使用遗传算法优化带储能的独立混合可再生能源系统设计：风能与太阳能时间互补性影响分析

Abstract This study analyzes the impact of temporal complementarity between wind and solar sources on the optimal design of stand-alone hybrid renewable energy systems with storage (HRES). A model was developed in GNU Octave that uses a fixed-seed genetic algorithm to ensure reproducibility and compare scenarios. The objective is to minimize the Net Present Cost (NPC) while complying with a reliability constraint defined by the LPSP (Loss of Power Supply Probability). Constant and variable load profiles are evaluated under different levels of complementarity, showing that their influence depends on the type of demand. Furthermore, a sensitivity analysis is performed on the LPSP, battery cost, and discount rate, demonstrating how these parameters affect the optimal configuration. The results indicate that high complementarity can significantly reduce the NPC, especially in contexts with strict reliability requirements. In environmental terms, an HRES supplying 1470 kWh per day would avoid between 108 and 375 tons of CO 2 per year, compared to a fossil source. These findings are key to energy planning in countries moving toward decarbonization, supporting investment decisions in distributed generation and mitigating the effects of curtailment on centralized systems.