotted versusfiring temperature. As firing temperature isreduced below about 1500°F/816°C the carbon

monoxide emissions increase quickly. Thischaracteristic curve is typical of all heavy-dutymachine series.



1/4 Load

1/2 Load

Full Load

3/4 Load

No. 2 Oil

Natural Gas

280

240

200

160

120

80

40

0

ISO Conditions

Firing Temperature

NO

(pp

mvw

)X

1000 1200 1400 1600 1800 2000

GT

25056

540 650 760 870 980 1090

(°F)

(°C)

Figure 1. MS7001EA NOx emissions

ISO Conditions

Firing Temperature

NO

(pp

mvw

)X

1/4 Load

1/2 Load

Full Load

3/4 Load

No. 2 Oil

Natural Gas

320

280

240

200

160

120

80

40

01000 1200 1400 1600 1800 2000

GT

25057

540 650 760 870 980 1090

(°F)

(°C)

Figure 2. MS6001B NOx emissions

Unburned Hydrocarbons Unburned hydrocarbons (UHC), like carbonmonoxide, are associated with combustion inef-ficiency. When plotted versus firing tempera-ture, the emissions from heavy-duty gas turbine

combustors show the same type of hyperboliccurve as carbon monoxide. (See Figure 6.) At allbut very low loads, the UHC emission levels forNo. 2 distillate and natural gas are less than7 ppmvw (parts per million by volume wet).



Firing Temperature

ISO Conditions

1/4 Load

1/4 Load

1/2 Load

Full Load

3/4 Load

No. 2 Oil

Natural Gas

NO

(pp

mv

w)

X

200

160

120

80

40

01000 1200 1400 1600 1800

GT

25

05

8

540 650 760 870 980

(°F)

(°C)

Figure 3. MS5001P A/T NOx emissions

GT

25059

ISO Conditions

Firing Temperature

NO

(pp

mvw

)X

1/4 Load

1/2 Load

Full Load

3/4 Load

No. 2 Oil

Natural Gas

160

120

80

40

01000 1200 1400 1600 1800

540 650 760 870 980

(°F)

(°C)

Figure 4. MS5001R A/T NOx emissions

Sulfur Oxides

The gas turbine itself does not generate sulfur,which leads to sulfur oxides emissions. All sulfuremissions in the gas turbine exhaust are caused

by the combustion of sulfur introduced into theturbine by the fuel, air, or injected steam orwater. However, since most ambient air andinjected water or steam has little or no sulfur,the most common source of sulfur in the gas



GT

25060

Ga

sTu

rbin

eM

ac

hin

eE

xh

au

st

CO

(pp

mv

d)

Firing Temperature

1/4 Load

1/2 Load 3/4 Load Full LoadDistillate

Oil

Natural Gas

200

160

120

80

40

0800 1000 1200 1400 1600 1800 2000 2200

430 540 650 760 870 980 1090 1200

(°F)

(°C)

Figure 5. CO emissions for MS7001EA

0

Firing Temperature

Ga

sTu

rbin

eM

ac

hin

eE

xh

au

st

UH

C(p

pm

vw

)

1/2 Load

1/4 Load

Full Load3/4 Load

Natural Gas

DistillateOil

120

100

80

40

60

20

600 800 1000 1200 1400 1600 1800 2000 2200

GT

25

06

1

320 430 540 650 760 870 980 1090 1200

(°F)

(°C)

Figure 6. UHC emissions for MS7001EA

turbine is through the fuel. Due to the latest hotgas path coatings, the gas turbine will readilyburn sulfur contained in the fuel with little orno adverse effects as long as there are no alkalimetals present in the hot gas.

GE experience has shown that the sulfur in thefuel is completely converted to sulfur oxides. Anominal estimate of the sulfur oxides emissionsis calculated by assuming that all fuel sulfur isconverted to SO2. However, sulfur oxide emis-sions are in the form of both SO2 and SO3.Measurements show that the ratio of SO3 toSO2 varies. For emissions reporting, GE reportsthat 95% of the sulfur into the turbine is con-verted to SO2 in the exhaust. The remainingsulfur is converted into SO3. SO3 combines withwater vapor in the exhaust to form sulfuric acid.This is of concern in most heat recovery appli-cations where the stack exhaust temperaturemay be reduced to the acid dew point tempera-ture. Additionally, it is estimated that 10% byweight of the SOx generated is sulfur mist. By

using the relationships above, the various sulfuroxide emissions can be easily calculated fromthe fuel flow rate and the fuel sulfur content asshown in Figure 7.

There is currently no internal gas turbine tech-nique available to prevent or control the sulfuroxides emissions from the gas turbine. Controlof sulfur oxides emissions has typically requiredlimiting the sulfur content of the fuel, either bylower sulfur fuel selection or fuel blending withlow sulfur fuel.

Particulates Gas turbine exhaust particulate emission ratesare influenced by the design of the combustionsystem, fuel properties and combustor operat-ing conditions. The principal components ofthe particulates are smoke, ash, ambient non-combustibles, and erosion and corrosion prod-ucts. Two additional components that could beconsidered particulate matter in some localitiesare sulfuric acid and unburned hydrocarbonsthat are liquid at standard conditions.



TYPICAL BASE LOADFUEL FLOW:

100 80 60 40 20 4

1.0

0.8

0.6

0.4

0.2

8 12 16 20

1600

1200

800

400

40

80

120

160

51P 4.7 lb/sec61B 6.2 lb/sec71EA 13.0 lb/sec91E 18.5 lb/sec

SO (lb/hr)3

SO /SO3 20.0658 by Weight

Total Fuel Flow Rate (lb/sec)

% Sulfur by Weight

Su

lfu

rM

ist

Em

iss

ion

Ra

te(l

b/h

r)S

O(l

b/h

r)2

GT25062

Figure 7. Calculated sulfur oxide and sulfur emissions

Smoke Smoke is the visible portion of filterable partic-ulate material. The GE combustor design cou-pled with air atomization of liquid fuels hasresulted in a nonvisible plume over the gas tur-bine load range for a wide variety of fuels. TheGE smoke-measuring unit is the Von BrandReflective Smoke Number (GEVBRSN). If thisnumber is greater than 93 to 95 for theMS7001E, then the plume will not be visible.For liquid fuels, the GEVBRSN is a function ofthe hydrogen content of the fuel. For naturalgas fuel, the smoke number is essentially 99 to100 over the load range and visible smoke is notpresent.

Dry Emissions Estimates at Base LoadThe ISO non-abated full load emissions esti-mates for the various GE heavy-duty gas turbinemodels are provided in Table 3. The natural gasand #2 distillate fuel emission estimates shownare for thermal NOx, CO, UHC, VOC, and par-ticulates. For reporting purposes, all particu-

lates are also reported as PM-10. Therefore PM-10 is not shown in the tables. The nominal fullrated firing temperature for each gas turbinemodel is also shown in Table 3.

As can be easily seen in the table, at base loadwithout NOx abatement, the emissions of CO,UHC, VOC, and particulates are quite low. Theestimated values of NOx vary between gas tur-bine designs and generally increase with theframe size firing temperature.

Dry Emissions Estimates at Part Load

Simple-Cycle Turbines At turbine outputs below base load the emis-sions change from the values given in Table 3.These changes are affected by the turbine con-figuration and application and in some cases bythe turbine controls.

Single-shaft gas turbines with non-modulatinginlet guide vanes operating at constant shaftspeed have part load emissions characteristicswhich are easily estimated. For these turbines



MS5001P 1730/943 128 195 25 42MS5001P-N/T 1765/963 142 211 25 42

MS6001B 2020/1104 161 279 25 65/42

MS7001B 1840/1004 109 165 25 42MS7001B Option 3 1965/1074 124 191 25 42MS7001B Option 4 2020/1104 132 205 25 42

MS7001EA 2020/1104 160 245 25 42

MS9001B 1940/1060 109 165 42 65MS9001B Option 3 1965/1074 124 191 42 65MS9001B Option 4 2020/1104 132 205 42 65

MS9001E 2020/1104 157 235 42 65MS9001E 2055/1124 162 241 42 65

6FA 2350/12887FA 2400/13167FA 2420/13279FA 2350/1288

MS3002F 1575/1625/857/885 115 201 42 50MS3002J 1730/943 128 217 42 50

MS3002J-N/T 1770/968 140 236 42 50

MS5002 1700/927 125 220 42 50MS5002B-N/T 1770/966 137 255 42 50

* S.C. = Simple Cycle and R.C. = Regenerative Cycle** Two-Shaft NOx Levels Are All on Gas Fuel

Single Shaft UnitsModel

Firing Temp.F/C

Gas(FG1A/FG1B)

Gas(FG1C/FG1F)Gas

Dry (Non-Abated) H2O/Steam Inj.

Dist.

Two Shaft Units*Model

Firing Temp.F/C S.C. S.C.S.C.

Dry (Non-Abated) H2O/Steam Inj.

R.C.**

GT

2328

9E

Table 3. NOx emission levels @ 15% O2 (ppmvd)



the NOx emissions vary exponentially with fir-ing temperature as shown previously in Figures1–4. The load points for each turbine are alsomarked on these figures. Due to the conver-sions used in the various NOx reporting meth-ods, the information in Figures 1–4 has beenredrawn in Figures 8–11. This information showsthe estimated ISO NOx emissions on a ppmvd@ 15% O2, ppmvw, and lb/hour basis forMS7001EA, MS6001B, MS5001P and MS5001R.In these figures, the nominal peak load firingtemperature point is also given. It should benoted that in some cases the NOx ppmvd@15%O2 reporting method can cause number valuesto increase as load is reduced (e.g., see theMS5001P A/T in Figure 10.) Since the GEMS9001E gas turbine is a scaled version of theMS7001E gas turbine, the MS7001E gas turbinefigures can be used as an estimate of MS9001Egas turbine part load emissions characteristics.

Many gas turbines have variable inlet guidevanes that are modulated closed at part loadconditions in order to maintain higher exhaust

temperatures for waste heat recovery equip-ment located in the gas turbine exhaust. Asshown in Figure 12, closing the inlet guide vaneshas a slight effect on the gas turbine NOx emis-sions. Figure 12 shows the effect on NOx ppmvd@ 15% O2 and Figure 13 shows the effect onNOx lb/hr. The figures show both MS5001Pand MS7001E characteristics. They also shownormalized NOx (% of base load value) vs. %base load. Curves are shown for load reductionsby either closing the inlet guide vanes whilemaintaining exhaust temperature control andfor load reductions by reducing firing tempera-ture while keeping the inlet guide vanes fullyopen.

Mechanical drive gas turbines typically vary theoutput load shaft speed in order to adjust theturbine output to match the load equipmentcharacteristic. Single-shaft gas turbines operat-ing on exhaust temperature control have a max-imum output NOx emissions characteristic vs.turbine shaft speed, as shown in Figure 14 for anMS5001R Advanced Technology uprated tur-

Firing Temperature

Peak Load

Full Load

1/4 Load1/2 Load

3/4 Load

600

500

400

300

200

100 100

400

600

800

1000

1200

800 1000 1200 1400 1600 1800 2000 22000 0

1. NO ppmvd @ 15% O - Chaindashed Curve

2. NO lb/hr - Dashed Curve

3. NO ppmvw - Solid Curve

x 2

x

x

NOTES:D - No. 2 DistillateG - Methane Natural GasISO Conditions

1

2

3

D

G

DGDG

NO

(pp

mv)

x

NO

(lb

/hr)

x

430 540 650 760 870 980 1090 1200

(°F)

(°C)

GT25063




bine. The characteristic shown is primarily dueto the gas turbine exhaust temperature controlsystem and the turbine thermodynamics. Asseen in Figure 14, as the turbine output shaft

speed is reduced below 100%, NOx emissionsdecrease directly with turbine shaft speed. Asthe speed decreases, the exhaust temperatureincreases till the exhaust component tempera-

Firing Temperature

Peak Load

Full Load

1/4 Load1/2 Load

3/4 Load

400

350

300

250

200

150

100

50 100

200

300

400

500

600

700

800

800 1000 1200 1400 1600 1800 2000 22000 0




x 2

x

x


D

G

G

D

NO

(pp

mv)

x

NO

(lb

/hr)

x

430 540 650 760 870 980 1090 1200

(°F)

(°C)

GT25064


8000

D

G

DD

GG

NO

(lb

/hr)

X

Firing Temperature

NO

(pp

mv)

X

3/4 Load

1/2 Load

1/4 Load

Peak Load

Full Load

250

200

500

400

300

200

100

01000 1200 1400 1600 1800

150

100

50




x 2

x

x


GT25065

430 540 650 760 870 980

(°F)

(°C)




ture limit is reached. Once the exhaust isother-mal limit is reached, the variation of NOx emis-sions with speed will become greater. In Figure 16this exhaust isothermal temperature limit isreached at approximately 84% speed. Two-shaftgas turbines also vary the output turbine shaft

speed with load conditions. However the gas tur-bine compressor shaft and combustor operatingconditions are controlled independent of theoutput shaft speed. On a two-shaft gas turbine, ifthe gas turbine compressor shaft speed is heldconstant by the control system while on exhaust

NO

(lb

/hr)

X

Firing Temperature

NO

(pp

mv)

X3/4 Load

1/2 Load1/4 Load

D

G

D

DGG

Peak Load

Full Load

200

160

400

360

320

280

240

200

160

120

80

40

0800 1000 1200 1400 1600 1800

120

80

40

0




x 2

x

x


GT25066

430 540 650 760 870 980

(°F)

(°C)


75 80 85 90 95 100

% Base Load

%N

O@

Base

Lo

ad

-p

pm

vd

@15%

OX

2 105

100

95

90

85

80

75

ISO Conditions

1. 51P Closing IGV’s2. 51P Dropping Firing Temperature3. 71E Closing IGV’s4. 71E Dropping Firing Temperature

3

1

2 4

GT

25067

Figure 12. Inlet guide vane effect on NOx ppmvd @ 15% O2 vs. load


temperature control, the NOx emissions are notaffected by the load turbine shaft speed.

Exhaust Heat Recovery Turbines Regenerative cycle and waste heat recovery two-shaft gas turbines are normally controlled tooperate the gas turbine compressor at the min-imum speed allowable for the desired load out-put. As load is increased from minimum, the

gas turbine compressor speed is held at mini-mum until the turbine exhaust temperaturereaches the temperature control curve. Withfurther increase in load, the control system willincrease the gas turbine compressor speedwhile following the exhaust temperature con-trol curve. If the turbine has modulated inletguide vanes, the inlet guide vanes will open firstwhen the exhaust temperature control curve is


75 80 85 90 95 100

% Base Load

%N

O@

Ba

se

Lo

ad

-L

b/H

ou

rX

105

100

95

90

85

80

75

70

65

ISO Conditions

1. 51P Closing IGV’s2. 51P Dropping Firing Temperature3. 71E Closing IGV’s4. 71E Dropping Firing Temperature

3

1

24

GT

25

06

8

Figure 13. Inlet guide vane effect on NOx lb/hour vs. load

75 80 85 90 95 100 105 110 115 120

Percent Compressor Speed

NO

Va

lue

sX

110

100

90

80

70

60

50

40

30

Lb/Hr

ppmvw

NOTES:

ISO Conditions

100% Compressor Speed = 5100 rpm

Natural Gas Fuel

Assumes Exhaust Isothermal LimitReached at 84 Percent Speed and Below

ppmvd @ 15% O2

GT

25069

Figure 14. MS5001R A/T NOx emissions vs. shaft speed

reached, and then, once the inlet guide vanesare fully open, the gas turbine compressorspeed will be increased.

Figure 15 shows the NOx characteristic of aregenerative cycle MS3002J gas turbine at ISOconditions. Initially, as load is increased, NOxincreases with firing temperature while the gasturbine compressor is operating at minimumspeed. For the turbine shown, the exhaustisothermal temperature control is reached at

approximately 48% load. The gas turbine com-pressor shaft speed is then increased by the con-trol system for further increases in load up tothe 100% load point. At approximately 96%load, the gas turbine exhaust temperature con-trol curve begins to limit exhaust temperaturebelow the isothermal exhaust temperature dueto the increasing airflow through the turbineand the NOx values are reduced by the charac-teristic shown.

For a typical regenerative cycle MS5002BAdvanced Technology gas turbine with modu-lated inlet guide vanes, the curve of NOx vs.load at ISO conditions is shown in Figure 16.

The NOx vs. load characteristic is similar to theMS3002J. However, this design turbine willoperate at low load with the inlet guide vanespartially closed and at minimum operating gasturbine compressor shaft speed. During initialloading, NOx increases with firing temperature.When the exhaust temperature control systemisothermal temperature limit is reached theinlet guide vanes are modulated open as load isincreased. At approximately 90% load the gas

turbine exhaust temperature control curvebegins to limit exhaust temperature below theisothermal exhaust temperature due to theincreasing airflow through the turbine and theNOx values are reduced. At approximately91.5% load for this turbine calculation, the inletguide vanes are fully open and further increas-es in load are accomplished by increasing thegas turbine compressor speed resulting in theNOx reduction as shown.

Other NOx Influences The previous sections of this paper consider theinternal gas turbine design factors which influ-



Percent Load

NOTES:ISO ConditionsConstant LP Shaft Speed

20100

120

140

160

180

200

220

30 40 50 60 70 80 90 100

NO

(pp

mv

d)

@1

5%

Ox

2

GT

25

07

0

Figure 15. MS3002J regenerative NOx vs. load

ence emissions generation. There are manyexternal factors to the gas turbine which impactthe formation of NOx emissions in the gas tur-bine cycle. Some of these factors will be dis-cussed below. In all figures under this topic, theNOx is presented as a percentage value where100% represents the thermal ISO NOx value forthe turbine operating on base temperature con-trol. For all figures except for the regeneratorchanges discussed, the curves drawn representa single “best fit” line through the calculatedcharacteristics for frame 3, 5, 6, 7, and 9 gas tur-bines. However, the characteristics shape that isshown is the same for all turbines.

Ambient Pressure. NOx ppm emissions varyalmost directly with ambient pressure. Figure 17provides an approximation for the ambientpressure effect on NOx production on a lb/hrbasis and on a ppmvd @ 15% O2 basis. This fig-ure is at constant 60% relative humidity. Itshould be noted that specific humidity varieswith ambient pressure and that this variation isalso included in the Figure 18 curves.

Ambient Temperature. Typical NOx emissionsvariation with ambient temperature is shown in

Figure 18. This figure is drawn at constant ambi-ent pressure and 60% relative humidity with thegas turbine operating constant gas turbine fir-ing temperature. For an operating gas turbinethe actual NOx characteristic is directly influ-enced by the control system exhaust tempera-ture control curve, which can change the slopeof the curves. The typical exhaust temperaturecontrol curve used by GE is designed to holdconstant turbine firing temperature in the59°F/15°C to 90°F/32°C ambient temperaturerange. The firing temperature with this typicalcurve causes under-firing of approximately20°F/11°C at 0°F/–18°C ambient, and approxi-mately 10°F/6°C under-firing at 120°F/49°Cambient. Factors such as load limits, shaft out-put limits, and exhaust system temperature lim-its are also not included in the Figure 18 curves.Based on the actual turbine exhaust tempera-ture control curve used and other potential lim-itations that reduce firing temperature, the esti-mated NOx emissions for an operating gas tur-bine are typically less than the values shown inFigure 18 at both high and low ambients.

Relative Humidity. This parameter has a very



x2

NO

(pp

mvd

@15%

O)

Percent Load

NOTES:ISO ConditionsConstant LP Shaft Speed

40 50 60 70 80 90 100

250

200

150

100

GT

25071

Figure 16. MS5002B A/T regenerative NOx vs. load

strong impact on NOx. The ambient relativehumidity effect on NOx production at constantambient pressure of 14.7 psia and ambient tem-peratures of 59°F/15°C and 90°F/32°C isshown in Figure 19.

The impact of other parameters such asinlet/exhaust pressure drops, regenerator char-acteristics, evaporative/inlet coolers, etc., aresimilar to the ambient parameter effectsdescribed above. Since these parameters are



1. NO (lb/hr)/(ISO lb/hr)

2. NO (ppmvd @ 15% O )/(ISO ppmvd @ 15% O )22

x

x

9 10 11 12 13 14 15

Ambient Pressure

NO

Perc

en

tag

eX

100

90

80

70

60

50

40

Curve Drawn at 59°F/15°C, 60% Relative Humidity

100% - Base Load Value at ISO Conditions

2

1

GT

25073A

0.62 0.68 0.75 0.82 0.89 0.96 1.03

psia

bar

Figure 17. Ambient pressure effect on NOx Frames 5, 6 and 7

Ambient Temperature

NO

Perc

en

tag

eX

120

110

100

90

80

70

60

50

400 20 40 60 80 100 120

2

1

Curve Drawn at 14.7 psia/1.013 bar, 0% Relative Humidity100% = Base Load Value at 59°F ambient

x

x

1. NO (lb/hr)/(ISO lb/hr)

2. NO (ppmvd @ 15% O )/(ISO ppmvd @ 15% O )2 2

GT

25074A

-18 -7 4 16 27 38 49

(°F)

(°C)

Figure 18. Ambient temperature effect on NOx Frames 5, 6 and 70% Relative Humidity

usually unit specific, customers should contactGE for further information.

Power Augmentation Steam Injection. The effectof power augmentation steam injection on gasturbine NOx emissions is similar to NOx steaminjection on a ppmvw and lb/hr basis. However,only approximately 30% of the power augmenta-tion steam injected participates in NOx reduc-tion. The remaining steam flows through dilu-tion holes downstream of the NOx producingarea of the combustor. 100% of the power aug-mentation steam injected is used in the conver-sion from ppmvw to ppmvd @ 15% O2.

Emission Reduction TechniquesThe gas turbine, generally, is a low emitter ofexhaust pollutants because the fuel is burnedwith ample excess air to ensure completecombustion at all but the minimum load condi-tions or during start-up. The exhaust emissionsof concern and the emission control techniquescan be divided into several categories as shown inTable 4. Each pollutant emission reduction tech-nique will be discussed in the following sections.

Nitrogen Oxides Abatement The mechanism on thermal NOx productionwas first postulated by Zeldovich. This is shownin Figure 20. It shows the flame temperature ofdistillate as a function of equivalence ratio.This ratio is a measure of fuel-to-air ratio inthe combustor normalized by stoichiometricfuel-to-air ratio. At the equivalence ratio ofunity, the stoichiometric conditions arereached. The flame temperature is highest atthis point. At equivalence ratios less than 1, wehave a “lean” combustor. At the values greaterthan 1, the combustor is “rich.” All gas turbinecombustors are designed to operate in the leanregion.

Figure 20 shows that thermal NOx production

rises very rapidly as the stoichiometric flame

temperature is reached. Away from this point,

thermal NOx production decreases rapidly.

This theory then provides the mechanism

of thermal NOx control. In a diffusion flame

combustor, the primary way to control thermal

NOx is to reduce the flame temperature.



x

x

x

x

1. NO (lb/hr)/(lb/hr at 59°F/15°C)

2. NO (ppmvd @ 15% O )/(ppmvd @ 15% O at 59°F/15°C)

3. NO (lb/hr)/(lb/hr at 90°F/32°C)

4. NO (ppmvd @ 15% O )/(ppmvd @ 15% O at 90°F/32°C)2

2 2

2

Percent Relative Humidity

NO

Perc

en

tag

eX

150

140

130

120

110

100

90

80

700 20 40 60 80 100

Curves Drawn at 14.7 psia/1.013 bar

100% - Base Load Value at ISO Conditions4

3

2

1

GT

25

07

5

Figure 19. Relative humidity effect on NOx Frames 5, 6 and 7

Lean Head End (LHE) Combustion Liners Since the overall combustion system equiva-lence ratio must be lean (to limit turbine inlettemperature and maximize efficiency), the firstefforts to lower NOx emissions were naturally

directed toward designing a combustor with aleaner reaction zone. Since most gas turbinesoperate with a large amount of excess air, someof this air can be diverted towards the flameend, which reduces the flame temperature.



NOx Lean Head End LinerWater or Steam InjectionDry Low NOx

CO Combustor Design Catalytic Reduction

UHC & VOC Combustor Design

SOx Control Sulfur in Fuel

Particulates & PM-10 Fuel Composition

Smoke Reduction Combustor Design - Fuel Composition- Air Atomization

Particulate Reduction Fuel Composition- Sulfur- Ash G

T25

092

Table 4. Emission control techniques

x

Equivalence Ratio

Fla

me

Tem

pera

ture

High SmokeEmissions

Temperature

NO

High COEmissions

No. 2 Oil, 10 ATMAir Preheat 590 K (600°F)

0.5 1.0 1.5

40002500

2000

300

200

100

1500

1000

3000

2000

RichLean

(K)

(°F)

x

Rate of Productionof Thermal NO

d NOdt

(ppmv/MS)

GT

11

65

7B

Figure 20. NOx production rate

Leaning out the flame zone (reducing theflame zone equivalence ratio) also reduces theflame length, and thus reduces the residencetime a gas molecule spends at NOx formationtemperatures. Both these mechanisms reduceNOx. The principle of a LHE liner design isshown in Figure 21.

It quickly became apparent that the reductionin primary zone equivalence ratio at full oper-ating conditions was limited because of thelarge turndown in fuel flow (40 to 1), air flow(30 to 1), and fuel/air ratio (5 to 1) in industri-al gas turbines. Further, the flame in a gas tur-bine is a diffusion flame since the fuel and airare injected directly into the reaction zone.Combustion occurs at or near stoichiometricconditions, and there is substantial recircula-tion within the reaction zone. These parametersessentially limit the extent of LHE liner tech-nology to a NOx reduction of 40% at most.Depending upon the liner design, actual reduc-tion achieved varies from 15% to 40%.

Figure 22 compares an MS5001P LHE liner to astandard liner. The liner to the right is the LHE

liner. It has extra holes near the head (flame)end and also has a different louver pattern com-pared to the standard liner. Table 5 summarizesall LHE liners designed to date. Field test dataon MS5002 simple-cycle LHE liners andMS3002J simple-cycle LHE liners are shown inFigures 23–25.

One disadvantage of leaning out the head endof the liner is that the CO emissions increase.This is clear from Figure 24, which compares CObetween the standard and LHE liner for aMS5002 machine.

Water/Steam Injection Another approach to reducing NOx formationis to reduce the flame temperature by introduc-ing a heat sink into the flame zone. Both waterand steam are very effective at achieving thisgoal. A penalty in overall efficiency must bepaid for the additional fuel required to heat thewater to combustor temperature. However, gasturbine output is enhanced because of the addi-tional mass flow through the turbine. By neces-sity, the water must be of boiler feedwater qual-



• LHE Liner has same diameter and length as

standard liner shown at left.

• The number, diameter, and location of the

mixing and dilution holes is different in the

LHE liner.

•As a result,

– more air is introduced in the head endof the LHE combustor

– NOx emissions decrease

mixingholes

dilutionhole

Figure 21. Standard simple-cycle MS5002 combustion liner

ity to prevent deposits and corrosion in the hotturbine gas path area downstream of the com-bustor.

Water injection is an extremely effective meansfor reducing NOx formation; however, the com-bustor designer must observe certain cautionswhen using this reduction technique. To maxi-mize the effectiveness of the water used, fuel

nozzles have been designed with additional pas-sages to inject water into the combustor headend. The water is thus effectively mixed with theincoming combustion air and reaches the flamezone at its hottest point. In Figure 26 the NOxreduction achieved by water injection is plottedas a function of water-to-fuel ratio for anMS7001E machine. Other machines have simi-lar NOx abatement performance with waterinjection.

Steam injection for NOx reduction followsessentially the same path into the combustorhead end as water. However, steam is not aseffective as water in reducing thermal NOx. Thehigh latent heat of water acts as a strong ther-mal sink in reducing the flame temperature.

In general, for a given NOx reduction, approxi-mately 1.6 times as much steam as water on amass basis is required for control.

There are practical limits to the amount ofwater or steam that can be injected into thecombustor before serious problems occur. Thishas been experimentally determined and mustbe taken into account in all applications if thecombustor designer is to ensure long hardwarelife for the gas turbine user.



Figure 22. Louvered low NOx lean head end combustion liners

Turbine Model Laboratory Development Completed First Field Test

S/C MS3002F December-98 Fall 1999S/C MS3002G December-98 to be determinedS/C MS3002J April-97 March-99

S/C MS5002B, C, & D April-97 September-97S/C MS5001 (All Models) 1986 Over 130 operating in field

R/C MS3002J February-99 to be determinedR/C MS5002B & C February-99 to be determined

GER 3751-19

GT

2563

4

Table 5. Lean head end (LHE) liner development

Injecting water/steam in a combustor affectsseveral parameters:

1. Dynamic Pressure Activity within theCombustor. Dynamic pressures can bedefined as pressure oscillations withinthe combustor driven by non-uniform

heat release rate inherent in anydiffusion flame or by the weakcoupling between heat release rate,turbulence, and acoustic modes. Anexample of the latter is selectiveamplification of combustion roar by



NO

Em

issio

ns

(pp

mvd

@15%

O)

x2

Combustor Exit Temperature

650

1200 1400 1600 1800 2000

760 870 980 1090

(°F)

(°C)

• Symbols are field testpoints collected inAlaska, September1997

• Solid lines areexpectations, fromscaled lab NOemissions

• Field test confirmed~40% NO reduction atbase load

• Good agreementbetween lab and field

x

x

Standard

LHE

140

120

100

80

60

40

20

0

Figure 23. Field test data: simple-cycle MS5002 NOx

CO

Em

iss

ion

s(p

pm

vd

)

Combustor Exit Temperature

300

250

200

150

100

50

1200

650

1400

760

1600

870

1800

980

2000

1090

0

Standard, Field

Standard, Lab

LHE, Field

LHE, Lab

(°F)

(°C)

• Field test confirmedsmall increase inCO at base load,larger increase atpart load conditions

• Good agreementbetween lab andfield

Figure 24. Field test data: simple-cycle MS5002 CO

the acoustic modes of the duct.Frequencies range from near zero toseveral hundred hertz. Figure 27 showsdynamic pressure activity for bothwater injection and steam injection foran MS7001E combustor. Water

injection tends to excite the dynamicactivity more than steam injection.The oscillating pressure loads on thecombustion hardware act as vibratoryforcing functions and therefore mustbe minimized to ensure long hardware



NO

xE

mis

sions

(ppm

v,dry

,15%

O2)

CO

Em

issi

on

s(p

pm

v,d

ry,1

5%

O2)

Standard

LHE (steam Off)

Standard

LHE (steam Off)

1800 1900170019001700 1800

25

0

5

10

15

20

0

25

50

75

100

125

Combustor Exit Temperature (°F)

• 30% reduction in NOx with negligible increase in CO.

• Injecting steam further reduces NOx.

NO

Em

issi

on

s(p

pm

vd

@1

5%

O)

x2

Combustor Exit Temperature Combustor Exit Temperature

927 982 1038

(°F)

(°C)

• 30% reduction in NO with negligible increase in CO.• Injecting steam further reduces NO .

x

x

CO

Em

issi

on

s(p

pm

vd

@1

5%

O) 2

927 982 1038

(°F)

(°C)

Figure 25. Field test data: simple-cycle MS3002J with steam injection for power augmentation

xx

Ra

tio

of

NO

Wit

hIn

j.to

NO

Wit

ho

ut

(pp

mv

d/p

pm

vd

)

Water-to-Fuel Mass Ratio

Distillate Oil

Natural Gas

0 0.2 0.4 0.6 0.8 1.0

1.00.90.80.7

0.6

0.5

0.4

0.3

0.2

0.1

GT

25

10

8

Figure 26. MS7001E NOx reduction with water injection

life. Through combustor designmodifications such as the addition of amulti-nozzle fuel system, significantreductions in dynamic pressure activityare possible.

2. Carbon Monoxide Emissions. As moreand more water/steam is added to thecombustor, a point is reached at whicha sharp increase in carbon monoxideis observed. This point has beendubbed the “knee of the curve”. Oncethe knee has been reached for anygiven turbine inlet temperature, onecan expect to see a rapid increase incarbon monoxide emissions with thefurther addition of water or steam.Obviously, the higher the turbine inlettemperature, the more tolerant thecombustor is to the addition of waterfor NOx control. Figure 28 shows therelationship of carbon monoxideemissions to water injection for aMS7001B machine for natural gas fuel.Figure 29 shows the effect of steam

injection on CO emissions for a typicalMS7001EA. Unburned hydrocarbonshave a similar characteristic with NOxwater or steam injection as carbonmonoxide. Figure 30 shows theMS7001EA gas turbine unburnedhydrocarbon versus firing temperaturecharacteristic with steam injection.

3. Combustion Stability. Increasingwater/steam injection reducescombustor-operating stability.

4. Blow Out. With increasingwater/steam injection, eventually apoint will be reached when the flamewill blow out. This point is theabsolute limit of NOx control withwater/steam injection.

Carbon Monoxide Control There are no direct carbon monoxide emissionreduction control techniques available withinthe gas turbine. Basically the carbon monoxideemissions within the gas turbine combustor canbe viewed as resulting from incomplete com-



2.01.81.61.41.21.00.80.60.4

1.8

0

1.2

1.0

2.2

2.0

1.6

1.4

0.2

Water

60-70% Load

Baseload

Peakload

Steam

Load

Distillate Fuel

Steam

Water

Water/Fuel Mass Flow Ratio

Rati

oof

RM

SD

yn

am

ic

Pre

ssu

reL

evel

s-

Wet

Ov

erD

ry

Figure 27. MS7001E combustor dynamic pressure activity

bustion. Since the combustor design maximizescombustion efficiency, carbon monoxide emis-sions are minimized across the gas turbine loadrange of firing temperatures. Reviewing Figure 5shows that the carbon monoxide emission levelsincrease at lower firing temperatures. In some

applications where carbon monoxide emissionsbecome a concern at low loads (firing tempera-tures), the increase in carbon monoxide can belowered by:

■ reducing the amount of water/steam



Water Injection – % of Compressor Inlet Air Flow

Carb

on

Mo

no

xid

e(p

pm

vd

)

1260/682

1460/793

1665/907

1870/1021

Fuel is Natural Gas

Firing Temperatures (°F/°C)

50

40

30

20

10

00 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6

GT10284A

Figure 28. Carbon monoxide vs. water injection effect of firing temperature – MS7001B

2

Natural Gas With Steam Injectionto 42 ppmvd @ 15% O

2

Distillate Oil withSteam Injection

to 65 ppmvd @ 15% O

Distillate Oil withNo Diluent Injection

Firing Temperature

Gas

Tu

rbin

eM

ach

ine

Exh

au

st

CO

(pp

mvd

)

Natural Gas With No Diluent Injection

800 1000 1200 1400 1600 1800 2000 2200

200

160

120

80

40

0

GT

25

08

0

430 540 650 760 870 980 1090 1200

(°F)

(°C)


injection for NOx control (if allowed)

– or –

■ closing the inlet guide vanes, whichwill increase the firing temperature forthe same load.

Unburned Hydrocarbons Control Similar to carbon monoxide, there are also nodirect UHC reduction control techniques usedwithin the gas turbine. UHCs are also viewed asincomplete combustion, and the combustor isdesigned to minimize these emissions. Thesame indirect emissions control techniques canbe used for unburned hydrocarbons as for car-bon monoxide.

Particulate and Smoke Reduction Control techniques for particulate emissionswith the exception of smoke are limited to con-trol of the fuel composition.

Although smoke can be influenced by fuel com-position, combustors can be designed whichminimize emission of this pollutant. Heavy fuels

such as crude oil and residual oil have lowhydrogen levels and high carbon residue, whichincrease smoking tendencies. GE has designedheavy-fuel combustors that have smoke per-formance comparable with those which burndistillate fuel.

Crude and residual fuel oil generally containalkali metals (Na, K) in addition to vanadiumand lead, which cause hot corrosion of the tur-bine nozzles and buckets at the elevated firingtemperatures of today's gas turbine. If the fuel iswashed, water soluble compounds (alkali salts)containing the contaminants are removed.Filtration, centrifuging, or electrostatic precipi-tation are also effective on reducing the solidcontaminants in the combustion products.

Contaminants that cannot be removed from thefuel (vanadium compounds) can be controlledthrough the use of inhibitors. GE uses additionof magnesium to control vanadium corrosion inits heavy-duty gas turbines. These magnesiumadditives always form ash within the hot gaspath components. This process generally



2

Natural Gas with Steam Injectionto 42 ppmvd @ 15% O

Firing Temperature

Ga

sTu

rbin

eM

ac

hin

eE

xh

au

st

UH

C(p

pm

vd

)

Natural Gas with No Diluent Injection

Distillate Oil Withand Without

Diluent Injection

600 800 1000 1200 1400 1600 1800 2000 2200

120

100

80

60

40

20

0

GT

25

08

1

320 430 540 650 760 870 980 1090 1200

(°F)

(°C)


requires control and removal of added ashdeposits from the turbine. The additional ashwill contribute to the exhaust particulate emis-sions. Generally, the expected increase can becalculated from an analysis of the particular fuelbeing burned.

In some localities, condensable compoundssuch as SO3 and condensable hydrocarbons areconsidered particulates. SO3, like SO2, can bestbe minimized by controlling the amount of sul-fur in the fuel. The major problem associatedwith sulfur compounds in the exhaust comesfrom the difficulty of measurement. Emissionsof UHCs, which are a liquid or solid at roomtemperature, are very low and only make aminor contribution to the exhaust particulateloading.

Water/Steam Injection HardwareThe injection of water or steam into the com-bustion cover/fuel nozzle area has been the pri-mary method of NOx reduction and control inGE heavy-duty gas turbines since the early1970s. The same design gas turbine equipment

is supplied for conversion retrofits to existinggas turbines for either injection method. BothNOx control injection methods require a micro-processor controller, therefore turbines witholder controls need to have their control sys-tem upgraded to Mark V or Mark VISPEEDTRONIC™ controls conversion. Thecontrol system for both NOx control injectionmethods utilizes the standard GE gas turbinecontrol philosophy of two separate independ-ent methods for shutting off the injection flow.

The NOx water injection system is shownschematically in Figure 31 and consists of a waterpump and filter, water flowmeters, water stopand flow control valves. This material is sup-plied on a skid approximately 10 x 20 feet insize for mounting at the turbine site. The waterfrom the skid is piped to the turbine base whereit is manifold to each of the fuel nozzles usingpigtails. The water injection at the combustionchamber is through passages in the fuel nozzleassembly. A typical water injection fuel nozzleassembly is shown schematically in Figure 32.For this nozzle design there are eight or twelve



Figure 31. Schematic piping – water injection system

water spray nozzles directing the water injectionspray towards the fuel nozzle tip swirler. Whilethis design is quite effective in controlling theNOx emissions, the water spray has a tendencyto impinge on the nozzle tip swirler and on theliner cap/cowl assembly. Resulting thermalstrain usually leads to cracks, which limits thecombustion inspections to 8000 hours or less.To eliminate this cracking, the latest designwater-injected fuel nozzle is the breech-loadfuel nozzle. (See Figure 33.) In this design thewater is injected through a central fuel nozzlepassage, injecting the water flow directly intothe combustor flame. Since the water injectionspray does not impinge on the fuel nozzleswirler or the combustion cowl assembly, thebreech load fuel nozzle design results in lowermaintenance and longer combustion inspec-tion intervals for NOx water injection applica-tions.

The NOx steam injection system is shownschematically in Figure 34, and consists of asteam flowmeter, steam control valve, steam

stop valve, and steam blowdown valves. Thismaterial is supplied loose for mounting nearthe turbine base by the customer. The steam-injection flow goes to the steam-injection mani-fold on the turbine base. Flexible pigtails areused to connect from the steam manifold toeach combustion chamber. The steam injectioninto the combustion chamber is throughmachined passages in the combustion cancover. A typical steam-injection combustioncover with the machined steam-injection pas-sage and steam injection nozzles is shown inFigure 35.

Water quality is of concern when injecting wateror steam into the gas turbine due to potentialproblems with hot gas path corrosion, andeffects to the injection control equipment. Theinjected water or steam must be clean and freeof impurities and solids. The general require-ments of the injected water or steam quality areshown in Table 6. Total impurities into the gasturbine are a total of the ambient air, fuel, andinjected water or steam. The total impurities



Fuel Gas Connection

Oil Connection

Water Injection Inlet

AtomizingAir Connection

GT

25

08

5

Figure 32. Water injection fuel nozzle assembly

requirement may lower the water or steam-injection quality requirements. It is importantto note that the total impurities requirement isprovided relative to the input fuel flow.

Minimum NOx LevelsAs described above, the methods used to reducethermal NOx inside the gas turbine are by com-bustor design or by diluent injection. To see



Fuel Gas Connection

Distillate Fuel Inlet

Water Injection Inlet

Atomizing Air Connection

GT

25

08

6

Figure 33. Breech-load fuel nozzle assembly

Figure 34. Schematic piping – steam injection system



GT25088

NOTE: This drawing is not to be used for Guarantees

Figure 35. Combustion cover – steam injection

• WATER/STEAM QUALITYTotal Dissolved Solids 5.0 ppm Max.Total Trace Metals 0.5 ppm Max.

(Sodium + Potassium+ Vanadium + Lead)

pH 6.5 – 7.5

NOTE: Quality requirements can generallybe satisfied by demineralized water.

Max. EquivalentConcentration

Contaminant (ppm – wt)

Sodium + Potassium 1.0Lead 1.0Vanadium 0.5Calcium 2.0

• TOTAL LIMITS IN ALL SOURCES(Fuel, Steam, Water, Air)

Table 6. Water or steam injection quality requirements

NOx emissions from each frame size withoutany control, refer to Table 3. With the LHE linerdesign, dry (no water/steam injection) NOxemissions could be reduced by 15–40% relativeto standard liner. This is the limit of LHE linertechnology.

With water or steam injection, significant reduc-tion in NOx is achieved. The lowest achievableNOx values with water/steam injection from GEheavy-duty gas turbines are also shown in Table3. The table provides the current minimumNOx levels for both methane natural gas fueland #2 distillate fuel oil.

Maintenance EffectsAs described previously, the methods used tocontrol gas turbine exhaust emissions have aneffect on the gas turbine maintenance intervals.Table 7 provides the recommended combustioninspection intervals for current designAdvanced Technology combustion systems usedin base load continuous duty gas turbines with-out NOx control systems and the recommendedcombustion inspection intervals with the vari-

ous NOx control methods at the NOx ppmvd @15% O2 levels shown. Both natural gas fuel and#2 distillate fuel recommended combustioninspection intervals are included. Review ofTable 7 shows that the increased combustiondynamics (as the combustor design goes fromdry to steam injection) and then to water injec-tion results in reductions in the recommendedcombustion inspection intervals.

Performance EffectsAs mentioned previously the control of NOxcan impact turbine firing temperature andresult in gas turbine output changes.Additionally, the injection of water or steam alsoimpacts gas turbine output, heat rate, andexhaust temperature. Figure 36 shows theimpact of NOx injection on these gas turbineparameters when operating at base load for allsingle shaft design gas turbines. Note that theinjection rate is shown as a percentage of thegas turbine compressor inlet airflow on a weightbasis. The output and heat rate change is shownon a percent basis while exhaust temperature is



Natural Gas/ Natural Gas Fired Hours No. 2 Distillate Fired HoursNo. 2 Distillate of Operation of Operation

ppmvd @ 15% O2 Water/Steam Injection Water/Steam Injection

MS5001P N/T Dry 142/211 12,000/12,000 12,000/12,000NSPS 87/86 12,000/12,000 6,000/6,000

42/65 6,000/6,000 6,000/6,00042/42 6,000/6,000 1,500/4,000

MS6001B Dry 148/267 12,000/12,000 12,000/12,000NSPS 94/95 8,000/8,000 6,000/6,000

42/65 8,000/8,000 8,000/8,00042/42 8,000/8,000 4,000/4,000

MS7001E Dry 154/228 8,000/8,000 8,000/8,000NSPS 96/97 8,000/8,000 8,000/8,000

42/65 6,500/8,000 6,500/8,00042/42 6,500/8,000 1,500/3,000

MNQC 25/42 8,000/8,000 6,000/6,000

MS9001E Dry 147/220 8,000/8,000 8,000/8,00042/65 6,500/8,000 6,500/8,000

Inspection Intervals reflect current hardware. Older units with earlier vintage hardware will have lower Inspection intervals.

The above values represent initial recommended combustion inspection intervals. The intervals are subject to change based on experience.

Base Load Operation.NSPS NOx levels are 75 ppm with heat rate correction included.

GT

2509

3

Table 7. Estimated ISO NOx level effects on combustion inspection intervals

shown in degrees F. Review of Figure 36 showsthat turbine output is increased when NOxinjection is used. The gas turbine load equip-ment must also be capable of this outputincrease or control changes must be made inorder to reduce the gas turbine output.

SummaryThe emissions characteristics of gas turbineshave been presented both at base load and partload conditions. The interaction of emissioncontrol on other exhaust emissions as well as

the effects on gas turbine maintenance and per-formance have also been presented. The mini-mum controllable NOx levels using LHE andwater/steam injection techniques have alsobeen presented. Using this information, emis-sions estimates and the overall effect of the var-ious emission control methods can be estimat-ed.

It is not the intent of this paper to provide site-specific emissions. For these values, the cus-tomer must contact GE.



% O

utpu

t In

crea

se%

Hea

t R

ate

Incr

ease

Cha

nge

in E

xhau

st T

emp

0

5

10

1 2

1 2

1 2

Diluent Injection (% Compressor Inlet Flow)

0

-2

-4

-6

-8

0

-1.1

-2.2

-3.3

-4.4

0

-4

4

-2

2

Solid Line = Water Inj for 5001Dashed Line = Steam Inj for 5001Chaindashed Line = Water Inj for 61, 71, 91Dotted Line = Steam Inj for 61, 71, 91

Figure 36. Performance effects vs. diluent injection

List of FiguresFigure 1. MS7001EA NOx emissions






Figure 7. Calculated sulfur oxide and sulfur emissions





Figure 12. Inlet guide vane effect on NOx ppmvd @ 15% O2 vs. load

Figure 13. Inlet guide vane effect on NOx lb/hour vs. load

Figure 14. MS5001R A/T NOx emissions vs. shaft speed

Figure 15. MS3002J regenerative NOx vs. load

Figure 16. MS5002B A/T regenerative NOx vs. load

Figure 17. Ambient pressure effect on NOx frame 5, 6 and 7

Figure 18. Ambient temperature effect on NOx frame 5, 6 and 7

Figure 19. Relative humidity effect on NOx frame 5, 6 and 7

Figure 20. NOx production rate

Figure 21. Standard simple cycle MS5002 combustor liner

Figure 22. Louvered low NOx lean head end combustion liners

Figure 23. Field test data: simple-cycle MS5002 NOx

Figure 24. Field test data: simple-cycle MS5002 CO

Figure 25. Field test data: simple-cycle MS3002J with steam injection for power augmentation

Figure 26. MS7001E NOx reduction with water injection

Figure 27. MS7001E combustor dynamic pressure activity

Figure 28. Carbon monoxide vs. water injection effect of firing temperature – MS7001B



Figure 31. Schematic piping – water injection system

Figure 32. Water injection fuel nozzle assembly

Figure 33. Breech-load fuel nozzle assembly

Figure 34. Schematic piping – steam injection system

Figure 35. Combustion cover – steam injection

Figure 36. Performance effects vs. diluent injection



List of TablesTable 1. Gas turbine exhaust emissions burning conventional fuels

Table 2. Relative thermal NOx emissions

Table 3. NOx emission levels at 15% O2 (ppmvd)

Table 4. Emission control techniques

Table 5. Lean head end (LHE) liner development

Table 6. Water or steam injection quality requirements

Table 7. Estimated ISO NOx level effects on combustion inspection intervals



Documents

GER 4211 - Gas Turbine Emissions and Control Turbine Emissions and Control.pdfWorldwide interest in gas turbine emissions and the enactment of Federal and State regulations in the