This technology develops a multi-time-scale voltage collaborative regulation technique, integrating multi-agent reinforcement learning with attention mechanisms to establish a threelayer risk assessment system. Through UML modeling and decision analysis methods, it achieves dynamic optimization of discrete/continuous equipment, enabling precise voltage control and power flow regulation in distribution networks. The technology addresses the challenge of heterogeneous equipment coordination, significantly enhancing grid safety, stability, and economic efficiency.

Technology provider:Nanjing University of Posts and Telecommunications