Abstract
Nowadays, with a large of complicated geography of Distributed Energy Sources (DES), how to integrate distributed renewable energy source and reduce the operational costs by Virtual Power Plant (VPP) becomes a mainstream problem in Internet of energy. The traditional method of energy integration and operational cost optimization utilizes the cloud computing technology to centralized control the computational task, which increases the burden of computing. According with the development of information communication technology, such as Internet of Things and 5G, edge computing technology is an effective way to offload computational task to the edge side of 5G networks. Moreover, with the increase of collected data, it becomes a key point to effectively improve the computing power of edge nodes in edge computing. Currently, machine learning is an effective way to process the big data. Based this situation, it leads the combination of machine learning and edge computing. In this paper, the Edge Intelligence (EI) structure is proposed to solve the Economic Dispatch Problem (EDP) in VPP of Internet of Energy. Compared with the traditional edge computing, the proposed EI structure inherits its original features which reduce the burden of cloud computing, and also the proposed EI structure improves the computational power of edge computing. Through the splitting model and deploying the particle model in the terminal, it is facility to real-time control and take the less costs of VPP. Due to the transmission between the splitting models with counterpart, it transmits the part information and gradient information, which effectively reduces the consumption of communication. The proposed method has verified the effectiveness and feasibility through the numerical experiments of real application data sets.