Adaptive Neural Network Based Control of PV Connected Distribution System

Abstract

This work presents an adaptive neural network-based control approach for maximum power extraction and injection of active power in a photo-voltaic (PV) array connected distribution system. The control algorithm based on adaptive neural network (ANN), improves the performance of a PV connected distribution system. The ANN based control algorithm comprises of the sigmoidal and Gaussian functions for linearizing the load currents within a unity band to improve the performance of PV connected distribution system. This control algorithm calculates the amplitude of active and reactive components of nonlinear load currents to mitigate harmonics and to compensate reactive power for improving the power quality of proposed system. The ANN-based control algorithm has good dynamic response as compared to other ANN-based control algorithm. This algorithm requires only few training layers and unknown weights, which simplifies the control algorithm and reduces its execution time. It also enhances the performance of the system by estimating real time weight and specifies its merits.