Two-stage robust optimal scheduling framework applied for microgrids: Combined energy recovery and forecast

Abstract

To raise the power supply reliability of microgrids in remote areas, it's necessary to establish a multi-energy system to improve energy efficiency. However, the uncertainties in load fluctuation and renewable energy will hinder the implementation of planning schemes, resulting in scheduling plans mismatching the actual demand, and ultimately failing to ensure the economic operation of microgrids. Therefore, the two-stage robust optimal scheduling framework that combines energy recovery, demand response, and forecast is proposed for scheduling in microgrids. Uncertainties and economic costs are fully considered in the framework to make the planning scheme for microgrids more in line with the actual project. Simulation consequences indicate that the BP neural network could efficiently forecast the load with an error within 5%. Additionally, considering thermodynamics and economy, 0.3 is the optimal split ratio for recompression supercritical CO2 Brayton cycle to recover waste heat from the gas turbine. Finally, Case 4 with energy recovery and demand response is economically optimal, with an annual operating cost of $1.71 million lower than the original Case 1 and a total annual cost 13.92% lower than Case 1. By comparing different cases, it is proved that the proposed Case 4 optimization framework has advantages for scheduling in microgrids.