Cloud-Edge Collaboration Based Operation Strategy in Medium-Low Distribution Network

Abstract

The escalating demand for electricity presents significant challenges to the safe and cost-effective operation of distribution networks. To address these challenges, this paper proposes a framework that leverages cloud edge collaboration to optimize the operation of medium and low voltage distribution networks and improve their power flow. The framework is structured into two levels: the centralized reconstruction level of the medium-voltage distribution network, and the distributed control level of the low-voltage distribution network demand response. The sequential reconfiguration of distribution networks is evaluated using the discrete soft behavior method, and the network radial problem is effectively solved through the action mask technique. The edge gathers resources by collecting vast amounts of local user-side demand response data. With the aid of the cloud's powerful computing resources and the edge's rapid response speed, the deep reinforcement learning model is trained in the cloud center, and a real-time demand response strategy is created at the edge. The proposed framework utilizes cloud-edge collaboration to achieve a refined regulation of distribution networks. The effectiveness of the proposed method and model is assessed in a 445-node distribution network, and the results confirm the potential of the framework to optimize distribution network operations.