Floating Offshore Wind Farm Yaw Control Via Model-based Deep Reinforcement Learning

This paper proposes a model-based deep reinforcement learning (DRL) framework to maximize the total power output of a floating offshore wind farm subject to wake effects. Recognizing the extensive interactions required for DRL training and the partially observable nature of wind farm dynamics, we first develop a physics-based simplified model that captures the aerodynamic interactions among floating wind turbines. The high computational efficiency of this model enables it to support DRL training and deployment. Subsequently, using this model, a model-based DRL scheme has been established. The key feature of the scheme is to use the simplified model to significantly improve the training efficiency and estimate the system states. Finally, simulation results with a utility-scale floating wind farm validate the effectiveness of the proposed model-based DRL scheme.