Lefebvre A.H., Ballal D.R. Gas Turbine Combustion: Alternative Fuels and Emissions

Подождите немного. Документ загружается.

420 Gas Turbine Combustion: Alternative Fuels and Emissions, Third Edition

lean-zone equivalence ratios, as well as by combustor inlet temperature. A

minimum lean-zone equivalence ratio of 0.6 was needed to achieve satisfactory

smoke levels. In more recent work, Rizk and Mongia [83] have applied three-

dimensional emissions modeling, using well-established reaction mechanisms,

to RQL combustion. Their results generally conrm the previous ndings of

Novick et al. in regard to the importance of rich-zone equivalence ratio to NO

emissions, but they also stress the contribution to NO

formation of residence

time and combustion pressure.

In Japan, Nakata et al. [79,84] have designed an RQL combustor for a 150 MW

class, stationary gas turbine with the double objective of maintaining stable

combustion when burning LHV gas and reducing the NO

emissions that are

produced from the NH

in fuel. An interesting feature of the design is that

strong swirl is imparted to the air and fuel as they enter the fuel-rich primary

zone with an equivalence ratio of 1.6. Tests carried out at atmospheric pres-

sure gave very satisfactory results in terms of good combustion stability, and

low NO

emissions (3 ppm for combustor exit temperatures up to 1500°C),

albeit at the expense of fairly high CO emissions.

The GE RQL2 combustor for LHV gas also features a swirl-stabilized

fuel-rich primary zone. This zone terminates in a converging section that

serves both to prevent the swirling ow from drawing lean-stage gases back

upstream into the primary zone and to reduce the ow area to a reason-

able size for proper quenching [85]. Rapid quenching is achieved by injecting

the air through holes of different sizes to obtain a uniform distribution of

quench air across the hot gas stream. When operating at 1670 K combustor

exit temperature, the NO

and CO emissions were 50 and 5 ppmv, respec-

tively. Perhaps of greater signicance is that the conversion of NH

to NO

was only 5%.

The RQL concept is being actively studied for aircraft applications by the

Pratt and Whitney company and other laboratories in the United States as

part of NASA’s HSCT (High Speed Civil Transport) program. The aim of this

program is to demonstrate the feasibility of attaining NO

levels of 3–8 g/kg

fuel (i.e., around 40–100 ppmv) at supersonic cruise conditions with kerosine

fuel [86].

Most of the work carried out so far on the RQL concept has conrmed its

potential for ultralow NO

combustion and low conversion of FBN into NO

With LHV fuels, the conversion of NH

into NO

is also greatly reduced. In

comparison with conventional combustors, RQL combustors have inherently

better ignition and lean blowout performance. In comparison with staged

combustors, they have the important practical advantage of needing fewer

fuel injectors. However, in order to fully exploit these assets, signicant

improvements in quench mixer design are needed. With liquid fuels, other

potential problems include high soot formation in the rich primary zone,

which could give rise to high ame radiation and exhaust smoke. These prob-

lems are exacerbated by long residence times, unstable recirculation patterns,

and nonuniform mixing.

Emissions 421

9.10 Catalytic Combustion

Catalytic combustion is a process that employs a catalyst to initiate and pro-

mote chemical reactions in a owing premixed fuel–air mixture at leaner

conditions than are possible in homogeneous gas-phase combustion. This

allows stable combustion to proceed at equivalence ratios that are below the

normal lean ammability limit of the fuel–air mixture. Combustion at such

reduced temperatures can be expected to dramatically decrease the produc-

tion of thermal NO

The principle of catalytic combustion is shown schematically in Figure 9.38.

Fuel is injected upstream of the reactor to vaporize and mix with the inlet

air. The fuel–air mixture then ows into a catalyst bed, or reactor, which may

consist of several stages, each made of a different kind of catalyst. For the

rst stage, it is desirable to use a catalyst that is active at low temperatures,

whereas subsequent stages need to be selected for good oxidation efciency.

Downstream of the catalytic bed, a thermal reaction zone is usually provided

to raise the gas temperature to the required turbine entry value and to reduce

the concentrations of CO and UHC to acceptable levels.

The potential of catalytic reactors for very low pollutant emissions has been

recognized since 1975 [87], but the harsh environment in a gas turbine com-

bustor and its wide range of operating conditions pose formidable problems

that must be overcome to secure the implementation of catalytic combustion

in operational gas turbines [88]. It is difcult to design a catalyst that will

ignite a fuel–air mixture at the low compressor exit temperatures corre-

sponding to crank lighting and engine operation at low load. Moreover, com-

bustor exit temperatures are usually in the range from 1450 to 1770 K, which

are well above the stability limits of most catalyst substrate materials. Even

ceramics that can withstand high combustion temperatures are susceptible

to thermal shock failure during engine transients. The doubts concerning

Fuel

Air

Fuel-air

preparation

Catalytic

reactor

Thermal

reactor

Dilution

Figure 9.38

Schematic representation of catalytic combustor.

422 Gas Turbine Combustion: Alternative Fuels and Emissions, Third Edition

the long-term durability of catalyst substrates and sustained high catalytic

activity for periods up to several thousand hours constitute a signicant bar-

rier to the development of viable catalytic combustors for gas turbines.

Current research and development activities in catalytic combustion

for gas turbines are mainly in the form of small-scale laboratory tests and

combustor rig tests. Many different concepts have been examined. Some of

the designs tested are fairly complex, and require the use of a preburner

along with fuel staging and/or variable geometry to achieve a satisfactory

operating range. However, the results obtained from these various stud-

ies fully conrm the potential of catalytic combustion for ultralow NO

(<5 ppmv) and provide the incentive to overcome the formidable challenges

posed by durability issues and the limited temperature range capability of

catalytic reactors.

9.10.1 Design Approaches

The various approaches to the design of catalytic combustors for gas turbines

have been described in a number of publications [89–99]. They fall broadly

into three main categories:

1. Traditional systems, as illustrated in Figure 9.38, in which the cata-

lyst is fed with a fuel–air mixture whose adiabatic ame tempera-

ture is equal to the required combustor outlet temperature, T

. This

approach is satisfactory for low values of T

but, on modern gas

turbines where T

could be as high as 1570 K, it would give rise to

problems such as sintering and vaporization of the active catalyst

components and thermal shock fracturing of ceramic supports.

2. Systems where only a part of the fuel is injected upstream of the

catalyst to limit catalyst temperatures below 1270 K, and the rest

of the fuel is injected downstream of the catalyst to achieve the

desired combustor outlet temperature. Examples of this approach

may be found in Cowell and Larkin [95], Ozawa et al. [96], and Fujii

et al. [97].

3. Systems where all of the fuel is injected upstream of the catalyst, but

is only partially reacted within the catalyst bed. Combustion then

proceeds to completion via homogeneous gas-phase reactions in a

postcatalyst combustion zone. The designs of Vortmeyer et al. [98],

Dalla Betta et al. [89,91], Schlatter et al. [92], and Dutta et al. [93,94]

use this approach.

The main objective of Approaches 2 and 3 is to keep substrate tempera-

tures low to prevent problems of thermal sintering and catalyst deactivation.

Limiting the maximum catalyst temperature in this way not only extends

catalyst life, but also broadens the selection of suitable catalyst components.

Emissions 423

9.10.2 Design Constraints

The problems facing the designer of catalytic combustors tend to focus on

temperature and the need to reconcile the broad range of temperatures

required by the engine with the relatively narrow range of temperatures

over which the catalyst bed can satisfy the conicting requirements of high

catalytic activity and mechanical integrity.

Any given catalytic combustor has a certain range of operating conditions

over which it can achieve stable combustion and low emission levels. This

range of conditions, which is often referred to as an operating “window,” is

determined primarily with regard to three important temperatures [91–94].

1. The inlet mixture temperature, which must be high enough to acti-

vate the catalyst. Below a certain minimum temperature, known as

the “lightoff’ temperature, the exothermic oxidation reactions occur-

ring on the reactor walls are too slow to generate the heat needed to

sustain the reactions. In general, temperatures in excess of 700 K are

necessary for catalyst lightoff. For example, noble metals, such as

platinum and palladium, require 617–783 K lightoff temperatures,

whereas metal oxide catalysts with active elements, such as nickel

and cobalt, require 866–1367 K temperatures [99].

2. The gas temperature leaving the catalyst must be high enough to

allow the catalytically initiated reactions to proceed to completion

in the available residence time or, if additional fuel is injected down-

stream of the catalyst, to promote rapid combustion of this fuel,

thereby reducing the emissions of CO and UHC to acceptable levels.

3. Catalyst wall temperatures must be low enough to provide the dura-

bility needed for stable long-term reactor operation.

9.10.3 Fuel Preparation

Successful operation of the catalyst bed is highly dependent on its entry ow

conditions, which should be very uniform in terms of mixture strength,

velocity, and temperature to ensure effective use of the entire catalyst area

and prevent damage to the substrate because of local high temperatures.

Failure to achieve a completely uniform mixture will detract from perfor-

mance in a number of ways that follow directly from the concept of an oper-

ating window, as discussed above. For any given inlet temperature, a low

fuel/air ratio may result in low catalyst temperatures and high emissions

of CO and UHC. On the other hand, too high a fuel/air ratio may cause the

catalyst to overheat. Clearly, if there are local variations in temperature and

fuel/air ratio in the mixture entering the reactor, some regions of the catalyst

will experience conditions on the low temperature (high CO and UHC emis-

sions) side of the operating window, whereas other regions will be exposed

to potential damage to the substrate due to local high gas temperatures

424 Gas Turbine Combustion: Alternative Fuels and Emissions, Third Edition

[91,93]. Tests carried out by Dutta et al. [94] showed that inhomogeneities in

fuel concentration greater than 10% (peak to peak) can lead to catalyst dam-

age and high CO and UHC emissions. Nonuniformities in inlet conditions

also make it more difcult to maintain the catalyst temperature within the

operating window, which curtails the degree of exibility in responding to

changes in engine operating conditions [91].

Mechanical mixers are sometimes used to promote more uniform ow

conditions at inlet to the catalyst bed [89,96–98]. The advantages and draw-

backs of mixers in premixed combustion applications have already been

discussed in the context of LPP combustors. Their main drawback is the

obvious one—the introduction of additional pressure loss that penalizes

engine performance. Also in common with LPP combustors, the residence

time in the fuel-preparation zone must be short enough to avoid the risk of

spontaneous ignition.

9.10.4 Catalyst Bed Construction

A common catalyst specication is one where the major active ingredient

(usually a metal, such as palladium, platinum, or magnesium, or a metal

oxide) is applied to a stabilized alumina washcoat on a honeycomb-type

ceramic monolith, such as cordierite. The washcoat is porous, to provide a

large surface area, and may contain a dispersion of active catalyst particles,

such as aluminum oxide or zirconium oxide.

One problem with ceramic monoliths is that as they become larger, their

mechanical and thermal reliability is reduced. This problem can be overcome

by replacing a large catalyst bed with a number of small segments that are

clustered together without cementing [96,97]. Although this method of con-

struction eases the problem, the durability of ceramic monoliths is always a

cause for concern, especially in heavy-duty gas turbines. A more robust form

of construction is one whereby each stage of the catalyst reactor is formed

by corrugating a strip of oxidation-resistant metal foil, and coiling the strip

in such a way as to form a channeled monolith structure through which

the fuel–air mixture can ow and react on the channel walls [91,92,100]. The

active catalytic material is deposited as a coating on the foils.

The reactor employed by Beebe et al. [100] and Dalla Betta et al. [91] con-

sisted of three such stages, each designed to deliver gas at the appropriate

temperature to the next stage or to a nal homogeneous combustion section.

An interesting feature of the design [90] is the manner in which it takes advan-

tage of the unique thermodynamics of palladium oxidation and reduction

to control surface temperatures. Palladium oxide decomposes to palladium

metal at temperatures between 1050 and 1200 K, depending on the pressure.

This transition between oxide and metal can be exploited to limit the catalyst

temperature, which allows a corrugated metal (Fecralloy) support to be used

instead of the less durable ceramic substrate.

Emissions 425

9.10.5 Postcatalyst Combustion

The main function of the postcatalyst combustion zone is to reduce the con-

centrations of CO and UHC to the required levels and to achieve the desired

combustor exit temperature. Placing the zone of maximum gas temperature

downstream of the catalyst allows the catalyst bed to operate at a relatively

low temperature (<1270 K) with consequent advantages in terms of reliability

and extended life. It is of interest to note that the temperature gradient in a

catalytic combustor is in the opposite direction to that in a conventional gas

turbine combustor. In the latter, maximum temperatures are attained in the

primary combustion zone and gas temperatures then decline in the down-

stream direction as more air is injected into the hot gas stream to reduce its

temperature to the required combustor exit value.

Dalla Betta et al. [89] have pointed out the following advantages of post-

catalyst combustion:

It allows the catalyst to operate at low temperatures that can reduce •

or eliminate many of the deactivation mechanisms that curtail cata-

lyst life

It permits the use of a wide variety of catalyst and substrate materials•

The problems associated with thermal shock fracture of substrates •

during startup, shut down, and transient operations are alleviated

The combustor can be developed to higher outlet temperatures with-•

out changing the catalyst material

9.10.6 Design and Performance

Two different types of catalytic combustors for gas turbines are described

below to illustrate the main features of current designs. Fujii et al. [97] used

Approach 2 in designing a catalytic combustor to operate with an outlet

temperature of 1570 K while keeping the catalyst bed temperature down to

around 1270 K. The system, shown schematically in Figure 9.39, consists of an

annular preburner, six fan-shaped catalyst segments, six premixing fuel noz-

zles, and a premix combustion section downstream of the catalyst. The inlet

air is heated to 720 K by the preburner and is distributed to both the catalytic

segments and the premixing nozzles, which are arranged alternately in the

form of a circle, as shown in Figure 9.39. The major active ingredient of the

catalysts is palladium, which is supported on a stabilized alumina washcoat

on a honeycomb-type monolith made of cordierite [96]. At its downstream

end, each nozzle is shaped to inject the premixed gas–air mixture owing

between the catalyst segments into the catalyst efux at an angle close to 90°.

The mixture is ignited by the hot combustion products emanating from the

catalyst segments, and LPM combustion then occurs to raise the combustor

outlet temperature to 1570 K.

426 Gas Turbine Combustion: Alternative Fuels and Emissions, Third Edition

When operating on natural gas over a pressure range from 0.10 to 1.35

MPa (1–13.5 bar), NO

emissions were always below 10 ppmv, of which a

signicant portion came from the preburner. The maximum overall combus-

tor pressure loss was about 2.5%, which is well below the values normally

associated with conventional combustors.

The alternative method (Approach 3) for avoiding exposure of the catalyst

to excessive temperatures is illustrated schematically in Figure 9.40, which

Igniter

Preburner

CatalystMixer

Baffle

Fuel nozzle

Fuel

Igniter

Air

Liner

Premixing

nozzle

Premixing

nozzle

Fuel nozzle

Combustion

gas

Transition piece

Catalyst

A — A section

Figure 9.39

High-pressure catalytic combustor. (From Ozawa, Y., Hirano, J., Sato, M., Saiga, M., and

Watanabe, S., Journal of Engineering for Gas Turbines and Power, 116, 511–16, 1994; Fujii, T.,

Ozawa, Y., Kikumoto, S., and Sato, M., ASME Paper 96-GT-382, 1996. With permission.)

Preburner

fuel inlet

Preburner

Main

fuel inlet

Main fuel

injector

Catalyst

Video

camera

Post catalyst

reaction volume

Transition piece

Nozzle box

(turbine inlet)

Air inlet

Perforated plate

Figure 9.40

GE catalytic combustor in test stand. (From Schlatter, J.C., Dalla Betta, R.A., Nickolas, S.G.,

Cutrone, M.B., and Beebe, K.W., ASME Paper 97-GT-57, 1997. With permission.)

Emissions 427

shows a catalytic combustor now under development by the General Electric

Company for its natural gas-red MS9001E gas turbine [92]. The technology

used to design this combustor has been demonstrated in a number of sub-

scale and full scale tests, e.g., References [89,91]. A basic feature of the concept

is that all the fuel required to achieve the combustor exit temperature of

1380 Κ is supplied to the catalytic reactor, which comprises two or more sep-

arate stages. The rst stage operates at a relatively low temperature, which

makes possible the high catalytic activity needed for catalyst operation at

the compressor outlet temperature. The last stage can operate successfully

with a lower catalytic activity because its inlet temperature is higher, but its

substrate temperature must also be high to provide the outlet gas tempera-

tures needed to initiate the homogeneous combustion reactions downstream

of the catalyst. Typically, about half of the fuel is reacted within the catalyst

stages and the remainder is consumed in the postcatalyst combustion zone.

The catalytic reactor for the GE MS9001E consists of three individually sup-

ported stages, each 508 mm in diameter. The catalyst stages are formed by

corrugating strips of very thin (50 µm) oxidation-resistant metal foil and then

depositing the active catalytic material (palladium oxide) as a coating on the

strips [92]. The strips are coiled in order to form channels through which the

fuel–air mixture ows while reacting on the channel walls. The overall length

of the catalyst container is 305 mm of which the catalyst occupies 230 mm.

The fuel-injection system consists of 93 individual venturi tubes arranged

across the ow path, with four fuel-injection orices at the throat of each

venturi [101]. The objective is to deliver to the catalyst a fuel–air mixture that

is uniform in composition, temperature, and velocity. The target for fuel-air

uniformity is a maximum range of 10% between the highest and lowest

concentration. By suitable tailoring of the individual fuel-injector orices, a

range of 12% has been achieved so far, which is close to the target value.

Results of tests carried out in June 1996 at the simulated base load operat-

ing point (P

= 1.25 MPa, T

= 714 K, T

= 1465 K), indicated NO

levels below

5 ppmv and even lower concentrations of CO and UHC [94].

9.10.7 use of Variable geometry

Depending on the type of catalyst employed and the degree of homogene-

ity achieved in the fuel–air mixture entering the reactor, stable combustion

can be sustained over a temperature range of only a few hundred degrees

Kelvin without serious loss of performance. This corresponds to a range of

overall fuel/air ratios of around 1.4–1, as opposed to the 5–1, which is readily

achieved with conventional combustors. It is essential, therefore, that cata-

lytic combustion be combined with variable geometry and/or fuel staging to

maintain stable combustion over the broad range of conditions encountered

in modern gas turbines.

One example of the use of variable geometry to broaden the operating

range of an engine tted with a catalytic combustor is the Solar design for

428 Gas Turbine Combustion: Alternative Fuels and Emissions, Third Edition

the Advanced Turbine Systems (ATS) program. This program is aimed at the

development of advanced recuperative gas turbines with pollutant emissions

at single digit levels. The Solar catalytic combustion system is modular in

design and includes a fuel–air mixer upstream of the reactor and a postcatalyst

homogeneous gas-phase reaction zone downstream of the catalyst bed to com-

plete the combustion process. Startup is accomplished using a conventional

(LPM) low-emissions fuel injector. The system transitions to catalyst operation

using a variable geometry valve that diverts airow into the catalyst at loads

greater than 50% of full load. This arrangement overcomes the limitations cre-

ated by the narrow turndown ratio of the catalyst and allows the catalyst outlet

temperature to remain fairly constant over the 50–100% load range [93].

9.10.8 Future

All the laboratory and combustor rig tests carried out on catalytic combus-

tors for gas turbines have amply demonstrated their capabilities in regard to

ultralow pollutant emissions. Considerable progress has also been made in

increasing the life expectancy of catalysts under typical high-temperature con-

ditions. Although most combustor rig testing to date has involved relatively

short test periods of less than 100 hours, one test has been reported in which

a catalyst system operated satisfactorily for 4500 hours on natural gas fuel at

atmospheric pressure and an adiabatic combustion temperature of 1570 K

without any apparent damage to the catalyst and its support structure [89].

Most of the work now in progress on catalytic combustion for gas turbines

is directed toward stationary applications and is aimed at meeting the broad

operating range and arduous operating conditions of gas turbines. Work is

continuing on catalyst development to lower the required reactor inlet temper-

ature and system development to enlarge the operating window. Engine tests

over long periods are now called for to fully validate the technology. The cur-

rent dearth of experience on stationary engines certainly eliminates catalytic

combustion from serious consideration for aircraft applications at this time.

9.11 Correlation and Modeling of Oxides of Nitrogen

and Carbon Monoxide Emissions

Many empirical and semi-empirical models are now in widespread use for

correlating experimental data on pollutant emissions in terms of all the rel-

evant parameters. These include combustor dimensions, design features,

and operating conditions, as well as fuel type and fuel spray characteristics.

Empirical models can play an important role in the design and development

of low-emissions combustors. They serve to reduce the complex problems

associated with emissions to forms that are more meaningful and tractable

Emissions 429

to the combustion engineer. They also permit more accurate correlations of

emissions for any one specic combustor than can be achieved by more com-

prehensive numerical models.

As the chemical reactions governing the formation of UHC and smoke are

highly complex, most of the empirical models developed so far have been

conned to NO

and CO. Lefebvre [102] assumed for both these species that

their exhaust concentrations are dependent on three terms, which are selected

to represent the following: (1) mean residence time in the combustion zone,

(2) chemical reaction rates, and (3) mixing rates. Expressions for these three

parameters were derived in terms of combustor size, liner pressure loss, air-

ow proportions, and operating conditions of inlet pressure, temperature, and

air mass ow rate. This approach led to the development of semi-empirical

expressions for NO

and CO, which are presented below.

9.11.1 Oxides of Nitrogen Correlations

According to Lefebvre [102], we have

= 9 × 10

−8

1.25

exp(0.01T

)/m˙

g/kg fuel. (9.7)

The values of the constant and exponents in this equation were obtained

from analysis of experimental data on NO

emissions from several different

aero-engine combustors. Equation 9.7 takes account of the fact that in the

combustion of heterogeneous fuel–air mixtures, it is the stoichiometric ame

temperature that determines the formation of NO

, not the average ame tem-

perature. However, for the residence time in the combustion zone, which is also

signicant to NO

formation, the appropriate temperature term is the average

value T

, as indicated in the denominator of Equation 9.7. The excellent cor-

relation of experimental data on NO

provided by Equation 9.7 for GE J79-17A

and F101 combustors is illustrated in Figures 9.41 and 9.42, respectively.

Equation 9.7 is suitable for conventional spray combustors only. For LPP

combustors, in which the maximum attainable temperature is T

, it may still

be used, provided that T

is substituted for T

Many other semi-empirical models for predicting NO

emissions have

been derived. For a critical review of the models developed before 1980, ref-

erence should be made to Mellor [103]. Some of the more recent expressions

for NO

emissions include the following.

9.11.1.1 Odgers and Kretschmer [104]

= 29 exp −(21,670/T

0.66

× [1 − exp −(250τ)]g/kg fuel. (9.8)

Odgers and Kretschmer’s recommended NO

formation times for aircraft

combustors are 0.8 ms (airblast atomizers) and 1.0 ms (pressure atomizers).

Quoted formation times for industrial combustors burning liquid fuels range

from 1.5 to 2.0 ms.