Coordinated operation of multi-energy microgrids considering green hydrogen and congestion management via a safe policy learning approach

Multi-energy microgrids (MEMGs) with green hydrogen have attracted significant research attention for their benefits, such as energy efficiency improvement, carbon emission reduction, as well as line congestion alleviation. However, the complexities of multi-energy networks coupled with diverse uncertainties may threaten MEMG's operation. In this paper, a data-driven methodology is proposed to achieve effective MEMG operation, considering the green hydrogen technique and congestion management. First, a detailed MEMG modelling approach is developed, coupling with electricity, green hydrogen, natural gas, and thermal flows. Different from conventional MEMG models, hydrogen-enriched compressed natural gas (HCNG) models, and weather-dependent power flow are thoroughly considered in the modelling. Meanwhile, the power flow congestion problem is also formulated in the MEMG operation, which could be mitigated through HCNG integration. Based on the proposed MEMG model, a reinforcement learning-based method is designed to obtain the optimal solution of MEMG operation. To ensure the solution's safety, a soft actor-critic (SAC) algorithm is applied and modified by leveraging the Lagrangian relaxation and safety layer scheme. In the end, case studies are conducted and presented to validate the effectiveness of the proposed method.