FURTHER ANALYSIS OF MULTIDISCIPLINARY OPTIMIZED METALLIC AND COMPOSITE JETS

Transcription

1 FURTHER ANALYSIS OF MULTIDISCIPLINARY OPTIMIZED METALLIC AND COMPOSITE JETS Antoine DeBlois Advanced Aerodynamics Department Montreal, Canada 6th Research Consortium for Multidisciplinary System Design Workshop Ann Harbor, Michigan July 26 th 27 th 2011

2 Presentation Plan Theoretical Framework Disciplinary modules High-Speed Aerodynamics Low-Speed Aerodynamics Subspace Structural Optimization Metallic Wingbox Composite Wingbox MDO results of a Business jet Analysis of the final designs Conclusion 2

3 Current MDO Ingredients Step Cruise Initial Cruise En route Climb Descent 1500 ft Start-up & Taxi-out Takeoff & Initial Climb Sector Distance Flight Time & Fuel Block Time & Fuel Approach & Landing Taxi-in Mission Requirements Systems High- Performance Computing Load factor 2.5 V A V C KEAS V D Loads Multi- Disciplinary Optimization Optimizer Aerodynamics High-speed Low-speed Icing Structures Materials Composite Metal 3

4 Conflicting Requirements Design Variable Aerodynamics High-Speed DISCIPLINES Aerodynamics Low-Speed Structures Systems Buffet Loads & Dynamics OUTBOARD INBOARD INBOARD INBOARD CASE DEPENDANT Spanload Thickness distribution Leading edge thickness LOW HIGH HIGH LOW CASE DEPENDANT THIN THICK THICK THICK CASE DEPENDANT Span Sweep HIGH LOW HIGH LOW LOW HIGH CASE DEPENDANT CASE DEPENDANT 4

5 Optimization Challenges Aerodynamics and Structural optimization differ in terms of scope Need to develop an MDO architecture that suits every disciplines Bi-level formulation [1] Metallic Wingbox [2] Aero: Transonic small Disturbance code, KTRAN; Structure: Full 3D FEM, NASTRAN TM 5

6 min w.r.t OPTIMIZER wiobj i i si Planform, Profiles DESIGN CASES M UPDATE GEOMETRY CL LOW-SPEED CFD LOW-SPEED PRESSURES VALAREZO CHECK CL max, critical STRUCTURES MATERIALS METAL / COMPOSITE BUILD / UPDATE SOLVE STRUCTURAL FEM SUBSPACE OPTIMIZATION Min w.r.t W struct t s, h s M. S s.t. 0.0 W struc Twist (deg) f span (in) u HIGH-SPEED CFD AERODYNAMIC SOLVER SPANLOAD - TWIST - CAMBER W fuel FIELD PERFORMANCE FLIGHT PERFORMANCE FUEL VOLUME MULTI-OBJECTIVE CONSTRAINTS 6

7 High-Speed Aerodynamics MDO environment is linked to CFD code KTRAN Solved modified Transonic Small Disturbance, TSD, equations Embedded cartesian grid generation Despite low-order formulation: KTRAN provides accurate pressure distribution and Aerodynamic loads Enables full aircraft configurations including engine, winglet, H-tail and nacelles Computes accurate drag with a mixture of semi-empirical and CFDbased routines KTRAN allows trimming of aircraft, hence Symmetrical maneuvers 7

8 High-Speed Aerodynamics: CFD challenges Challenger CL-601 Cruise : Mach 0.82, = 1.5º n Dive: Mach 0.90, C L = 0.45 (2.5g) 2.5 n -1.0 V A V C V D KEAS -1.0 V A V C V D KEAS CFD remains a challenge at the edges of the flight envelope due to: Mach close to 1.0 ( M d M 0.07 ) Possible flow separation Large deformations mo 8

9 LOW-SPEED AERODYNAMICS Traditionally, manual iterations alter the high-speed optimized design to meet low-speed requirements Lengthy process Final design not guaranteed to be true optimum 9

10 LOW-SPEED REQUIREMENTS Lift curve Stall Cfx over the wing Lift Coefficient (C L) Pitching Moment Coefficient (C M) Inboard stall Alpha (deg) 10

11 SLAT-LESS DESIGN: LOW-SPEED REQUIREMENTS / OBJECTIVES CL MAX Lift Coefficient (C L) CL MAX _ CLEAN CL MAX _ req max DTO icing conditions C L max bucket stall Pitching Moment Coefficient (C M) crit crit _ req max [1] Cp TE Cp stall TE outbd max stall warning Alpha (deg) min C M stall crit sw [1] Valarezo, et al., Maximum Lift Prediction for Multi-element Wings" 30th Aerospace Sciences Meeting and Exibit,

12 Low-Speed Aerodynamics An automatic isolated wing VSAERO mesh generator AUTOMATIC VSAERO 1 DLR F-6 Isolated wing 10 sec (1 CPU) FANSC 2 DLR F-6 Wing-body 94 min! (32 CPU s P5-575) [1] Analytical Methods, Inc., A Code for Calculating the Nonlinear Aerodynamic Characteristics of Arbitrary Configuration [2] Full Airfcraft Navier-Stokes Code, Laurendeau, E., Development of the FANSC Full Aircraft Navier-Stokes Code," 12

13 Low-Speed Aerodynamics An automatic isolated wing VSAERO mesh generator AUTOMATIC VSAERO 1 DLR F-6 Isolated wing FANSC 2 DLR F-6 Wing-body [1] Analytical Methods, Inc., A Code for Calculating the Nonlinear Aerodynamic Characteristics of Arbitrary Configuration [2] Full Airfcraft Navier-Stokes Code, Laurendeau, E., Development of the FANSC Full Aircraft Navier-Stokes Code," 13

14 Structures AWSOM [1] automatically generates a 3D FEM All principal Structural Elements are modelized The FEM methodology is the same as the one used for certification models [1] DeBlois, A, et al., AIAA , 13th AIAA/ISSMO Multidisciplinary Analysis Optimization Conference 14

15 Structures module: Load transfer Whole Flight envelope to be analyzed: Speed, altitude and weight BALANCED MANEUVER C L = 2.5g W = MTOW M = M d BALANCED MANEUVER C L = -1.0g W = MTOW M = M c LANDING C L = 1.0g W = MLW R LOOP CLqS L n 2. 5g n CLqS L n 1. 0g n V = V a L W W(envelope) W(envelope) Ref BM

16 Structures module: Wing Box Sizing Procedure Optimization procedure is decomposed into: Multiple, Sequential; Reduced scope; Skin-stringer panels are Geometrically dependant chord-wise Geometrically independent span-wise (different for composite ) Algorithm: MPI_LoadBalance(); for rib = 0 Number of rib bays for str = 0 N str [rib]-1 Optimize Skin w.r.t skin constraints end loop str for str = 0 N str [rib] Optimize Stringer w.r.t skin-stringer constraints end loop str end loop rib Section A-A CPU rib + 1 CPU CPU rib rib -1 A A 16

17 Wing Box Sizing Program: Composite vs Metallic Same applied loads Same boundary conditions Optimized Metallic W sk-str = lbs Optimized Composite W sk-str = lbs 17

18 DLR F-6 Wingbox Weight and Stiffness Comparison Assumed Structural Layout Structural Weight Comparison Assumptions: MTOW = 100, 000 lbs MZFW = 60, 000 lbs No applied gear loads Bending Stiffness Comparison Torsional Stiffness Comparison 18

19 MDO Application on a Business Jet Problem statement: Flight and Field performance Optimization of a business jet aircraft through: Optimization of the wing planform shape; Wing sectional profiles; Optimization of the wing-box structure The MDO environment is ISIGHT Wing Sectional Wing Planform airfoils Analytical Descriptor 13 shape function parameter for each airfoil (x 7 airfoil) Lower Bound Upper Bound Wing Aspect Ratio Wing Leading edge sweep 32.0 o 40.0 o Root Trailing edge sweep 0.0 o 15.0 o Wing Taper Ratio Spanwise Break Location Wing Reference Area Decomposition method is used: 19

20 Objective Function MTOW Weighting Factor Minimize MTOW CL max Weighting Factor Maximize CL max [1] MTOW Scaling Factor CL max Scaling Factor Objective function analysed: OBJ 0.50 mass 0.50CLmax [1] The purpose of the objective function chosen is to validate the MDO setup, but does not reflect an actual wing design formulation 20

21 Metallic MDO results 21

22 Pareto front post-processed Mass Fuel Initial Point CL MAX 22

23 Pressure Coefficient Comparison Mid Cruise flight conditions: MACH = mid-cruise CL = mid-cruise INITIAL 23

24 Pressure Coefficient Comparison Mid Cruise flight conditions: MACH = mid-cruise CL = mid-cruise INITIAL 24

25 Composite MDO results 25

26 Analysis of final design: Stall characteristics Stall margin inexistent Outboard stall 26

27 Analysis of final design: Sensitivity to contamination Clean characteristics Wing Design Contaminated characteristics CLopt CL ini Optimization formulation: Min CL clean-dto s.t CL max CL max_ini w.r.t 2D Leading edge Results: CL clean-dto reduced by 29% Conclusion: There exist profiles that yield the same CL max, with less sensitivity to contamination 27

28 Analysis of final design: Aeroelastic effect on drag Initial Mid Final Drag Rigid Drag elastic Jig 0.5% Therefore, jig twist design allows proper aeroelastic twists at any condition, thus optimal off-design performance DESIGN POINT Initial cruise Mid Cruise Final Cruise Drag under estimated Drag over estimated 28

29 Analysis of final design: Aeroelastic effect on structural weight Wing deformed 2.5g C L Cl*c/Cav Structural Weight (lbs) Weight Rigid 2.5% over estimated Weight Elastic 29

30 Conclusion An industrial multi-fidelity MDO framework was presented Allows to compute the best compromise between High-speed Aerodynamics Low-speed Aerodynamics Structure Material choice A DLR F-6 was sized and composite wingbox shown to be: 20% lighter compared to metallic wingbox MDO results were presented on a generic business jet: Composite structures allows lighter designs better Aerodynamics characteristics Manufacturing constraints were ignored for composites, which gave them an unfair advantage Blending on stacking sequence typically increases weight 30

31 Final thoughts The design obtained from MDO process is not final The framework has some inherent assumptions and simplifications that force the engineer to fine tune the optimized design Nonetheless, the simplifications do not invalidate the design MDO does not replace the job of experienced engineers Future work: Improve the fidelity of the CFD code Introduce stall progression characteristics (slat-less and slat designs) Introduce icing contamination Introduce manufacturing constraints for composite design 31

32 Questions? Acknowledgements: Cedric Kho Awot Berhe Temesgen Mengistu 32