Application of Reinforcement Learning Algorithm to Parameter Identification for MPPT Control of PMSG Wind Energy Conversion Systems

Abstract

With the depletion of fossil fuels as a typical background, the introduction of renewable energy sources has been increasing in many parts of the world. In particular, the introduction of wind energy conversion systems (WECS) has increased significantly in recent years, and research and development of offshore wind power generation systems is being actively conducted. Permanent magnet synchronous generators (PMSGs) are suitable for wind power generation systems because of their high efficiency compared to other generators. WECS are adopted Maximum Power Point Tracking (MPPT) control. This is because depending on the wind speed, there is an optimum speed at which the maximum output can be obtained. There are three types of MPPT control: mountain climbing method, optimum speed ratio control, and optimum torque control. However, the optimum speed ratio control is a direct MPPT control method that directly changes the rotational speed, which may cause transient adverse effects on the WECS. Therefore, the optimal torque control, which is an indirect MPPT control method, has attracted much attention. This paper proposes a method of reinforcement learning for parameter identification in indirect MPPT control of PMSG-WECS. Optimal torque control is applied to the MPPT control, and the parameters are identified by reinforcement learning. Some case studies demonstrate the effectiveness of the proposed method.