Bi-level capacity optimization of a wind–solar–hydrogen microgrid based on an improved pelican optimization algorithm

• A bi-level capacity–operation co-optimization framework is proposed for wind–solar–hydrogen microgrids considering electrolyzer start–stop degradation. • An optimized electrolyzer start–stop scheduling strategy significantly reduces cycling frequency while maintaining high hydrogen output. • An improved Pelican Optimization Algorithm (IPOA) enhances global search capability and convergence speed for capacity configuration problems. • Case studies verify reduced wind–solar curtailment, lower levelized hydrogen production cost, and improved overall system performance. In wind–solar renewable hydrogen production systems, power fluctuations and electricity shortages can trigger frequent alkaline electrolyzer (AEL) start-stop cycles, accelerating lifespan degradation. Existing capacity configuration methods alleviate wind–solar power fluctuations and improve renewable energy utilization by increasing energy storage, introducing operational scheduling, or applying optimization algorithms. However, insufficient coupling between planning and operational stages makes it challenging to simultaneously balance economic viability, AEL start-stop behavior, and lifespan constraints. To address this, this paper constructs a bi-level capacity optimization model for a wind–solar hydrogen production DC microgrid including wind turbines, photovoltaic units, energy storage systems, and AEL, based on capacity configuration and operational characteristics. The upper-level model maximizes annualized revenue and minimizes wind–solar curtailment to achieve coordinated optimization of system resource allocation and operational strategies. The lower-level model minimizes AEL start-stop cycles and maximizes hydrogen production, proposing an optimized start-stop scheduling strategy to realize synergistic optimization of capacity allocation and power distribution. During solution, an improved Pelican optimization algorithm based on Tent–Levy mapping is adopted, yielding a capacity configuration that balances economic performance, energy efficiency, and AEL lifespan under practical conditions. Case study analysis demonstrates that under identical wind–solar output, the proposed bi-level scheme outperforms a direct wind–solar supply scheme and conventional capacity configuration with operational scheduling, ensuring hydrogen production, improving economic performance and equipment lifespan, reducing wind–solar curtailment by 3.89%, and significantly enhancing overall microgrid performance.