arXiv cs.AI· Yuning Wang, Pol Suarez, Mathis Bode, Ricardo Vinuesa·· 6 天前AI 评分27
物理引导代理学习实现湍流机翼零样本控制
Physics-guided surrogate learning enables zero-shot control of turbulent wings
AI 导读
研究者利用壁面湍流的局部结构，在匹配机翼边界层统计特征的湍流槽道流中训练控制策略，无需额外训练即零样本部署到 Re_c=2×10^5 的 NACA4412 翼型上。迁移后的策略降低总阻力 11.1%，优于当前最优的 opposition control（8.9%）；计入驱动成本后净节能分别为 10.9% 和 8.8%。代理训练相比直接在机翼上训练，探索效率估计提升四个数量级。
正文 · 原文
Abstract:Turbulent boundary layers over aerodynamic surfaces are a major source of aircraft drag, yet their control remains challenging due to multiscale dynamics and spatial variability, particularly under adverse pressure gradients. Reinforcement learning has outperformed state-of-the-art strategies in canonical flows, but its application to realistic geometries is limited by computational cost and transferability. Here we show that these limitations can be overcome by exploiting local structures of wall-bounded turbulence. Policies are trained in turbulent channel flows matched to wing boundary-layer statistics and deployed directly onto a NACA4412 wing section at $Re_c=2\times10^5$ without additional training on the wing, constituting zero-shot deployment. The transferred policies reduce total drag by 11.1\%, while state-of-the-art opposition control reduces it by 8.9\%. With the actuation cost taken into account, the corresponding net-energy savings are 10.9\% and 8.8\%, respectively. Surrogate training provides an estimated improvement of four orders of magnitude in exploration efficiency over direct on-wing training, enabling scalable flow control.
| Subjects: | Fluid Dynamics (physics.flu-dyn); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2604.09434 [physics.flu-dyn] |
| (or arXiv:2604.09434v3 [physics.flu-dyn] for this version) | |
| https://doi.org/10.48550/arXiv.2604.09434 arXiv-issued DOI via DataCite |
Submission history
From: Yuning Wang [view email]
[v1]
Fri, 10 Apr 2026 15:50:44 UTC (43,645 KB)
[v2]
Mon, 13 Apr 2026 02:40:35 UTC (31,127 KB)
[v3]
Fri, 9 Oct 2026 14:16:19 UTC (34,226 KB)
来源：arXiv cs.AI · arxiv.org