基于网络拓扑的改进图卷积神经网络的电力通信毁伤韧性量化评估方法OA

The current assessment methods for the damage resilience of power networks have problems such as not considering the functional characteristics of the power network itself and not effectively utilizing data information,which makes it difficult to guarantee the reliability of the assessment results.Therefore,a network topology based method for predicting power communication risks and quantitatively evaluating damage resilience is proposed.Firstly,a power communication risk prediction model based on improved graph convolutional neural network and multi head self attention is adopted to achieve quantitative risk warning evaluation.At the same time,a business based power super net-work is constructed to evaluate node importance;Secondly,develop recovery strategies based on the damage form,and calculate the overall performance of the power communication network through simulation of the recovery process;Finally,the PRF curve is used to accurately evaluate the damage resilience of the power network.The experimental results showed that compared with the optimal baseline,the recall rate and F1 score of the proposed model increased by 9.87%and 5.96%,respectively.This model can effectively improve the risk prediction perfor-mance of power communication and achieve accurate quantitative evaluation of the damage resilience of power communication networks.