EUVIMEDEUVIMED entdecken
Crossref

Optimization of duty-cycled plasma actuation parameters over a square cylinder using surrogate-assisted evolutionary reinforcement learning

Muhammad Hamza, Fahad Nawaz, Kashif Ali, Xuanshi Meng

Physics of Fluids · 01.11.2026 · 38 (11)

DOI: 10.1063/5.0352993 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Physics of Fluids, online erschienen am 13.11.2026. Verfasst von Muhammad Hamza und 3 weiteren Autor:innen. Die Arbeit nennt 55 Literaturangaben. Gefördert von: National Natural Science Foundation of China, The Foundation of National Key Laboratory of Aircraft Configuration Design, 111 Project.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Plasma-based flow control has emerged as a promising approach for suppressing wake instabilities and reducing aerodynamic drag in bluff-body flows; however, its effectiveness depends on complex nonlinear interactions between actuation parameters and vortex dynamics. In this study, a surrogate-assisted evolutionary reinforcement learning (SA-ERL) framework is used to optimize duty-cycled plasma actuation over a square cylinder at Reynolds number Re=200. The control space consists of duty-cycle frequency f, duty-cycle ratio τ, and phase difference ϕ between symmetric trailing-edge actuators, forming a coupled, three-dimensional, multi-objective optimization problem. A hybrid Gaussian process-Transformer (GP-Transformer) surrogate is trained on computational fluid dynamics samples and refined through uncertainty-guided active learning, coupled with an evolutionary reinforcement learning (ERL) optimizer that combines a genetic algorithm (GA) for global search and proximal policy optimization for local refinement. The surrogate achieves high predictive accuracy, while the ERL optimizer outperforms standalone GA and deep reinforcement learning by converging to a higher and more stable terminal reward. Compared with the baseline, the optimized actuation (f=20 Hz, τ=0.9, ϕ=0) achieves about 52% reduction in mean drag and 97% suppression of lift oscillations, while reducing normalized actuation power nearly 10% relative to continuous forcing. Flow-field analysis shows the optimized frequency dominates over natural vortex shedding, reducing resonance risk, weakening coherent wake structures, and weakening the recirculation region. These results demonstrate that SA-ERL offers a computationally efficient, interpretable framework for multi-parameter plasma control, extendable to higher Reynolds numbers and three-dimensional flows.

Förderung

  • National Natural Science Foundation of China
  • The Foundation of National Key Laboratory of Aircraft Configuration Design
  • 111 Project

Zitieren

APA

Hamza, M., Nawaz, F., Ali, K., & Meng, X. (2026). Optimization of duty-cycled plasma actuation parameters over a square cylinder using surrogate-assisted evolutionary reinforcement learning. Physics of Fluids, 38(11). https://doi.org/10.1063/5.0352993

Vancouver

Hamza M, Nawaz F, Ali K, Meng X. Optimization of duty-cycled plasma actuation parameters over a square cylinder using surrogate-assisted evolutionary reinforcement learning. Physics of Fluids. 2026;38(11). doi:10.1063/5.0352993

BibTeX
@article{Hamza2026,
  author = {Hamza, Muhammad and Nawaz, Fahad and Ali, Kashif and Meng, Xuanshi},
  title = {Optimization of duty-cycled plasma actuation parameters over a square cylinder using surrogate-assisted evolutionary reinforcement learning},
  journal = {Physics of Fluids},
  year = {2026},
  volume = {38},
  number = {11},
  doi = {10.1063/5.0352993}
}
Kostenlos starten →