Coordinated dispatch of alkaline and PEM electrolyzers: A stochastic optimization approach under renewable energy uncertainty

To enhance the efficient utilization of wind and photovoltaic (PV) power, this study proposes a stochastic optimal dispatch strategy integrating alkaline (ALK) and proton exchange membrane (PEM) electrolyzers. Differentiated coordination mechanisms are designed for two typical scenarios of renewable energy, namely “surplus” and “scarcity”: in surplus scenarios, both types of electrolyzers produce hydrogen simultaneously to accommodate curtailed electricity; in scarcity scenarios, ALK units maintain hot-standby while PEM units shut down to optimize power demand. To address the uncertainty of renewable energy, a two-stage method of “generating candidate scenarios via Conditional Generative Adversarial Network (CGAN) — reducing scenarios via improved K-means clustering” is adopted, where the Mahalanobis depth function is used to adaptively determine the number of clusters, and the weighted Euclidean-Cosine distance is employed to ensure scenario representativeness. A mixed-integer linear programming (MILP) model is constructed to minimize the total operating cost, which includes the costs of electrolyzer operation, battery charging/discharging, grid power purchase, renewable energy operation and maintenance as well as curtailed electricity, while accounting for hydrogen sales revenue. Verified with actual data from a power system in Northwest China, the results show that: compared with single-electrolyzer schemes, the proposed strategy increases the renewable energy utilization rate by more than 1.2% and reduces the total operating cost by 7.4%–10.8%; the characteristic complementarity of ALK and PEM effectively enhances the flexibility and resilience of power systems with high-proportion renewable energy penetration.