Walt Silva Aeroelasticity Branch NASA Langley Research Center October 14 2020 Aerospace Flutter and Dynamics Council Outline Benchmark SuperCritical Wing BSCW Aeroelastic Prediction Workshop ID: 1047190
Download Presentation The PPT/PDF document "BSCW Computations Using KESTREL and AERO..." is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
1. BSCW Computations Using KESTREL and AEROMWalt SilvaAeroelasticity BranchNASA Langley Research CenterOctober 14, 2020Aerospace Flutter and Dynamics Council
2. OutlineBenchmark SuperCritical Wing (BSCW), Aeroelastic Prediction Workshop (AePW)BackgroundAEROMResultsAlpha=0, 1, 3, and 5 degs (skip alpha=1 deg; similar to alpha=0 deg)Amplitudes = 0.04, 0.01, 0.08, and 0.001 (modal)Comparison of full solution GAFs vs ROM GAFs (error quantification)Root locusComparison of full solution transients vs ROM transients (at a specified Q)
3. Benchmark Supercritical Wing (BSCW)3Chosen as a challenging test case, flow-wise, but simple geometryStrong shock with suspected shock-induced separated flowSome preliminary assessments from AePW Computational methods had difficulty producing converged solutions due to flow field complexityComplex flow field also observed in experimental data; Largest magnitude of dynamic behavior appears to represent shock oscillationsCFD solutions vary widely, even for static solution; Likely plan of action: Form technical working group of BSCW analystsExtensive study of available experimental data; characterize different flow phenomenaBenchmark against more benign cases- lower Mach number, lower angle of attackAnalyze the static (unforced) problem using time-accurate evaluation methodsStudy of time convergence criteriaM=0.85, Rec=4.49 million, test medium: R-134a, α = 5°, θ = 1°, freq 1 & 10 HzFrom AePW-1 Debrief
4. AePW-2 Summary of ResultsAt Mach 0.74, a = 0° Small variation in computational results Match the experimental results well Match the linear results wellAt Mach 0.85, a = 5°Significant variation in predictionsNo experimental data for comparison4
5. BackgroundApplication of KESTREL and KESTREL/AEROM to BSCW wing At M=0.74, alpha=0 deg, KESTREL, KESTREL/AEROM, FUN3D, and FUN3D/AEROM all compare nearly exactly with dataAt M=0.85, alpha=5 deg, large flow separation/unsteadiness; very challenging problemLarge variance from one code to another at this condition; differences between KESTREL and FUN3DGoal of this analysis: evaluate application of AEROM using KESTREL at M=0.80 and alpha=0, 1, 3, and 5 deg (similar unsteadiness to M=0.85) using multiple amplitude excitations (Walsh functions)Follow-up with investigation of code-to-code comparisons”Characterizing Aerodynamic Damping of a Supersonic Missile with CFD”, Shelton (AF), Martin (AF), Silva (NASA), SciTech 2018
6. 6System Identification for Multiple Input-Output (cont)System/Observer/Controller Identification Toolbox(SOCIT)Computational Fluid Dynamics(Kestrel or FUN3D)
7. 7Pitch Moment DampingWalsh Functions vs Harmonic Motion and Reference
8. AEROM: ROM Development ProcessUnsteady Aerodynamic State-Space ROMWalsh Functions(modal)Structural state-space ROM8KESTREL1234
9. BSCW Application (KESTREL), Two modesKESTREL BSCW*.fmh file*.modes fileTwo Walsh input functionsTwo GAFs output functions
10. AEROM: ROM Development ProcessError Minimization (Unsteady Aerodynamics)LowError ?YesNoPULSEERAContinue10
11. AEROM: ROM Development ProcessAeroelastic Simulation ROM (Simulink)11Simulink model that includesUnsteady Aero ROM and Structural Dynamic ROM usedto compare ROM responses at anyDynamic pressure with full KESTRELSolutions.
12. AEROM: ROM Development ProcessAeroelastic Root Locus Plot Using Aeroelastic Simulation ROMModeled as a closed-loopsystem with gain (dynamicpressure) in MATLABEigenvalues Z-PlaneRoot Locus S-PlaneRoot LocusFrequency vsDamping Ratio Root Locus12
13. Walsh Function Amplitudes and AlphasModal Amp/Alpha (deg)0 (K)13 (K)5 (K)0.04M8A0a M8A1aM8A3a M8A5a 0.01M8A0bM8A1bM8A3bM8A5b0.08M8A0cM8A1cM8A3cM8A5c0.001M8A0dM8A1dM8A3dM8A5dK = comparison with full KESTREL solution available or underway; using ‘Venkat’ limiterKESTREL Solutions: Alpha = 0 deg, Q = 100 psf, Initial Modal Velocity = 5 Alpha = 3 deg, Q = 100 psf, Initial Modal Velocity = 5 Alpha = 5 deg, Q = 50 psf, 130 psf, Initial Modal Velocity = 5CPU Cost (NAS)ROM solution: 10K steps, 240 cores, 9hrs 45min (set up for high alpha case; shorter duration for benign cases)Full Solution per Q: 15K steps, 240 cores, 53 hrs 04min (~5.5 cycles)
14. Alpha = 0 deg
15. M8A0a – ROM/KESTREL GAFsa.04b.01c.08d.001
16. M8A0a – ROM/KESTREL GAFs Error
17. M8A0a – ROM/KESTREL Root LocusZoom
18. M8A0a – ROM Responses (SIMULINK)Q = 0.9 psiQ = 1.3 psi
19. M8A0a – ROM/KESTREL Responses, Q=100 psfMode 1Mode 2
20. M8A0a – ROM/KESTREL Responses, Q=100 psfMode 1Mode 2
21. M8A0b – ROM/KESTREL GAFsa.04b.01c.08d.001
22. M8A0b – ROM/KESTREL Root LocusZoom
23. M8A0b – ROM Responses (SIMULINK)Q = 1.3 psiQ = 1.2 psi
24. M8A0c – ROM/KESTREL GAFsa.04b.01c.08d.001
25. M8A0c – ROM/KESTREL GAFs Root Locus
26. M8A0c – ROM Responses (SIMULINK)Q = 1.2 psiQ = 1.3 psi
27. M8A0d – ROM/KESTREL GAFsa.04b.01c.08d.001
28. M8A0d – ROM/KESTREL Root Locus
29. M8A0d – ROM Responses (SIMULINK)Q = 1.2 psiQ = 1.3 psi
30. Alpha = 1 deg
31. M8A1a – ROM/KESTREL GAFs
32. M8A1a – ROM/KESTREL Root Locus
33. M8A1a – ROM ResponsesQ = 1.2 psiQ = 1.3 psi
34. M8A1b – ROM/KESTREL GAFs
35. M8A1b – ROM/KESTREL Root Locus
36. M8A1b – ROM ResponsesQ = 1.2 psiQ = 1.3 psi
37. M8A1c – ROM/KESTREL GAFs
38. M8A1c – ROM/KESTREL Root Locus
39. M8A1c – ROM ResponsesQ = 1.2 psi
40. M8A1d – ROM/KESTREL GAFs
41. M8A1d – ROM/KESTREL Root Locus
42. M8A1d – ROM ResponsesQ = 1.2 psiQ = 1.3 psi
43. Alpha = 3 deg
44. M8A3a – ROM/KESTREL GAFs
45. M8A3a – ROM/KESTREL Root Locus
46. M8A3a – ROM ResponsesQ = 1.2 psiQ = 1.3 psi
47. M8A3b – ROM/KESTREL GAFs
48. M8A3b – ROM/KESTREL Root Locus
49. M8A3b – ROM ResponsesQ = 1.2 psiQ = 1.3 psi
50. M8A3b – ROM ResponsesQ = 1.5 psi
51. M8A3c – ROM/KESTREL GAFs
52. M8A3c – ROM/KESTREL Root Locus
53. M8A3c – ROM ResponsesQ = 1.2 psiQ = 1.3 psi
54. M8A3d – ROM/KESTREL GAFs
55. M8A3d – ROM/KESTREL Root Locus
56. M8A3d – ROM ResponsesQ = 1.2 psiQ = 1.3 psi
57. M8A3d – ROM ResponsesQ = 1.5 psi
58. Alpha = 5 deg
59. M8A5a – ROM/KESTREL GAFs
60. M8A5a – ROM/KESTREL Root Locus
61. M8A5a – ROM ResponsesQ = 1.2 psiQ = 1.3 psi
62. M8A5a – ROM and KESTREL ResponsesQ = 50 psf, Mode 1TimeFrequency
63. M8A5a – ROM and KESTREL ResponsesQ = 50 psf, Mode 2TimeFrequency
64. M8A5a – ROM and KESTREL ResponsesQ = 130 psf, Mode 1TimeFrequency
65. M8A5a – ROM and KESTREL ResponsesQ = 130 psf, Mode 2TimeFrequency
66. M8A5b – ROM/KESTREL GAFs
67. M8A5b – ROM/KESTREL Root Locus
68. M8A5b – ROM ResponsesQ = 1.2 psi
69. M8A5c – ROM/KESTREL GAFs
70. M8A5c – ROM/KESTREL Root Locus
71. M8A5c – ROM ResponsesQ = 1.2 psiQ = 1.3 psi
72. M8A5c – ROM ResponsesQ = 1.5 psi
73. M8A5d – ROM/KESTREL GAFs
74. M8A5d – ROM/KESTREL GAFs
75. M8A5d – ROM/KESTREL Root Locus
76. M8A5d – ROM ResponsesQ = 1.2 psi
77. ROM/KESTRELRoot Locus Plot ComparisonM=0.80, alpha=0, 5 deg
78. Alpha=0 deg Alpha=5 deg
79. Conclusions (Thus Far)At benign conditions (alpha=0 deg), all ROMs provide same result; gratifying, as it indicates linearized flow dynamicsAs alpha is increased, variations between the different ROMs (amplitudes) start to become obviousAt alpha=5 deg, there are noticeable differences between the ROMs, indicating that the higher-amplitude ROMs do a better job (at this condition) of predicting the aeroelastic responses at multiple dynamic pressuresSo far, ROM Flutter Q=179 psf; KESTREL Flutter Q=168 psf (~6.5%)Will start analyses at M=0.85, similar alpha ranges.