In many practical circumstances in aeronauticalengineering and in related industries, flows arisewhere concentrated streamwise vorticity is embeddedwithin a turbulent shear flow. The juxtaposition ofthese features makes such flows particularlychallenging to compute using CFD methods. Inparticular, one may be forced to abandon strategiesfor modelling the turbulent stresses that proveadequate in two dimensional shear flows (i.e. inflows where the vorticity vector is orthogonal tothe mean velocity). The paper reviews developmentsin modelling turbulent stresses from their transportequations (rather than by appeal to an eddyviscosity hypothesis) and demonstrates, by way of arange of examples, the capabilities of this approachfor handling flows with concentrated streamwisevorticity.

The Royal Aeronautical Society Aerodynamics Grouporganised a CEAS European Forum on High Lift andSeparation Control which was held at the Universityof Bath on 29-31 March 1995. This paper gives anidea of the scope of the Forum and of the mainconclusions that emerged and it is followed in thisspecial issue of The AeronauticalJournalby four papers that were invitedfor the Forum.

Windtunnel tests on a two-dimensional model of a threeelement high lift system in a takeoff mode indicatethat airjet vortex generators can produce asignificant increase in lift at a given angle ofincidence and a substantial increase inCLmax. An associatedreduction in profile drag is inferred frommeasurements of momentum defect on the uppersurface.

The improvements are much greater than those typicallyachieved with vane vortex generators and cannotsimply be associated with the suppression ofboundary layer separation. Instead, the airjets andthe vortices which they generate promote enhancedmixing and momentum transfer across the complexshear layers above the main wing, from the externalflow right through to the surface. Detailed surveysreveal, for example, that the slat wake is dispersed— or absorbed into a region of greater shear — andthat the growth of the main wing boundary layer issignificantly reduced.

It is suggested that neither the relatively lowReynolds numbers . and Mach numbers of the tests,nor the effects of windtunnel interference, detractfrom the general relevance of the novel principleinvolved. Attention is drawn, however, to the needfor further work to elucidate and quantify the flowmechanisms and to establish the balance between thebenefits from the improved aerodynamics and theperformance costs of installing and supplying thejets.

Measurements of wing buffeting, using root straingauges, were made in the Nasa Langley 0-3 mcryogenic windtunnel to refine techniques which willbe used in larger cryogenic facilities such as theUnited States National Transonic Facility (NTF) andthe European Transonic Windtunnel (ETW). Thequestions addressed included the relative importancevariations in frequency parameter and Reynoldsnumber, the choice of model material (consideringboth stiffness and damping) and the effects ofstatic aeroelastic distortion.

The main series of tests was made on three half modelsof slender 65° delta wings with a sharp leadingedge. The three delta wings had the same planformbut widely differing bending stiffnesses andfrequencies (obtained by varying both the materialand the thickness of the wings). It was known thatthe steady flow on this configuration would beinsensitive to variations in Reynolds number. Onthis wing at vortex breakdown the spectrum of theunsteady excitation is unusual, having a sharp peakat particular frequency parameter.

Additional tests were made on one unswept half-wing ofaspect ratio 1·5 with an NPL 9510 aerofoil section,known to be sensitive to variations in Reynoldsnumber at transonic speeds. The test Mach numberswere M = 0·21 and 0·35 for the delta wings and to M= 0·30 for the unswept wing. On this wing theunsteady excitation spectrum is fairly flat (as onmost wings). Hence correct representation of thefrequency parameter is not particularlyimportant.

The program VICONOPT is used to find the optimum (leastmass) dimensions of a range of stiffened wing panelswhich are subject to buckling and material strengthconstraints and are loaded in axial compression witha sinusoidal manufacturing imperfection. Designplots are presented to show the effects that variousrib spacings and stiffener types have on optimumdesign mass. A simplified model of a complete wingbox is used to illustrate the design of a full wingpanel and plots of optimum values of designvariables at various stations along the wing havebeen obtained. The results were chosen to illustratethe practicality of optimisation with reference tomanufacture of a full wing panel and to show theeffect of changing the sophistication of modellingand theory used for the range of panels considered.The important aspects of the choice of designvariables and design concepts are highlighted andpercentage savings in mass, compared with an optimummetal panel design, are given for the various(global) optima found along with some examples of(rejected) local optima.

The performance of a helicopter rotor has always beendogged by the dissymmetry of air velocity over theblades caused by its translation in a directionsubstantially parallel to the plane of the rotordisc. Furthermore, in order to make the rotor evenusable, this dissymmetry of lift has caused theintroduction of major mechanical complications tothe rotor hub. In forward flight the rotor isaerodynamically split into two halves either side ofthe flight direction. The advancing side, where therotor rotation is in the same sense as the forwardvelocity, will see a greater dynamic head than theretreating side, where the rotor rotation andforward speed are in opposition. A rigid rotor wouldconsequently suffer a major rolling moment and if aviable proposition is to be achieved the rollingmoment must be avoided.

An experimental investigation into the wall jetscreated by a single inclined, twin inclined and afour inclined jet impingement arrangement isdescribed. Wall jet velocity profiles were recordedat various azimuthal and radial locations for arange of nozzle heights at jet pressuresrepresentative of current V/Stol technology. Use wasmade of wall jet properties established for a normaljet impingement to provide a means of comparing theattributes of the various configurations.Non-splayed twin inclined impinging jets generatesymmetry plane reinforcement and weakening inregions remote from the interaction. Splayed jetgeometries show significant height effects on groundplane flow development. The four-jet configurationalso demonstrates velocity reinforcement at thelines of wall jet interaction, together with aweakening of the flows remote from the interactionplanes. Reinforcement velocities for the four-jetcase are significantly lower than those expectedfrom the constituent twin-jet flows.

Recent advances in aeronautical technology arereviewed, with particular reference to work in theUnited Kingdom, and the process is outlined wherebythese have progressed from fundamental researchthrough to exploitation in marketable products.Aerodynamic, materials, propulsion and avionicaspects are discussed and successful applicationsdescribed in commercial and combat aircraft andhelicopters. Current and anticipated advances intechnology are considered in the context of possiblefuture applications, including brief considerationof the possibility of a new generation of supersonictransport aircraft. Lessons are drawn from pastexperience on actions necessary now if theseadvances are to be exploited successfully inimproving the performance, or reducing theacquisition or operating cost, of the nextgeneration of aircraft, without introducing anunacceptable level of risk.

A computational study is presented, which examines theperformance of variants of second-moment closure andnon-linear eddy-viscosity models when used topredict attached and separated flows over a highlift aerofoil for a range of incidence angles. Thecapabilities of both model types, especially inrespect of resolving the onset of suction-sideseparation at high incidence, are contrasted withthose of a low-Re k-ε model basedon the linear Boussinesq stress-strain relationship.The second-moment model contains a conventionallinear approximation of the pressure strainingprocess; a cubic (realisable) variant has beeninvestigated in an earlier study and found to offerno advantages. The quadratic eddy-viscosity modelfeatures coefficients which are sensitised to thestrain and vorticity invariants. While both models,in the form originally proposed, are superior to thelinear eddy-viscosity variant, neither performs wellin respect of resolving separation, unless modifiedso as to return the requisite low level of shearstress in the boundary layer approaching separation.Once separation is resolved with sufficient realism,the near wake aft of the trailing edge is also wellrepresented. All models return poor representationsof the far wake which is characterised by low levelsof turbulence production to dissipation ratio.

This paper describes an experimental investigation intothe effect of miniature vortex generators (VGs) onthe longitudinal aerodynamic characteristics of ahighly-swept wing with drooped leading edges. Theexperiments were performed in a low-speed windtunnelon a wing of the same delta planform as that used ina previous study. The latter wing was of fixedcamber, while the subject of the present study was awing with three different leading edge droop angles.The maximum reduction in drag due to the VGs wasfound to be about half that for the fixed-camberwing. This is reconciled with the differentbehaviour of the flow on the upper surface for thetwo types of wing. Without control, the drag of thevariable-droop wings is much lower than that of thefixed-camber wing. As a result, with control, thevariable droop wings have lower drag than that ofthe fixed-camber wing. Compared to that of thevariable-droop wings without control, thefixed-camber wing, with control, has lowerlift-dependent drag for lift coefficients between0·4 and 0·7. However, at higher lift coefficients,the reverse applies.

The design of control systems for helicopters in hoverand at low speed is a basic requirement for theextension of mission profiles and new missiondemands. A special task for various applications isthe position hold under wind and gust conditionsabove a ground fixed or moving target, like ashipboard reference, or a small vessel or lifeboatin rescue missions. For the solution of this problema controller concept was developed and thefeasibility was proven and successfully demonstratedin flight tests.

The in-flight helicopter simulator ATTHeS of the DLRhas been equipped by the Institute of FlightMechanics with an innovative measurement system forthe hover position above a target. A video camera incombination with a highly parallel computer systemfor processing the optical information was used asan integra ted sensor system for the measurement ofthe relative position of the aircraft to a target.Based on the existing well-proven flight controllaws of ATTHeS for the forward flight condition,which are implemented for handling qualitiesinvestigations, these control laws were modified andadapted to fulfil the special requirements of theposition hold task, including altitude hold andheading hold capabilities. The integrated system ofoptical position sensor and control computer enablesthe helicopter to hover automatically above adefined target in constant altitude and withconstant heading. Flight tests above a moving carunder wind and gust conditions underline the futurepotential of the overall system to be used underoperational conditions.

A rigid, 55° sweep, half delta wing has been oscillatedin pitch at subsonic speeds, and the unsteadypressures on both the upper and lower surfacesrecorded for pre-stalled conditions. Theoscillations were of low amplitude and atfrequencies equivalent to a typical wing firstbending mode for full scale applications.

When compared to a quasi-steady approximation, theunsteady pressures on the upper surface of the winglag the steady case along the line of the primaryattachment. The lag represents a constant convectivetime from the leading-edge with increasing frequencyof oscillation. A further localised area of laggingflow is observed beneath the vortex burst point, thelocation of which is a function of mean angle ofattack.

The magnitude of the unsteady pressure variations wasseen to increase linearly with the amplitude of thepitching oscillation while the phase lag was seen toincrease linearly with frequency parameter.

The process of designing the Boeing 777 high liftsystem is described. Particular attention is givento design constraints and examples are given of thetrade studies completed. The theoretical designtools and experimental test program are described. Asurvey of the high lift elements is presented andflight test results are compared withpredictions.

An analysis technique capable of simulating the effectof engine failure during takeoff from offshoreplatforms is presented. Use is made of inversesimulation whereby the control actions required fora subject helicopter to follow a particulartrajectory can be established. A mathematicalrepresentation of the towering takeoff procedure,and details of the modified inverse simulationtechnique needed to cope with the modelling of anengine failure are described. Detailed pilotingaccounts of the strategy used to fly the toweringtakeoff (with and without engine failures) are alsogiven and are used to give qualitative validation ofthe analytical approach. Simulation results for asingle engine failure of a transport helicopterduring critical phases of a towering takeoff arepresented. Finally, some directions for future workand potential applications of the technique arediscussed.

Two numerical methods are presented for the computationof steady and unsteady Euler flows. These areapplied to steady and unsteady flows about the NACA0012 aerofoil, using structured grids generated bythe transfinite interpolation technique. An explicitcentral-difference scheme is produced based on thecell-vertex method of Ni modified by Hall. Themethod is second-order accurate in time and space,and with flow quantities stored at boundaries theboundary conditions are simple to apply. This is adefinite advantage over the cell-centred approach ofJameson, where extrapolation of the flow quantitiesis required at the boundaries, making unsteadyboundary conditions difficult to apply. An explicitupwind-biased scheme is also produced, based on theflux-vector splitting of van Leer. The method adoptsa three stage Runge-Kutta time-stepping scheme and ahigh-order spatial discretisation which is formallythird-order accurate for one-dimensionalcalculations. The upwind scheme is shown to beslightly more accurate than the central-differencescheme for steady aerofoil flows, but it is notclear which is the more accurate for unsteadyaerofoil flows. However, the central-differencescheme requires less than half the CPU time of theupwind-difference scheme, and hence is attractive,especially when considering three-dimensional flows.The transfinite interpolation technique is ideal forgenerating moving structured grids due to itssimplicity, and grid speeds are availablealgebraically by the same interpolation as gridpoints. The method is also ideal for use in amulti-block approach.

Although mini-tufts are considered generally to showthe surface flow direction on windswept surfaces,there is some uncertainty about their correctinterpretation, particularly as three-dimensionalflows approach separation. This note provides someguidance on this controversial question, based onexperience with mini-tufts on wings, fins andcanards in the DRA 13 ft x 9 ft low speedwind-tunnel at Bedford.

An illustration of the value of the technique is given.Mini-tuft photographs are used to identify thetrailing edge condition which relates to theclassification of the steady and fluctuationpressure distributions observed a swept wing with aconical, separated flow.

The optimal takeoff performance of a vectored thrustV/Stol aircraft is analysed by a nonlinearprogramming method based on a sequentialunconstrained minimisation technique. The study isfocused on ship operations which involve VTO andski-jump STO launches. Once the optimal rampgeometry is determined, minimum-fuel andminimum-time problems are solved. The optimalcontrol sequences are evaluated and the effects ofship motion and atmospheric winds are discussed. Amicroburst encounter following an STO launch is alsodealt with in terms of aircraft survival.

The aerodynamic interaction between a wing and a coupleof propellers in tractor configuration isinvestigated by means of a model based on thelifting line concept and under the assumption ofquasi-steady incompressible motion of inviscidfluid. The ways the propellers influence the wingperformances, particularly the induced drag, areanalysed. It turns out that the lift increase andthe possible drag reduction depend strongly on thedirection of the blade rotation and on the propellerdistances from midspan, and result from two maineffects: the direct propeller induction on the wingand the modification of the lift distribution alongthe span. An additional nonlinear (mixed)contribution affects the drag but has onlynegligible influence on the lift. The describedmodel allows one to evaluate separately all theseeffects and helps understanding their influence onthe global modification of the wingperformances.

A model of the viscous flow past a flexible sailsection operating near the ideal incidence isdescribed. Viscous effects are calculated via weakviscous-inviscid interaction of a panel method andan integral boundary layer method, and a new modelfor the leading edge separation bubble isintroduced. The flexible section is allowed todeform in response to the pressure and shearstresses acting on it. Results are presented for theeffect of incidence, excess length and Reynoldsnumber on the development of the boundary layers oneach side of the section and the consequences forthe lift and drag of the section are described. Thenumerical model is compared with experimentalresults, giving in general good agreement andshedding light on the physics of the viscous flowpast flexible membranes.