Artificial Neural Network-Based Modeling and Stability Assessment of Solar PV-Integrated Microgrids

Abstract

The characteristics of photovoltaic (PV) distributed generation and its modeling using Artificial Neural Networks (ANN) have been presented. The proposed NARX-ANN control-based feedback design incorporates a recurrent feedback loop, allowing the weights matrix to adapt within the range of error variations. This controller design is also applied to the dynamic behavior of the Phase-Locked Loop (PLL) between the converter station and the Point of Coupling (PC). The proposed controller model is derived from a set of differential equations. The performance of this control system is validated through a small-signal model, and dynamic oscillation damping is verified in terms of load variations and stability, assessed using a root locus plot with the aid of MATLAB/Simulink software.