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

Подождите немного. Документ загружается.

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

Most of the research carried out on effervescent atomization [36–43] has

used atomizers of the plain-orice type shown in Figure 6.22. A drawback

to this simple concept is that the spray cone angle is fairly small, typically

around 20° [43]. Most gas turbine combustors require injectors that distrib-

ute the fuel in the form of a conical spray of approximately 90° included

angle. Whitlow et al. [41] have studied several different types of efferves-

cent atomizers designed to produce wide-angle sprays. One design was

essentially the same as the atomizer shown in Figure 6.22, except that the

single-hole orice was replaced with four equispaced holes drilled at an

angle of 40° from the axis of the mixing tube. Tests carried out on this four-

hole design showed that the total liquid ow was uniformly distributed

between the four holes to within a few percent. Using four holes instead of

one had no deleterious effect on atomization quality, as Figure 6.23 clearly

shows.

Whitlow et al. also found that a wide-angled spray could be produced

by replacing the normal circular discharge orice with a suitably angled

annular passage. With this arrangement, the two-phase mixture is ejected

from the atomizer in the form of a hollow-cone spray, with most of the

droplets concentrated around the outer periphery. One of the advantages

of this simple conguration is that the annular discharge passage can be

designed to produce a hollow-cone spray having virtually any desired

spray angle. The performance of this type of conical-sheet effervescent

atomizer was investigated over wide ranges of pressure, air/liquid ratio,

and annular gap width. A satisfactory and stable spray was observed at all

operating conditions.

More detailed information on the design aspects of various types of

effervescent atomizers, including single-hole, multi-hole, conical-spray, and

annular spray, may be found in References [42] and [44].

Air injection

holes

Region of

bubbly flow

Liquid

Air

gas

Figure 6.22

Plain-orice effervescent atomizer.

Fuel Injection 251

6.12 Vaporizers

Apart from the various atomization methods already discussed, an alterna-

tive method of preparing a liquid fuel for combustion is by heating it above

the boiling point of its heaviest hydrocarbon ingredient, so that it is entirely

converted to vapor before combustion. This method is, of course, applicable

only to such high-grade fuels as can be completely vaporized, leaving no

solid residue (see Chapter 1).

An alternative and much simpler method of vaporization is to inject the

fuel, along with some air, into tubes that are immersed in the ame. The

injected fuel–air mixture is heated by the tube walls and, under ideal condi-

tions, emerges as a mixture of vaporized fuel and air. The remainder of the

combustion air is admitted through apertures in the liner wall and reacts

with the fuel–air mixture issuing from the tubes.

Operating pressure, kPa

4-hole orifice

bubbly flow regime

689

140

120

SMD, µm

100

0 0.01 0.02

Air/liquid mass ratio

0.03 0.04

552

414

276

Figure 6.23

Inuence of injection pressure and air/liquid ratio on mean drop size. (From Whitlow, J.D.,

Lefebvre, A.H., and Rollbuhler, J.R., Fuels and Combustion Technology for Advanced Aircraft

Engines, AGARD Conference Proceedings, 536, 38, 1993.)

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

Some of the early designs, one of which is illustrated in Figure 6.24, were

generally known as “walking stick” or “candy cane” vaporizers. They

were used on the Mamba, Sapphire, and Viper engines in the UK, and

on the Curtiss-Wright J65 and Westinghouse J46 in the United States. The

Lycoming Τ vaporizer, shown in Figure 6.25, incorporates a splitter that runs

down the center of the inlet leg and, in effect, converts the vaporizer into two

back-to-back “walking sticks” sharing a common inlet. SNECMA and Rolls

Royce developed this concept further and made it suitable for application

to high-performance engines. The RR design for the RB199 engine is shown

in Figure 6.26. Note in this gure that the fuel tube has a bifurcated end to

ensure that equal amounts of fuel are supplied to both arms of the Τ vapor-

izer. The AFR within the tubes varies from around 6 at idling conditions to

between 2 and 3 at maximum power.

Casing

Liner

Vaporizing

tube

Turbulence

Air “hat”

Fuel

Air

Figure 6.24

“Walking stick” vaporizing system.

Exit leg

Splitter

Fuel jet

Figure 6.25

Lycoming Τ vaporizer. (Reproduced by permission of Lycoming Engines, a div. of Avco

Corporation.)

Fuel Injection 253

Vaporizing systems have useful advantages in terms of low cost, mod-

est fuel-pump pressure requirements, and fairly low soot formation. Their

drawbacks include risk of thermal damage to the vaporizing elements and

sensitivity to variation in fuel type. Moreover, during the starting cycle, the

tubes are too cold to effect vaporization and a torch igniter is needed to initi-

ate combustion. Usually, this takes the form of a plain-orice, pressure-jet

atomizer adjacent to an igniter plug. A further drawback is that during rapid

engine acceleration the sudden addition of more fuel can overcool the tubes,

thereby lowering evaporation rates and reducing combustion efciency.

It is now widely recognized that the term vaporizer is largely a misnomer

because at high-power conditions the heat transferred to the tubes is insuf-

cient to vaporize more than a small fraction of the fuel. Thus, only at the lowest

fuel ows can the system be regarded as a true vaporizer. Where vaporizers are

used on modern engines, their main function appears to be that of providing a

satisfactory distribution of fuel throughout the primary combustion zone.

Vaporizing systems are now in service on a number of Rolls Royce engines,

including the Pegasus, Olympus, and RB199. Useful descriptions of these can

be found in papers by Parnell and Williams [45], Low [46], Sotheran [47], and

Jasuja and Low [48]. A typical modern vaporizing combustor is shown in

Figure 6.27.

In regard to future applications, it is important to bear in mind that vapor-

izer elements survive only because they are fuel-cooled. This can be a prob-

lem during rapid engine decelerations when the fuel is cut off quickly and

the only available coolant is the combustor inlet air, which, temporarily at

least, is still at a high temperature. As engine pressure ratios continue to rise,

and combustor inlet air temperatures along with them, the cooling effective-

ness of this air will diminish correspondingly. This clearly has important

implications for the mechanical integrity of any future vaporizer design.

Fuel

Air

Figure 6.26

RB 199 vaporizer. (Courtesy of Rolls Royce plc.)

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

6.13 Fuel Nozzle Coking

This problem is by no means new, but it is becoming especially serious for

advanced turbojet engines because of the growing use of fuel as a heat sink

for cooling the airframe, avionics, and engine lubricating oil. It is further

exacerbated by the fact that the fuel feed arm is immersed in the compressor

efux air. This high-pressure, high-velocity airow causes convective heat-

ing, which further raises the temperature of the fuel before it ows into the

fuel injector. The combined effect of all these various inputs is that by the

time the fuel is sprayed into the combustion zone, its temperature is appre-

ciably higher than when it left the fuel tank.

From a combustion viewpoint, this elevation in fuel temperature is not

altogether undesirable because it reduces fuel viscosity, thereby promoting

ner atomization. Unfortunately, high fuel temperatures stimulate oxidation

reactions, which lead to the formation of gums and other insoluble materi-

als (including carbon) that tend to deposit on the walls of the passages and

metering orices within the nozzle. The rate of deposition is governed mainly

by fuel temperature, but is also enhanced by an increase in wall temperature

[49,50]. These deposits can distort the fuel spray and create appreciable non-

uniformities in spray patternation [51,52].

The problems created by the deposition of carbonaceous materials, gener-

ally referred to as “coke,” within the fuel nozzle are of special importance

for pressure-swirl nozzles because they contain small internal passages that

Scoop

Fuel feed pipe

Compressor

outlet guide vane

Prediffuser

Shoulder ring

Combustion

chamber

Primary

ports

Dilution

ports

Hotshot

injector

Turbine entry

duct (TED)

Nozzle

guide

vane

Combustion chamber

outer casing

Baseplate

Vaporiser

“Blown ring”

Inner casing

Figure 6.27

Modern vaporizing combustor. (Courtesy of Rolls Royce plc.)

Fuel Injection 255

are especially prone to plugging and blockage. Coke agglomerates, formed

either upstream of the nozzle tip or within the nozzle itself, can break off

and be carried into the metering passages. Airblast atomizers are inherently

less susceptible than pressure-swirl atomizers to the problems of fuel coking

because they employ much larger fuel passages in the nozzle tip. However,

the inability of airblast atomizers to always meet the requirements of cold

day starting has prevented pure airblast systems from completely displacing

pressure-swirl atomizers from engine designs.

The effects of partially or totally blocked fuel metering passages on the

fuel-air distributions produced by an airblast atomizer have been examined

by McCaldon et al. [52]. They found that as more and more fuel metering holes

are obstructed with increasing operating time, more fuel is forced through

the remaining nozzles. Consequently, engine damage may be caused by

those injectors that, if tested individually, still ow within tolerances.

In recent years, the U.S. Naval Air Propulsion Center has sponsored an

Innovative High-Temperature Fuel Nozzle Program with the objective of

designing and evaluating fuel nozzles capable of operating satisfactorily

despite extreme fuel and air inlet temperatures. As part of this program,

Stickles et al. [53] evaluated 27 different nozzle designs, all of which were

based on the production GE F404 fuel nozzle. Heat transfer analysis high-

lighted the following design rules for reducing wetted-wall temperatures:

1. Reduce fuel ow passage area to increase fuel velocity

2. Add air gaps

3. Substitute ceramics for metal parts

4. Avoid bends and steps in the fuel ow path

Sample tube coking test results showed the importance of surface nish on the

fuel coking rate. Reducing the surface roughness from 3.1 to 0.25 µm reduced

the deposition rate by 26%. In summary, Stickles et al. found that reduced

passage ow area, reduced surface roughness, additional insulating air gaps,

and replacement of metallic tip components with ceramics, minimized the

wetted-wall temperature, thereby reducing the rate of deposition.

Thermal modeling studies carried out by Myers et al. [54] as part of the same

U.S. Navy program showed that the two major sources of heat absorption into

the fuel nozzle are the air swirler vanes and any surface exposed to the ame.

At an altitude cruise condition, for example, the predicted heat ux enter-

ing the nozzle face from ame radiation is more than 20 times that absorbed

by conduction and convection through the burner feed arm. The frontal area

exposed to the ame is thus a key element in nozzle thermal loading.

Myers et al. concluded that substantial reductions in wetted-wall tempera-

tures can be realized at extreme fuel and air inlet temperatures by using

simple air gaps as thermal barriers. Detailed thermal analysis and simple

thermal barriers, rather than exotic cooling schemes, can produce dramatic

improvements in thermal protection.

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

The problem of fuel coking and its strong adverse effects on spray uni-

formity and pollutant emissions is one of growing concern because of the

anticipated gradual deterioration in fuel quality and the continuing trend

toward higher temperature engines.

6.14 Gas Injection

Provided that their energy density is reasonably high (say not less than 6 MJ/

), gaseous fuels present no special problems, at least from a combustion

standpoint. They are usually characterized by clean combustion, with low rates

of formation of soot and nitric oxides. The main problem is that of achieving the

optimal level of mixing in the combustion zone. A mixing rate that is too high

produces narrow stability limits, but a mixing rate that is too low may make

the system prone to combustion-induced pressure oscillations. On engines

designed to operate on both gaseous and liquid fuels, it is important that the gas

ow pattern be matched to that of the liquid fuel; otherwise, some variation in

the temperature distribution of the outlet gases could occur during the change-

over from one fuel to the other. During this period, careful control over the liq-

uid and gas ow rates is required to avoid ame blowout or overtemperature.

Many different methods have been used to inject gas into conventional

combustion chambers, including plain orices, slots, swirlers, and venturi

nozzles. Good descriptions of these methods can be found in Winterfeld

et al. [55]. The various methods of gas injection employed in modern low-

emissions combustors are described in Chapter 9.

6.15 Equations for Mean Drop Size

For any given atomizer type, mean drop sizes are largely dependent on

atomizer size, design features, and operating conditions. Atomization qual-

ity is also highly dependent on the physical properties of the fuels employed

and on the properties of the surrounding gaseous medium.

The three fuel properties of relevance to atomization are density, surface

tension, and viscosity. In practice, the signicance of density for atomization

performance is diminished by the fact that most gas turbine fuels exhibit

only minor differences in this property. Surface tension is important in atom-

ization because it resists the formation of new surface area, which is fun-

damental to the atomization process. Whenever atomization occurs under

conditions where surface tension is important, the Weber number is a use-

ful dimensionless parameter for correlating drop-size data. From a practical

Fuel Injection 257

standpoint, viscosity is the most important fuel property. An increase in vis-

cosity lowers the Reynolds number of the ow inside the atomizer, thickens

the fuel sheet produced at the atomizer exit, opposes the development of

instabilities in the fuel jet or sheet, and generally delays the onset of atomi-

zation. This delay causes atomization to occur further downstream from the

nozzle where conditions are less conducive to the production of small drops.

Another important practical consideration is that although the variations

normally encountered in surface tension are only about 15%, the correspond-

ing variations in viscosity are more than an order of magnitude.

The most important air property inuencing atomization is density.

With air-assist and airblast atomizers, if breakup occurs by the classical

mechanism, an increase in air density generally improves atomization by

raising the Weber number. With pressure-swirl atomizers, the effect of an

increase in ambient air density on atomization is more complex. The con-

comitant increase in the Weber number is again benecial, but this effect is

opposed by a decrease in spray cone angle, which reduces the interaction

between the developing sheet and the surrounding air, and increases the

initial sheet thickness (see Figure 6.28). Furthermore, an increase in air den-

sity reduces the breakup length, so that breakup occurs closer to the nozzle

where the fuel sheet is thicker. As SMD is proportional to the square root of

the sheet thickness at breakup, the increase in sheet thickness produced by

these two separate effects acting together must result in larger drops. Thus,

an increase in air density can either raise or lower the SMD depending on

whether the benecial effect of increasing the Weber number outweighs the

adverse effect of the increase in sheet thickness. Usually, it is found that if the

ambient air density is increased continuously from its normal atmospheric

level, the cone angle gradually falls until a value of density is reached beyond

which there is no further reduction in the cone angle [56,57]. Moreover, as the

breakup length declines with increase in air density, a condition is eventually

reached where breakup occurs directly at the nozzle exit, or even within the

nozzle itself. Beyond this point, sheet thickness has little or no effect on mean

drop size. The net effect of all these separate inuences is that drop sizes

generally increase with ambient air density up to a maximum value (which

= t cosθ

Figure 6.28

Relationship between sheet thickness and spray cone angle.

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

roughly corresponds to the condition at which the breakup length becomes

zero) and then slowly decline with further increases in air density [57].

Unfortunately, the physical processes involved in atomization are not

sufciently well understood for mean diameters to be expressed in terms

of equations derived from rst principles. In consequence, the majority of

investigations into the drop-size distributions produced in atomization have

been empirical in nature and have resulted in empirical equations for mean

drop size. The most authentic of these equations are those in which mean

drop size is expressed in terms of dimensionless groups, such as Reynolds

number, Weber number, or Ohnesorge number.

Most of the mean drop-size equations published before the 1970s should be

regarded as suspect because of deciencies in the methods available for drop-

size measurements. Even equations based on accurate experimental data

should only be used within the ranges of air properties, liquid properties, and

atomizer operating conditions employed in their derivation. Extrapolation to

other conditions is fraught with risk because changes in any of these vari-

ables could produce a change in the mode of atomization, which could have

a signicant effect on the manner and extent to which variations in the rel-

evant ow parameters affect the drop-size distributions in the spray.

The following equations for SMD have been selected from the large

number available in the literature. More detailed information on drop-size

equations for all types of atomizers may be found in Lefebvre [58].

6.16 SMD Equations for Pressure Atomizers

6.16.1 Plain Oriﬁce

With this device, a simple circular orice is used to inject a round jet of high-

velocity liquid into the surrounding air or gas. Finest atomization is achieved

with small orices but, in practice, the difculty of keeping liquids free from

foreign particles usually limits the minimum orice size to around 0.3 mm.

Due to the formidable problems involved in making drop-size measure-

ments in the dense sprays produced by plain-orice nozzles, few equations

for mean drop size have been published. According to Elkotb [59]

SMD3.08

LLAL

(

)

∆

−

0 385

0 737

006054.

.σρ ρ P

(6.20)

6.16.2 Pressure Swirl

In this type of nozzle, a circular outlet orice is preceded by a swirl chamber

that causes the liquid to emerge from the nozzle as an annular sheet, which

spreads radially outward to form a hollow conical spray. Despite its appar-

ent simplicity, the various physical phenomena involved in pressure-swirl

Fuel Injection 259

atomization are highly complex. For most of the past half century, mean

drop sizes have been correlated using empirical equations of the form

SMDconstant

LL L

−

σµ

ab cd



∆ .

(6.21)

For example, Radcliffe’s equation [60] is

SMD

0.2

LLL

−−

06 02 02504

....

σµρ



m ∆

(6.22)

whereas subsequent work by Jasuja [61] yielded the expression

SMD

LL LL

−−

06 016016 022043

.. .. .

σµ ρ



mP∆

(6.23)

It is noteworthy that in the experiments of Radcliffe and Jasuja, the variation

in surface tension was quite small and was accompanied by wide variations

in viscosity. Thus, the surface tension exponent of 0.6 has no special signi-

cance in Equations 6.22 and 6.23.

Another example of this type of equation, which has an advantage over

most others in that it is dimensionally correct, is the following [58]:

SMDP

0.25

LL LA

=∆

−−

225

025025 05 025

.. ..

σµ



m ρ

(6.24)

It is now generally accepted that for both pressure and airblast nozzles, the

relative velocity between the liquid and the surrounding air has a profound

effect on atomization. It generates the protuberances on the liquid surface

that are a prerequisite to atomization and also furnishes the energy needed

to convert these protuberances into ligaments and then drops. However,

another important factor in atomization, as discussed above, is the contribu-

tion made to sheet or jet disintegration by the instabilities created within the

liquid itself, which are very dependent on liquid velocity. In airblast atomi-

zation, where high-velocity air impacts on a slow-moving liquid, the only

factor promoting atomization is the relative velocity between the air and the

liquid. This is equally important in pressure atomization but, by achieving

this relative velocity through liquid motion instead of air motion, an impor-

tant advantage is gained in that the liquid now makes an additional and

independent contribution to its own disintegration, an effect that is either

absent or negligibly small in airblast and air-assist atomization.

These arguments highlight the special importance of velocity in pressure

atomization. The velocity at which the liquid is discharged from the noz-

zle has two separate effects on atomization. One important effect, which is

dependent on the absolute velocity U

, is in generating the turbulence and

instabilities within the liquid stream that contribute to the rst stage of the

atomization process. The other effect, which depends on the relative velocity

, is in promoting the atomization mechanisms that occur on the liquid

surface and in the adjacent ambient air.