Research on the Rectangular Lobed Exhaust Ejector /Mixer Systems* 1

By Hui HU,*2 Toshio KOBAYASHI,*2 Tetsuo SAGA,*2 Nobuyuki TANIGUCHI,*2Huoxing LIU*3 and Shousheng WU*3

Key Words Aero-Engine Exhaust System, Exhaust Ejector/Mixer System, Rectangular Lobed Nozzle,Jet Mixing Enhancement

Abstract

Experimental research of seven rectangular exhaust ejector/mixer systems, which were the combinations offour rectangular lobed nozzles with three rectangular mixing tubes, has been conducted to investigate the effectof the geometry of rectangular lobed nozzles on the aerodynamic performances of exhaust ejector/mixer systems.The experimental results showed that: rectangular aligned lobed nozzles have better pumping and mixingenhancement abilities than rectangular staggered lobed nozzles, while they also cause bigger pressure losses.Scalloping treatment on the lobe structure can improve the pumping ability and mixing enhancement performanceof a lobed nozzle, but it will also cause a big extra pressure loss. Among the seven tested exhaust ejector/mixersystems, the exhaust ejector/mixer system which is a combination of the staggered lobed nozzle B and therectangular mixing tube III has the best aerodynamic performances.

Nomenclature inlet of the mixing tubeexit of the mixing tube

ADL

cross sectional areadiameter of the mixing tubethe required pressure recovery length along themixing tubemass flow rate

pressuretemperaturevelocitydensity

Introduction

MpT

p

cjPPT*

T*

An exhaust ejector is a device, which converts a highvelocity fluid flow of given mass flow rate into a fluidflow of lower velocity. This conversion is achieved bythe transfer of momentum and energy through viscousinteraction of the high velocity (primary) fluid flow witha lower velocity (secondary) fluid flow within a mixingtube (Fig. I ). During the past decades, this fluid dynamicdevice has been further utilized to improve aircraftperformance in a variety of ways, including enginecomponent cooling, thrust augmentation, and exhaustnoise and infrared radiation reduction.

The conventional mixing of the primary and secondaryflow in a mixing tube occurs very slowly, which isperformed mainly by a small scale viscous mixing in ashear layer. Thus, a conventional ejector requires a longmixing tube to entrain the secondary flow, and a longmixing tube results in large wall friction loss, extra weightand higher cost. For this reason, a lobed exhaustejector/mixer system (Fig. 2), in which a lobed nozzlewas used as a primary nozzle, was proposed in the pastseveral years. It was found that a lobed nozzle can causelarge scale streamwise vortices to be shed at the trailingedge of lobe structures, so the downstream of the flowfield is embedded with arrays of large scale streamwisevortices, and a rapid exchange of momentum and energy

cp

IF

L1P:T

)1 + rI>T*)(l + rI»parameter (

Subscriptsa: ambient

P: primary flow

S : secondary flow

T: total

*1 Received May 12th, 1998.*2 Institute of Industrial Science, The University of Tokyo, Tokyo, Japan.*3 Jet Propulsion Laboratory, Department of Jet Propulsion, Beijing University of Aeronautics and Astronautics, Beijing, 100083, P. R. China

188 Trans. Japan Sac. Aera. Space Sci Vol. 41, No. 34

research, seven rectangular lobed ejector/mixer systems,which were the combinations of four rectangular lobednozzles with three rectangular mixing tubes, were studiedexperimentally to investigate the effect of the geometryof rectangular lobed nozzles on the aerodynamic per-formances of rectangular lobed exhaust ejector/mixer

systems.

pr;"",rynozzle -

(PrT

\.TPT.

2. Experimental Set-UpPsr

Ts1

'ig.

:ig. 2. The concept of a lobed exhaust ejector /mixer

system.

An experimental research was conducted on the lowspeed exhaust ejector system test rig (corresponding tothe case of helicopter used aero-engine) in the JetPropulsion Laboratory ofBeijing University of Aeronau-tics and Astronautics (BUAA). The primary flow issupplied by a compressor with a combustor, and itsflow rate and temperature can be adjusted. Figure 3 isa schematic of the test section. The flowrate of theprimary flow is measured before the test section. Thetotal temperature and total pressure of the primary floware obtained by the thermocouples and total pressuresensor located at the inlet of the tested exhaust ejector/mixer system. The temperature and pressure fields at theexit of the mixing tube are measured by a 12 pointtemperature/pressure rake with a traverse mechanism.The flowrate of the secondary flow can be obtained bythe flow nozzle installed at the top of the settling chamber(Fig. 3) or/and by the above measured pressure andtemperature fields at the exit of the mixing tube. Thestatic pressure distributions along the mixing tube areobtained by two rows of six static pressure tabs alongthe mixing tube. One row is at the top of mixing tubeand the other row is at the side of mixing tube (seeFig. 5). The signals of thermocouples and pressuresensors are transferred to an IBM PC computer for dataacquirement and procession. During the test, the flow-rate of the primary flow is about 1.Okg/s, and the exitvelocity of the primary flow is around 70 m/s. The detailinformation of the experimental set-up can be obtainedfrom the Ref. 13.

It had been found that the geometrical parameters ofthe lobed structures, such as the lobe height, the lobepenetration angle, the form of the lobed structure, thearea ratio of the primary nozzle, mixing tube etc, canaffect the aerodynamic performance of a lobed ejector/mixer system very much. The detail information canbe found in the papers of Skebe et al.,l) Presz et al.,3)Eckerle et aC) and WU.11) Based on these previousresearches, all the rectangular lobed nozzles used inpresent study are designed to have the same lobepenetrate angle, lobe height and lobe form. Only theeffects of the lobed width, the way of lobe structurearrangement and the scalloping treatment of the lobedstructures on the aerodynamic performance of a lobedejector /mixer system are discussed in the current paper .

The four rectangular lobed nozzles used in the pres-~

was achieved by means of intense mixing within a shortdownstream with very little loss (Skebe et al.l).

On account of the above advantage, lobed exhaustejector/mixer systems have received great attention bymany researchers in the recent years (Refs. 1-10) andthey have been widely applied to the aero-engine area.For example, on the commercial aero-engines, such asJT8D-217C, CFM56-5C and RB211-524G/H, lobed ex-haust ejector/mixer systems had been used to reducetake-off jet noise and Specific Fuel Consumption (SFC)(Presz6J). On the military aero-engines, such as militaryhelicopter "Tiger" developed by German-France, Amer-ican RAH-66 "Comanche" and stealth fighter F-117(WU11J), lobed exhaust ejector/mixer system had beenused to reduce the infrared radiation signals of theexhaust system to improve the survivability of themilitary aircraft in modem war .

It is well known that, a rectangular nozzle has bettermixing performance than a circular nozzle. A rectangularlobed nozzle, which is formed by the combination of lobestructures with a rectangular nozzle, seems to be muchmore promising for mixing enhancement (Liu et al.12J).Meanwhile, circular lobed exhaust ejector/mixer systemshave been widely studied in the past several years.However, few researches were found on the rectangularlobed ejector/mixer systems in which rectangular lobednozzles were used as primary nozzles. In the present

Feb. 1999 189Ho Hu et alo : Rectangular Lobed Exhaust Ejector/Mixer Systems

total

pressure/

temperature

rake

Fig. 3. Schematic of the test section.

!-+--

~~r lobe width

l outer lobe width

I

A. rectangular staggered lobed nozzle with bigger inner lobe width B. rectangular staggered lobed nozzle with equal lobe width

--":I=.~e

c. rectangular aligned lobed nozzle with equal lobe width D. rectangular aligned scalloped lobed nozzle with equal lobe width

Fig. 4. Tested rectangular lobed nozzles.

are simple rectangular tubes which have the same aspectratio (AR) as the rectangular lobed nozzle A (AR = 1.0)and nozzle B (AR = 1.2). The rectangular mixing tube III(AR = 1.2) is designed to have a multiple-stage cooling

structure and a 20 degree diffuser at the end of the tube.The length of these three mixing tubes is the same, whichis 450 mm. All the mixing tubes have the same crosssection area (19,200mm2), therefore, the area ratiosbetween the rectangular lobed nozzles and mixing tubesof the seven tested ejector/mixer systems are the same,which is I: 3.

3. Experimental Results and Discussions

During the experiment, cold tests (the temperature ofthe primary flow is the same as the secondary flow, whichwas 295 K) were conducted firstly to determine theaerodynamic performances of seven rectangular lobedexhaust ejector/mixer systems listed in Table I. Then,

ent research are shown in Fig. 4. Nozzle A and nozzleB are the rectangular lobed nozzles with a staggeredarrangement of the two rows of the lobe structure. Thedifference between the nozzle A and the nozzle B is that,for the lobed nozzle A, the inner lobe width is two timesthe outer lobe width, while the lobed nozzle B has theequal value of inner lobe width and outer lobe width.Nozzle C and nozzle D are rectangular lobed nozzleswith an aligned arrangement of the two rows of thelobe structure. The rectangular lobed nozzle D is designedthe same as the rectangular lobed nozzle C but with ascalloping treatment (Yu et al.14J) of the lobe structure.All these four rectangular lobed nozzles have the sameexit area (6400mm2, equivalent diameter is about 90mm). Rectangular lobed nozzle A has an aspect ratio(AR) of 1.0, while the aspect ratios (AR) of lobed nozzleB, lobed C and lobed D are same, which is 1.2.

The three mixing tubes used in the present study areshown in Fig. 5. The mixing tube I and mixing tube II

190 Trans. Japan Sac. Aera. Space Sci. Vo1. 41, No 34

I

rectangular simple mixing tube I

( AR=1.0)

rectangular simple mixing tube n rectangular mixing tube m with

(AR=1.2) multiple-ring cooling structure (AR=1.2)

Fig. 5. Tested rectangular mixing tubes.

Table Aerodynamic performance of seven tested rectangular lobed exhaust ejector/mixer systems.

Cold test Hot test Tp=573K

Rectangularexhaust

ejector/mixersystem

Static

pressurerecoverydistance

LID (at side)

Average Highesttemperature temperature atat the exit of the exit of the

the mixing mixing tube (K )tube (K)

Combined

aerodynamicparameter

IF

Pumpingcoefficient

4i

Pressure losscoefficient

JPPT*

Combinations

A+I

A+II

B+I

B+II

B+111

C+111

D+111

1.251

1.231

1.245

1.226

1.245

1.342

2.298

0.870.840.870.870.981.021.34

0.3580.3640.3560.3510.3180.3290.420

.3

.8

.8

.8

.8

.8

.8

0.70.70.7

0,71.30.7

4

429.3

425.2

401.3

492.3488.3431.7

6

'7

Note: During the test, the used thermocouples have the range of300-1 073 K, accuracy for the full range is 0.1 %, and the used pressure transducershave the range of 0-0.2 MPa, accuracy for the full range is 0.02%.

three representative rectangular lobed exhaust ejectorsystems were selected from the seven for further ex-periment at hot test condition (the primary flow tem-perature at the inlet of the lobed nozzle was 573:!: 5 K,while the temperature of the secondary flow was 295 K).The studied aerodynamic performances of the exhaustejector/mixer systems include pumping coefficient,pressure loss coefficient, combined aerodynamic param-eter, static pressure recovery characteristics along themixing tube and the velocity and temperature distribu-tions at the exit of the mixing tube.3.1. Pumping coefficient (J) The pumping coefficient!1> was defined as the ratio of secondary flow rate to theprimary flow rate. The pumping coefficient of a con-ventional circular ejector with the same area ratio (I: 3)of the primary nozzle and mixing tube is about 0.45(Hu et al.13»). The pumping coefficient of a circular

lobed ejector/mixer system with the same area ratio(1 : 3) and the same lobe configuration as used in currentstudy is about 0.81 (Hu et al.14»). However, from theexperimental results listed in Table 1, it can be seen thatthe pumping coefficients of the seven tested rectangularlobed ejector/mixer systems (0.84--1.34) are much higherthan that of a conventional circular ejector. That is, thepumping abilities of the rectangular lobed ejector systemsare about 1 to 2 times that of the conventional ejector.It can also be seen that a rectangular lobed ejector/mixersystem will have a much bigger pumping ability than acircular lobed ejector/mixer system does for the samelobe configuration. This may come from the differencebetween the rectangular geometry and the circulargeometry of the primary nozzle.

Among the tested exhaust ejector/mixer systems, itcan also be seen, there is no apparent difference be-

Static

pressurerecoverydistance

LID (at top)

Feb. 1999 H. Hu et a! Rectangular Lobed Exhaust Ejector/Mixer Systerr

100un

{jf)jstaggered arrangement aligned arrangement

Fig. 6. The effect of the way of lobe structure arrangement.

a) Conventionallobe b) Scalloped lobe

Fig. 7. Scalloping effects of lobe structure.

tween the ejector/mixer systems using nozzle A andnozzle B as primary nozzles (system 1, system 2, system3 and system 4) in terms of pumping coefficient. However ,considering along with the pressure loss coefficient andcombined aerodynamic parameter (to be discussed later),the ejector/mixer systems using nozzle B as primarynozzle seems to be a bit better than the one usingnozzle A. Among the exhaust ejector/mixer systems withnozzle B as primary nozzle (system 3, system 4 and system5), the pumping coefficient of the exhaust ejector/mixersystem 5 (combination B + Ill) is the best. This can beexplained by presence of a diffuser at the end of therectangular mixing tube Ill, which can improve thepumping ability of the ejector system (Skebe et al.1J).Meanwhile, the mixing tube III also has a multiple-stagecooling structure (Fig. 5), which makes the exhaustejector/mixer system 5 a multiple stage ejector system,this also resulting in a bigger pumping coefficient.

From the comparison of the exhaust ejector/mixersystems 5, 6, and 7 (combinations B+III, C+III andD + Ill), it can be said that, the pumping abilities ofthe aligned lobed nozzles (nozzle C and nozzle D) arehigher than that of the staggered lobed nozzle (lobednozzle B). The reason may be that the scale of thestreamwise vortices induced by aligned lobed nozzles willbe bigger than that of the staggered lobed nozzle B (Fig.6) and larger streamwise vortices will give a higherpumping ability (Presz et al.3J). From the comparison ofthe exhaust ejector/mixer systems 6 and 7 (combina-tions of C + III and D + Ill), it can be suggested thatscalloping treatment of the lobe structure can improvethe pumping ability of the lobed nozzle greatly. The

reason may be that additional vortices can be generatedat the parallel sides of lobes by the scalloping treatmentof the lobe structure (Fig. 7 and Fig. 6 of the Ref. 15,Yu et al.). This will result in a larger vortices roll-up andenhances the "stir up effect" (Presz et al.3») of thelarge-scale streamwise vortices, hence the lobed nozzleD has the highest pumping ability.3.2. Pressure loss coefficient JPPT* The pressureloss coefficient LlPPT*' which is defined as the ratio of thepressure difference between the total pressure at theinlet of the primary nozzle and the ambient pressure to thedynamic pressure of primary flow, can indicate the powerloss of an aero-engine caused by the installation of anejector system. The bigger the pressure loss coefficientis, the higher the engine power loss will be. From theexperimental results listed in Tablel, it can be seenthat; the exhaust ejector/mixer systems with rectangularaligned lobed nozzles as primary nozzles (exhaustejector/mixer systems 6 and 1) have the bigger pressure1oss coefficients than those with rectangular staggeringlobed nozzles as primary nozzles (exhaust ejector/mixersystem 1 to 5). The explanation of this is that the rect-angular aligned lobed nozzles can induce larger scalestreamwise vortices as mentioned above (Fig. 6), whichwill also cause bigger mixing loss. The exhaust ejector/mixer system 7, which uses rectangular aligned scal-loped lobed nozzle D as primary nozzle and can generateadditional vortices at the parallel sides of lobe struc-ture and has the highest pumping coefficient among thetested exhaust ejector/mixer systems, also has the big-gest pressure loss coefficient.3.3. Combined aerodynamic parameter III An exhaust

192 Trans. Japan Soc. Aero. Space Sci Vo1. 41, No.134

LID

ejector was always expected to have higher pumpingcoefficient and smaller pressure loss coefficient. How-ever, from the above experimental results, it can besaid that, the exhaust ejector/mixer system with higherpumping coefficient always has bigger pressure loss. So,a combined aerodynamic parameter 'P (Hu et al.16J),which indicates the ratio of the momentum of exhaustedflow from an exhaust system with and without an exhaustejector/mixer system instillation, is introduced to evalu-ate the overall aerodynamic performance of an exhaust

ejector system.

JPPT*

+ <1>T*)(l + <1»

From the definition of the parameter IJI , it can be seenthat, a bigger pumping coefficient and a less pressure losscoefficient will cause a smaller combined aerodynamicparameter IJI. So, the smaller the parameter IJI is, thebetter aerodynamic performance will be given by theejector system. From the combined aerodynamic param-eter IJI values listed in Table 1, It can be seen that,the exhaust ejector/mixer systems 6 and 7, which usesrectangular aligned lobed nozzles (nozzle C and nozzleD) as primary nozzles, have bigger combined aero-dynamic parameter values than the one with rectangularstaggered lobed nozzles as primary nozzles (the exhaustejector/mixer systems I to 5).

Among the exhaust ejector/mixer systems with rect-angular staggered lobed nozzles as primary nozzles(the exhaust ejector/mixer systems I to 5), the one withnozzle A (which has bigger inner lobe width) as primarynozzle has a bigger combined aerodynamic parameter IJIthan the one with the rectangular lobed nozzle B asprimary nozzle. Meanwhile, the exhaust ejector systemswith the same aspect ratio (AR) of rectangular lobednozzle and mixing tube, i. e. system I (combination A+I) and system 3 (combination B + II ), are better than thehybrid systems (system 2 (combination A+II ) andsystem 4 (combination B+I)). Among the seven tested

exhaust ejector/mixer systems, the system 5 (combina-tion B + III), has the smallest combined aerodynamic pa-rameter value, so it has the best combined aerodynamicperformance, while, the system 7 (combination D +III),which has the highest pumping coefficient and the biggestpressure loss coefficient, has the worst combined aero-

dynamic performance.3.4. Characteristics of static pressure recovery perfor-mance The static pressure recovery performance ischaracterized by the static pressure distribution alongthe mixing tube of an ejector system. It is well knownthat, the faster the static pressure recovery is, the shorterthe mixing tube can be. This will be beneficial to thesize and weight of ejector systems. Two typical mea-sured static pressure recovery characteristics are shownin Fig. 7. Based on the figures, the required pressurerecovery distance (L/D) at which mixing flow reachesnear uniform condition for each ejector/mixer systemcan be obtained, and the results are listed in Tablel.

From the data listed in Tablel, it can be seen that: therequired pressure recovery distance L/ D of the testedcombinations is about 0.7-1.8, which is much shorterthan that required by a conventional ejector (which isabout 4-6 (Presz et al.9». This means that, unlike theconventional ejector, the required mixing tube lengthcan be reduced to 1/2 or 1/3 with a lobed nozzle as theprimary nozzle of the exhaust ejector /mixer system.Waitz et al.17) also got a similar result.3.5. Exit velocity and temperature distributions Thevelocity and temperature distributions at the exit of amixing tube can directly indicate the mixing efficiency ofthe primary gas and pumped ambient air in the mixingtube. Three typical measured velocity and temperaturedistributions at the exit of the mixing tubes for the exhaustejector systems 5, 6 and 7 (combinations B + III, C + IIIand D + III) are given in Figs. 9 and 10.

From the figures, it can be seen that the region of thehigh speed and high temperature flow is bifurcated anddeviates from the center at the exit of mixing tube due

H. Hu et al. : Rectangular Lobed Exhaust Ejector/Mixer SystemFeb. 1999

a. B+ ill combination b. C+ ill combination c. D+ ill combination

Fig. 9. Velocity distributions (V2/Vp) at the exit of mixing tubes (cold test).

a. B+ m combination b. C+ m combination c. D+ m combination

Fig. 10. Temperature distributions (T 2T/T PT) at the exit of mixing tubes (hot test).

mixer system (combination D + III) has the smallestaverage and highest temperatures at the exit of the mix-ing tube.

4. Conclusion

to the "stir up effect" of the streamwise vortices inducedby the lobed nozzles. Since the rectangular aligned lobednozzles (lobed nozzle C and lobed nozzle D) can generatebigger scale streamwise vortices and has higher pumpingability than the rectangular staggered lobed nozzles(lobed nozzle A and lobed nozzle B), the size of the highspeed and high temperature at the exit of the mixing tubeof the exhaust ejector/mixer system 6 (combinationC+III, Fig. 9(b) and Fig. IO(b)) is less than that of theexhaust ejector/mixer system 5 (combination B + III,Fig. 9(a) and Fig. IO(a)). Furthermore, since the rect-angular aligned scalloped lobed nozzle D can generateadditional vortices at the parallel sides of lobe structureand enhance the roll-up and "stir up effect" of thestreamwise vortices, the bulk of low speed and cold flowis engulfeds into the center of the mixing flow (Fig. 9(c)and Fig. IO(c)).

From the comparison of the average and highesttemperature values at the exit of the mixing tube for theexhaust ejector/mixer systems 5, 6 and 7 (combinationsB + III, C + III and D + III) listed in Table I, it also canbe seen that: the average and highest temperature valuesof the systems with aligned lobed nozzles (combinationsC + III and D + III) as primary nozzle are smaller thanthat of the combination B + III. This indicates that therectangular aligned lobed nozzle has one step higherpumping ability and better mixing enhancement abilitythan the rectangular staggered lobed nozzle. The rect-angular aligned scalloped lobed nozzle D can generateadditional vortices and enhance the "stir up effect"of the streamwise vortices. Thus, the exhaust ejector/

From the above analysis and discussions, it can besaid that: compared with a conventional circular ejector ,the exhaust ejector/mixer systems with rectangular lobednozzle as primary nozzle can improve pumping ability200%-300%, and reduce the required mixing tube length(L/D) to 1/2-1/3. The geometry of the lobed nozzles,such as the lobe configuration and the way of lobearrangement can influence the aerodynamic perfor-mances of rectangular lobed exhaust ejector/mixer

systems.Through the experimental research, the following

conclusions can be obtained:(I) The way of the arrangement of the lobe structures

can affect the pumping ability and mixing enhancementperformance of the rectangular lobed nozzles very much.Compared with the rectangular staggered lobed nozzles,rectangular aligned lobed nozzles have higher pumpingcoefficients and mixing enhancement abilities, but theyhave bigger pressure losses. Thus, they are very fit forthe area where the higher pumping coefficient and mixingenhancement are mainly required and the pressure lossdoes not call for severe consideration.

(2) The scalloping treatment of the lobe structurescan improve the pumping ability greatly and mixingenhancement of th.e lobed nozzle, but it will suffer a big

194 Trans. Japan Sac. Aera. Space Sci Vol. 41, No.134

extra pressure loss.(3) Among the seven tested systems, the exhaust

ejector /mixer system 5 (combination B + III) has thesmallest combined aerodynamic parameter value, i.e. hasthe best combined aerodynamic performance of the test~dcombinations. The exhaust ejector/mixer system 7(combination D + 111), which has the highest pumpingcoefficient and mixing enhancement ability, but suffersthe biggest pressure loss, has the worst combinedaerodynamic performance of the tested combinations.

Acknowledgements

The authors wish to thank Prof. Wei Fuqing and Ms. Li Liof Beijing University of Aeronautics and Astronautics for help

in conducting the present study. The research fellowshipprovided by Japan Society for Promotion of Science (JSPS) isalso gratefully acknowledged.

References

Measurement of the Vortex Development Downstream of a

Lobed Forced Mixer, J. Eng. Gas Turbines and Power, 114 (1992),

pp.63-71.8) McCormick, D. C. and Bennett, J. C., Jr. : Vortical and Turbulent

Structure of a Lobed Mixer Free-shear Layer, AIAA Paper

93-0219 (1993).

9) Presz, W. M., Jr., Reynolds, G. and McCormick, D. : Thrust

Augmentation Using Mixer-Ejector-Diffuser Systems, AIAA

Paper 94-0020(1994).10) Hu, H., Kobayashi, T., Saga, T., Taniguchi, N., Segawa, S. and

Ono, A. : Research on the Mixing Enhancement Performance of

Lobed Nozzle by Using PIV and LIF, Proceedings of 1998 ASM E

Fluids Engineering Division Summer Meeting, FEDSM98-4994,

1998.

11) Wu, S. : Lobed Mixing Enhancement Structure and Its Ap-

plication to the Aeroengine, CSAA 93-118(P), The Third Con-

ference of the Chinese Society of Aeronautics and Astro-

nautics, 1993.

12) Liu, H. and Wu, S. : Investigation on a Rectangular Lobed

Nozzle, ASME GT-95-223 (1995).

13) Hu, H. : A Study of Mixing Enhancement of Jet Flow by Using

Lobed Nozzle and Mechanical Tabs, Ph. D. Thesis, Beijing

University of Aeronautics and Astronautics, 1996.

14) Hu, H., Li, L. and Wu, S. : Experimental Research on the

Aerodynamic Performances of Lobed Exhaust Ejector Systems,

Proceeding of 1995 Yokohama International Gas Turbine Congress,

Yokohama, Japan, 1995.15) Yu, S. C. M., Yip, T. H. and Liu, C. Y. : Mixing Char-

acteristics of Forced Mixers with Scalloped Lobes, J. Propul.

Power, 13,2 (1997), pp. 306-311.

16) Hu, H. and Wu, S. : Theoretical Research on the Aerodynamic

Performance of a Low Speed Exhaust Ejector System, CSAA

95-PH-007, The First Conference of Helicopter Power, 1995.

17) Waitz, I. A., Qiu, Y. J., Manning, T. A., Fung, A. K. S., Kerwin,

J. M., Krasnodebsk, J. K., O'Sullivan, M. N., Tew D. E.,

Greitzer, E. M., Marble, F. E., Tan, C. S. and Tillman, T. G. :

Enhance Mixing With Streamwise Vorticity, Prog. Aerospace Sci.,

33 (1997), pp. 323~351.

I) Skebe, S. A., McCornlick, D. C. and Presz, W. M., Ir. : Parameter

Effects on Mixer-Ejector Pumping Performance, AIAA Paper

88-0188 (1988).

2) Paterson, R. W. : Turbofan Mixer Nozzle Flow Field: A

Benchmark Experimental Study, J. Eng. Gas Turbines and

Power, 106 (1984), pp. 692-698.3) Presz, W., Ir., Gousy, R. and Morin, B. : Forced Mixer Lobes in

Ejector Design, AIAA Paper 86-1614 (1986).

4) Werle, M. I., Paterson, R. W. and Presz, W. M., Ir. : Flow

Structure in a Periodic Axial Vortex Array, AIAA Paper 87-0610

(1987).5) Presz, W. M., Ir., Blinn, R. F. and Morin, B. : Short Efficient

Ejector Systems, AIM Paper 87-1837 (1987).

6) Presz, W. M., Ir. : Mixer/Ejector Noise Suppresser, AIAA Paper

91-2234 (1991).7) Eckerle, W. A., Sheibani, H. and Awad, I. : Experimental