Digital twin based multi-objective energy management strategy for energy internet

扫码查看

原文链接

NETL
NSTL
Elsevier

外文摘要：Energy management problem (EMP) has been a widely researched topic in optimal operation of Energy Internet (EI). However, the rapid growth in energy network scale and penetration of distributed renewable generations (DRGs) bring new challenges to energy management. Therefore, a digital twin (DT) based parallel energy management strategy is proposed for the large-scale EI which consists of We-energy (WE). Firstly, a parallel energy management framework is proposed. By establishing this triple parallel structure, states of energy networks can be observed realtimely, which enables flexible responses to fluctuations of DRGs and energy plug-and-play. Abandoned renewable energy is taken into account in the optimization model, which promotes the utilization of renewable energy. Then, a multi-timescale optimization strategy is proposed to handle different timescales of multi-energy networks. Furthermore, for better obtaining and processing information and avoiding dimensional curse, a DT based deep Q-learning algorithm (DQN) is proposed. Eventually, compared with the traditional benefit consensus based strategy, the simulation verifies the effectiveness of the DT based parallel energy management strategy.

外文关键词：

Digital twinParallel systemMulti-objective energy managementDeep reinforcement learning

作者：

Danlu Wang、Ruyi Fan、Yushuai Li、Qiuye Sun

展开 >

作者单位：

College of Information Science and Engineering, Shenyang, 110819, Liaoning Province, China

CRRC Nanjing Puzhen Co., Ltd., Nanjing, 210032, Jiangsu Province, China

College of Information Science and Engineering, Shenyang, 110819, Liaoning Province, China, Department of Informatics, University of Oslo, 0316 Oslo, Norway

出版年：

2023

DOI：

10.1016/j.ijepes.2023.109368

International journal of electrical power and energy systems

ISSN：0142-0615

年,卷(期)：2023.154(Dec.)

参考文献量35