通过构网型混合光伏电站提升电力系统短期电压稳定性

Short-term voltage stability problems are a major concern in power systems with high penetration of inverter-based resources (IBRs). Fault-induced delayed voltage recovery (FIDVR) is one such problem that can lead to short-term voltage instability. This paper investigates FIDVR mitigation using dynamic voltage support from hybrid PV plants and load control. We propose a deep reinforcement learning (DRL) based optimization framework for optimal dynamic reactive power injection from hybrid PV plants and optimal load control. The Q-injection and amount of load tripping depend on system operating conditions, so the data-driven framework gives optimal setpoints that are adaptable to the varying operating conditions. The numerical simulations show that the optimal dynamic reactive power injection and load control improve the voltage recovery and decrease the load trip amount to comply with the voltage recovery criterion.