As global industrialization and energy demands rise, excessive reliance on fossil fuels escalates carbon emissions, making clean energy alternatives an urgent priority for sustainable development. As a key transition pathway, wind and solar power can be converted into hydrogen via electrolyzers for electricity generation, thermal supply, or natural gas synthesis. This enables flexible multi-energy coordination and improves overall renewable energy utilization efficiency. However, conventional electrolyzer scheduling approaches typically assume fixed hydrogen production efficiency, failing to account for dynamic variations in operating conditions, efficiency attenuation, and lifetime degradation under fluctuating renewable inputs. This inadequacy compromises the long-term sustainability of green hydrogen systems. To address these challenges, this paper proposes a hybrid AEL-PEM electrolyzer power allocation and operating condition array rotation strategy. Piecewise linear models are established to characterize the efficiency and full life cycle degradation of both electrolyzer types across normal operation, overload, and start–stop transitions. A mixed-integer linear programming (MILP) model is formulated with an objective function incorporating energy purchase costs, start–stop penalty costs, and electrolyzer lifetime degradation costs, and is solved using the Gurobi solver. Simulation validation is conducted using a 24 h typical summer day dataset with a 15 min resolution. Three comparative schemes are evaluated to verify the strategy’s effectiveness in minimizing total system operation costs and enhancing renewable energy utilization efficiency through optimized operating condition management. Results demonstrate that the proposed strategy reduces total system costs by 23%, entirely eliminates renewable energy curtailment, and balances electrolyzer lifespan degradation across all units, collectively advancing the economic efficiency, asset sustainability, and long-term operational reliability of green hydrogen systems.
Ma et al. (Mon,) studied this question.