Hostname: page-component-5c6d5d7d68-vt8vv Total loading time: 0.001 Render date: 2024-08-15T21:57:03.492Z Has data issue: false hasContentIssue false
  • English
  • Français

Discovering optimal flapping wing kinematics using active deep learning

Published online by Cambridge University Press:  07 November 2023

Baptiste Corban Open the ORCID record for Baptiste Corban [Opens in a new window] ,
Michael Bauerheim  and
Thierry Jardin Open the ORCID record for Thierry Jardin [Opens in a new window]
Baptiste Corban*
Affiliation:
ISAE-SUPAERO, Université de Toulouse, 31055 Toulouse CEDEX 4, France
Michael Bauerheim
Affiliation:
ISAE-SUPAERO, Université de Toulouse, 31055 Toulouse CEDEX 4, France
Thierry Jardin
Affiliation:
ISAE-SUPAERO, Université de Toulouse, 31055 Toulouse CEDEX 4, France
*
†Email address for correspondence: baptiste.corban@gmail.com
Get access Rights & Permissions [Opens in a new window]

Abstract

This paper focuses on the discovery of optimal flapping wing kinematics using a deep learning surrogate model for unsteady aerodynamics and multi-objective optimisation. First, a surrogate model of the unsteady forces experienced by a 3-D flapping wing is built, based on deep neural networks. The model is trained on a dataset of randomly generated kinematics simulated using direct numerical simulation (DNS). Once trained, the neural networks can quickly predict the unsteady lift and torques experienced by the wing, using sparse information on the kinematics. This fast surrogate model allows multi-objective optimisation to be performed. The resulting Pareto front consists of new kinematics that may be very different from the kinematics of the initial dataset. A few arbitrarily chosen kinematics on the Pareto front are thus simulated using DNS and used to enhance the database. The new dataset is used to train again the networks, and this active deep learning/optimisation framework is performed until convergence, obtained after only two iterations. Overall, this method reduced the cost of optimisation by 83 %. Results reveal two distinct families of motions. Kinematics promoting high efficiency are characterised by large stroke amplitudes and relatively low angles of attack, as observed for fruit flies, honeybees or hawkmoths. For those, lift production is driven by quasi-steady effects and the formation of a stable leading edge vortex. Kinematics promoting high lift are characterised by small stroke amplitudes and high angles of attack, reminiscent of mosquitoes. Lift production is driven by the rapid generation of vorticity at the trailing edge.

JFM classification

Mathematical Foundations: Machine learning Vortex Flows: Vortex dynamics Biological Fluid Dynamics: Swimming/flying
Type
JFM Papers
Information
Journal of Fluid Mechanics , Volume 974 , 10 November 2023 , A54
Check for updates
Copyright
© The Author(s), 2023. Published by Cambridge University Press

Access options

Get access to the full version of this content by using one of the access options below. (Log in options will check for institutional or personal access. Content may require purchase if you do not have access.)

References

Ajuria-Illaramendi, E., Bauerheim, M. & Cuenot, B. 2022 Performance and accuracy assessments of an incompressible fluid solver coupled with a deep convolutional neural network. Data-Centric Engng 3, e2.CrossRefGoogle Scholar
Alguacil, A., Bauerheim, M., Jacob, M.C. & Moreau, S. 2021 Predicting the propagation of acoustic waves using deep convolutional neural networks. J. Sound Vib. 512, 116285.CrossRefGoogle Scholar
Alguacil, A., Bauerheim, M., Jacob, M.C. & Moreau, S. 2022 Deep learning surrogate for the temporal propagation and scattering of acoustic waves. AIAA J. 60 (10), 5890–5906.CrossRefGoogle Scholar
Altshuler, D.L., Dickson, W.B., Vance, J.T., Roberts, S. & Dickinson, M.H. 2005 Short-amplitude high-frequency wing strokes determine the aerodynamics of honeybee flight. Proc. Natl Acad. Sci. 102 (50), 18213–18218.CrossRefGoogle ScholarPubMed
Aono, H., Liang, F. & Liu, H. 2008 Near-and far-field aerodynamics in insect hovering flight: an integrated computational study. J. Expl Biol. 211 (2), 239–257.CrossRefGoogle ScholarPubMed
Baqué, P., Remelli, E., Fleuret, F. & Fua, P. 2018 Geodesic convolutional shape optimization. In Proceedings of the 35th International Conference on Machine Learning, vol. 80, pp. 472–481. PMLR.Google Scholar
Bayiz, Y.E. & Cheng, B. 2021 State-space aerodynamic model reveals high force control authority and predictability in flapping flight. J. R. Soc. Interface 18 (181), 20210222.CrossRefGoogle ScholarPubMed
Bhat, S.S., Zhao, J., Sheridan, J., Hourigan, K. & Thompson, M.C. 2020 Effects of flapping-motion profiles on insect-wing aerodynamics. J. Fluid Mech. 884, A8.CrossRefGoogle Scholar
Blank, J. & Deb, K. 2020 pymoo: multi-objective optimization in Python. IEEE Access 8, 89497–89509.CrossRefGoogle Scholar
Bomphrey, R., Nakata, T., Phillips, N., et al. 2017 Smart wing rotation and trailing-edge vortices enable high frequency mosquito flight. Nature 544, 92–95.CrossRefGoogle ScholarPubMed
Bos, F.M., Lentink, D., Van Oudheusden, B.W. & Bijl, H. 2008 Influence of wing kinematics on aerodynamic performance in hovering insect flight. J. Fluid Mech. 594, 341–368.CrossRefGoogle Scholar
Bury, Y. & Jardin, T. 2012 Transitions to chaos in the wake of an axisymmetric bluff body. Phys. Lett. A 376 (45), 3219–3222.CrossRefGoogle Scholar
Busch, C., Gehrke, A. & Mulleners, K. 2012 On the parametrisation of motion kinematics for experimental aerodynamic optimisation. Exp. Fluids 63, 10.CrossRefGoogle Scholar
Cai, X., Kolomenskiy, D., Nakata, T. & Liu, H. 2021 A CFD data-driven aerodynamic model for fast and precise prediction of flapping aerodynamics in various flight velocities. J. Fluid Mech. 915, A114.CrossRefGoogle Scholar
Calado, A., Poletti, R., Koloszar, L. & Mendez, M.A. 2023 A robust data-driven model for flapping aerodynamics under different hovering kinematics. Phys. Fluids 35 (4), 047122.CrossRefGoogle Scholar
Campet, R., Roy, P.T., Cuenot, B., Riber, E. & Jouhaud, J.-C. 2020 Design optimization of an heat exchanger using Gaussian process. Intl J. Heat Mass Transfer 150, 119264.CrossRefGoogle Scholar
Colombo, M., Bauerheim, M. & Morlier, J. 2023 Prediction of gust aeroelastic dynamics of HALE using graph neural networks. AIAA SCITECH 2023 Forum, 2570.Google Scholar
Deb, K., Pratap, A., Agarwal, S. & Meyarivan, T. 2002 A fast and elitist multiobjective genetic algorithm: NSGA-II. IEEE Trans. Evol. Comput. 6 (2), 182–197.CrossRefGoogle Scholar
Diaz-Arriba, D., Jardin, T., Gourdain, N., Pons, F. & David, L. 2021 Numerical investigation of three-dimensional asymmetric hovering flapping flight. Phys. Fluids 33 (11), 111907.CrossRefGoogle Scholar
Dickinson, M.H. & Götz, K.G. 1993 Unsteady aerodynamic performance of model wings at low Reynolds numbers. J. Expl Biol. 174 (1), 45–64.CrossRefGoogle Scholar
Dickinson, M.H., Lehmann, F.O. & Sane, S.P. 1999 Wing rotation and the aerodynamic basis of insect flight. Science 284 (5422), 1954–1960.CrossRefGoogle ScholarPubMed
Eldredge, J.D. & Jones, A.R. 2019 Leading-edge vortices: mechanics and modeling. Annu. Rev. Fluid Mech. 51, 75–104.CrossRefGoogle Scholar
Ellington, C.P. 1984 The aerodynamics of hovering insect flight. Phil. Trans. R. Soc. Lond. 305 (1122), 1–181.Google Scholar
Ellington, C.P., Van Den Berg, C., Willmott, A.P. & Thomas, A.L. 1996 Leading-edge vortices in insect flight. Nature 384 (6610), 626–630.CrossRefGoogle Scholar
Gehrke, A. & Mulleners, K. 2021 Phenomenology and scaling of optimal flapping wing kinematics. Bioinspir. Biomim. 16 (2), 026016.CrossRefGoogle ScholarPubMed
Gupta, R. & Jaiman, R. 2022 Three-dimensional deep learning-based reduced order model for unsteady flow dynamics with variable Reynolds number. Phys. Fluids 34, 033612.CrossRefGoogle Scholar
Hochreiter, S. & Schmidhuber, J. 1997 Long short-term memory. Neural Comput. 9 (8), 1735–1780.CrossRefGoogle ScholarPubMed
Jardin, T. & David, L. 2017 Root cutout effects on the aerodynamics of a low-aspect-ratio revolving wing. AIAA J. 55 (8), 2717–2726.CrossRefGoogle Scholar
Jardin, T., David, L. & Farcy, A. 2009 Characterization of vortical structures and loads based on time-resolved PIV for asymmetric hovering flapping flight. Exp. Fluids 46, 847–857.CrossRefGoogle Scholar
Jardin, T. & Doué, N. 2019 Influence of pitch rate on freely translating perching airfoils. J. Fluid Mech. 873, 49–71.CrossRefGoogle Scholar
Jardin, T., Farcy, A. & David, L. 2012 Three-dimensional effects in hovering flapping flight. J. Fluid Mech. 702, 102–125.CrossRefGoogle Scholar
Krügener, M., Zapata Usandivaras, J., Bauerheim, M. & Urbano, A. 2022 Coaxial-injector surrogate modeling based on Reynolds-averaged Navier–Stokes simulations using deep learning. J. Propul. Power 38, 1–16.CrossRefGoogle Scholar
Lee, Y.J. & Lua, K.B. 2018 Optimization of simple and complex pitching motions for flapping wings in hover. AIAA J. 56 (6), 2466–2470.CrossRefGoogle Scholar
Lee, Y.J., Lua, K.B., Lim, T.T. & Yeo, K.S. 2016 A quasi-steady aerodynamic model for flapping flight with improved adaptability. Bioinspir. Biomim. 11 (3), 036005.CrossRefGoogle ScholarPubMed
Lentink, D. & Dickinson, M.H. 2009 Rotational accelerations stabilize leading edge vortices on revolving fly wings. J. Expl Biol. 212 (16), 2705–2719.CrossRefGoogle ScholarPubMed
Li, J., Du, X. & Martins, J.R.R.A. 2022 Machine learning in aerodynamic shape optimization. Prog. Aerosp. Sci. 134, 100849.CrossRefGoogle Scholar
Lighthill, M. 1973 On the Weis-Fogh mechanism of lift generation. J. Fluid Mech. 60 (1), 1–17.CrossRefGoogle Scholar
Liu, H. & Aono, H. 2009 Size effects on insect hovering aerodynamics: an integrated computational study. Bioinspir. Biomim. 4 (1), 015002.CrossRefGoogle ScholarPubMed
Liu, L., Du, G. & Sun, M. 2020 Aerodynamic-force production mechanisms in hovering mosquitoes. J. Fluid Mech. 898, A19.CrossRefGoogle Scholar
Lye, K., Mishra, S., Ray, D. & Chandrashekar, P. 2021 Iterative surrogate model optimization (ISMO): an active learning algorithm for PDE constrained optimization with deep neural networks. Comput. Meth. Appl. Mech. Engng 374, 113575.CrossRefGoogle Scholar
Medina, A. & Jones, A.R. 2016 Leading-edge vortex burst on a low-aspect-ratio rotating flat plate. Phys. Rev. Fluids 1 (4), 044501.CrossRefGoogle Scholar
Nakata, T., Liu, H. & Bomphrey, R.J. 2015 A CFD-informed quasi-steady model of flapping-wing aerodynamics. J. Fluid Mech. 783, 323–343.CrossRefGoogle ScholarPubMed
Pandi, A., et al. 2022 A versatile active learning workflow for optimization of genetic and metabolic networks. Nat. Commun. 13, 3876.CrossRefGoogle ScholarPubMed
Paszke, A., et al. 2019 PyTorch: an imperative style, high-performance deep learning library. In Advances in Neural Information Processing Systems 32, pp. 8024–8035. Curran Associates, Inc.Google Scholar
Roy, P.T., Segui, L.M., Jouhaud, J.-C. & Gicquel, L. 2018 Resampling strategies to improve surrogate model-based uncertainty quantification: application to LES of LS89. Intl J. Numer. Meth. Fluids 87 (12), 607–627.CrossRefGoogle Scholar
Sane, S.P. & Dickinson, M.H. 2001 The control of flight force by a flapping wing: lift and drag production. J. Expl Biol. 204 (15), 2607–2626.CrossRefGoogle ScholarPubMed
Shyy, W., Aono, H., Chimakurthi, S.K., Trizila, P., Kang, C.-K., Cesnik, C.E.S. & Liu, H. 2010 Recent progress in flapping wing aerodynamics and aeroelasticity. Prog. Aerosp. Sci. 46 (7), 284–327.CrossRefGoogle Scholar
Sorteberg, W.E., Garasto, S., Pouplin, A.S., Cantwell, C.D. & Bharath, A.A. 2018 Approximating the solution to wave propagation using deep neural networks. In 32nd Conference on Neural Information Processing Systems - NeurIPS, arXiv:1812.01609. Springer.CrossRefGoogle Scholar
Sun, M. & Tang, J. 2002 Unsteady aerodynamic force generation by a model fruit fly wing in flapping motion. J. Expl Biol. 205 (1), 55–70.CrossRefGoogle Scholar
Tompson, J., Schlachter, K., Sprechmann, P. & Perlin, K. 2017 Accelerating Eulerian fluid simulation with convolutional networks. In Proceedings of the 34th International Conference on Machine Learning, in Proceedings of Machine Learning Research, vol. 70, pp. 3424–3433.Google Scholar
Van Veen, W.G., Van Leeuwen, J.L., Van Oudheusden, B.W. & Muijres, F.T. 2022 The unsteady aerodynamics of insect wings with rotational stroke accelerations, a systematic numerical study. J. Fluid Mech. 936, A3.CrossRefGoogle Scholar
Wang, Z.J., Birch, J.M. & Dickinson, M.H. 2004 Unsteady forces and flows in low Reynolds number hovering flight: two-dimensional computations versus robotic wing experiments. J. Expl Biol. 207, 449–460.CrossRefGoogle Scholar
Wang, A., Liang, H., McDannald, A., Takeuchi, I. & Kusne, A.G. 2022 Benchmarking active learning strategies for materials optimization and discovery. Oxford Open Mater. Sci. 2 (1), itac006.CrossRefGoogle Scholar
Wang, P., Wang, Y., Liu, T. & Zhang, D. 2022 Active learning-based multi-objective optimization for aerodynamic performance of a supercritical carbon dioxide turbine. Struct. Multidiscipl. Optim. 65, 267.CrossRefGoogle Scholar
Wu, J.C. 1981 Theory for aerodynamic force and moment in viscous flows. AIAA J. 19 (4), 432–441.CrossRefGoogle Scholar
Wu, J.H. & Sun, M. 2004 Unsteady aerodynamic forces of a flapping wing. J. Expl Biol. 207 (7), 1137–1150.CrossRefGoogle ScholarPubMed
Zheng, L., Hedrick, T.L. & Mittal, R. 2013 A multi-fidelity modelling approach for evaluation and optimization of wing stroke aerodynamics in flapping flight. J. Fluid Mech. 721, 118–154.CrossRefGoogle Scholar
Zheng, H., Xie, F., Ji, T. & Zheng, Y. 2020 Kinematic parameter optimization of a flapping ellipsoid wing based on the data-informed self-adaptive quasi-steady model. Phys. Fluids 32 (4), 041904.CrossRefGoogle Scholar
Zhu, H.J. & Sun, M. 2017 Unsteady aerodynamic force mechanisms of a hoverfly hovering with a short stroke-amplitude. Phys. Fluids 29, 081901.CrossRefGoogle Scholar

Corban et al. Supplementary Movie 1

Q-criterion isosurfaces for the most efficient flapping motion

Download Corban et al. Supplementary Movie 1(Video)
Video 7 MB

Corban et al. Supplementary Movie 2

Q-criterion isosurfaces for the highest-lift-generating flapping motion

Download Corban et al. Supplementary Movie 2(Video)
Video 5.5 MB