Benders Decomposition Algorithm To The Corrective Security-Constrained Optimal Power Flow Problem

The Optimal Power Flow (OPF) was in the 1960s and 1970s, with the blackout in America and other developed countries, attached great importance to the research and practice of the Security-Consrtained Optimal Power Flow (SCOPF). Nowadays. OPF has already been an basic tool in power systems'control, dispatch, overhaul and operational planning, whose application field is very widespread.To solve the Corrective Security-Constrained Optimal Power Flow(CSCOPF) problem which was repeatedly involved in power system operation and economic dispatch, this thesis proposed a practical approach based on Benders Decomposition. So that the solution of complex and large power systems could be carried out.Firstly, according to the characteristics of the CSCOPF problem, its constraints were divided into three categories:constraints of the pre-contingency state, constraints of the post-contingency states and coupling constraints between the base case and post-contingency states. On the basis of the relationship analysis of these three categories of constraints, we obtained three types of relationship between post-contingency states and the optimum of base case, which provided theory basis for the decomposition algorithm and contingency filtering techniques in this thesis.Secondly, CSCOPF problem was decomposed and coordinated by Benders Decomposition algorithm based on the above classification of constraints. Therefore, constraints of the pre-contingency state were dealed with Benders master problem, constraints of the post-contingency states were dealed with Benders sub-problems and the coupling constraints were reflected by Benders cut. Consequently, the CSCOPF model was divided into a series of OPF models and the connecting information was coordinated by Benders cut.Thirdly, Benders Decomposition method needed to form sub-problem model for every post contingency, to solve these optimization models would consume huge calculating amount especially for a large power system. However, in real life applications, most contingencies did not constrain the optimum, hence, a contingency filtering technique based on selecting dominant contingencies was adopted in this thesis to avoid the invalid contingencies and to reduce the scale of the sub-problem set.Finally, Benders master problem model and sub-problem models were all classical OPF models, there were many redundant constraints (mainly were transmission components security constraints), so we solved these optimization models based on selecting the dominant component under the operating point which was not fixed. By doing this, we formed the minimum set of constraints to improve computational efficiency.