Reinforcement Learning-Based Optimal Scheduling for Hydrogen-Enriched Integrated Energy Systems

Hydrogen-enriched regional integrated energy systems (HRIES) have garnered significant attention in addressing climate-related issues. Incorporating Power-to-Hydrogen-Heat (P2HH) and Underground Hydrogen Storage (UHS) technologies into Hybrid Renewable Integrated Energy Systems (HRIES) is a viable solution for addressing seasonal fluctuations in energy supply. This study utilizes a deep deterministic policy gradient (DDPG) methodology, which is based on empirical modeling, for effectively resolving the challenges in HRIES. The optimization procedure is defined through a Markov decision process, highlighting the dynamics between the agent and its surrounding environment. The research conclusively demonstrates the efficiency of the proposed scheduling technique and the increased adaptability of HRIES by integrating P2HH and UHS.