Application of machine learning decision tree based power switching control for a standalone PV system with grid...

Area of Science:

• Electrical Engineering
• Renewable Energy Systems
• Artificial Intelligence in Energy

Background:

• Standalone photovoltaic (PV) systems face challenges in reliable power delivery due to intermittent solar irradiance and dynamic load variations.
• Conventional switching control schemes in PV systems often result in reduced efficiency and suboptimal performance under fluctuating conditions.
• Integrating machine learning for real-time switching control in grid-connected standalone PV systems is an emerging research area.

Purpose of the Study:

• To address the limitations of conventional control schemes in standalone PV systems by developing an ML-based switching strategy.
• To enhance power management and system efficiency through intelligent, real-time switching decisions.
• To enable autonomous operation of PV systems, utilizing grid backup only when solar energy is insufficient.

Main Methods:

• Development of a machine learning Decision Tree (DT) based switching strategy for standalone PV systems.
• Training the DT model using real-time monitored features: PV array voltage, current, irradiance, and load demand.
• Simulation of the proposed control framework using MATLAB/Simulink to evaluate performance.

Main Results:

• The proposed ML-DT switching strategy automates the switching operation between PV and grid supply.
• Demonstrated reduction in switching losses compared to conventional rule-based methods.
• Achieved improved voltage regulation in the standalone PV system.

Conclusions:

• The ML-DT based switching strategy effectively enhances power management and efficiency in standalone PV systems.
• The developed framework offers a promising solution for reliable power delivery by intelligently managing energy sources.
• The study highlights the potential of machine learning in optimizing autonomous renewable energy systems.