Performance Evaluation of Neural Network Controlled Grid-Connected Photovoltaic System for Power Quality Enhancement

This paper studies a neural network (NN) control for three-phase grid-connected Photovoltaic (PV) system. This study aims to improve the performances of the system, ensure a high quality of the Total Harmonic Distorsion and extract the maximum power while ensuring unity power factor. A neural network controller is used instead of the two PI regulation loop for grid currents. The Levenberg Marquardt (LM) algorithm is used to train the NN controller and the validation and training data of the utilized neural controller are obtained by simulation of the entire system with the PI controllers calculated for variation of solar irradiance. The simulation study shows that the neural controller indicates significantly more enhanced performance than that of the PI controller, including faster response times, lower overshoot and Total Harmonic Distortion (THD).

Paper

Full text

PDF

Performance Evaluation of Neural Network Controlled Grid-Connected Photovoltaic System for Power Quality Enhancement

Semantic Scholar · Engineering · 2019

Abstract

This paper studies a neural network (NN) control for three-phase grid-connected Photovoltaic (PV) system. This study aims to improve the performances of the system, ensure a high quality of the Total Harmonic Distorsion and extract the maximum power while ensuring unity power factor. A neural network controller is used instead of the two PI regulation loop for grid currents. The Levenberg Marquardt (LM) algorithm is used to train the NN controller and the validation and training data of the utilized neural controller are obtained by simulation of the entire system with the PI controllers calculated for variation of solar irradiance. The simulation study shows that the neural controller indicates significantly more enhanced performance than that of the PI controller, including faster response times, lower overshoot and Total Harmonic Distortion (THD).

Similar papers

© 2026 NYSGPT2525 LLC