Metaheuristic strategies for solving the Optimal Reactive Power Dispatch with discrete variables

Optimal power flow (OPF) is a classical nonlinear, large-scale, non-convex problem. It is used as a tool to determine the optimal operating point of the electric power system, improving its performance. Usually its resolution complexity is rather increased in the presence of discrete control variables, which poses as a challenge for the application of general classical non-linear optimization algorithms. In this paper, the Evolutionary Particle Swarm Metaheuristic algorithm that combines features of classical optimization methods and non-deterministic methods is proposed to solve a specific class of OPF problems, the Optimal Reactive Power Dispatch with discrete control variables such as in-phase transformer taps. Two strategies have been used to initialize the control variables. In order to validate the proposed method, tests have been performed on IEEE 14, 30, 57, 118 and 300 bus systems.