City Research Online

Abstract

Active flow control (AFC) of bluff bodies is a critical concept in a variety of engineering applications, ranging from structural safety to clean energy harvesting. The current study reviews the application of deep reinforcement learning (DRL) in the AFC of bluff bodies with a focus on key studies between 2017 and 2024. DRL has improved AFC capabilities by enabling real-time optimization and adaptive control. The review showed a focus on a single circular cylinder, generally simulated two-dimensionally at a low Reynolds number of 100, using CFD tools like OpenFOAM and FEniCS. Synthetic jets are favored actuators and Proximal Policy Optimization (PPO), paired with deep neural networks, is the most widely used DRL algorithm. Despite significant advances, key challenges still remain in the field. Scalability to higher Reynolds numbers and turbulent flows, high computational cost, and sensor optimization issues are found as the main challenges. The review also identifies inconsistent reporting of hyperparameters and varied training procedures, which may hinder reproducibility of works for further steps. Future research should address these gaps to extend the applicability of DRL in practical AFC systems. This review also proposes a comprehensive framework to connect initial simulation results with real-world applications.

Publication Type:	Article
Additional Information:	© 2025. This manuscript version is made available under the CC-BY-NC-ND 4.0 license https://creativecommons.org/licenses/by-nc-nd/4.0/
Publisher Keywords:	Active flow control, Machine Learning, Deep reinforcement learning, Bluff Body
Subjects:	H Social Sciences > HN Social history and conditions. Social problems. Social reform Q Science > QA Mathematics > QA75 Electronic computers. Computer science T Technology > TJ Mechanical engineering and machinery
Departments:	School of Science & Technology School of Science & Technology > Engineering
SWORD Depositor:	Symplectic Administrator

Export

Downloads