vspaerowith transonic and viscous flow methods applied to...

1Advanced Air Vehicles ProgramAdvanced Air Transport Technology Project

VSPAero with Transonic and Viscous Flow Methods Applied to Flexible Wing Transport Aircraft

Daniel ChaparroMORI Associates/ NASA Ames Research Center

Gustavo E.C. FujiwaraUniversity of Washington

Eric TingNASA Ames Research Center

Nhan NguyenNASA Ames Research Center

Jason FugateNASA Ames Research Center

OpenVSP Workshop 2017September 1st, 2017


Outline

• Introduction• Generic Transport Model (GTM) Overview• Static Aeroelastic Model

– Aerodynamic Models– Modeling Framework– Transonic and Viscous Correction Methodology

• Results– Rigid Clean Wing GTM - Inviscid– Rigid Clean Wing GTM - Viscous– Aeroelastic Clean Wing GTM - Viscous– Rigid GTM Wing Equipped with Variable Camber Continuous Trailing edge Flap (VCCTEF)

• Conclusion


• Modern commercial aircraft technology trends indicate a direction towards lightweight, highly flexible wing structures.

• Increased flexibility increases aeroelastic interactions that can negatively impact performance.

• Active wing shaping control can be used to improve aerodynamic efficiency by tailoring aeroelastic wing deformation across the flight envelope.

• Active wing shaping design requires rapid analytical tools with reasonable accuracy in order to scan large design spaces.

Introduction


• Scope of this presentation: – Develop computationally efficient transonic and viscous flow vortex lattice modeling

framework for analysis and optimization of flexible wing transport aircraft

Introduction

Multi-FidelityModeling• Multi-fidelityaeromodeling(Cart3D,OVERFLOW,Panair,Vorlax,

VSPAero)• AeroelasticFEM(Beam3D,NASTRAN)• Capabilitiestocoupleaeroelastic FEMwithaerocodes• Flutteranalysis

Multi-DisciplinaryOptimization• Aerodynamicoptimizationfor

dragreduction(Cart3D,OVERFLOW,Panair,Vorlax,VSPAero)

• Multidisciplinaryoptimizationwithcoupledaeroelasticity

FlightDynamics• Dynamicsofcontrol

actuation• Dynamicaeroelastic FEM

coupledwith6-dofrigid-bodyflightdynamics

Aeroservoelastic Control• Aeroservoelastic control(flutter

suppression,loadalleviation)• Real-timedragminimization• Multi-objectiveflightcontrolfor

distributedflapsystem

ControlActuation• VCCTEF• Distributedpropulsion• Others(distributedcontrol,activeflow

control,etc…)

PerformanceAnalysis• Trajectoryoptimizationto

minimizefuelburn• Missionanalysisfor

energyefficiency

MDAO framework includes:


• Notional single aisle, midsize, 200 passenger aircraft developed by NASA Langley Research Center– Cruise Mach: 0.797– Cruise Altitude: 36,000 ft– Cruise CL: 0.51 (Baseline Stiffness)/ 0.497 (Half Stiffness)

• Variable wing stiffness to represent trends towards lightweight/ flexible wing designs

• Equipped with the Variable Continuous Trailing Edge Flap (VCCTEF)

Half stiffness

Baseline stiffnessRigid wing

Half Stiffness Wing

Baseline Stiffness Wing

Rigid Wing

Generic Transport Model (GTM)


Finite Element Aeroelasticity

• BEAM3D is a rigorously developed research program that employs equivalent beam models to construct an aeroelastic finite element model (FEA/FEM) for static and dynamic aeroelastic analysis.

• Key features of BEAM3D:– Current development in MATLAB– Multi-element beam/frame modeling– Modular design with upgradability– Capable of dynamic aeroelasticity– Nonlinear analysis capability

W

-Wx

QV

Vx


Aerodynamic Models

• Solver:VSPAero byKinney• Pre/PostProcessor:Matlab• PotentialFlowSolver– Inviscid SubsonicorFullySupersonicFlow• MeanCamberLiftingSurfaces

VortexLatticeModel

• Solver:TSFOIL(2D)byMurman,BaileyandJohnson• ComputationallyEfficient• SequentialMeshRefinement• ComparesWellwithEulerSolversatModerateAnglesofAttack(Chaparroetal1)

TransonicSmallDisturbanceModel(TSD)

• LaminarBoundaryLayer:CompressibleFormofThwaites’method• TransitionCriteria:Michel’sMethod• TurbulentBoundaryLayer:CompressibleFormofHead’sMethod• ComparesWellwithXfoil forSubsonicFlowandRANSforTransonicFlow(Fujiwaraetal2)

IntegralBoundaryLayerModel(IBL)

• Solver:MSESbyDrela• CapableofAnalyzingWideRangeofMachandReynoldsNumbers• AutomatedCoarsening/RefinementforRobustnessandComputationalEfficiency

Euler-IBLModel

1Chaparro, D., Fujiwara, G.E.C., Ting, E., and Nguyen, N., "Aerodynamic Modeling of Transonic Aircraft Using Vortex Lattice Coupled with Transonic Small Disturbance for Conceptual Design", AIAA Aviation 2016 Conference, Washington, D.C., AIAA paper 2016-3418, June 2016.2Fujiwara, G. E. C., Chaparro, D., and Nguyen, N., "An Integral Boundary Layer Direct Method Applied to 2D Transonic Small-Disturbance Equations," AIAA Aviation 2016 Conference, Washington, D.C., AIAA paper 2016-3568, June 2016.


Static Aeroelastic Framework

• Geometry deformer deflects the VCCTEF in Matlab environment and writes input files for the aerodynamic codes

• TSD/IBL loop is placed outside of the aeroelastic loop to improve computational efficiency• The Half Stiffness FEA (Lebofsky et al3) 3D beam model is coupled to the VLM model via

translation and rotation of the VLM panels

3Lebofsky, S., Ting, E., Nguyen, N., “Multidisciplinary Drag Optimization of Reduced Stiffness FlexibleWing AircraftWith Variable Camber Continuous Trailing Edge Flap,” 56th AIAA/ASCE/AHS/ASC Structures, Structural Dynamics, and Materials Conference, AIAA-2015-1408, January 2015






Wall-clock time is comparable if lift coefficient or angle of attack is specified







2D TSD/IBL loop is replaced with MSES for comparison



Transonic and Viscous Correction Methodology

1. The virtual twist angle due to transonic and viscous corrections, γ(y), is initialized to zero

γ(y)

Spanwise coordinate (y)



1. The virtual twist angle due to transonic and viscous corrections, γ(y), is initialized to zero2. The VLM model is coupled to the transonic and viscous corrections and the FEA model via

the panel incidence angle

Jig-shape Incidence

Angle

Elastic Twist (Pitch Axis)

Virtual Twist Angle

Effective Incidence

Angle




the panel incidence angle3. The VLM and FEA models iterate until the tip twist converges

Φtip




the panel incidence angle3. The VLM and FEA models iterate until the tip twist converges4. The effective 2D angle of attack is calculated for each airfoil

Airfoil Zero Lift Angle of

Attack

Virtual Twist Angle

Effective Airfoil Angle

of Attack

Sectional lift coefficient

2D lift slope

Prandtl-GlauertTransformation




the panel incidence angle3. The VLM and FEA models iterate until the tip twist converges4. The effective 2D angle of attack is calculated for each airfoil 5. Each airfoil is analyzed by the TSD/IBL or MSES model at the effective angle of attack




the panel incidence angle3. The VLM and FEA models iterate until the tip twist converges4. The effective 2D angle of attack is calculated for each airfoil 5. Each airfoil is analyzed by the TSD/IBL or MSES model at the effective angle of attack6. The virtual twist angle is updated for each section

Sectional lift coefficient

Airfoil lift coefficient from TSD/IBL or MSES




the panel incidence angle3. The VLM and FEA models iterate until the tip twist converges4. The effective 2D angle of attack is calculated for each airfoil 5. Each airfoil is analyzed by the TSD/IBL or MSES model at the effective angle of attack6. The virtual twist angle is updated for each section7. Repeat Steps 2 through 6 until the sectional lift and the 2D lift from the TSD/IBL or MSES

model converge for all sections




the panel incidence angle3. The VLM and FEA models iterate until the tip twist converges4. The effective 2D angle of attack is calculated for each airfoil 5. Each airfoil is analyzed by the TSD/IBL or MSES model at the effective angle of attack6. The virtual twist angle is updated for each section7. Repeat Steps 2 through 6 until the sectional lift and the 2D lift from the TSD/IBL or MSES

model converge for all sections8. Wave and friction drag are calculated by the TSD/IBL or MSES model. Lift, pitching

moment, and induced drag are calculated by the VLM model.


Results

• Rigid Clean Wing GTM – Inviscid

• Rigid Clean Wing GTM – Viscous

• Aeroelastic Clean Wing GTM – Viscous

• Rigid GTM Wing Equipped with VCCTEF - Viscous


Rigid Clean Wing GTM – Inviscid

• 3D Euler (Cart3D) is compared with transonic flow VLM models

*Coarse Mesh: exploratory studies**Fine Mesh: higher fidelity drag studies

VLM VLM+TSD VLM+MSES Euler

CL0 0.2479 0.2826 0.2782 0.292*

CLα 6.048 7.021 7.148 7.879*

45.7 35.5 37.4 36.2**LD

cruise( )

VLM VLM+TSD VLM+MSES

CL0 15.0% 3.1% 4.6%

CLα 23.3% 10.9% 9.3%

-26.2% 1.8% -3.4%LD

cruise( )

Percent Difference with Euler


VLM+TSD/IBL VLM+MSES RANS

CL0 0.266 0.266 0.259CLα 6.713 6.681 6.990

20.1 19.6 19.5

Rigid Clean Wing GTM - Viscous

• RANS (LAVA) is compared with transonic flow VLM models

LD

cruise( )

VLM+TSD/IBL VLM+MSES

CL0 -2.5% -2.4%

CLα 4.0% 4.4%

-3.3% -0.8%LD

cruise( )

Percent Difference with RANS


Aeroelastic Clean Wing GTM - Viscous

• RANS (LAVA) is compared with transonic flow VLM models

VLM+TSD/IBL VLM+MSES RANS

CL0 0.204 0.204 N/A

CLα 5.676 5.472 N/A

20.062 19.336 19.219LD

cruise( )

VLM+TSD/IBL VLM+MSES

-4.4% -0.6%LD

cruise( )

Percent Difference with RANS


Aeroelastic Clean Wing GTM - Viscous

Cp DistributionforAeroelastic Wing

Triangular load distribution due to wing deformation

Transonic and viscous flow VLM with TSD/IBL and MSES models are in excellent agreement


• Total of 85 candidate VCCTEF deflections are analyzed with TSD/IBL and MSES rigid wing frameworks at cruise CL = 0.51

• Linear combination of Chebyshev polynomials parameterize the spanwise flap deflections subject to the following constraints– Adjacent flap sections have relative deflections of less than or equal 2 degrees

– Minimum flap deflection is -1.0 degree– Maximum flap deflection is +6.0 degrees

– Linear coefficients are between -10 and +10

• The three VCCTEF camber segments are constrained to a circular deflection profile

VCCTEF Design Exploration


Best Performing VCCTEF Deflections

TSD/IBL and MSES agree on three of the five VCCTEF candidates with lowest drag

TSD/IBL and MSES agree on the shape of the 5th best candidate, but the magnitude is offset by 0.75 degrees


Best Performing VCCTEF Deflections

• The lift distribution for the minimum drag VCCTEF deflection moves away from elliptic near the root to lower wave drag

• Lift distribution from the TSD/IBL and MSES models are in good agreement

• Average drag reduction by the TSD/IBL model is 2.3 drag counts larger than the MSES model

Best performing VCCTEF deflection reduces drag by 7.04% and 5.71% according to the TSD/IBL and MSES models, respectively

Minimum Drag VCCTEF Deflection


Clean Wing

VCCTEF Deployed

Best Performing VCCTEF Deflection

Shock Weakening Due to VCCTEF

VLM+TSD/IBL Model


Wall-Clock Time Comparison

• All calculation are conducted on a Macbook Pro:– 16gb Memory, 2.5GHz Intel i7 processor (4 cores)– Not practical to run RANS

• VLM +TSD/IBL model is two and four orders of magnitude faster than Euler and RANS respectively.

• VLM +TSD/IBL model is ~3x faster than VLM+MSES model with similar modeling fidelity

Wall-clock Comparison of Multiple Aerodynamic Codes for a Single Fixed CL Aeroelastic Solution

2.2

3.3

8.7

12.1

22.4

163.9

12380.4

1 10 100 1000 10000WallClockTime[minutes]

RANS(LAVA)*EstimateCart3D(Coarse)VLM+MSES(Viscous)VLM+MSES(Inviscid)VLM+TSD/IBLVLM+TSDVLM


Lessons Learned/ Observations

• For VLM, removing the fuselage greatly simplifies modeling in VSPAero– Motivator to transition to panel method

• Excessive airfoil tessellation can corrupt the solution

• Executing VSPAero at transonic Mach numbers required disabling 2nd order Cp correction (Karman-Tsien Correction)

• .fem file only outputs the last condition in a sweep of solutions


Conclusion

• Developed transonic and viscous flow VLM framework for rapid design exploration and optimization of flexible wing transport aircraft – Modeled the GTM equipped with the VCCTEF– Demonstrated close agreement with LAVA RANS solver with both rigid and flexible wings– Frameworks shows excellent agreement between clean wing solutions using MSES and TSD/IBL– Efficiently explored large design space of VCCTEF deflections

• Demonstrated the ability of the framework to capture drag reduction due to the VCCTEF • Validated agreement between the TSD/IBL and MSES models for the top drag reducing flap deflections

– Wall-clock time was two orders of magnitude faster than Euler and four orders of magnitude faster than RANS

• The computational efficiency of the transonic and viscous flow VLM and its close agreement with RANS make it ideal for design optimizations that would be impractical with RANS or Euler solvers.

• Future work– Extend framework to capture transonic and viscous physics on fuselage and tail – Extend framework to use panel method instead of VLM– Extend framework to analyze more complex configurations including the truss braced wing


Publications

• Chaparro, D, Fujiwara, G. E. C., Ting, E., and Nguyen N., "Transonic and Viscous Potential Flow Method Applied to Flexible Wing Transport Aircraft", 35th AIAA Applied Aerodynamics Conference, AIAA 2017-4221, June 2017

• Ting, E., Chaparro, D., and Nguyen, N., "Aero-Structural Optimization of Variable Camber Continuous Trailing Edge Flap Configurations Using Transonic and Viscous Potential Flow Method", 35th AIAA Applied Aerodynamics Conference, AIAA 2017-4220, June 2017

• Ting, E., Chaparro, D., and Nguyen, N., “Development of an Integrated Nonlinear Aeroservoelastic Flight Dynamic Model of the Truss-Braced Wing Aircraft,” 58th AIAA/ASME/ASCE/AHS/ASC Structures, Structural Dynamics, and Materials Conference, AIAA-2017-1815, January 2017

• Fujiwara, G. E. C., Chaparro, D., and Nguyen, N., "An Integral Boundary Layer Direct Method Applied to 2D Transonic Small-Disturbance Equations," 34th AIAA Applied Aerodynamics Conference, AIAA 2016-3568, June 2016.

• Nguyen, N. and Ting, E., “Inertial Force Coupling to Nonlinear Aeroelasticity of Flexible Wing Aircraft,” 15th Dynamics Specialists Conference, AIAA 2016-1094, January 2016


Acknowledgments

• Marie Denison - Science and Technology Corporation in the NASA Ames Computational Aerosciences Branch for RANS and CART3D Modeling Support.

• Advanced Air Transport Technology Project under NASA ARMD for funding support

Thank You

vspaerowith transonic and viscous flow methods applied to...

Documents