Multi-stage cooperative planning among shared energy storage operator and multiple prosumers in regional integrated energy system considering long-term uncertainty

Jiejie Liu,Xianyang Meng,Jiangtao Wu

DOI: https://doi.org/10.1016/j.est.2024.114244

IF: 9.4

2024-10-25

Journal of Energy Storage

Abstract:The regional integrated energy system (RIES) incorporating energy sharing and transaction provides an attractive pathway to reduce energy consumption and emission. However, the long-term uncertainty with region development and conflict of benefit between multiple participants hinder the rational design and optimization for RIES. Therefore, this work proposes a multi-stage cooperative planning framework to deal with long-term uncertainty and profit balance. Firstly, the hierarchical cooperative optimization model is developed, in which the upper layer handles centralized capacity design and lower layer focuses on decentralized operation. Secondly, the multi-stage optimization is formulated as a Markov decision process, in which deep reinforcement learning is employed to learn optimal planning strategy and Stackelberg game-based dynamic pricing method is used to solve the energy trading optimization between multiple participants. Finally, the data-driven scenario generation method, named gate recurrent unit-improved generative adversarial network-self organizing maps (GRU-IGAN-SOM) algorithm, is proposed to depict long-term uncertainty and generate multi-stage stochastic scenario sets. The developed optimization methodology is adopted for a case study of community. The results show that the proposed GRU-IGAN-SOM method could describe the possible future pathways of uncertainty well. Case 1 considering multi-stage optimization and shared energy storage has the best performance compared with other baseline cases.

energy & fuels