Control of A Multifunctional PV-Integrated SAPF with Simpler Neural Harmonic Extraction

This study presents a new, efficient Neural Synchronous Reference Frame technique (New-Neural-SRF) for the control of a three-phase PV-integrated a Shunt Active Power Filter system (PVSAPF). This multifunctional model guarantees an uninterrupted power injection from the photovoltaic panels, and/or the AC source according to the irradiance profile variation. Further, the model offers additional services as harmonic attenuation, reactive current compensation and DC-link voltage regulation. The effectiveness of the PV-SAPF compensation is principally related to the control method which first extracts the load harmonic currents, and then generates the compensating reference current. Therefore, in order to enhance the harmonic extraction algorithm and to deal with the limitation of the classical SRF technique, mainly under online varying system parameters, a novel and a simple Neural-SRF controller using a feed forward backpropagation artificial neural network is proposed. The suggested method of control is tested under a MATLAB/Simulink environment. The results validate the robustness of this Neural technique to ensure an accurate and a fast harmonic-extraction, fewer overshoot, a quick response time and lower value of the Total harmonic distortion THD, in accordance with the IEEE-519 standard requirements, under different simulation cases.

I. Introduction

Power quality improvement in electrical power systems and electrical devices has acquired utmost attention and has become an increasing concern over the eras [1, 2].