Deep Reinforcement Learning for Stability Enhancement of a Variable Wind Speed DFIG System

Low-frequency oscillations are a primary issue for integrating renewable source into the grid. The objective of this study was to find sensitive parameters that cause low-frequency and design Twin Delayed Deep Deterministic Policy Gradient (TD3) agent controller damp without requiring an accurate system model. In work, Q-learning (QL)-based model-free wind speed DFIG designed on rotor-side converter (RSC), QL-based DC-link voltage regulator grid-side (GSC) enhance stability system. next step, TD3 trained learn dynamics by replacing inner current controllers RSC, which replaced first stage, conventional PSS Proportional–Integral (PI) were introduced both RSC GSC. Then, become PI with QL algorithm under very small variations. second introducing large variations in speed. results reveal can sustain assuming detailed control structure beforehand, while stabilize doubly fed induction generator (DFIG)-equipped energy conversion (WECS)