Comparative study between Proportional Integral and Neural Networks Controllers of a Shunt Active Power Filter Powered by a Photovoltaic System

Abstract

The objective of this paper is the study of the implementation of compensation algorithms using shunt active power filter with voltage structure, for the mitigation of harmonic currents generated by nonlinear loads in distorted electrical networks. This work presents an approach to reduce the harmonic distortion rate (THD) of the source current and other approaches to improve the power factor by a parallel active power filter. For this, I will present a structure of inverter of two levels on the one hand, and the different control techniques between classical and intelligent on the other hand. To do this, a miniature prototype simulation of a control of a SAPF by (PI) and (ANN) controllers powered by photovoltaic source has been established. These controllers are used with instantaneous active and reactive powers method (PQ). The simulation results allow us to understand the importance of increasing the level of the inverter and the control strategy used for the improvement of the THD and the power factor at the same time.