design synthesis and optimization of an advanced …

ICAS 2000 CONGRESS

162.1

Abstract

The development of the baseline configuration,the design synthesis and the optimization of thebaseline configuration of an advanced shorttake-off and vertical landing (ASTOVL) combataircraft are presented. The evaluated baselineconfiguration is an advanced technology,supersonic STOVL combat aircraft design withan internal weapon bay and powered by amilitary turbofan with a remote lift system (RLS)for landing. The RLS provides lift for the frontof the aircraft and is made of three ductsstarting just after the turbine exit position andending in a nozzle at the undersurface of theaircraft, aft of the nose undercarriage bay. Atlanding, thrust in the rear part of the aircraft isprovided by a pair of swivelling duct nozzlessituated either side of the fuselage, aft of theengine position. The design synthesis describesthe aircraft baseline configuration inmathematical form, by means of semi-empiricaland sometimes analytical design .relationships.The optimization was carried out with the use ofan available optimization code. This wasdesigned to solve constrained optimizationproblems, that is to minimize a function. subjectto a number of constraints and upper and lowerbounds imposed on designated variables. Thecombined synthesis and optimization studyindicated that the best objective function for thereduction of the aircraft take-off mass waseither the aircraft empty mass or the enginemass. The optimized baseline configuration wasa 19.8 m, 18.9 t take-off weight combat aircraft.Comparison with other ASTOVL designsindicated considerable similarities. Thecomparison suggested that, from the design

point of view, a separate lift engine might be abetter solution than a RLS.

1. Introduction

The need of short take-off and vertical landingcapability for future advanced combat aircrafthas led to a number of propulsion liftconfigurations being proposed [1].

In this paper, a remote lift system (RLS)concept has been selected and subsequentlyused in the design synthesis and optimization ofan advanced short take-off and vertical landing(ASTOVL) combat aircraft with an internalweapon bay [2].

This study is based on a design synthesisphilosophy developed in the late 1970's by theDefense and Evaluation Research Agency(DERA) [3,4].

The DERA design synthesis philosophydefines an aircraft mathematically, which is thensubjected to an optimization under specifiedconstraints. Following the DERA philoshopy awide range of aircraft types, including combataircraft [2,5,6] , have been studied.

The objectives of this work were:

a) The evaluation of the aircraft baselineconfigurationb) The design synthesis of the baselineconfigurationc) The coding of the design synthesis andits interfacing with an optimization coded) The optimization of the baselineconfigurationc) The comparison of the optimizedconfiguration with similar designs

DESIGN SYNTHESIS AND OPTIMIZATION OF ANADVANCED SHORT TAKE-OFF AND VERTICAL

LANDING (ASTOVL) COMBAT AIRCRAFT

J.P. Fielding, College of Aeronautics, Cranfield University, Cranfield, Bedford MK43 OAL, UK

N. Kehayas, Consultant, 5 Gelonos St., 115 21 Athens, Greece

J.P. Fielding, N. Kehayas

162.2

2. Description

2.1. Baseline ConfigurationThe design of an ASTOVL combat aircraft is acomplicated issue where various schools ofthought propose markedly different approaches[7,8,9,10]. The configuration that was specifiedfor this study, that of a supersonic ASTOVLcombat aircraft with an internal weapon bay, isa very complex one and adds in the difficultiesenvisaged in the STOVL field.

STOVL configurations are broadly dividedinto vectored thrust and remote lift concepts[7,9]. Some, to a certain degree, mixed conceptsexist, in the sense that remote lift systems canbe vectored. Vectored thrust concepts arerelatively straightforward. The engine thrust isvectored for landing, and sometimes for take-off, and in forward flight [7]. Remote lift refersto lift provided by some means to a distant, inrelation to the engine, position. The remote liftconcepts encompass a variety of schemes; fromthe use of separate engines for landing (andtake-off), perhaps a category of its own, tovarious arrangements for remote lift, either plainor augmented.

From the engine point of view, vectoredthrust is favoured. Ward and Lewis [11] suggestthat a vectored thrust scheme is much moreefficient and offers a better performance than aremote lift one. However, vectored thrust is notalways the best choice for supersonic STOVLcombat aircraft. Vectored thrust requires theengine to be placed near the centre of gravity ofthe aircraft and this requirement results inaircraft with big fuselages in the middle; asituation in contrast to the area-ruling needed insupersonic aircraft. Therefore, a remote liftsystem (RLS) was selected.

The baseline configuration that evolvedafter several exploratory configurations isshown in Fig. 1.

The wing is swept back, in a medium-to-high position, and trapezoidal in planform. It isa low tapered, low aspect ratio wing and isbased on a thin aerofoil section of constantthickness-to-chord ratio along the span. Thewing does not have any dihedral or twist. The

tail is conventional and is placed low. The finand tailplane aerofoil sections are identical tothe wing.

The engine is a modern military reheatedturbofan of the remote lift type when operatingin the landing mode. A system of three ducts,used in landing to provide thrust in the front partof the aircraft, starts just after the turbine exitposition and ends in a nozzle at the undersurfaceof the aircraft after the nose undercarriage bay.It is a hot system without any reheat at thenozzle. It is, therefore, a simply remote liftsystem (RLS) without any augmentation. Threeducts have been adopted, instead of one, inorder to reduce the enormous size a single ductwould have. The three ducts, one bigger in themiddle and two smaller either side (Fig.2),recombine at the front nozzle. This front nozzlehas a limited vectoring capability. At landingthe thrust in the rear part of the aircraft is notprovided by vectoring the engine rear nozzle,but by a pair of swivelling duct nozzles situatedeither side after the engine position. These twoside duct nozzles are positioned verticallyduring landing and are stowed horizontally inthe fuselage during take-off and forward flight.The system of front ducts and nozzle and theside duct nozzles are used only for landing. Theengine has a conventional variable geometryrear nozzle for forward flight.

Two S-shaped intake diffusers withrectangular intakes are used. The intakes have afixed geometry since the aircraft is specified notto exceed Mach 1.6. The cross-section of theintake diffusers slowly changes, especially afterthey join, to become circular and meet theengine face.

Due to the operation of the RLS nozzle inthe front part of the aircraft during landing, theintake diffusers incorporate auxiliary inlets intheir upper surface (Fig.1). These are put intooperation at landing, when the usual intakes areclosed, to avoid hot ingestion.

The nose undercarriage bay is located afterthe radome. The main undercarriage bay lies flatunder the engine gas generator (Fig. 1).

The weapons of the aircraft, comprisingfour medium and two short range air-to-airmissiles or a number of bombs, are carried

DESIGN SYNTHESIS AND OPTIMIZATION OF AN ADVANCED SHORT TAKE-OFF AND VERTICAL LANDING (ASTOVL) COMBAT AIRCRAFT

162.3


162.4

internally in a bay, placed between the RLSnozzle and the engine axially, and between thetwo intake diffusers laterally (Figs 1, 2). Theyare jettisoned by means of ejectors situated onthe top of the bay. An internal gun is located ina bay on the left side of the weapons bay.

In addition to the wing fuel tanks, threefuselage tanks have been specified. A centralfuel tank located above the internal weaponsbay, ahead of the engine and between the intakediffusers, and two side tanks in either side of theengine gas generator (Fig.3). The side tankshave the same height as the engine at thatposition and blend with the side duct nozzlesthat follow them.

The initial baseline configuration definesan aircraft of around 19 t take-off weight and 20m in length. The wing span is 13 m and thewing surface area is nearly 40 m2. The enginemaximum gross thrust at sea level conditions is185 kN.

2.2. Design SynthesisThe design synthesis describes the aircraftbaseline configuration in mathematical form, by

means of semi-empirical and sometimesanalytical design relationships.

Initially, using design input data, the sizingof the aircraft basic items is performed. Basicitems are all the "standard" parts of the aircraftsuch as the radome, the cockpit, theundercarriage, the internal weapons bay, theinternal gun bay, the intake diffusers, theengine, the RLS, the basic wing and theempennage. These parts are determined by inputdesign data, and, in this sense, are consideredbasic items within the design synthesis. In theoptimization, as it will be discussed later, manyof these basic items may vary.

Next, the geometry of the aircraft isevaluated. Use of input design data is madetaking into consideration the related basic itemsso that the aircraft can contain all the specifiedbasic items. The fuselage geometry is evaluatedat several stations along the fuselage with thehelp of a fairing curve, in order to define anacceptable aerodynamic shape, and area rule ifrequired.

From the basic items and the geometry, themass and the centre of gravity are calculated.


162.5

The aerodynamics follow, based on thegeometry and the input aerodynamic data.

At the end, after having established all thecharacteristics of the aircraft, and using thesortie profile specifications (Table 1), theaircraft point performance is calculated.

2.3. OptimizationThe optimization of the ASTOVL combataircraft baseline configuration is carried out - byinterfacing the aircraft design synthesis with anoptimization code. The optimization code thatwas used is RQPMIN. This was developed bySkrobanski [12] and is one of a series ofoptimization codes used by DERA [13].

RQPMIN is a general numerical multi-variate optimization code. It is designed to solveconstrained optimization problems, that is tominimize or maximize a function, subject to anumber of constraints and upper and lowerbounds imposed on designated variables.

The operation of RQPMIN is based on theLagrange-Newton optimization. A stationarypoint of the Langrangian function is calculatedby Newton's method. RQPMIN differs fromsimilar codes in that it does not use a penalty ora criterion function to force global convergence.Instead, the concept of pseudo-feasibility isapplied. A trial point is rejected if the square

root of the sum of the squares of the constraintsis greater than the radius of pseudo-feasibility.

RQPMIN involves in its operation externalvariables (EVs), independent variables (IVs),dependent variables (DVs), an objectivefunction (OF), equality constraints (ECs) andinequality constraints (ICs). RQPM1N variablesare the variables of the design synthesis code.EVs are design input variables. IVs arevariables that are given an initial value and thenare varied by RQPMIN. DVs are dependent onEVs and IVs. The OF is a design synthesisfunction chosen to be optimized by RQPMIN.ECs are specified equalities, involving designsynthesis variables, that must be satisfied by theend of the optimization. ICs are similar to ECswith the difference that involve inequalities;they become active only if these inequalities arenot satisfied. The selected IVs for this study arepresented in Table 2.


162.6

2.4. Objective Function (OF) DevelopmentAlthough it has been suggested in a number ofreferences that the aircraft take-off mass or theaircraft empty mass is the best OF, a number ofother OFs were tried. This investigation aimedfor an improved aircraft and this does notalways mean an aircraft with lower empty mass.For example, an OF other than the empty massmay give an aircraft with the same empty massbut lower fuel mass. In search for a better OF,and since this type of aircraft is built "aroundthe engine", the engine mass was nextconsidered. In addition to the aircraft emptymass and the engine mass, four more OFs weretested; the engine scale factor, the wing area, thewing mass and the fuel mass.

It should be noted that the results of the OFinvestigation are not related to the baselineconfiguration optimization. This is because, forthe purpose of the OF investigation, very lightconstraints were set which facilitate aconvergence. Light constraint conditions werechosen to produce a more substantial reductionin OF investigation time and effort than wouldotherwise be needed.

3. Discussion of the Baseline ConfigurationStudy Results

STOVL combat aircraft configurations areamong the challenges in conceptual aircraftdesign, especially if a supersonic capability isalso specified. The difficulties in evaluatingsuch configurations are mainly due to the thrust-to-weight ratio needed, the matching of thrustand centre of gravity and the internal packagingconstraints.

The requirements of a high thrust-to-weight ratio that is necessary for verticallanding - giving a long engine-and of an internalweapon bay with very long medium-rangemissiles, are conflicting. Several attempts weremade to place the internal weapon bay in such away as not to have the length of it added to thelength of the engine. These failed, and therefore,the weapon bay remained ahead of, and in linewith the engine.

Another major problem was the size of theRLS front duct. With a configuration that

dictated an approximately 50 % thrust splitbetween the front RLS duct and the rear engineside duct nozzles, a RLS duct of enormouscross-section became necessary. It was evidentthat a RLS duct of this size could not beaccommodated in the aircraft fuselage. Twocategories of alternatives were considered; thefirst was the non-circular cross-section duct andthe second the multiple duct. Although manynon-circular cross-sectional shapes providesatisfactory volume efficiency, they suffer fromsubstantial pressure losses. In the multiple ductcategory the volume efficiency is almost asgood, but at a small penalty in terms ofincreased frictional losses and weight. Theselected solution was a RLS front ductcomposed of three circular ducts, one bigger inthe middle and two smaller ones on either side.This three-duct arrangement is much smaller inheight than a single circular duct of the samecross-section and, therefore, can fit in the upperpart of the aircraft fuselage between the cockpitand the tail (Fig. 1). However, the matterrequires further investigation.

Finally, the well known V/STOVL combataircraft problem of engine hot gas ingestion[141, while hovering, was very brieflyinvestigated with reference to the intake diffuserinlet position. Positioning of the inlet furtherback and away from the RLS front nozzlewould either reduce the length of the intakediffuser unacceptably from the intake flow pointof view, or would unnecessarily increase theaircraft length. The chosen alternative was touse auxilliary intakes in the upper surface of thediffusers, to be used in the hover. The forwardflight intakes were chosen to be fixed in view ofthe Mach 1.6 maximum speed of the aircraft.

3.1. Discussion of the Design SynthesisAlthough the design synthesis methoddeveloped by Lovell [4] was, in general,followed, significant changes were made. Mostchanges resulted from differences in thetechniques used, and additions related to thespecific ASTOVL configuration chosen. Thebasic item additions due to the ASTOVLconfiguration were related to the engine and theRLS. The engine thrust-to-weight ratio at


162.7

landing was put to 1.15 (excluding bleed air forthe reaction control system (RCS)). To this wereadded a 7 % allowance for the RCS [15] and anestimated 3 % due to RLS nozzle and duct, andengine swivelling duct losses. This brought thetotal thrust-to-weight ratio to 1.25. However, itshould be pointed out that considerablevariations exist for thrust losses; an example isthe suckdown losses given as 13 % in Ref.15and as 3 % in Ref.16.

The RLS presented many difficulties.Estimates had to be made to account not onlyfor its structural but for its insulation mass aswell. The structural mass calculation was basedon comparisons with RALS (remote augmentedlift system) concepts[17,18]. Willis, Konarskiand Sutherland [19] being particularly useful.Although they dealt with ejector RALS, theyprovided useful size and mass information onducts and nozzles.

The RLS nozzle position was examined inrelation to the engine front/rear split. As verylittle room was available for nozzle movementdue to the configuration (internal weapon bay),a 50 % thrust split was decided to provide anozzle position between the cockpit frontbulkhead and the internal weapon bay (Figs 1and 2). A constant thrust split poses a limitationon the code, but a variable one would requiretransition stability limitation considerations thatwere beyond the scope of this work.

The fuselage geometry was evaluated withdue care for the definition of the cross-sectionalareas in question. The number of fuselagestations was greater than Lovell's [4], mainlybecause of the complicated rear fuselage. A lotof emphasis was placed on the determination ofthe fairing curve. It was made very clear that thefairing curve relates to the net cross-sectionalarea of the fuselage.

In general, Lovell's [4] tendency to usedensities in mass calculations was not followed.Instead, actual mass values were preferred. Itwas felt that for items far away from the aircraftcg, an improved accuracy was particularlyimportant for an ASTOVL aircraft. Lovell's [4]internal fuel tank solution was not adopted. Theevaluation of the internal fuel volume capacityof the fuselage by way of subtracting the

volume of the basic items from the fuselagevolume has some drawbacks. The reason is thatmany parts of the fuselage are not suitable forfuel storage. Instead, three fuselage fuel tankswere designated, their location dictated bysafety, reliability and aircraft cg limitations.

Only minor alterations have been made toLovell's [4] drag estimation, but his lift-curveslope estimation was enhanced with interferenceeffects. Preliminary calculations showed theseeffects to amount to around 10 % . For body andwing-body interference, the informationavailable in Ref.20 was used, after beingtransformed into analytical form. An analyticalexpression for the tail interference downwash,based on finite wing theory, was derived fromfirst principles [21].

The available engine performance modulein data form took too much computing time andshould be reconsidered. Perhaps a formula-based engine performance approach might be abetter solution in the future.

3.2. Discussion of the Optimization and theDevelopment of the Objective FunctionA number of difficulties were encountered withthe application of the optimization codeRQPMIN, 1986 version, used in this study. Thefirst is related to the tolerance limits of the code.RQPMIN [12] was designed for mathematicalapplications, and, its tolerance limits, thecriterion as to whether the constraints have beensatisfied or convergence achieved, were toosmall. Even their upper (maximum) value wasinappropriate for engineering purposes.Numerous runs ended unsuccessfully becausethese tolerance limits could not be satisfied,even though they produced substantialreductions in the OF.

Another major problem was connectedwith the RQPMIN condition that all parameters,with the use of scale factors, should have beenkept at values of around one. This was notalways possible since during the optimizationsome parameters changed considerably.

During its feasibility steps, RQPMINvaried the IVs in every direction until it found afeasible path. When very tight constraints werespecified, only four IVs exhibited variation


162.8

during the optimization. They were the enginescale factor, the wing quarter-chord position,and the front and rear spar positions. Since theaircraft is built "around the engine" it is evidentthat the engine scale factor was the decisiveparameter for the aircraft mass reduction. Theaircraft mass reduction produces a shift in cg,and the wing quarter-chord position was used tocontrol it. The spar position variation could notbe explained.

It should be mentioned that RQPMIN leadto a much better optimization - empty massreduction when convergence was not achieved;this occurred with the constraints beingsatisfied.

The constraint which drove theconfiguration was the thrust-to-weight ratio atlanding. This constraint resulted in a large

engine, and consequently, a large RLS. Bothadded substantially to the aircraft mass. Anothersignificant constraint, deriving from the aircraftconfiguration, was the in-line positioning of theinternal weapon bay and the engine. Again, thisconstraint added to the aircraft mass,, due to theresulting longer fuselage.

The most important performanceconstraints were those of the sustained turn rate,the specific excess power and the maximumMach number (Table 3). The second sustainedturn rate constraint proved impossible toachieve.

The closest to the constraint value of 9.5deg/s at 9,000 m, M 0.9 and full engine powerthat could be achieved was around 4. In general,the specified performance constraints make aconvergence very difficult. It seems that the


162.9


162.10

optimizer needs some freedom in order toconverge. Furthermore, it is evident that someof the performance constraints contradict eachother.

From the purely optimizational point ofview the most difficult constraints were theECs, because the tolerance problem discussedearlier was much more acute than in the ICs.The explanation lies in the fact that an equalityis approached from both sides; an inequalityfrom one. A way to by-pass this difficulty mightbe to replace every EC with two ICs, thusallowing whatever tolerance is consideredappropriate. The results of the optimization ofthe baseline configuration are shown in Table 4.Both the initial and the optimized configurationsatisfied the specified constraints.

In the search for a better OF, the aircrafttake-off mass was used as a measure ofcomparison. In the course of the investigation,six aircraft parameters - empty mass, enginemass, engine scale factor, wing surface area,wing mass, and fuel mass - were tried as OFs.The results are shown in Table 5. Starting withthe engine mass as OF, and after a number ofdifferent sets of initial conditions., onlymarginally better results were achieved. At best,the aircraft take-off mass was reduced by 73 kgor 0.54 % (Table 5, column 7). This slightimprovement derived mainly from fuel massreduction,,- while the empty mass is 0.05 %worse than the empty mass OF case. Next. theengine scale factor was tried as OF, asanalternative to engine mass, but with similar


162.11

results (Table 5, column 8). During these trials,,a tendency for wing mass to increase wasobserved. Thus. wing surface area and wingmass were set as OFs. Wing surface area gavesimilar results, but wing mass was much worse(Table 5, columns 9 and 10). The last attemptwas with fuel mass but without success (Table5, column 11).

Consequently, engine mass, engine scalefactor and wing surface area gave, as OFs,slightly better results than empty weight, butfrom the engineering point of view thisimprovement could only be regarded asnominal. A more realistic approach would be toinvestigate the minimum acquisition or life-cycle costs, but this was beyond the scope ofthis study.

4. Comparison with Similar ASTOVLDesigns

The optimized baseline configuration wascompared with three other ASTOVL designs, asshown in Table 6. Two are case studies; the oneby Cox and Roskam [15] and the other theCranfield College of Aeronautics design projectS-95 [22]. The third is the Russian experimentalASTOVL fighter, the YAK-141 [23].

All four designs are of the remote lift type,with Cox and Roskam and YAK-141 havingseparate lift engines. The YAK-141 figures arefor STO operation.

From what it can be deduced, especiallyfor the YAK-141, all four designs have similarmission profiles, performance and weapons. Asshown in Table 6. three out of the four designs(the YAK-141, the CoA S-95 and the present)have comparable dimensions and weights.

Some differences exist in the wing area,fuel and empty weight figures. The CoA S-95has a larger wing area and the present design ahigher empty weight. The higher empty weightof the present design can be attributed to itscomplex powerplant configuration. A reason forthe larger S-95 wing area was its blended natureand edge-alignment, to improve its stealthcharacteristics. The YAK-14 and S-95 exhibit

significantly higher fuel consumption.YAK-141 fuel consumption must be a result of

its much higher payload, and, to a lesser extent,combat radius.

The Cox and Roskam design is muchsmaller in every respect. The most obviousexplanation may be that the structural weight ofthis design was underestimated [15].

Separate lift engine designs tend to havelower empty weight.

5. Concluding Remarks

The baseline configuration, the design synthesisand the optimization of the baselineconfiguration of an ASTOVIL combat aircrafthave been performed.

The baseline configuration represents anadvanced supersonic STOVL combat aircraftwith an internal weapon bay and a remote liftsystem for landing.

The design synthesis code describes thebaseline configuration and provides a fastcomputation of the aircraft geometry, mass.aerodynamics and performance. It is modular inform and can be easily modified to incorporatedifferent methods and techniques, and toaccount for other configurations. Every efforthas been made to strike a good balance betweenthe various parts of the code in relation to thedetail of the techniques used, but furtherconsideration is needed.

The interfacing of the design synthesiscode with the optimizer RQPMIN and theoptimization of the baseline configuration weresuccessfully achieved. The optimizer RQPMINexhibits some functional problems, for whichfurther attention is required. The specified pointperformance constraints were very tight, and asa result the second sustained turn rate constraintcould not be satisfied. The best objectivefunction for the reduction of the aircraft take-offmass was found to be either the aircraft emptymass or the engine mass.. The optimizedbaseline configuration is a 19.8 m long, 18.9 ttake-off weight aircraft.

Comparison with other ASTOVL designsindicated considerable similarities. However.the comparison suggested that separate liftengine designs tend to have lower emptyweight. Therefore. from the design point of


162.12

view, without taking into account operational,maintenance or logistic aspects, a separate liftengine might be a better solution than a RLS.

References

[1] Tamplin G C, Hammond D L and Fredette R E.Next-generation STOVL fighter. AerospaceEngineering, pp 15-20, January 199 1.

[2] Kehayas N. ASTOVL combat aircraft designsynthesis and optimization. Ph.D. Thesis. CranfieldInstitute of Technology, 1992.

[3] Collingbourne J. Multivariate optimization applied tothe initial design of transport aircraft. RAE TechnicalReport 84044, 1984.

[4] Lovell D A. The application of multivariateoptimization to combat aircraft design. RAETechnical Report 88003, January 1988.

[5] Serghides V C. Design synthesis for canard-deltacombat aircraft. Ph.D. Thesis, Cranfield Institute ofTechnology, 1988.

[6] Siegers F and Smith H. Numerical synthesismethodology for conceptual design of newgenerations of combat aircraft. AIAA Paper 95-3881,1995.

[7] Fozard J W. Tactical jet V/STOVL Its future ina CTOL world. SAE Paper 861637, October1986.

[8] Brown D M. Supersonic STOVL conceptualdesign of a fighter attack aircraft. AIAA Paper89-2112, 1989.

[9] Levine J and Inglis M. US/UK advanced shorttake-off and vertical landing (ASTOVL)program. AIAA Paper 89-2039, 1989.

[10] Schrage D P and Meyer S A. VTOL operationalconsiderations and their impact on futuremilitary design requirements. AIAA Paper 86-2649, 1986.

[11] Ward B D and Lewis W J. Advantages of thrustvectoring for STOVL. AIAA Paper 87-1708,1987.

[12] Skrobanski J J. User guide for RQPMIN. RAETechnical Memorandum AERO 2059, January1986.

[13] Smith J S and Lee C A. The RAE combataircraft multi-variate optimization. AIAA Paper89-2080, 1989.

[14] McLemore H C and Smith C C. Hot gasingestion of large-scale jet VTOL fighter typemodels. NASA TN D-4609, 1968.

[15] Cox B and Roskam J. Preliminary design of asupersonic short take-off and vertical landing

(STOVL) fighter aircraft. AIAA Paper 90-3231,1990.

[16] Raymer D P. The impact of VTOL on theconceptual design process. AIAA Paper 88-44798, 1988.

[17] Sammuels J J and Hahn A S. A parametricapproach to hover balance analysis of twoSTOVL fighter concepts. SAE Paper 861631,October 1986.

[18] Scott M L. S-80 roll control and remoteaugmented lift systems. MSc. Thesis, CranfieldInstitute of Technology, 19 8 1.

[19] Willis W S, Konarski M and Sutherland M V.Conceptual design, evaluation and researchidentification for remote augmented propulsivelift systems (RALS) with ejectors for VTOLaircraft. NASA CR-167906, .1983.

[20] USAF stability and control. McDonnell DouglasCorporation, 1975.

[21] Hoerner S F and Borst H V. Fluid-dynamic lift.Published by the author, 1975.

[22] Smith H. Systems design for an advanced shorttake-off and vertical landing combat aircraft.Aircraft Engineering and AerospaceTechnology, Vol 69, No 2, pp 129-144,1997.

[23] World Aircraft and Systems Directory1996/1997, Brassey's, 1996.

design synthesis and optimization of an advanced …

Documents