Design optimization of a multi-source renewable energy system using a novel method based on selective ensemble learning

This study conducts a comparative assessment to optimize the design of a Hybrid Renewable Energy System (HRES) consisting of PV panels, wind turbines (WTs), and hydrogen storage (PV/WT/FC). The study focuses on Dakhla City, leveraging its favorable weather conditions to evaluate the potential of renewable energy sources. A novel approach, Selective Ensemble Marine Predators (SEMPA), is introduced. SEMPA utilizes a selective ensemble learning strategy to enhance the performance of the Marine Predators Algorithm (MPA) and determine the optimal system size with the lowest Total Net Present Cost (TNPC) and improved reliability. To benchmark SEMPA's performance, we compare it against established methods, including MPA, Particle Swarm Optimization (PSO), and Harmony Search (HS). Results demonstrate that the integrated PV/WT/FC system provides a reliable and cost-effective energy solution for the southwest region of Morocco, particularly in Dakhla. SEMPA outperforms HS, PSO, and traditional MPA, achieving stable convergence after 30 iterations. The system exhibits a superior Cost of Energy (COE) at 0.3127 $/kWh, a TNPC of 0.9689 M$, and a low Loss of Power Supply Probability (LPSP) at 0.04747.