Two Stage Robust Economic Dispatching of Microgrid Considering Uncertainty of Wind, Solar and Electricity Load Along with Carbon Emission Predicted by Artificial Intelligence Method

Abstract

Distributed microgrid has become significant carriers to solve energy security and user requirements while the randomness and volatility characteristics would have a significant impact on the operation of microgrid. In this paper, the wind, photovoltaic power and electric load are forecasting based on BP neural network algorithm with average error of 3.42%、7.26% and 3.92% respectively. Based on forecasting fluctuation intervals, an economic dispatching model is established considering uncertainty and carbon emissions, and it is solved by two-stage robust optimization model and C&CG algorithm. The power output of gas turbine is closely related to the time-use-price strategy while the operation of energy storage system depends on the D-value of peak and valley price and charging/discharging cost. As the uncertainty intervals increases, the electricity purchase cost increases significantly where microgrid refers to purchase more electricity to improve conservatism while the valley electricity price is lower than that of gas turbine. Robust optimization scheduling has the advantage of resisting real-time market fluctuations, and has lower costs compared to intraday scheduling on the basis of higher electricity prices than day-head which could reach 13.15% decreasing. Considering carbon constraints dispatching is more inclined to utilize gas turbine with lower carbon emission factor while the total cost would rise by 13.33% with the high operating cost of gas turbine. With the relaxation of carbon allocations, the purchase and sale of electricity are both increasing along with carbon emissions, and the net profit of grid interaction would decrease and the total operation cost would lead to a minimum value and keep balanced.

Keywords: Load forecasting, Robust optimization, Economic operation, Carbon allocations, C&CG algorithm, Day-ahead dispatching