考虑源荷多尺度预测的风光储氢系统优化运行

Abstract The randomness of renewable energy output and the uncertainty of multiple load demands significantly increase the complexity of optimization and scheduling in integrated energy systems, rendering traditional optimization methods inadequate for practical operational needs. To address this challenge, this paper proposes an optimal operation method for a wind–solar–storage–hydrogen system based on multi-scale forecasting of both energy sources and loads. The multi-scale forecasting model is constructed based on correlation and periodicity analyses, and TimeGAN is introduced for data augmentation to enhance forecast accuracy. Then, a more forward-looking day-ahead operation plan is formulated using day-ahead forecasting data, followed by hour-level intra-day adjustments based on updated forecasts to reduce power deviations of equipment. Simulation results on typical days show that the proposed strategy reduces operational costs by 16.22 % compared to traditional methods, while the utilization rates of renewable energy sources on the three typical daysare improved to 80.63 %, 91.54 %, and 95.83 %, respectively, significantly enhancing both the economic performance and renewable energy accommodation of the system.