MADDPG Based Power Coordination Control Strategy for Hydrogen Energy Storage Units

To address the issues of limited output power of individual electrolyzer and fuel cell, as well as the dynamic energy losses during hydrogen-electricity conversion in the operation of electric-hydrogen coupled system, this paper proposes a power coordination and allocation strategy for hydrogen energy storage system (HESS) units based on a Multi-agent Deep Deterministic Policy Gradient (MADDPG) algorithm. The strategy aims to minimize the energy losses during hydrogen-electricity conversion, reduce the operating costs of HESS units, and minimize the cost of purchasing spinning reserve from the upper-level power grid. Considering the power and energy constraints during the hydrogen-electricity conversion process, the MADDPG algorithm is employed to coordinate the output of multiple HESS units to address power imbalances in the wind-solar-load system. The effectiveness of the proposed strategy is validated through comparison with traditional control strategy.