Gaussian Activation Function Based Neural Network Control Strategy for Single Stage PV - Battery Integrated Three Phase Distribution System

This paper proposes gaussian activation function based neural network control strategy of solar photovoltaic (PV) battery system assimilated to a distribution system. At the DC link of voltage source converter (VSC), the battery is incorporated through a bidirectional converter, which is charged and discharged during base load and high load demand. The active current constituent of nonlinear load current, is evaluated with neural network based control strategy, to make the grid current harmonic free and improves the power factor. The incremental conductance (I&C) method is utilized, for the extraction of maximum power point tracking (MPPT) of a solar PV array. The enhancement in dynamic performance of system in adaptable atmospheric conditions, is realized by utilizing the feed-forward constituent of solar PV power (FFSPV). The system with a neural filter control, operates successfully at solar insolation change and load perturbation. The operation of scheme under constant power mode and adaptable power mode, for steady state and dynamic states,