A Genetic-Reinforcement Learning Algorithm based Fuzzy Distributed Frequency Regulation Strategy With Virtual Power Plant Under PMU Missing Data

This paper investigates the frequency regulation problem on power system with virtual power plant (VPP) under phasor measurement unit (PMU) missing data. First, the distributed load frequency control (LFC) incorporating VPP is established, in which a unified dynamic model based on VPP is introduced with the consideration of PMU missing data. Then, a genetic-reinforcement learning algorithm is developed to optimize the power output of various frequency regulation resources. Furthermore, a fuzzy dynamic event-triggered mechanism (FDETM) is proposed to achieve the balance between bandwidth resource utilization and frequency regulation performance. Finally, theoretical analysis is conducted using Lyapunov theory. Sufficient conditions for ensuring the system stability are derived. Evaluations on IEEE 9-bus test system in MATLAB/Simulink and IEEE 39-bus test system on hardware-in-the-loop (HIL) experiment platform demonstrate the effectiveness and superiority of the proposed strategy.