A contingency partitioning approach for preventive-corrective security-constrained optimal power flow computation

• Published in: Electric power systems research 2016-03, Vol.132, p.132-140
• Main Authors: Xu, Yan, Yang, Hongming, Zhang, Rui, Dong, Zhao Yang, Lai, Mingyong, Wong, Kit Po
• Format: Article
• Language: English
• Subjects: Benders decomposition; Contingency partitioning; Corrective control; Optimal power flow
• ISSN: 0378-7796; 1873-2046
• DOI: 10.1016/j.epsr.2015.11.012

•A novel and improved approach is proposed for preventive-corrective security-constrained optimal power flow (PCSCOPF) problem.•The contingencies are optimally partitioned into two exclusive sets: one is secured in the preventive control (PC) stage and the other in the corrective control (CC) stage.•The PC corresponds to an ordinary SCOPF model and solved by the Benders decomposition method; the CC corresponds to an ordinary OPF model.•The optimal partitioning of the contingencies is determined using an evolutionary algorithm (EA).•The searching dimension only depends on the number of insecure contingencies, hence tends to lead to a much higher solution speed. Security-constrained optimal power flow (SCOPF) is an important tool to optimize power system's operating state while satisfying security requirements with respect to credible contingencies. As a necessary extension to conventional SCOPFs, the recently proposed preventive–corrective SCOPF (PCSCOPF) aims to achieve the best coordination between the preventive control (PC) and corrective control (CC) considering the probabilistic nature of the contingencies and the cost of CC as well as other binding constraints. However, PCSCOPF renders a large-scale, multi-stage, non-linear programming problem which is difficult to directly solve. This paper proposes a novel and improved approach to the PCSCOPF problem. We aim to partition the contingencies into two exclusive sets: one is secured in the PC stage and the other in the corrective control CC stage. The PC corresponds to an ordinary SCOPF model and solved by the Benders decomposition method; the CC corresponds to an ordinary OPF model. The optimal partitioning of the contingencies is determined using an evolutionary algorithm (EA). Compared with the existing methods, the proposed approach is advantageous in that its searching dimension only depends on the number of insecure contingencies, hence tends to lead to a much higher solution speed. The proposed method is verified on the IEEE 118-bus system and compared with other two existing methods. Simulation results show that the proposed method can provide high-quality solutions with much higher computation speed.