Q-Learning-Based Frequency and Voltage Restoration for Parallel Grid-Forming Inverters

This study addresses the critical challenge of frequency and voltage restoration in parallel Inverter-Based Resources (IBRs) while maintaining precise control over active and reactive power distribution. The proposed approach integrates advanced voltage and frequency restoration technique to enhance system stability. To optimize controller parameters during dynamic state transitions, particularly during load connection and disconnection events, a Q-learning algorithm is implemented. Through the presented studies, the performance of the proposed Q-learning-based control strategy is evaluated and compared against conventional PI-based controllers and Particle Swarm Optimization (PSO) based approaches. The results demonstrate the effectiveness of the proposed Q-learning-based restoration strategy, achieving the most stable frequency response during different local load events.