基于RBF神经网络的主动配电网通信过程安全态势感知方法

A Security Situation A wareness Method of Active Distribution Network Communication Process Based on RBF Neural Network

为提前拦截主动配电网通信过程中的恶意攻击、保证电网稳定运行,提出基于RBF神经网络的通信过程安全态势感知方法。综合考虑电压越限、失负荷、供电能力等因素,筛选主动配电网通信安全态势评价指标;利用非线性基函数实现从输入到输出的非线性变换,采用改进ReliefF算法选取网络通信入侵行为特征,并将其作为RBF神经网络的输入量;将主动配电网络节点部署的安全态势数据分割为若干分量,通过参数调整平滑处理安全态势感知数据,再利用双层滤波检测法输出RBF神经网络的感知结果。由试验结果可知:采用该方法时,特征识别准确率为97.61%,安全态势值为0.894,态势感知时间为0.361 s,能有效感知攻击行为,精准呈现网络态势变化,且感知过程用时较短。

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.