A Regional Voltage Control Strategy for Distribution Network Based on Multi-Agents

Distributed generation not only increases the power supply but also raises the complexity of operating and controlling the distribution network. To fully utilize the regulatory capabilities of various types of regulatory resources in the distribution network and enhance the regional autonomy of the distribution network, particularly in the context of high new energy integration, a regional collaborative voltage control strategy for the distribution network is proposed based on multi-agent deep reinforcement learning (MADRL). Firstly, a distribution network zoning method that is easy to implement is designed to achieve the division of distribution network zones. Then, multi-agents are employed to model the problem of regional collaborative voltage control of the distribution network, and deep reinforcement learning is used to solve the problem of regional collaborative voltage control of the distribution network. Finally, a test network is used to verify the advantages of the proposed zoning method and regional voltage control method in promoting the voltage control performance of the distribution network.