Safe Reinforcement Learning for Grid-Forming Inverter Based Frequency Regulation with Stability Guarantee

This letter investigates a safe reinforcement learning strategy for grid-forming (GFM) inverter based frequency regulation. To guarantee stability of the inverter based resource (IBR) system under the learned control policy, a model based reinforcement learning (MBRL) technique is combined with Lyapunov approach which determines safe region of states and actions. To obtain near optimal control strategy, the control performance is safely improved by approximate dynamic programming (ADP) using data sampled from the region of attraction (ROA). Moreover, to enhance the control robustness against parameter uncertainty in the inverter, a Gaussian process (GP) model is adopted by the proposed MBRL to effectively learn system dynamics from measurements. Numerical simulations validate the effectiveness of the proposed method.