A Comparative Evaluation of Hybrid GEO-RBFNN and GBDT-BCMO Methods for Power Quality Improvement using Renewable Energy Resources in Distributed Generation

Abstract

This research presents a hybrid strategy to improving power quality via the combination of distributed systems that use solar Photovoltaic (PV) and battery storage. The Golden Eagle Optimizer (GEO) and the Radial Basis Functional Neural Network (RBFNN) are both part of the suggested hybrid technique. This method is often referred to as GEO-RBFNN. Here, the RBFNN approach is used to train the inputs using the target reference power of the prior instantaneous energy sources and the present load needs. In order to provide the best control signal and keep the Hybrid Renewable Energy Sources (HRES) running, the suggested technique uses load fluctuation to determine the PI controller gain settings. In GBDT-BCMO, we aim to achieve the objectives mentioned before. Applying the BCMO approach improves the GBDT machine learning technology. Inputs used to train the GBDT technique include the goal reference power, the present time necessary load demand, and the instantaneous energy of available sources in the past. When making predictions, the GEO-RBFNN approach takes into account all possible fluctuations in system characteristics, including active and reactive power, DC voltage, and more. Consequently, the suggested GEO-RBFNN technique produces the specified line voltage for reactive power compensation while minimizing power system damping. After running the suggested technique on the MATLAB/Simulink environment, its performance is evaluated in comparison to other ways that are currently available. By generating line voltage and enhancing power system damping, the GBDT-BCMO system offers reactive power compensation.