Reinforcement Learning-Based Voltage Control for Power Systems with Static Var Generator Integration

Voltage stability is a critical aspect of power system operation, where sudden load or generation fluctuations can lead to voltage instability, impacting end-user equipment and potentially triggering cascading failures. This paper presents a novel approach for voltage control using Static Var Generators (SVGs) optimized through Multi-Agent Deep Deterministic Policy Gradient (MADDPG), a reinforcement learning (RL) algorithm. The proposed method integrates Lyapunov stability constraints within the RL training process to ensure stability, addressing a key challenge where conventional RL policies lack formal guarantees for system stability. The results validate the method’s effectiveness in maintaining voltage within acceptable ranges under dynamic conditions, offering a viable solution for real-world power system applications.