Proximal Policy Optimization-Based Stability Control Method for Permanent Magnet Direct-Drive Wind Power Generation System

With the increasing proportion of wind power grid connection, the stable control performance of wind turbines is crucial for the safety and stability of the power system. This paper proposes a proximal policy optimization (PPO) algorithm-based adaptive stability control framework for low voltage ride-through (LVRT) in permanent magnet direct-drive wind power generation system. Considering phase-locked loop (PLL) impact on system performance, this method real-time adjusts PLL parameters to enhance LVRT. A Markov decision process (MDP) model for LVRT control is established, and reward functions for distinct fault stages are designed. The PPO algorithm solves for the adaptive control strategy. Simulation results validate the method's effectiveness across various faults, demonstrating significant improvements in system LVRT capability, adaptability and stability control capability.