High-speed wind tunnel test of the CAE aerodynamic validation model

Roy GEBBINK,Ganglin WANG,Min ZHONG

aGerman-Dutch Wind Tunnels,Marknesse 8316 PR,The Netherlands

bChinese Aeronautical Establishment,Beijing 100012,China

1.Introduction

The global demand and visions1for more sustainable means of air transportation bring forth many challenges for all aeronautical disciplines,from the level of aircraft design up to aircraft operations.In the field of aerodynamics,design innovation for commercial civil aircraft commonly2–4focuses on the increase of the aerodynamic efficiency Ma×L/D,where Ma is the Mach number,L the lift,and D the drag.Optimizing Ma×L/D in the transonic flight regime is often challenging,as the drag associated with shockwave formations generally increases progressively when increasing the flight Mach number.In the framework of a design study for an Ma=0.87 long-haul business jet aircraft concept,5the Chinese Aeronautical Establishment(CAE)initiated a project to conduct a dual-purpose wind tunnel experiment.The first purpose of the experiment was to establish the aerodynamic cruise performance for multiple configurations of the aircraft.The second purpose was to establish a reliable dataset for validation of an in-house developed CFD code.To serve both purposes within a single wind tunnel campaign,the CAE Aerodynamic Validation Model(CAE-AVM)was developed,and an experiment was designed that consisted of parallel deployments of various measurement techniques.This paper addresses the design features of the CAE-AVM,along with a description of the wind tunnel experiment and the typical test results at a cruise Mach number.

2.AVM design and manufacture

2.1.Configuration

The Ma=0.87 long-haul business jet aircraft concept is a configuration that consists of a narrow fuselage with a low mounted swept wing of a high aspect ratio,two aft-fuselage mounted engines,and a T-tail(see Fig.1).The overall length of the aircraft is 33 m and the wing span is 33.5 m,including the winglets.The aircraft targets a range of 11000–13000 km with a cruise Mach number of 0.87 at an altitude of 13 km.As the aircraft comprises a slender design,it is anticipated that a conventional metallic scale model would be severely affected by aeroelastic effects when tested in a pressurized wind tunnel.Refs.6–8have shown that for the transonic speeds of interest,aeroelastic model deformation(in particular wing twist)is an important uncertainty in a direct comparison between experimental and computational results.

In order to establish a comprehensive experimental dataset for CFD validation purposes,the goal was defined to keep the wing deformation of the model within a predefined limit of no more than-1°twist at the wingtip for the envisioned cruise lift condition.To realize this target,the cruise Mach number for the wind tunnel test model was reduced from 0.87 to 0.85,which allowed the new wing to be thickened and the sweep angle to be reduced.In parallel,the winglets were omitted,and modifications were made to the baseline wing sections,in an attempt to remove trailing edge separations at the lower model scale Reynolds number attainable in the envisioned wind tunnel.Similarly,the shape of the engine mounting pylons was adapted to remove shock-induced separations at the model scale Reynolds number.The engines themselves were represented by through- flow-nacelles.The resulting geometry was denominated as the CAE-AVM.

In collaboration with the German-Dutch Wind Tunnels(DNW)and the Netherlands Aerospace Centre(NLR),a project was defined for building an accurate wind tunnel model of the AVM,suitable for testing in the DNW High-Speed Tunnel(HST).Given the dimensions of the HST test section,the size of the AVM was selected so as to keep the ratio of the model frontal area to the test section’s cross-sectional area less than 1%and to keep the ratio of the wing span to the test section’s width less than 70%.As a result from these two wall interference considerations,9the CAE-AVM has a reference wing area of 0.208 m2,a reference span of 1.37 m,and a mean aerodynamic chord of 0.199 m.The wing has an aspect ratio of 9.0 and a quarter chord sweep angle of 35°.The aerodynamic design point,also referred to as the cruise lift condition,is Ma=0.85,CL=0.50,and Re=4.7×106(where CLis the lift coefficient based on the reference wing area,and Re the Reynolds number based on the mean aerodynamic chord).

Fig.1 Artistic impression of the long-haul business jet,with its cruise Mach number of 0.87.

2.2.Model features

The wind tunnel model is designed for simultaneous measurements of the overall forces and moments,wing pressure distributions,and model attitude.The model design allows for testing of two main aircraft configurations:the fuselage/wing(BW)combination and the complete configuration(BWNVH)that consists of the fuselage/wing/pylon-nacelles/vertical-tail/horizontal-tail(see Fig.2).

The core of the model is designed as a single central body,which houses an internal strain gauge balance developed by the NLR.The balance features an end-face flange that provides a model-balance interface with minimal hysteresis through the principle of a pre-stressed connection.The design of the model is such that the balance can be mounted on either a ventral z-sting support or a dorsal blade sting support.

For the assessment of the wing load distribution,a total of 180 pressure taps were incorporated.The taps were distributed in six span-wise stations amongst the left and right wings,at η={0.20,0.35,0.45,0.55,0.65,0.75}.Because the inclusion of the pressure taps weakened the construction,special design considerations were implemented to keep stiffness losses to a minimum.For example,the pockets in the wings(required for installation of the pressure tubes)were provided with load-bearing covers;in the wing root area,special attention was given to the tube and cable routing whilst maintaining a firm fixation interface between the wings and the central balance house.For reasons of strength and stiffness of the wings,the central balance housing and the empennage were all built from Ramax stainless steel.All other model parts were built from an aluminum alloy.

To verify that the design meets the predefined twist requirement,a finite element analysis(FEA)was conducted.As input for the FEA,a wing load distribution was used that was obtained from a CFD simulation for the nominal,nondeformed,aircraft geometry.The use of a load distribution for a non-deformed wing is judged to yield a conservative estimate of the wing twist,as the loading of the outboard wing sections is expected to be higher compared to a load distribution associated with a deformed wing shape.The FEA results confirmed that the wing twist should not exceed-1°at the tip,at the cruise lift condition.

Fig.2 As-built CAE-AVM full-model assembly,with its cruise Mach number of 0.85.

Upon completion of design,the actual manufacturing took place.During the manufacturing process,the accuracy of the model contour was closely monitored at the level of individual parts,as well as in-assembly.Hand finishing was applied to ensure smooth surfaces and no gaps and/or any forward facing steps between joints.After manufacturing,a Coordinate Measuring Machine(CMM)was used as independent verification that the as-built model was within tolerances.In terms of the as-built wing twist,the CMM confirmed that it was within±0.06°along the span.

3.Test approach

3.1.Wind tunnel facility

The test facility is the DNW-HST10in Amsterdam,the Netherlands.The HST is a continuous- flow pressurized wind tunnel that allows for aircraft configuration testing at subsonic and transonic speeds:the Mach number range is from 0.2 up to 1.3.The shell structure of the wind tunnel(Fig.3)allows the facility to be pressurized or evacuated,enabling testing at a constant Reynolds number at variable Mach numbers,or variable Reynolds numbers at a constant Mach number.The attainable total pressure range is from 25 to 390 kPa,and the operational total temperature ranges between 275 and 320 K.The air in the circuit is driven by an axial compressor which comprises of several rows of rotor vanes with adjustable blade pitch;a control system allows the Mach number to be kept constant within±0.002 during operation.The compressor itself is powered by a 20 MW electrical drive system.

The test section is of rectangular shape with a fixed width of 2.0 m and an adjustable height of either 1.6 m or 1.8 m.The side walls are solid,whilst the top and bottom walls are slotted with an openness ratio of 12%per wall.For the AVM experiment,the 1.8 m by 2.0 m slotted wall test section configuration was used.The test section is surrounded by a plenum chamber to accommodate ventilation of the air through the slotted walls.Downstream in the test section,a permanently installed vertical strut provides interfacing for several types of model support systems.

3.2.Test set-ups

The primary test set-up for the CAE-AVM consisted of a ventral z-sting with the so-called straight boom which was connected to the boom-base on the vertical strut downstream in the test section(Fig.4(a)).The boom-base is traversable in height and provides a pivot point for pitch angle variation.

Fig.3 Artistic impression of the DNW-HST circuit.

The kinematics was such that the Moments Reference Point(MRP)of the model could be kept on the tunnel centerline during pitch angle sweeps up to 10°.In this ventral set-up,the AVM was mounted upside-up on the z-sting via an internal balance.The z-sting was specifically selected for the primary set-up,because of its modest interference on the model aerodynamics,whilst providing the flexibility of testing the tail-off and tail-on model configurations.

Aerodynamic interference of the primary z-support(i.e.,the ventral z-sting with the straight boom support),a secondary test set-up was used.For the secondary set-up,the AVM was mounted via the internal balance attached to the dorsal blade sting shown in Fig.4(b).In this set-up,the model was mounted upside-down to allow the dorsal sting to connect to a pitching mechanism underneath the test section floor.Between the dorsal sting and the pitching mechanism,an xtraverse was installed.The x-traverse and pitching mechanism could be operated so as to match the exact though inverted model position and orientation,as occurred during pitch angle sweeps with the primary ventral set-up.

The aerodynamic influence of the primary set-up was established with two measurements in the dorsal set-up:one measurement with the model solely on the dorsal sting(Fig.4(b));the other measurement with the model mounted to the dorsal sting,whilst a dummy of the ventral z-support was present(Fig.4(c)).The latter consisted of the straight boom plus the so-called dummy z-sting,which entered the ventral model cavity without mechanical contact.The model loads were measured with the internal balance connected to the dorsal sting.In order to prevent any model-internal flow from the dorsal sting cavity to the z-sting cavity,the model was provided with a seal downstream from the balance.For the configuration without the dummy support,the ventral z-sting cavity was closed with an aft-fuselage plug so as to represent the nominal fuselage geometry.Hereby,the difference in balance readings between measurements with and without the dummy zsupport include the external interference effects due to the dummy z-support,plus any potential cavity(pressure)effects.This whole procedure relies on the assumption that the interference from the dorsal blade sting on the test object has no interaction with the(dummy)ventral z-support.This is deemed a reasonable assumption,thanks to the streamlined shape of the dorsal sting and its placement on the opposite side of the model so that there is no direct wake-interaction between the two support systems.

Fig.4 Side views of the test set-ups.

3.3.Instrumentation

For measurements of the overall forces and moments,a sixcomponent strain gauge balance was used.The balance’s longitudinal load ranges are:±14600 N for normal force,±1,360 N for axial force,and ±930 N·m for pitching moment.The balance was calibrated in advance of the test,by means of a multi-component loading scheme,and has a calibrated uncertainty(i.e.,a 95%confidence interval11,12)better than±0.3%of the range of each respective component.

The model pitch and roll orientations were measured with two Q- flex inclinometers.The inclinometers have an uncertainty of ±0.02°.The output of the pitch angle inclinometer was compensated for model vibrations by means of supporting signals from two linear accelerometers plus two angular rate accelerometers.13

The 180 wing pressures were measured with three Electronically Scanned Pressure(ESP)modules with digital temperature compensation.The ESP modules have a range of±105 kPa and an uncertainty of±0.15%of their range.Besides the wing pressures,three more pressures were measured in the aft-fuselage cavity where the z-sting enters the model.Furthermore,two rows of 56 pressure taps were measured along the test section’s side wall using ESP modules with a range of±17.5 kPa and an uncertainty of±0.15%of their range.

For the measurement of wing deformation,the so-called Stereo-Pattern Recognition(SPR)technique was used.SPR is an optical technique that makes use of a stereoscopic camera set-up to record images of markers on the wind tunnel model.For the AVM,ultraviolet fluorescent markers were applied on the wings and fuselages(Fig.5(a)).The markers were illuminated with several light emitting diodes to obtain highcontrast marker images,whilst keeping reflections to a minimum.The SPR technique allowed the position of an individual marker to be determined with an uncertainty of±0.1 mm.

For the application of infrared thermography(IRT),14,15the port wing was provided with a coating.The coating provided sufficient insulation from conductivity effects,so that surface heat radiation due to aerodynamic phenomena(such as boundary layer transition and/or shock waves)could be detected with infrared cameras.Thermal images were acquired with two FLIR SC3000 cameras,set-up so as to view the top and bottom surfaces of the wing.A 0.1 mm coating was applied so that all pressure taps remained open.On top of the coating,three rows of silver painted tick marks were applied(one mark each 10%of the chord,as shown in Fig.5(b))merely for an easy chord-wise position reference in the raw thermal images.

Fig.5 Views of the lower sides of the wings.

3.4.Test conditions and data acquisition

Primary database measurements for the BW and BWNVH model configurations were conducted on the ventral z-sting support.The measurements were conducted as sweeps of angle of attack at different Mach numbers,whilst the Reynolds number was kept constant.The Mach numbers ranged from 0.4 up to 0.9,and the angle of attack was varied between-4°≤α≤20°depending on the Mach number.The Reynolds number was kept constant at either 4.7×106or 2.0×106,and the total temperature ranged between 290 K and 310 K.

The majority of the test program was conducted with forced laminar-turbulent transition tripping,using selfadhesive dots.The tripping dots were applied on all relevant surfaces.On the wing,dots with a height of 0.089 mm were applied at 7%chord on the upper surface and at 5%chord on the lower surface.On the horizontal tail,0.079 mm dots were used at 7%chord on both sides.Similarly,0.089 mm dots were used at 7%chord of the vertical tail.The fuselage was provided with 0.114 mm dots applied at 10 mm downstream from the nose,whilst the pylons and nacelle inner and outer surfaces had 0.089 mm dots located 15 mm downstream from their respective leading edges.The sizing of the dots was based on the method in Ref.16.

Wind-on testing took place within runs,with one or more polars per run.In the ventral set-up,measurements were conducted in a continuous sweep test mode with an α-sweep rate of 0.1(°)/s and a data increment of 0.25°.For polars where wing deformation was measured,a step-by-step approach was used.During these polars,the flow conditions,balance loads,model attitude,wing surface and cavity pressures,wing deformation,thermal images,and wall pressures were all acquired in a parallel and synchronized mode.Fig.6 shows how various techniques were arranged,so as to allow all information to be acquired simultaneously during the primary database test.In addition,colored oil flow visualizations were conducted during separate runs,with the three views indicated in Fig.6.

Fig.6 Schematic of the optical arrangement.

In the dorsal set-up,only the model attitude,balance loads,and cavity pressures were measured.Wing surface pressures were not measured,as space requirements associated with the dorsal sting hampered the inclusion of the ESP modules in the narrow fuselage of the AVM.With the dummy zsupport installed,measurements were conducted in a stepby-step test mode,in the range from-2°≤α≤6°with a step of 0.5°.Without the dummy z-support,measurements were conducted in a continuous sweep testing mode again,in a range from-3°≤ α ≤ 7°with an increment of 0.25°.

4.Data reduction and corrections

4.1.General

In the data reduction,the conversion of the measured quantities to relevant engineering units and dimensionless aerodynamic coefficients was handled.In the process,the following corrections were applied to account for fundamental wind tunnel interference effects:

(1)Blockage correction: a Mach number correction based onresults of dedicated test section centerline calibrations.

(2)Empty test section buoyancy correction:a correction on the drag to account for the effect of the empty test section’s axial pressure gradient.The buoyancy force is proportional to the product of the pressure gradient and the effective volume of the body.17For our purposes,the correction is formulated as

where S is the reference wing area,?Cp/?x the centerline axial pressure gradient,A(x)the model cross-sectional area distribution,x the coordinate along the test section centerline,and xmthe length of the model.

(3)Corrections for the in fluences of the z-support:corrections on the lift,drag,and pitching moment coefficients,as derived from dedicated dummy support measurements(Section 4.2).

The model’s angle of attack is not specifically corrected for flow angle upwash;neither is a correction applied for the wall interference of the slotted test section.Regular reproduction measurements with a reference model have shown that upwash in the HST is within ±0.01°.As this is smaller than the uncertainty band of the applied inclinometer it was concluded that an upwash correction is not necessary for this test.As for the wall interference,Ref.9suggests that for the specific openness ratio of the slotted test section,no corrections are required for civil aircraft models with a reasonable span.The justification if wall interference is indeed also negligible for the AVM is presently hampered,due to the lack of wall pressure information along the top and bottom walls.

4.2.Support interference

As the support system generally imposes one of the largest influences on the test article aerodynamics,18–20the method for assessing and correcting this effect is addressed in this section in some length.Support interference is here defined as the influence of the primary z-support plus the influence of the empty test section with its permanently installed vertical strut.

The effects due to the empty test section are accounted by means of the blockage and buoyancy corrections described in Section 4.1.The aerodynamic influence of the support is derived via two measurements on the dorsal set-up:one measurement with the model solely on the dorsal sting and the other measurement with the model mounted on the dorsal sting,whilst a dummy of the ventral z-support is present.For each test set-up,the reference flow conditions are set at the model MRP.The flow conditions at the MRP are known through calibrations of the static pressure distribution along the centerline of the test section.Such centerline calibrations are conducted for two situations: firstly,the situation with the straight boom present in the test section;secondly,the bare test section without the straight boom support present.To ensure identical flow conditions(i.e.,identical reference Mach number)between the two set-ups,ΔMa at the MRP due to the presence of the straight boom support is taken into account,on top of the Mach number due to the test section itself.The blockage due to the z-sting is neglected in this procedure,for the practical reason that the z-sting is not included in the centerline static pressure calibrations.The latter is deemed to be a reasonable approximation is deemed to be a reasonable assumption due to the small volume of the sting and its streamlined shape.

With an identical reference Mach number set for both test set-ups,the difference in forces and moment coefficients is calculated by subtracting the balance results of the situation without the dummy z-support(DOR)from the results of the situation with the dummy z-support(DUM)at an identical angle of attack as follows:

where X={L,D,m}for respectively the lift,drag,and pitching moment;ΔCXzrepresents the influence of the z-support.At this stage,the empty test section buoyancy correction of Eq.(1)is added,so as to end up with a correction ΔCXwhich covers all support interference effects as follows:

Through this bookkeeping,ΔCXcomprises of the combined effect of the z-sting plus the straight boom(i.e.,the z-support)and the test section.The interaction between the z-support and the walls is hereby implicitly taken into account.Moreover,because it is verified that the pressure level in the sting-entry cavity is equal between the Fig.4(a)and(c)situations,it can be stated that the in fluences of the z-sting cavity and/or any model-internal cavity pressure effects are included in the support interference correction as well.Because the support system interferes differently for different model configurations,angles of attack,and Mach numbers,the corrections are determined separately for multiple combinations.Corrected results are finally calculated as

where subscript SIC refers to the support interference corrected results,and subscript VEN refers to the uncorrected results on the Fig.4(a)ventral z-support.

4.3.Wing deformation

The SPR technique was used for assessment of the deformation of the wing under aerodynamic loading.Recorded stereo-images serve as input to derive the locations of model markers by means of a volumetric calibration,which links stereo-pairs of pixels to a unique position in a threedimensional space.The reconstruction of model markers in a three-dimensional space is done twice:once for a situation at a wind-off condition;the other for a situation at a wind-on condition.The wind-on results are aligned with the wind-off results by means of a subset of four reference markers on the wing root and fuselage.In turn,model deformation is defined as the difference between the wind-on and wind-off positions of the remaining markers in the model axis system.For each individual marker,the displacement in the z-direction is defined as its bending.To derive a value for the local twist,a pair of markers is used:one marker near the Leading Edge(LE)and the other near the Trailing Edge(TE).The vector connecting the LE and TE markers is determined twice:once for the wind-on situation;the other for the wind-off situation.Finally,the local twist is defined as the angular difference between these two vectors,projected in the zx-plane;nose-up twist is defined as positive.

5.Results and discussion

5.1.Tripping validation

Infrared thermography was used to verify the effectiveness of the applied tripping dots to force laminar-turbulent transition on the wing.Fig.7 shows exemplary IRT results for a freetransition configuration(i.e.without tripping dots),as well as results for a configuration where the tripping dots were installed.

Fig.7 IRT results for the upper wing surface at Ma=0.85.

For the free-transition case in Fig.7(a),a clear temperature step can be seen over the observable part of the wing span.This temperature step reflects a sudden change in the local heat transfer coefficient as the boundary layer flow transitions from laminar to turbulent.For the free-transition case in Fig.7(b),again a clear temperature step is visible,but in this case,the temperature step is caused by the upper surface shock wave.Fig.7(c)and(d)shows that with the dots,the radiation images are notably different,as transition is forced at the locations of the dots.This can be deduced from Fig.7(c)where wedge-like shapes occur at the two stations where the dots were removed locally.The wedges represent an area with locally a lower heat transfer coefficient as is typical for laminar boundary layer conditions whilst directly behind the tripping dots,the heat transfer coefficient is larger,implying turbulent boundary layer conditions.Based on Fig.7,along with all other IRT data acquired during the test,it was concluded that the tripping dots are effective in forcing laminar-turbulent transition along the wing span at all Ma,α conditions considered.

5.2.Data repeatability

The repeatability of the measurement chain was inspected at various points along the test.This is done with two repeat polars measured within a single run,as well as with four repeat polars that were measured in different runs which had multiple model configurations changes between those runs.Fig.8 presents the lift coefficient versus the differences in the angle of attack Δα,drag coefficient ΔCD,and pitching moment coefficient ΔCm.The presented differences are obtained via linear interpolation of the data points from one polar to the identical lift coefficient value of the second repeat polar.These differences,or residuals,represent the level of variation between the repeat polars.The dashed lines shown on each plot indicate the 95%confidence interval,based on the applied instrumentation.Fig.8 shows that essentially all residuals are within the uncertainty band,which is in compliance with the expected standards for this type of test.18

5.3.Support interference results

Fig.8 Repeatability of the angle of attack,drag,and pitching moment coefficients at an identical(interpolated)lift coefficient at Ma=0.85(Dashed lines are the 95%confidence interval based on the applied instrumentation).

Fig.9 Established support interference corrections for the BW and BWNVH configurations at Ma=0.85.(Error bars show the 95%confidence interval based on the applied instrumentation).

The established support interferences for the BW and BWNVH configurations at the cruise Mach number are presented in Fig.9.The figure shows the interference levels according to Eq.(3);the error bars reflect the 95%confidence interval,based on the underlying instrumentation.The first observation is that the established interferences for the lift and pitching moment are relatively small for BW compared to BWNVH.Near the design lift condition(α ～ 2.5°),the established interferences are ΔCL=-0.005 and ΔCm=0.001 for BW,compared to ΔCL=-0.015 and ΔCm=0.025 for BWNVH.The signs and magnitudes of these ΔCmvalues imply that the support interferes mostly on the aft-fuselage,pylon-nacelles,and empennage.This is in-line with the expectation that the presence of the support causes the flow to stagnate progressively towards the back-end of the model,causing a decrease of the horizontal stabilizer effectiveness.The axial pressure gradient as induced by the test section and the z-support is further expected to cause a buoyancy force that alleviates the measured drag.The positive sign of the ΔCDcorrection for BW(Fig.9(b))confirms that the support is indeed pushing the model in the upstream direction.For the more complex BWNVH configuration,the ΔCDcorrection is initially also positive(i.e.,at a negative α),but changes smoothly towards negative drag interference levels as α increases.In fact,all interferences are found to vary to some extent with α,with a noticeable trend change near α ～ 2.5°.This change of trend is found to correlate with the point where the flow separation starts on the wing.

Upon establishing the interference levels at several discrete angles of attack,the data is least-squared fitted with a polynomial so as to obtain a continuous correction function for the considered α-range.When correcting data at angles outside of this α-range(e.g., α≥ 5°),the value of the nearest border is maintained constant.

Fig.10 SPR results for two repeat α-polars at Ma=0.85(Re=4.7×106,dynamic pressure q=54 kPa).

5.4.Wing deformation

An impression of the established twist and bending response along two α-polars is shown in Fig.10.For illustration purposes,the results of the two repeats are combined to generate a least-square fitted polynomial surface response for respectively bending and twist,as a function of the semi-span η and the angle of attack α.The resulting response surfaces show that the wing deforms elastically as a function of the angle of attack.Wing deformation is most pronounced at the tip,where the bending response varies from-9 mm to+18 mm along the α-polar,whilst the tip twist varies between+0.5°and-1.4°.At the design lift condition CL=0.5,the established tip twist is-1.0°,which complies with the initial twist requirement.In terms of uncertainty,the 95%confidence interval of the established twist result is determined as±0.05°in the wing root area and±0.2°at the wing tip.This is primarily due to data-reduction procedures described in Section 4.3,which causes the aforementioned SPR measurement accuracy of±0.1 mm for an individual marker to translate into a twist uncertainty that increases progressively towards the tip of the tapered wing.Because of the swept wing,one may expect that the established twist variation translates into a progressive loss of the total lift and an increase of the nose-up pitching moment along the α-polar.Such effects may be quantified using a CFD approach that considers the adapted geometry,according to the deformation results determined in this wind tunnel experiment.

6.Conclusions

Through design and wind tunnel testing of the CAE-AVM,an experimental dataset is established for two configurations of the aircraft.The overall loads and the wing load distributions are established for a variety of flow conditions between Mach numbers of 0.4 and 0.9.The repeatability of the overall loads is found to be within the expected uncertainty band of the underlying instrumentation.The interference imposed by the model support system is quantified on the level of the overall loads.The effectiveness of the applied wing transition trips is verified,and the amount of twist and bending of the wing is determined.Thereby,a comprehensive dataset is established which can be used for validation of computational methods.

As part of potential future test work,additional wall pressure measurements can be considered for confirmation on the accuracy of the current dataset.

1.European Commission.Flight Path 2050:Europe’s vision for aviation.Munich:Public Office of the European Union;2011.

2.Poll DIA.Aerospace and the environment what contribution can aerodynamics make.The Aeronautical J 2000;104(1037):321–3.

3.Takahashi TT.Optimum aspect ratio for subsonic air vehicles.J of Aircraft 2011;48(6):1984–93.

4.Torenbeek E.Synthesis of subsonic aircraft design.Netherlands:Delft University Press;1982.p.155–66.

5.Hua J,Zheng S,Zhong M,Wang GL,Chu L,Liu FL,et al.Design and verification study of an aerodynamic validation model.7th Asia-Pacific international symposium on aerospace technology;2015.

6.Levy DW,La flin KR,Vassberg JC,Tinoco EN,Mani M,Rider B,et al.Summary of data from the fifth AIAA CFD Drag Prediction Workshop.Reston:AIAA;2013.Report No.:AIAA-2013-0046.

7.Eberhardt S,Benedict K,Hedges L,Robinson A,Tinoco EN.Inclusion ofaeroelastic twist into the CFD analysis of the twinengine NASA common research model.Reston:AIAA;2014.Report No.:AIAA-2014-0251.

8.Keye S,Brodersen O,Rivers MB.Investigation of aeroelastic effects on the NASA common research model.J of Aircraft 2014;51(4):1323–30.

9.Maseland JEJ,Laban M,Ven H,van der Kooi JW.Development of CFD-based interference models for the DNW-HST transonic wind-tunnel.Reston:AIAA;2006.Report No.:AIAA-2006-3639.

10.Philipsen I,Poel M,Elsenaar A.DNW-HST(High Speed Tunnel)50-year anniversary.Reston:AIAA;2006.Report No.:AIAA-2006-3639.

11.Bergmann R,Philipsen I.Some contemplations on a proposed definition of uncertainty for balances.Reston:AIAA;2010.Report No.:AIAA-2010-4546.

12.Coleman HW,Steel WG.Experimentation and uncertainty analysis for engineers.2nd ed.New York:Wiley&Sons;1989.p.24–8.

13.Fuijkschot PH.Vibration compensation of gravity sensing inclinometers in wind tunnel models.Proceedings ISA 42nd international instrumentation symposium;1996.p.493–503.

14.Kompenhans J,Agocs J,Egami Y,Engler R,Fey U,Frahnert H,et al.Recent developments of image based measurement methods for application to transonic flows in industrial wind tunnels.Chin J Aeronaut 2006;19(2):115–25.

15.Crawford BK,Duncan Jr GT,West DE,Saric WS.Laminarturbulent boundary layer transition imaging using IR thermography.Optics and Photonics J 2013;3(3):233–9.

16.Braslow AL,Knox EC.Simplified method for determination of critical height of distributed roughness for boundary layer transition at mach numbers 0 to 5.Langley Field:National Advisory Committee for Aeronautics,Langley Aeronautical Lab.;1958.Report No.:NACA-TN-4363.

17.Swanson RS,Gillis,CL.Wind-tunnel calibration and correction procedures for three-dimensional models.Langley Field:National Advisory Committee for Aeronautics,Langley Aeronautical Lab.;1944.Report No.:NACA-WR-L-1.

18.Wubben FJM,Takara E.Wind tunnel model support and wall interference corrections in DNW-HST-ensuring high data quality standards.5th CEAS air&space conference;2015.

19.Elsenaar A,Han SOTH.A break-down of sting interference effects.Model support corrections in wind tunnels symposium;1991.

20.Rivers MB,Hunter CA,Campbell RL.Further investigation of the support system effects and wing twist on the NASA common research model.Reston:AIAA;2012.Report No.:AIAA-2012-3209.