Proximal policy optimization–based noncausal control for wave energy conversion systems

This paper proposes a framework for wave energy converter (WEC) control that utilizes proximal policy optimization (PPO), which is a state-of-the-art policy gradient reinforcement learning algorithm. Unlike conventional model-based controllers that rely on precise models, the proposed framework is model-free and noncausal, utilizing wave prediction to enhance energy harvesting. By integrating PPO with a noncausal controller, the proposed control scheme adaptively adjusts the control parameters in real time based on the response of the WEC and wave predictions. To the best of the authors’ knowledge, this is the first attempt to utilize noncausal-based PPO in a WEC application. The proposed framework directly addresses the challenges of controlling emerging WEC systems, and it is particularly suited to soft-body devices, e.g., dielectric elastomer generators and dielectric fluid generators, which are difficult to model accurately. The proposed control scheme was evaluated with different wave prediction horizons and compared against a conventional reactive controller. The results demonstrate that the proposed control scheme obtains higher energy generation while maintaining stable operation. In addition, the proposed scheme exhibits robust performance under wave prediction error and uncertainty conditions.