A normalized neural network based controller for power quality improved grid connected solar PV systems

This paper is aimed at normalized neural network based control scheme for power quality improved integration of solar PV and utility grid. In the proposed scheme the solar PV and grid are integrated using a three leg voltage source converter consists of six IGBTs, three interfacing inductors and a DC bus capacitor. The neural network based scheme is used for estimating fundamental real and reactive power components of the load current in all three phases independently therefore the three phase grid current remains balanced and sinusoidal under all type of loading conditions including unbalancing in load currents of three phases. The proposed controller mitigates the harmonic current, compensates reactive power need of the system, improves system power factor and regulates the system voltage at the point of common coupling (PCC).