1.School of Electrical Engineering, Beijing Jiaotong University, Beijing 100044, China;2.NARI Technology Co., Ltd., Nanjing 211106, China

The uncertainty of distributed renewable energy output significantly affects the accurate characterization of the feasible region for virtual power plant aggregation. It is necessary to accurately calculate the adjustable capacity of virtual power plants to enhance the balancing capability of the power system. This paper proposes a probability model for the uncertainty of distributed renewable energy output in virtual power plants and a construction method of the credible probability-based feasible region based on multiple probability variables. Firstly, based on the ensemble learning and conformal quantile regression methods, an uncertainty probability model is proposed to describe the heteroscedastic time series of renewable energy output. Secondly, the characteristics of Minkowski sum and optimization methods for solving feasible region are studied, and the influences of renewable energy output constraints and network constraints on the feasible region of the virtual power plant are analyzed. Then, this paper proposes a construction method for the probability-based feasible region of the virtual power plant based on output probability intervals of distributed renewable energy. Finally, considering the coupling characteristics of wind and photovoltaic power, the construction method for the probability-based feasible region of the virtual power plant is optimized. By coordinating the confidence coefficients of the output probability intervals of each distributed renewable energy, an overall confidence level of the probability-based feasible region of the virtual power plant is obtained. The case verification results show that the proposed method can accurately characterize the adjustable capacity of virtual power plants.

This work is supported by National Key R&D Program of China (No. 2022YFB2403400).