Distributed multi-energy storage is critical for enhancing the operational flexibility of power systems with high renewable energy penetration. However, the heterogeneity between electric and thermal energy storage in dynamic response, multi-time-scale features and coupling interactions causes three key limitations of existing aggregation methods: conservative feasible-region characterization, excessive computational complexity, and inaccurate dispatch capability evaluation. To address these issues, this paper proposes a distributed multi-energy storage aggregation method based on support functions and the Minkowski sum. First, a unified convex polyhedral feasible-region model incorporating electro-thermal coupling constraints is established. Then, a distributed parallel aggregation strategy with constraint relaxation–reconstruction and adaptive optimal direction selection is developed to achieve high-precision approximation of the aggregated feasible region. Case studies on a park-level integrated energy system with 10 electric and 5 thermal storage units show an average support function error of 1.35%, a volume similarity ratio of 0.953, and a dispatch feasibility rate of 98.70%. Electro-thermal coupling reduces feasible-region volume by 32.50% with only 5.20% overestimation. Under fluctuating commands, the tracking deviation is 2.80% with no constraint violations. Furthermore, measured engineering data verify its linear scalability for large clusters and 1.27% cost deviation in day-ahead dispatch. The proposed method accurately characterizes the joint regulation capability of multi-energy storage clusters while ensuring favorable computational scalability.
Gao et al. (Thu,) studied this question.