Multi-time Scale Scheduling Strategy With Dynamic Adjustment Technology Of Park-level Microgrid Integrated Energy System

In order to realize the low-cost operation of park-level microgrids including photovoltaic power supply systems, electric boilers, refrigeration systems and energy storage systems, this paper proposes a multi-time scale strategy for microgrids. The strategy is divided into three layers: pre-day predictive scheduling, intraday scheduling, and real-time scheduling, and a more reasonable prediction model is obtained by adjusting the multi-time prediction of different units. With the optimization goal of minimizing the cost of the microgrid, the system data is gradually corrected by combining the ELMAN regression neural network with day-ahead and intraday rolling prediction schemes and multi-gradient prediction to make the microgrid operate at optimal power throughout the day. Finally, the park-level microgrid system is used as an example, and the system is verified by GUROBI solver through the MATLAB software platform, and the results show that the proposed algorithm is reasonable and effective.