Study on the Impact of Fuzzy PI and Artificial Neural Network on a Hybrid Indirect Matrix Converter with a Flying Capacitor Inverter Control in Active Power Filtering Application

Khadidja AMEUR, Fatima AMEUR, Takieddine AMEUR, Aissa AMEUR

Abstract

This paper investigates the implementation and comparative performance of a Fuzzy PI controller and an Artificial Neural Network (ANN)-based controller applied to a Hybrid Indirect Matrix Converter (HIMC) and to a Flying Capacitor Inverter (FCI) with its output RL filter, for active power filtering applications. The main objective is to achieve an accurate reference current tracking under DC voltage perturbations. A RLC circuit and a three-phase full-wave converter are employed for generating a fluctuating DC input voltage for the inverter. The FCI is designed to produce a compensating current equal in magnitude and opposite in phase to the harmonic components, while its semiconductor switches must withstand the applied voltage stress. The inverter’s dynamic response, governed by the semiconductor switching behavior, is inversely related to the supported voltage. In order to analyze and validate both control strategies, simulations were carried out in the MATLAB/Simulink environment using the Simscape and SimPowerSystems toolboxes. The results confirmed that both controllers effectively improved the filtering performance, with the neural controller providing a higher accuracy and the fuzzy PI controller enabling a simpler implementation.

Keywords

Flying Capacitor Inverter (FCI), Hybrid Indirect Matrix Converter (HIMC), Fuzzy PI Controller, Artificial Neural Network (ANN), Active Power Filter, Total Harmonic Distortion (THD), Intelligent Control, Harmonic Compensation.

View full article