Knowledge-driven multi-timescale optimization dispatch for hydrogen-electricity coupled microgrids

Abstract

Hydrogen-electricity coupled microgrids (HEMGs), which convert surplus renewable electricity into hydrogen, promise to reduce the asynchronous spatial and temporal distribution between renewable energy outputs and the load demands. However, given the complex system structure and different characteristics of renewable and hydrogen energies, optimizing the dispatch is a difficult task. In this paper, the source-storage-load dispatch problem of HEMGs is solved with a knowledge-driven multi-timescale optimization dispatch strategy. First, the multi-timescale dispatch strategy of seasonal hydrogen storage, day-ahead economic dispatch and intraday dynamic optimization is proposed to balance the supply-and-demand fluctuations. Second, an improved knowledge-driven optimization framework is designed and a knowledge network is established by using the knowledge accumulated from historical data and expert experience to enhance the optimization quality. Third, the system operation under fluctuating inputs is optimized using a variable neighborhood search solution algorithm based on the dynamic window approach. Case studies demonstrated that the proposed optimization strategy reduces the operating cost, carbon emissions and solution time by 23.58 %, 33.95 % and 25.35 % with the proposed optimization strategy.