System scheduling of the zero-carbon hydrogen-based industrial park via cGAN-distributionally robust optimization: A case study in ordos, China

In the context of China's dual carbon goals, hydrogen-based zero-carbon industrial parks have become a central focus of national strategies for energy conservation and carbon emission reduction. However, the operation of zero-carbon industrial park system (zero-carbon IPS) faces many challenges, such as multiple uncertainties that are difficult to accurately characterize and unclear multi-energy coupling mechanisms. To address this, a distributionally robust optimization approach incorporating data generation techniques is proposed to balance economic performance and operational risk in the scheduling of zero-carbon IPS. Firstly, a conditional generative adversarial network is employed to generate uncertain factor data within the zero-carbon IPS, effectively avoiding reliance on model assumptions and prior distributions, thereby enhancing scenario generation efficiency. Secondly, the deviation between the generated and real-world scenarios is quantified using ϕ-divergence, and a green hydrogen energy certificate market mechanism is introduced. Based on this, a distributionally robust chance-constrained optimization scheduling model is constructed to account for operational risks in the zero-carbon IPS. Finally, the proposed model is reformulated as a mixed-integer linear programming problem using the sample average approximation method and linearization techniques to improve computational efficiency. The zero-carbon IPS in Ordos, China, is used as a case study to verify the effectiveness of the proposed method.