recent progress on powerplant / airframe …

ICAS 2000 CONGRESS

283.1

Abstract

Over the past 20 years, Powerplant / AirframeIntegration has become an essential step in thedesign of any transport aircraft.

For the aircraft manufacturer, it isessential to reply to the program requirementsin term of performance of the aircraft, tominimize the risks linked with the aircraftbehavior in flight, and to reduce the costs andthe development cycle. To achieve these goals, itis important for the Airbus partner in charge ofpowerplant / airframe integration to master thedesign of all related parts as pylon and nacelleaerodynamic lines.

This paper describes the experienceacquired by AEROSPATIALE MATRA AIRBUSon their products, highlighting the recentprogress at each step of the design process. It isdeveloped around the A340-600 program andthe new project A3XX.

1 Aerodynamics: the stakes for theaircraft

In the Airbus organization, the AerodynamicsDepartment of AEROSPATIALE MATRAAIRBUS (AM Airbus) has the responsibility ofthe design of the external shape of the pylon, ofthe nacelle external lines, of the nacelle air inletand more globally is in charge of powerplant /airframe integration. The aerodynamic work isdirectly linked with some important stakes forthe aircraft:

• to reach an optimum performance of theaircraft, in order to reply to its generalspecifications,

• to reduce at a minimum level, the risks linkedwith the aircraft behavior in the completeflight domain,

• to minimize the costs and the time cycle ofthe design process.

Aircraft performance improvement leadsthe engine manufacturers to increase enginebypass ratio. This results in larger nacelleswhich makes it always more difficult to bothachieve the required cruise performances andcomply with low speed requirements. Nacellebecomes a significant part in the overall aircraftperformance assessment [1], requiring thedesigner to minimize isolated nacelle drag aswell as to optimize installed propulsion systemefficiency [2].

After a description of the aerodynamicaspects of powerplant / airframe integration, thispaper describes the recent progress achieved ateach step of the design process, and concludeson the replies brought to the 3 stakes abovementioned.

2 Aerodynamic aspects of powerplant /airframe integration

2.1 Global aircraft performance

2.1.1 High speed aircraft performanceThe contribution of the engine installation to thetotal drag of the aircraft is not negligible (seeFigure 1). It can be shared in two parts: thefriction drag due to the wetted surface of thedifferent components, and the interaction drag.It is mainly on this second part that theaerodynamic optimization can have a significantinfluence.

RECENT PROGRESS ON POWERPLANT / AIRFRAMEINTEGRATION AT AEROSPATIALE MATRA AIRBUS

Arnaud HUREZAerodynamic Department, AEROSPATIALE MATRA AIRBUS

316 route de Bayonne (M0142/3) F-31060 Toulouse Cedex 03

Keywords: aerodynamics, engine integration, nacelle, design process, wind tunnel test

Arnaud HUREZ

283.2

Figure 1: Typical airframe drag breakdown

The aerodynamic phenomenon at the originof this interaction can be explained by the localmodification of the flow leading to a substantialdegradation of the local lift on the wing aroundthe pylon, and so to a decreasing of the globallift coefficient (CL) for a same angle of attack.To recover the necessary lift for the aircraft, thisloss of CL needs to be balanced by an increaseof the angle of attack, with consequences on thewing shock and on the induced drag. Theobjective is to design an integrated set of lines(pylon, nacelle, wing) which minimizes theseconsequences.

2.1.2 High speed nacelle performanceThe nacelle drag divergence Mach number mustnot be lower than the aircraft one, and sets dragvariation limits within a specified range of massflow and angle of attack covering typical cruiseconditions.

2.1.3 Low speed aircraft performanceThe low speed performance of the aircraft islinked with the efficiency of the high liftdevices: flaps and slats. Due to the kinematicsof deployment, and to the sweep angle of thewing, a large gap between the side of thedeployed slat and the inboard side of the pylonmay occur if no special attention is paid in thedesign. It is important for the low speedperformance, and more particularly at take-off,to design an appropriate pylon shape to obtain aconstant and small gap (see Figure 2).

Figure 2: Gap between pylon and deployed slat

2.2 Operational riskLocalized aerodynamic impacts can haveconsequences on the flight domain limitation ofthe aircraft.

It is the case of the over-speed which iscreated on the lower surface of the wing on theinboard side of the outboard pylon. This over-speed is created when the flow skirts the pylonand wing leading edges, as the angle of attackdecreases (low CL). If this over-speed becomesimportant, it can induce a flow separation whichis at the origin of excitations which can lead towing vibrations, depending on aircraft damping.This phenomenon becomes more critical whenthe Mach number is more important and so canimpact the flight domain of the aircraft at highMach/ low CL (rapid descent), reducing theairlines interest for the aircraft.

2.3 Aerodynamic nacelle requirementsThese objectives, associated with sometimesconflicting requirements, ask for AM Airbus,who are in charge of powerplant / airframeintegration, to master the whole nacelle linesdefinition. This is done either directly by thelines definition (air inlet, fan cowl lines), orthanks to close relationship with enginemanufacturer (nozzle lines).

These nacelle flow lines must be designedto comply with various requirements andobjectives, relative to both the aircraft and theengine design.

2.3.1 Aircraft specificationThe inlet needs to satisfy engine mass flowdemand within acceptable distortion limitsthroughout the aircraft operating envelope, atlow speed. This is to be translated into highAngle of Attack, as well as inlet cross windrequirements. Taking into account the winginduced flow deflection, the aircraft low speedflight envelope translates into inlet high angle ofattack operating conditions. In order to preventadverse impact on wing maximum lift and on2nd segment drag, an other low speedrequirement is to avoid inlet external flowseparation at low engine mass flow (see Figure3).

283.3


Figure 3: Low speed / high incidenceaerodynamic requirements

2.3.2 Engine aerodynamic requirementsEngine requirements refer to engine cycle andperformance. The inlet needs to satisfy enginemass flow demand within acceptable distortionlimits (inlet/engine compatibility) throughoutthe whole aircraft and engine operatingenvelope. On the other side nozzle exhaustareas, in both direct and reverse modes aredriven by engine cycle. Engine performanceguarantees impose inlet pressure recovery andnozzle flow coefficients objectives (see Figure4).

Figure 4: Engine requirements

2.4 Geometrical constraints

2.4.1 PylonThe installation of an engine pod and a pylonunder a wing is first driven by multi-disciplinaryconstraints as primary structure, mounts andsystems to be enveloped in the pylonaerodynamic lines (see Figure 5).

Figure 5: Pylon geometrical constraints

2.4.2 NacelleNacelle hardware requirements refers to enginegeometry, i.e. flange diameter or engine length,as well as location and size of equipment’swhich usually drives nacelle maximum crosssection (see Figure 6). Acoustic treatment areas,dictated for noise attenuation, are also to betranslated into geometrical constraints, as theymay have an impact on both air intake andnozzle duct length.

Figure 6: Nacelle hardware requirements

Nacelle flow lines must be compatible withstructure design and manufacturing requirement,and are therefore strongly affected by nacelletechnology. In addition, the design must beintegrated with the rest of the aircraft. It is whynacelle flow lines are generally the result ofextensive cooperation between Aircraft, Engineand Nacelle manufacturers.

2.4.3 NozzleBasically, a nozzle can first be regarded as abody of revolution. However one cannot ignorethat real nozzle geometry is much morecomplex and, for mechanical reasons,substantial amount of nozzle flow is three-dimensional, as shown on Figure 7.

Arnaud HUREZ

283.4

Figure 7: Nozzle and reverser fairings

3 The means: numeric and experimentaltools

It can be seen that the quality of the design ofthe engine integration is essential for the globalaircraft performance and for the minimization ofthe risks. It is why, to complete the experienceof the aerodynamicists, a lot of efforts havebeen made in AM Airbus on tools in order toimprove the design result, the time cycle and thecosts.

The design process follows different stepswhich are:• the aerodynamic shape definition in a

Computational Aided Design (CAD) system,• the mesh generation of the domain around

the aircraft,• the Computational Fluid Dynamics (CFD)

analysis of the flow in this domain,• the wind tunnel test of the most promising

shapes,• the flight test which "crowns" all the process.

These different steps are linked by aniterative process which is described on Figure 8.

Figure 8: Design process

3.1 Progress in numeric tools

3.1.1 Surface generationThe use of a CAD system gives the possibilityto take into account the numerous multi-disciplinary constraints above described.

The ICEMSURF software provides a lot ofnew possibilities for external aircraft surfaces.The main advantages of this tool are the globaldeformation allowing the aerodynamic designerto modify locally the geometry of the surfaces,keeping continuity with the neighboring ones,and the ability to analyze in real time the surfacemodifications.

3.1.2 Mesh generationThe industrial software used at AM Airbus isthe result of a ten years co-operation with ICEMTechnologies, and makes possible the 3Dstructured mesh generation of complex aircraftconfigurations. In parallel to this co-operation,AM Airbus developed their own meshenvironment which allows to control, modify(geometry and topology), enrich (Navier-Stokes, adaptation) and publish the mesh.

Three creation modes are used forstructured multi-block mesh design. With theinteractive mode it takes less than 8 days tocreate the topology and the mesh around acompletely new four-engine aircraftconfiguration. For similar aircraft geometries,automatic mode can replay an existing mesh ona new geometry (less than one day for acomplete aircraft). For local surfacemodifications, immediate mode only projectsthe existing mesh done around first surfaces, onthe new surfaces (less than one hour for acomplete aircraft). The above described modesare used to generate meshes for Eulercalculations.

Navier-Stokes (N-S) calculations are alsocommonly used now for Aircraft design [3]. It isstill a challenging task to generate interactivelya structured multi-block mesh for N-Scalculation, especially to have a good control ofnode location in the boundary layer region. SoAM Airbus have developed their own industrialsoftware to convert fully automatically an Eulermesh into a N-S mesh. For example, theconversion of an Euler mesh around complete

283.5


four-engine aircraft (1.5 106 nodes) into a N-Smesh (5 106 nodes) takes less than 1.5 hour.

For more specific studies (aerothermics),internal aerodynamic flow computation are nowbased on unstructured meshes generated withthe required tetrahedrons.

3.1.3 CFD toolsThe Computational Fluids Dynamics (CFD) isthen used to analyze the flow around thecomplete aircraft or to modelize the flow inmore precise areas.

AM Airbus use in an industrial basis,Navier Stokes / viscous Euler codes developedat ONERA and CERFACS. Their rapidresponse, even for complex geometry (completeaircraft), gives the possibility to reduce thedesign cycle up to one cycle per day.

3.1.4 Post processingA post-processing tool named QUICKVIEW,developed internally, gives facilities for 3Dphenomena representation on screen (see Figure9). It is possible to investigate eachaerodynamic values computed in all the domainwith different types of representation (surfacevalues, flow lines, cut in the domain, isosurface, 2D cuts …) with an interactive way ofworking, as the exploitation is very rapid, evenfor "big" results data.

Figure 9: Typical QUICKVIEW views

In fact, this tool can also be qualified of"pre-processing" tool, as it gives to designerfacilities to prepare the mesh, to prepare thecalculation (boundary conditions definition), tofollow the calculation itself (convergence). It isreally integrated to all the design process, with

possible return links towards surface generationand mesh tools.

3.1.5 Design environmentAll tools used for numerical aerodynamicanalysis (surface generation, mesh generation,CFD computation, visualization, post-processing) are activated by the Aerodynamicdesigner through an environment namedAEROSTATION [4].

The architecture of this consistent workbench is compliant with CORBA standard fordistributed objects. It allows communicationbetween components of different process andmanagement data tools. The completetraceability of the design process is ensuredautomatically, and the users can work together,sharing data, working methods, and JAVAscenarios.

This frame work will make easier the workdone with partners, allowing the communicationbetween the different tools and the managementof data exchanges.

3.2 Progress in experimental toolsIn order to validate the most promisingconfigurations computed, the next step of thedesign process is the wind tunnel tests.

3.2.1 Air inlet testsLow speed tests of air inlet are carried out atONERA F1-Le Fauga wind tunnel, for bothhigh angle of attack and crosswindperformances. The corrected mass flow in theinlet is achieved by the flow between the highpressure air in tunnel and atmospheric air andcontrolled by venturis.

The model used for cruise speed tests atONERA S1-Modane is the same one as in LeFauga. The main objective of S1 wind tunneltest is: to measure inlet total pressure recoveryin cruise conditions, and to measure inlet dragvariation in a specified range of Mach number,Mass flow and angle of attack.

3.2.2 High speed drag of nacelleIn order to have access to absolute level of dragof the nacelle itself, a new kind of test isperformed in ONERA S2 Modane wind tunnel.

Arnaud HUREZ

283.6

The nacelle is mounted on a special long pylon,as it can be seen on Figure 10.

Figure 10: S2 set-up

3.2.3 Nozzle testsAfter the optimization of aerodynamic surfaces,thrust bench tests are run so as to validate theperformances of the nozzle (see Figure 11).These tests, can be performed either in Fluidyne(USA) or in ONERA Modane BD2 (France).Several configurations are usually tested inorder to quantify exhaust area modifications (forexample nozzle exhaust trim) effects on nozzlecoefficients.

Figure 11: Nozzle test bench set-up

3.2.4 global aircraft performanceFor a test dedicated to a global aircraftperformance analysis in presence of nacellesand pylons, a full span model is used (seeFigure 12) and the simulation of power-plant ismade by Through Flow Nacelles (TFN).

Figure 12: Full span model of aircraft

3.2.5 Engine installationIn the case of a test dedicated to engineinstallation analysis, a bigger model is used(half model, see Figure 13), and a more precisejet representation is required. The exactrepresentation of the jet effect on the complexflow behavior in the region of the pylon / wing

junction can be made with a Turbo PoweredSimulator (TPS) which is a little enginepowered with high pressure air (see Figure 14).This device gives the possibility to have a realjet exhausting the nacelle and interfering the restof the airframe. The difficulty resulting fromthis test technique is to have a conception ofmodel with high pressure air coming throughthe wing and the pylon to the TPS. After acalibration on a static bench of the internal dragof this TPS (in relation with its mass flow rate),it is possible to accurately measure theinteraction drag level.

Figure 13: half model for engine installation

In certain cases, AM Airbus also usesimple but still representative TFN nacelleswhich are in fact triple body ones: with fancowl, core cowl and plug (see Figure 14). Itgives relevant and complementary resultscompared to TPS nacelles, with reduction ofmodel complexity and so reduction ofmanufacturing cycle and costs.

Figure 14: TPS and TFN triple body nacelles

3.2.6 Analysis of flow separation riskAs it was described in § 2.2, one of the potentialproblems can be the flow separation in thepylon / wing junction region. To have aqualitative idea of the phenomenon, flowvisualization can be performed in wind tunnelwith oil put on the model surface. But theprecise understanding of this phenomenon canonly be done with the quantitative values ofexcitation. This is now deduced frommeasurements in wind tunnel with unsteadypressure sensors "Kulite". AM Airbus have now

283.7


developed dedicated experience in treatment ofthat information, which validates theaerodynamicists' design work in the process ofrisk mitigation.

AM Airbus master all these above describeddifferent tests, and the links between them, tohave a global experimental analysis ofaerodynamic phenomena.

4 Validation of results and respect ofconstraints

Before to see which level of performance andrisk minimization has been effectively reachedwithin the design process, it is important tovalidate the results, and to verify that theaerodynamic requirements are satisfied.

4.1 Global aircraft performanceAs described in § 2.1, one of the goals of theaircraft manufacturer is to achieve a level ofaircraft performance in line with the programrequirements.

4.1.1 High speed aircraft performanceThe flow around complete aircraft is firstanalyzed with CFD tools (see Figures 15).

Figure 15: Euler results on complete aircraft

Then the results can be compared withwind tunnel results: a good agreement betweenthe 2 kinds of results can be seen on Figure 16on pressure measurements on wing.

Figure 16: CFD and experimental results comparison

4.1.2 High speed nacelle performanceThe nacelle drag divergence Mach number ispredicted with an Euler code post-processing(see Figure 17).

Figure 17: Isolated nacelle shock wave

It is possible to predict drag variation (Cx)with respect to Mach number (Mo), mass flowratio (ε) and angle of attack (AoA). This goodcorrelation with S1 isolated inlet test results canbe seen on Figure 18.

Figure 18: S1 isolated inlet drag results

4.2 operational risksAs described in § 2.2, an important risk for theaircraft is to have vibrations in flight domain,because of flow separation. The lower surfacewing over-speed on the inboard side of outboardpylon can be at the origin of such flowseparation (see Figure 19). It is why a specialattention must be paid in the design of theshapes in this area.

Figure 19: High Mach / Low CL Euler computation ofover-speed.

Arnaud HUREZ

283.8

Wind tunnel analysis based on oilvisualizations (see Figure 20), and on unsteadypressure signal treatment (kulites) give the datanecessary for the validation of the design.

Figure 20: S1 oil visualization

4.3 Aerodynamic nacelle requirementsThe 2 aerodynamic goals above described, haveto be reached with some geometricalconstraints, but also with some aerodynamicrequirements, more particularly on nacellebehavior, as explained in § 2.3.

4.3.1 Inlet / engine compatibility at low speedand high angle of attackFor high angle of attack flight conditions,pressure distortion arises from thick boundarylayer in the internal keel region. At a specifiedMach (Mo)/ Angle of Attack (AoA) condition,maximum inlet flow capacity is limited by flowpressure gradient and shock that may induceseparation. Inlet performance prediction (seeFigures 21 and 22) results from viscous flowanalysis, either Euler/Boundary layer couplingor Navier-Stokes modelisation.

Figure 21: Local Mach numberHigh incidence, high mass flow

Figure 22: Inlet / Engine compatibility

Then, the final validation of the design(respect of specifications) is made using F1isolated nacelle test results (see Figure 23).

Figure 23: F1 results and specifications

4.3.2 Inlet / engine compatibility in crosswindconditionsGround operations (i.e. take-off in crosswindconditions) results in the same internal flowanalysis at the inlet side. However special caremust be taken at low mass flow when adversepressure gradient usually induce flow separationat typical crosswind conditions. Potential flowcoupled with boundary layer analysis hasproven to be a reliable design tool, substantiatedby wind tunnel test experience. Figure 24 showsthe local Mach number and the strong deviationbetween external flow and friction line inducedby the adverse pressure gradient.

Figure 24: CFD in crosswind conditions

Total pressure recorded in F1 in cross windconditions (see Figure 25) is in line with thespecification, and clearly shows:• Flow separation induced by adverse pressure

gradient at low engine mass flow.• Separation free air intake at intermediate

engine mass flow demand.

283.9


• Distortion induced by shock strength atmaximum engine mass flow requirement.

Figure 25: Distortion Coefficient (DC) versus enginemass flow (Wr) in cross wind conditions

4.3.3 External flow separationInlet external flow separation must be avoidedat low engine mass flow, and at idle airflow (upto the aircraft maximum angle of attackencountered for certification testing). In theseconditions, top stagnation point is well insidethe inlet, with a high Mach number peak closeto the top leading edge (see Figure 26). Themass flow ratio at the specified Altitude / Machnumber /AoA conditions and the local curvaturenear the crown hilite have the strongestinfluence on inlet performance.

Figure 26: Local Mach number,High incidence, low mass flow

F1 test results (see again Figure 23) allowto demonstrate that the design meets therequirements.

4.4 Nozzle designCFD codes used at AM Airbus havesuccessfully demonstrated their ability tosimulate nozzle flow and predict nozzle flow

coefficients. Figure 27 shows Navier-Stokesflow computation results.

Figure 27: Mach number in the nozzle (Navier-Stokes)

Three dimensional CFD analysis is alsocarried to minimize impact of all the fairings onnozzle performance (see Figure 28).

Figure 28: Euler results on 3D nozzle shapes

Figure 29 shows the good agreementbetween test results and industrial CFDpredictions on these shapes

Figure 29: CFD and experimental resultsin fan duct and core duct

5 The reply to the stakes

Thanks to all the progress made in tools, andassociated with expertise acquired on previousprograms, it is possible to reach theaerodynamic goals:

5.1 Aircraft performanceThe objective of all the work done, is toimprove the performance of the aircraft, and inthis case, to reduce the interaction drag due to

Arnaud HUREZ

283.10

the powerplant installation, by an optimumintegration work. On A340-600, compared toA340-300, the engine installation drag,measured in wind tunnel at Mach 0.82, has beenimproved by about 30% (see Figure 30).

Figure 30: Installation drag of A340-300 and A340-600

5.2 Operational risk minimizationWind tunnel analysis based on unsteadypressure signal treatment (kulites) has been usedfor the design of A340-600 pylon in order topush the risk associated with flow separation outof the flight domain (see Figure 31).

Figure 31: Separation risk control

5.3 Cost and cycle reductionIn order to reply to the previous technicalrequirements, but with an industrial logic, it isnecessary to master the cost and the design loopcycle. A large improvement has been made forthe A340-600 development, as, in comparisonwith A340-300 program, the design studieshours, the wind tunnel model manufacturingcosts and the test hours have been reduced byabout half (see Figure 32).

Figure 32: cost and cycle reduction

6 Conclusion

For some years, large progress have been madeat each step of the design process inAerospatiale Matra Airbus.

These improvements can be measured onthe quality of design: aircraft performance andminimization of risks, but also on the reductionof design loop cycle, of wind tunnel tests hoursand costs. The A340-600 development isalready a good example of these improvements.

AM Airbus have also demonstrated theircapability to design the Trent 500 nacelle linesfor this aircraft in the short timescale requiredby this program, with validation in Wind tunnelthat they meet engine and aircraft specificationsin terms of low speed characteristics and highspeed performances.

The new A3XX project facing the Airbuspartners, is even more challenging (closerposition, bigger engine, more rapid aircraft), andno doubt the processes and tools presented herewill be put to good use and further enhanced.

Acknowledgments

The author would like to express itsappreciation for the technical support providedby all the people involved in the AM AirbusAerodynamics Department.

References

[1] Mazat E, Pelagatti O and Surply T. Design of A340-600/Trent 500 nacelle lines. AAAF 7th Europeanpropulsion forum – Aspects of engine / airframeintegration, Pau, France, 10-12 March 1999.

[2] Hurez A. Recent advances on engine airframeintegration at Aerospatiale. AAAF 7th Europeanpropulsion forum – aspects of engine / airframeintegration, Pau, France, 10-12 March 1999.

[3] Gacherieu C, Collercandy R, Larrieu P, Soumillon S,Tourette L, Viala S. Navier-Stokes calculations atAerospatiale Matra Airbus for aircraft design. 22nd

congress of ICAS, Harrogate, United Kingdom, 27August- 1st September 2000.

[4] Casties C, Soulard A, Chaput E, Barrera L, HuchardJ. Aerostation - Corba component approach toaerodynamic design framework. 22nd congress ofICAS, Harrogate, United Kingdom, 27 August- 1st

September 2000.

recent progress on powerplant / airframe …

Documents