Research on the Collaborative Operation of Diversified Energy Storage and Park Clusters: A Method Combining Data Generation and a Distributionally Robust Chance-Constrained Operational Model

Energy storage is crucial for enhancing the economic efficiency of integrated energy systems. This paper addresses the need for flexible resources due to high renewable energy integration and the complexity of managing multiple resources. We propose a decentralized collaborative multi-stage distributionally robust scheduling method for electric-thermal systems, incorporating energy storage to mitigate renewable energy fluctuations. Firstly, we model the electric-thermal system with multiple flexible resources. Uncertain parameters of renewables are estimated using conditional generative adversarial networks (CGANs), assuming empirical probability distributions. Secondly, given the distinct operators of electric and thermal systems and information barriers, we develop a data-driven distributionally robust chance-constrained optimization model (DRCCO). This model ensures decentralized collaboration without compromising information security or fairness. Then, we introduce an Alternating Direction Method of Multipliers (ADMM) algorithm with parallel regularization to decouple the model. This approach facilitates rapid solution finding with minimal information exchange. Finally, numerical examples confirm the model’s effectiveness in enhancing system flexibility and ensuring wind power consumption.