MODELING A SHUNT ACTIVE FILTER USING DIFFERENTIAL EVOLUTION

Abstract

DOI: 10.14684/intertech.13.2014.105-110

This paper proposes the modeling of a Parallel Active Filter (PAF) using a search algorithm to define the optimum parameters of inductance (Lf), voltage (Vdc) and bandwidth (Bw), since these parameters are mandatory to improve the performance of the modulation using hysteresis. The design of FAP devices may be obtained through the search algorithm of optimal points. The simulation prevailing contributes to the security of the system that should receive the compensation device as the design considers the current load. After finding the parameters through the optimization technique (based on an evolutionary algorithm named Differential Evolution), the active filter performance was evaluated and some simulation results proved the validity of the proposal. The association of DE with instantaneous power theory berfore utilized in active parallel filters, was capable mitigate the currentes harmonics so traht eliminate the unbalance between phases, presenting satisfactory results. Index Terms - Differential evolution, parallel active filter, parameters optimization, instantaneous power theory.