Enhancing voltage stability in photovoltaic and wind micro grids with a hybrid optimization approach

Abstract

Renewable energy sources (RES) like photovoltaic (PV) and wind power integrate into micro grids, and maintaining stable DC link voltage is crucial for efficient power distribution. This paper presents a hybrid approach to optimize DC link voltage in PV and wind micro grids, addressing the challenges posed by the intermittent nature of RES. The proposed method combines Sooty Tern Optimization Algorithm (STOA) and Augmented Physics-Informed Neural Networks (APINN). The aim is to improve stable DC bus voltage and enhance power quality in micro grids. STOA optimizes gain parameters of the Fractional Order Proportional-Integral-Derivative (FOPID) controller, while APINN predicts optimal voltage, improving PV and wind system efficiency during load changes. The approach, implemented in MATLAB, is compared with Dandelion Optimization (DO), Artificial Neural Network (ANN), and Deep Residual Network (DRN) with Green Anaconda Optimization (GAO). Results show a maximum conversion efficiency of 97%, response time of 0.97 s, and improved error metrics, outperforming existing methods.