A Security Situation Awareness Method of Active Distribution Network Communication Process Based on RBF Neural Network
In order to intercept malicious attacks in the communication process of active distribution networks in advance and ensure the stable operation of the power grid,a communication process security situation awareness method based on RBF neural network is proposed in this paper.First,the evaluation indexes of communication security situation of active distribution networks are selected by considering the factors such as voltage off-limit,load loss and power supply capacity.Second,the nonlinear basis function is used to realize the nonlinear transformation from input to output,and the improved ReliefF algorithm is used to select the network communication intrusion behavior characteristics,which are taken as the input of RBF neural network.The security situation data of nodes deployed in active distribution networks are divided into several components,and the security situation awareness data are processed smoothly by parameter adjustment,and the perception results of RBF neural network are outputted with the double-layer filtering method.The test results show that with this method,the accuracy of feature recognition is 97.61%,the security situation value is 0.894,and the situation awareness time is 0.361s.It can effectively perceive the attack behavior,accurately present the changes of network situation,and the perception process takes a short time.
RBF neural networkactive distribution networksituational awarenessevaluation indexsmoothing
NETL
NSTL
万方数据
