Two-Stage Reactive Power Optimization for Distribution Network with Distributed Generation Based on Extreme Scenarios

Zhang Yipu,Ai Xiaomeng,Fang Jiakun,Yao Wei,Wen Jinyu

DOI: https://doi.org/10.19595/j.cnki.1000-6753.tces.L70664

2018-01-01

Abstract:Public awareness of fossil-fuel depletion and policy incentives has emulated the development and deployment of distributed renewable energy sources, such as wind power and solar energy. However, the uncertainties of fluctuating wind power and solar power challenge the traditional methods. In this paper, a two-stage optimization method based on extreme scenarios is proposed for optimal reactive power dispatch in the distribution system. The proposed approach deals with random variables by the extreme scenarios, and determines slow variables associating with the inflexible devices in the system in coordination with those flexible sources. It guarantees the security of system operation, and at the same time it achieves the goal of minimizing the loss and maximizing the renewable integration. The proposed method can be divided into 3 steps. First, the optimal reactive optimization problem is formulated as a mixed integer non-convex nonlinear optimization based on the relaxed branch flow model. Then the big M-approach and conic relaxation are used to convert the original problem into a quadratic programming. Third, the slow variables in the distribution network are decided by sampling the output power of the distributed generation using an extreme scenario method. Theoretical proof and simulation results demonstrate the effectiveness of the proposed method.