Hostname: page-component-857557d7f7-bvshn Total loading time: 0 Render date: 2025-12-08T21:38:00.888Z Has data issue: false hasContentIssue false
  • English
  • Français

Physics of the vortex gust–airfoil interaction under an optimal mitigation strategy learned through deep reinforcement learning

Published online by Cambridge University Press:  01 December 2025

Brice Martin Open the ORCID record for Brice Martin [Opens in a new window] ,
Thierry Jardin Open the ORCID record for Thierry Jardin [Opens in a new window] ,
Emmanuel Rachelson  and
Michaël Bauerheim Open the ORCID record for Michaël Bauerheim [Opens in a new window]
Brice Martin*
Affiliation:
ISAE-SUPAERO , 10 Av. Marc Pélégrin, Toulouse 31400, France
Thierry Jardin
Affiliation:
ISAE-SUPAERO , 10 Av. Marc Pélégrin, Toulouse 31400, France
Emmanuel Rachelson
Affiliation:
ISAE-SUPAERO , 10 Av. Marc Pélégrin, Toulouse 31400, France
Michaël Bauerheim
Affiliation:
ISAE-SUPAERO , 10 Av. Marc Pélégrin, Toulouse 31400, France
*
Corresponding author: Brice Martin, brice.martin@isae-supaero.fr
Get access Rights & Permissions [Opens in a new window]

Abstract

This paper aims to elucidate the physical mechanisms underlying airfoil–vortex gust interaction and mitigation. The vortex gust mitigation problem consists in finding the pitch rate sequence that minimises the gust-induced lift disturbance of an NACA0012 airfoil at Reynolds number 1000. The instantaneous flow fields and resulting lift are obtained from numerical resolution of the Navier–Stokes equations. The controller is modelled as an artificial neural network and trained to minimise the lift fluctuation using deep reinforcement learning (DRL). The paper shows that DRL-trained controllers are able to mitigate medium- and high-intensity vortex gusts by more than 80 % compared to the uncontrolled scenario. It then presents a comparative analysis of the controlled and uncontrolled lift generation mechanisms using the force partitioning method (FPM). The FPM provides a quantitative assessment of the amount of lift generated by each flow region. For medium-intensity gusts, the main phenomenon is the asymmetry in the airfoil boundary layer induced by the vortex. The control strategy mitigates the gust-induced lift by restoring the flow symmetry around the airfoil. For high-intensity gusts, the boundary layer asymmetry remains, but the gust interaction with the airfoil also triggers flow separation and the formation of a strong leading-edge vortex (LEV). Consequently, the control command balances several aerodynamic phenomena such as boundary layer asymmetry, flow detachment, LEV, and secondary recirculation regions to produce a net quasi-zero lift fluctuation. Thus this work highlights the potential of DRL control, enhanced by advanced post-processing such as FPM, to discover and interpret optimal flow control mechanisms.

JFM classification

Mathematical Foundations: Machine learning Vortex Flows: Vortex interactions

Information

Type
JFM Papers
Information
Journal of Fluid Mechanics , Volume 1024 , 10 December 2025 , A18
Check for updates
Copyright
© The Author(s), 2025. 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.)

Article purchase

Temporarily unavailable

References

Andreu-Angulo, I. & Babinsky, H. 2020 Negating gust effects by actively pitching a wing. In AIAA Scitech 2020 Forum. American Institute of Aeronautics and Astronautics.10.2514/6.2020-1057CrossRefGoogle Scholar
Beck, A. & Kurz, M. 2021 A perspective on machine learning methods in turbulence modeling. GAMM-Mitteilungen 44 (1), e202100002.10.1002/gamm.202100002CrossRefGoogle Scholar
Beckers, D. & Eldredge, J.D. 2024 Deep reinforcement learning of airfoil pitch control in a highly disturbed environment using partial observations. Phys. Rev. Fluids 9 (9), 093902.10.1103/PhysRevFluids.9.093902CrossRefGoogle Scholar
Beintema, G., Corbetta, A., Biferale, L. & Toschi, F. 2020 Controlling Rayleigh–Bénard convection via reinforcement learning. J. Turbul. 21 (9–10), 585–605.10.1080/14685248.2020.1797059CrossRefGoogle Scholar
Berger, S., Arroyo Ramo, A., Guillet, V., Lahire, T., Martin, B., Jardin, T., Rachelson, E. &Bauerheim, M. 2024 Reliability assessment of off-policy deep reinforcement learning: a benchmark for aerodynamics. Data-Centric Engng 5, e2.10.1017/dce.2023.28CrossRefGoogle Scholar
Bhagwat, M. & Leishman, G. 2002 Generalized viscous vortex model for application to free-vortex wake and aeroacoustic calculations. In 58th Annual Forum and Technology Display of the American Helicopter Society International. American Helicopter Society, Inc.Google Scholar
Biler, H., Sedky, G., Jones, A.R., Saritas, M. & Cetiner, O. 2021 Experimental investigation of transverse and vortex gust encounters at low Reynolds numbers. AIAA J. 59 (3), 786–799.10.2514/1.J059658CrossRefGoogle Scholar
Choi, J., Colonius, T. & Williams, D.R. 2015 Surging and plunging oscillations of an airfoil at low Reynolds number. J. Fluid Mech. 763, 237–253.10.1017/jfm.2014.674CrossRefGoogle Scholar
Corkery, S.J., Babinsky, H. & Harvey, J.K. 2018 On the development and early observations from a towing tank-based transverse wing–gust encounter test rig. Exp. Fluids 59, 1–16.10.1007/s00348-018-2586-0CrossRefGoogle Scholar
Fujimoto, S., Hoof, H. & Meger, D. 2018 Addressing function approximation error in actor–critic methods. In International Conference on Machine Learning, pp. 1587–1596. PMLR.Google Scholar
Fukami, K., Nakao, H. & Taira, K. 2024 Data-driven transient lift attenuation for extreme vortex gust–airfoil interactions. J. Fluid Mech. 992, A17.10.1017/jfm.2024.592CrossRefGoogle Scholar
Fukami, K. & Taira, K. 2023 Grasping extreme aerodynamics on a low-dimensional manifold. Nat. Commun. 14, 6480.10.1038/s41467-023-42213-6CrossRefGoogle ScholarPubMed
Goodfellow, I.J., Bengio, Y. & Courville, A. 2016 Deep Learning. MIT Press.Google Scholar
Guastoni, L., Rabault, J., Schlatter, P., Azizpour, H. & Vinuesa, R. 2023 Deep reinforcement learning for turbulent drag reduction in channel flows. Eur. Phys. J. E 46, 27 10.1140/epje/s10189-023-00285-8CrossRefGoogle ScholarPubMed
Gunnarson, P., Mandralis, I., Novati, G., Koumoutsakos, P. & Dabiri, J.O. 2021 Learning efficient navigation in vortical flow fields. Nat. Commun. 12 (1), 7143.10.1038/s41467-021-27015-yCrossRefGoogle ScholarPubMed
Helmbold, H.B. 1942 Der unverwundene Ellipsenflügel als tragende Fläche. Jahrbuch I111–I113.Google Scholar
Herrmann, B., Brunton, S.L., Pohl, J.E. & Semaan, R. 2022 Gust mitigation through closed-loop control. II. Feedforward and feedback control. Phy. Rev. Fluids 7, 024706.10.1103/PhysRevFluids.7.024706CrossRefGoogle Scholar
Jardin, T. 2017 Coriolis effect and the attachment of the leading edge vortex. J. Fluid Mech. 820, 312–340.10.1017/jfm.2017.222CrossRefGoogle Scholar
Jardin, T. & Doué, N. 2019 Influence of pitch rate on freely translating perching airfoils. J. Fluid Mech. 873, 49–71.10.1017/jfm.2019.421CrossRefGoogle Scholar
Jardin, T., Farcy, A. & David, L. 2012 Three-dimensional effects in hovering flapping flight. J. Fluid Mech. 702, 102–125.10.1017/jfm.2012.163CrossRefGoogle Scholar
Jardin, T., Ferrand, V. & Gowree, E.R. 2025 NACA0012 airfoil at Reynolds numbers between 50 000 and 140 000 – Part 1: Steady freestream. Intl J. Heat Fluid Flow 112, 109655.10.1016/j.ijheatfluidflow.2024.109655CrossRefGoogle Scholar
Jones, A.R. & Cetiner, O. 2021 Introduction to the virtual collection: unsteady aerodynamic response of rigid wings in large-amplitude gust encounters. AIAA J. 59 (3), 764–764.10.2514/1.J060379CrossRefGoogle Scholar
Jones, A.R., Cetiner, O. & Smith, M.J. 2022 Physics and modeling of large flow disturbances: discrete gust encounters for modern air vehicles. Annu. Rev. Fluid Mech. 54 (1), 469–493.10.1146/annurev-fluid-031621-085520CrossRefGoogle Scholar
Kaufmann, W. 1962 Über die ausbreitung kreiszylindrischer wirbel in zähen (viskosen) Flüssigkeiten. Ingenieur-Archiv 31 (1), 1–9.10.1007/BF00538235CrossRefGoogle Scholar
Kazarin, P., Golubev, V., MacKunis, W. & Moreno, C. 2021 Robust nonlinear tracking control for unmanned aircraft in the presence of wake vortex. Electronics 10 (16), 1890.10.3390/electronics10161890CrossRefGoogle Scholar
Kurz, M., Offenhäuser, P. & Beck, A. 2023 Deep reinforcement learning for turbulence modeling in large eddy simulations. Intl J. Heat Fluid Flow 99, 109094.10.1016/j.ijheatfluidflow.2022.109094CrossRefGoogle Scholar
Küssner, H.G. 1932 Stresses produced in airplane wings by gusts. Tech. Rep. TR 645. National Advisory Committee for Aeronautics (NACA), Washington, DC.Google Scholar
LeCun, Y., Boser, B., Denker, J.S., Henderson, D., Howard, R.E., Hubbard, W. & Jackel, L.D. 1989 Backpropagation applied to handwritten zip code recognition. Neural Comput. 1 (4), 541–551.10.1162/neco.1989.1.4.541CrossRefGoogle Scholar
Liu, Z., Beckers, D. & Eldredge, J.D. 2025 Model-based reinforcement learning for control of strongly-disturbed unsteady aerodynamic flows. AIAA J. 63, 4105–4125.10.2514/1.J064790CrossRefGoogle Scholar
Liu, Z. & Eldredge, J.D. 2025 Attention on flow control: transformer-based reinforcement learning for lift regulation in highly disturbed flows. arXiv:2506.10153.Google Scholar
Martin, B., Jardin, T., Rachelson, E. & Bauerheim, M. 2024 Vortex gust mitigation from onboard measurements using deep reinforcement learning. Data-Centric Engng 5, e47.10.1017/dce.2024.38CrossRefGoogle Scholar
Menon, K. & Mittal, R. 2021 Significance of the strain-dominated region around a vortex on induced aerodynamic loads. J. Fluid Mech. 918, R3.10.1017/jfm.2021.359CrossRefGoogle Scholar
Mnih, V. et al. 2015 Human-level control through deep reinforcement learning. Nature 518 (7540), 529–533.10.1038/nature14236CrossRefGoogle ScholarPubMed
Mousavi, H. & Eldredge, J.D. 2025 Low-order flow reconstruction and uncertainty quantification in disturbed aerodynamics using sparse pressure measurements. J. Fluid Mech. 1013, A41.10.1017/jfm.2025.10253CrossRefGoogle Scholar
Novati, G., Verma, S., Alexeev, D., Rossinelli, D., van Rees, W.M., Wim, M. &Koumoutsakos, P. 2017 Synchronisation through learning for two self-propelled swimmers. Bioinspir. Biomim. 12 (3), 036001.10.1088/1748-3190/aa6311CrossRefGoogle ScholarPubMed
Paris, R., Beneddine, S. & Dandois, J. 2021 Robust flow control and optimal sensor placement using deep reinforcement learning. J. Fluid Mech. 913, A25.CrossRefGoogle Scholar
Puterman, M.L. 2014 Markov Decision Processes: Discrete Stochastic Dynamic Programming, 2nd edn. John Wiley & Sons.Google Scholar
Quartapelle, L. & Napolitano, M. 1983 Force and moment in incompressible flows. AIAA J. 21 (6), 911–913.10.2514/3.8171CrossRefGoogle Scholar
Rabault, J., Kuchta, M., Jensen, A., Réglade, U. & Cerardi, N. 2019 Artificial neural networks trained through deep reinforcement learning discover control strategies for active flow control. J. Fluid Mech. 865, 281–302.10.1017/jfm.2019.62CrossRefGoogle Scholar
Rabault, J. & Kuhnle, A. 2019 Accelerating deep reinforcement learning strategies of flow control through a multi-environment approach. Phys. Fluids 31 (9), 094105.10.1063/1.5116415CrossRefGoogle Scholar
Ren, F., Rabault, J. & Tang, H. 2021 Applying deep reinforcement learning to active flow control in weakly turbulent conditions. Phys. Fluids 33 (3), 037121.10.1063/5.0037371CrossRefGoogle Scholar
Renn, P.I. & Gharib, M. 2022 Machine learning for flow-informed aerodynamic control in turbulent wind conditions. Commun. Engng 1 (1), 45.10.1038/s44172-022-00046-zCrossRefGoogle Scholar
Sedky, G., Gementzopoulos, A., Andreu-Angulo, I., Lagor, F.D. & Jones, A.R. 2022 Physics of gust response mitigation in open-loop pitching manoeuvres. J. Fluid Mech. 944, A38.10.1017/jfm.2022.509CrossRefGoogle Scholar
Smith, L., Fukami, K., Sedky, G., Jones, A. & Taira, K. 2024 A cyclic perspective on transient gust encounters through the lens of persistent homology. J. Fluid Mech. 980, A18.10.1017/jfm.2024.16CrossRefGoogle Scholar
Sonoda, T., Liu, Z., Itoh, T. & Hasegawa, Y. 2023 Reinforcement learning of control strategies for reducing skin friction drag in a fully developed turbulent channel flow. J. Fluid Mech. 960, A30.10.1017/jfm.2023.147CrossRefGoogle Scholar
Sutton, R.S. & Barto, A.G. 2018 Reinforcement Learning: An Introduction. MIT Press.Google Scholar
Theodorsen, T. 1935 General theory of aerodynamic instability and the mechanism of flutter. NACA Tech. Rep. TR 496. NACA.Google Scholar
Usandivaras, J.F.Z., Bauerheim, M., Cuenot, B. & Urbano, A. 2025 Data-driven multifidelity surrogate models for rocket engines injector design. Data-Centric Engng 6, e2.10.1017/dce.2024.56CrossRefGoogle Scholar
Verma, S., Novati, G. & Koumoutsakos, P. 2018 Efficient collective swimming by harnessing vortices through deep reinforcement learning. Proc. Natl Acad. Sci. USA 115 (23), 5849–5854.10.1073/pnas.1800923115CrossRefGoogle ScholarPubMed
Vignon, C., Rabault, J., Vasanth, J., Alcántara-Ávila, F., Mortensen, M. & Vinuesa, R. 2023 Effective control of two-dimensional Rayleigh–Bénard convection: invariant multi-agent reinforcement learning is all you need. Phys. Fluids 35 (6), 065146.10.1063/5.0153181CrossRefGoogle Scholar
Wagner, H. 1925 Über die entstehung des dynamischen auftriebes von tragflügeln. Z. Angew. Math. Mech. 5 (1), 17–35.10.1002/zamm.19250050103CrossRefGoogle Scholar
Zhong, Y. et al. 2023 Sparse sensor reconstruction of vortex-impinged airfoil wake with machine learning. Theor. Comput. Fluid Dyn. 37, 269–287.10.1007/s00162-023-00657-yCrossRefGoogle Scholar