Design of Intelligent Control Law Embedded With Dynamic Flow Field Perception Method for High‐Angle‐of‐Attack Maneuvers of Numerical Virtual Flight Agent.
Saved in:
| Title: | Design of Intelligent Control Law Embedded With Dynamic Flow Field Perception Method for High‐Angle‐of‐Attack Maneuvers of Numerical Virtual Flight Agent. |
|---|---|
| Authors: | Zhao, Dongyu1,2 (AUTHOR), Liu, Shuyuan1,2 (AUTHOR), Chang, Xinghua2 (AUTHOR) cxh_cardc@126.com, Zhang, Laiping2 (AUTHOR), Deng, Xiaogang2,3 (AUTHOR), Yan, Binbin (AUTHOR) yanbinbin@nwpu.edu.cn |
| Source: | International Journal of Aerospace Engineering. 4/24/2026, Vol. 2026, p1-22. 22p. |
| Subjects: | Angle of attack (Aerodynamics), Intelligent control systems, Reinforcement learning, Aerofoils, Computational aerodynamics |
| Abstract: | In numerical virtual flight (NVF) tasks, strong unsteady flow disturbances can change aerodynamic behavior and reduce the effectiveness of control surfaces, increase the difficulty of aerodynamics modeling and controlling at high‐angle‐of‐attack (high‐AoA). Therefore, an intelligent control law embedded with dynamic flow field perception module for high‐AoA maneuvering is proposed, enabling pitch‐up and recovery maneuvers across arbitrary AoA. First, a NACA 0012 airfoil equipped with an elevator is adopted as the geometric model. Flow fields of high‐AoA maneuvers are simulated using rigid dynamic grid and overset grid techniques to capture unsteady aerodynamic data. Then, a deep reinforcement learning environment is established under unsteady flow conditions, which can enforce physical constraints and accurately capture transient aerodynamic effects through an embedded dynamic flow field perception module. Finally, a physically consistent TD3‐based control law for surface deflection is established, 69.29% performance improvement is achieved compared with baseline model under high‐AoA conditions. In 1000 randomized NVF experiments, the intelligent control law can reliably control elevator deflection, enabling high‐AoA maneuvers and autonomous recovery at arbitrary angles, with an average angular error of 0.902° and mean response time (MRT) of 0.007 s. Its effectiveness under nonlinear and unsteady conditions has been demonstrated, the potential for engineering applications is highlighted. [ABSTRACT FROM AUTHOR] |
| Copyright of International Journal of Aerospace Engineering is the property of Wiley-Blackwell and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.) | |
| Database: | Engineering Source |
|
Full text is not displayed to guests.
Login for full access.
|
|
| Abstract: | In numerical virtual flight (NVF) tasks, strong unsteady flow disturbances can change aerodynamic behavior and reduce the effectiveness of control surfaces, increase the difficulty of aerodynamics modeling and controlling at high‐angle‐of‐attack (high‐AoA). Therefore, an intelligent control law embedded with dynamic flow field perception module for high‐AoA maneuvering is proposed, enabling pitch‐up and recovery maneuvers across arbitrary AoA. First, a NACA 0012 airfoil equipped with an elevator is adopted as the geometric model. Flow fields of high‐AoA maneuvers are simulated using rigid dynamic grid and overset grid techniques to capture unsteady aerodynamic data. Then, a deep reinforcement learning environment is established under unsteady flow conditions, which can enforce physical constraints and accurately capture transient aerodynamic effects through an embedded dynamic flow field perception module. Finally, a physically consistent TD3‐based control law for surface deflection is established, 69.29% performance improvement is achieved compared with baseline model under high‐AoA conditions. In 1000 randomized NVF experiments, the intelligent control law can reliably control elevator deflection, enabling high‐AoA maneuvers and autonomous recovery at arbitrary angles, with an average angular error of 0.902° and mean response time (MRT) of 0.007 s. Its effectiveness under nonlinear and unsteady conditions has been demonstrated, the potential for engineering applications is highlighted. [ABSTRACT FROM AUTHOR] |
|---|---|
| ISSN: | 16875966 |
| DOI: | 10.1155/ijae/4223020 |
