Deep Reinforcement Learning-Based Adaptive Coordination of DC Reactor Fault Current Limiters for Protection of Hybrid AC/DC Microgrids with High Renewable Penetration

The increasing penetration of renewable energy sources (RES) into hybrid AC/DC microgrids introduces highly dynamic and uncertain fault current profiles that severely challenge conventional protection coordination schemes. Existing fault current limiter (FCL) coordination strategies rely on fixed impedance thresholds or pre-defined rule-based controllers, which prove inadequate under varying grid topologies, islanding transitions, and diverse fault conditions. This paper proposes a Deep Reinforcement Learning (DRL)-based intelligent coordination framework for DC Reactor Fault Current Limiters (DCR-FCLs) deployed in a hybrid AC/DC microgrid. A Deep Q-Network (DQN) agent is trained to adaptively determine the optimal limiting impedance and switching sequence of DCR-FCLs in real time under diverse fault scenarios, including three-phase, line-to-ground, line-to-line, and double line-to-ground faults. The proposed framework integrates FCL coordination with directional overcurrent relay (DOCR) settings to achieve optimal protection coordination (OPC). The system is modelled and simulated in MATLAB/Simulink, and the results are validated against a benchmark hybrid microgrid test system. Comparative analysis demonstrates that the DRL-based approach significantly outperforms conventional fixed and rule-based FCL coordination strategies in terms of fault current reduction, bus voltage restoration, relay coordination time, and protection selectivity across all tested RES penetration levels.