Interior Point Method With Optimal Factor For Reactive Power Optimization

Optimal reactive power dispatch (ORPD) is one of the most important measure to improve the voltage profile and reduce power loss, which ensure the power system run securely and economically. The study of ORPD has great significance in practical application.Algorithms for ORPD include linear programming, nonlinear programming and intelligent optimization methods. Interior point method is an excellent one. It shows great efficience in solving large scale optimization problems with a great deal of equality and inequality constraints. A key drawback of the majority of traditional algorithms is that the solution time may grow exponentially for large scale problems. By the contrast, interior point is a polynominal time algorithm, which makes it is superior when applied to large scale problems.In this thesis, primal-dual interior method and its variants, including predictor- corrector interior point method, multi centrality corrections interior point method and fuzzy second correction interior point method are presented first. The numerical performance of the presented algorithms is evaluated on IEEE test systems. Extensive numerical results show that the interior point method is fast and robust.A interior point method with optimal factor for ORPD is proposed.The key to direct nonlinear interior point method is to solve the KKT conditions, a system of nonlinear equations. Newton method which may diverge from poor starting point is usually employed .To improve the robustnese of traditional interior point method, new interior point method with optimal factor is developed. ORPD model in rectangular coordinate, in which all the constraints become quadratic, is presented first. A merit function expressed by step length is established. Guided by the merit function, we get an optimal factor and use it as the step length. The proposed method is carried out on IEEE test system and results indicate the effectiveness.