Weighted Multiple Predictor-corrector Interior Point Method for Optimal Power Flow

The interior point method is considered one of the most efficient methods for solving different types of optimal power flow problems. In this article, a weighted multiple predictor-corrector interior point method is proposed and applied to solve two non-linear optimal power flow problems, which include the generation cost minimization problem and the active power loss minimization problem. A two-stage line-search strategy is employed to obtain the optimal composite direction in order to improve the convergence property of the predictor-corrector interior point method. The proposed method is evaluated on three IEEE test systems and three large-scale systems ranging in size from 57 to 2790 buses. Numerical results demonstrate that, compared with the original multiple predictor-corrector interior point method, the proposed method can converge to an optimal power flow solution with a fewer iterations and faster computational time. Moreover, comparison numerical studies show that the proposed method can be faster and more robust than that traditional predictor-corrector interior point method and its variants.