Two-Stage Full-Data Processing for Microgrid Planning With High Penetrations of Renewable Energy Sources

With the high penetrations of diverse renewable energy resources and energy storage devices, the optimal planning of microgrids based on selected representative operating periods (ROPs) are facing significant challenges. This paper proposes a full data-driven planning method to cope with such challenges. The proposed method takes the full data, which contain all the information on load and meteorological conditions in the planning horizon as input, and applies a two-stage data processing approach to obtain weekly and hourly ROPs, while preserving the chronological order of the raw data. Two time series with different resolutions are established in the planning model in order to satisfy coarse time intervals required for investment decisions and conform to finer time scales required for operating decisions. The proposed method reduces the scale of data, speeds up the solution process, and surmounts the computational burdens in solving the planning model in the designated horizon. Numerical results on an industrial park microgrid in Shanghai demonstrate that the proposed method can provide more accurate ROP and planning results than competitive methods.