Interior/exterior-point methods with inertia correction strategy for solving optimal reactive power flow problems with discrete variables

Abstract

Interior/exterior-point methods have been widely used for solving Optimal Reactive Power Flow problems (ORPF). However, the utilization of such methods becomes difficult when transformer taps and/or capacitor/reactor banks are more rigorously represented in the problem formulation by means of discrete control variables. This work investigates the solution of the ORPF problem when transformer tap ratios are modeled as discrete variables. The solution method proposed handles discrete variables by means of sinusoidal penalty function, while the penalized problems are solved by an exterior-point method. An inertia correction strategy is proposed in order to assure that only local minima are obtained for the penalized problems. New search directions are also investigated that combine predictor and corrector directions. Numerical simulations are performed involving the IEEE 14, 30 and 57 bus systems. The results show the efficiency of the proposed inertia correction strategy and also reveals that the proposed exterior-point method outperforms traditional interior-point methods in terms of the number of iterations and computation times.