Health-Aware Distributionally Robust Scheduling of Integrated Electro-Hydrogen Systems Considering Electrolyzer Degradation Inertia and Recovery

Vollständiger Abstract

Worum geht es in dieser Arbeit?

Renewable-driven operation exposes proton-exchange membrane (PEM) electrolyzers to ramps, starts, and partial-load conditions that accelerate degradation and weaken scheduling reliability. This paper develops a health-aware scheduling framework for an integrated electro-hydrogen system. It represents operating stress, delayed response, irreversible degradation, recoverable performance loss, and efficiency feedback. Scheduled rest partially relaxes the recoverable state, while per-unit health budgets yield degradation shadow prices that redirect load from health-scarce stacks. A Wasserstein distributionally robust model implements a health-conditioned risk-aversion policy by adjusting the protection radius with pre-horizon fleet health and filtered stress. The tractable finite-support formulation is evaluated through progressive ablations and five uncertainty treatments using chronological Liaoning wind, solar, and load data. Rotational recovery provides the main degradation mitigation, whereas shadow prices primarily improve allocation among heterogeneous stacks. Compared with fixed-radius DRO, the health-conditioned policy reduced the point estimates of CVaR95, energy-violation rate, and severe health-budget exceedance by 3.91%, 1.68 percentage points, and 3.36 percentage points, respectively. These results indicate the value of coordinating equipment health and uncertainty protection in short-term electro-hydrogen scheduling.