Deep Learning-Based Distributed Optimal Control for Wide Area Energy Internet

Abstract

The concept of energy Internet (EI) emphasizes comprehensive utilization of the whole energy system and has been considered as a new development stage of the smart grid. As the core device of an EI, the energy router (ER) is used to connect various electrical equipment. The access of ERs brings new challenges for the energy management issues in EI. In this paper, an improved ER’s structure is investigated, and the network structure of the wide area EI in an off-grid mode is introduced. In addition, in order to maximize the utilization of renewable energy sources (RESs) and to reduce power transmission loss, we propose a novel distributed energy management approach based on deep learning algorithm. Only the local information exchange among the neighboring ERs is required and the global optimization effect is achieved. Finally, simulation results within a 100-routers test system are provided to illustrate the effectiveness of the proposed approach.