基于自适应联邦学习的输配网动-静态综合状态估计方法研究OA

As the operational characteristics of power systems undergo profound changes,traditional segregated state estimation methods for transmission and distribution networks struggle to achieve effective coordination.Additionally,with the rapid development of the internet,the problem of data privacy breaches in power systems has become increasingly prominent.To address these challenges,this paper proposes a self-adaptive federated learning-based integrated dynamic-static state estimation method for transmission and distribution networks.This method enhances the coordinated control of transmission and distribution networks,enabling rapid state sensing and privacy protection.First,this paper proposes an integrated dynamic-static state estimation method for transmission and distribution networks.Based on Monte Carlo simulations,state estimation dataset is constructed.Then,a self-adaptive federated learning model is developed and trained on the power system state estimation dataset.Finally,the proposed method is validated by simulations on the IEEE 39-and 123-bus systems.The results show that,compared with the traditional federated learning method,the proposed method improves the accuracy of voltage angle and magnitude state estimation by 10.73%and 13.49%,respectively,demonstrating the effectiveness of the method.