Modeling and Simulation Using Artificial Intelligence Techniques to Analyze the Performance of Independent Hybrid Renewable Energy Systems

Abstract

This research analyzes the performance of an independent hybrid renewable energy system combining wind and solar photovoltaic power using artificial intelligence techniques to optimize its performance. Hybrid systems represent an ideal solution for meeting energy needs in remote or off-grid areas.

We modeled the various system components, including wind turbines, solar panels, DC-DC converters, and energy storage batteries, using Simscape boxes in the MATLAB/Simulink environment. We also incorporated variable data for solar irradiance and wind speed to simulate realistic operating conditions. Neural networks (ANNs) were used to estimate the duty cycle of the DC-DC converters to optimize performance. The impact of this value on the system was then analyzed, and the results were compared with the traditional Incremental Conductance (InCond) method without modifying the system architecture. The results showed that using neural networks to estimate the Duty Cycle value contributes to increasing the efficiency of the hybrid system, confirming its effectiveness as a tool for performance analysis and optimization. It provides a faster and more stable response to power curves compared to the traditional InCond method, while reducing ripple and increasing the system's actual efficiency. The results also indicated that artificial neural networks adapt better to sudden changes in solar irradiance and wind speed, enhancing system efficiency in remote and off-grid areas and ensuring long-term power generation stability.

These results confirm that integrating artificial intelligence techniques with traditional methods for maximizing power point (MPP) enhances the overall performance of the standalone hybrid system and provides an effective solution for sustainable energy applications in remote or off-grid areas.