Research Of Network Fault Analysis And Health Assessment Technology For Manufacturing Enterprises

In recent years,with the size of manufacturing enterprise network continues to expand,the network management becomes more and more difficult.As industrial network under the special environment,it is easy to cause equipment fault,which will bring economic losses to the enterprise or even casualties.How to manage the network effectively is the focus of the research in the field of network security.Network fault diagnosis and health assessment technology is the intelligent technology that can determine the type of fault type and assess the current health of the system based on the current and historical data of system.We can quickly find the cause of the fault and grasp the health status overall,so as to provide decision guidance to maintenance personnel.This paper focus on study the Industrial Ethernet network fault diagnosis and health assessment technology under the support of "863" Projects.The main research is as follows:Firstly,design the integrated monitoring management platform,this platform is responsible for data collection of device.The advantage of the platform is that it collections the parameter information of each device by means of the plugin installed in the management terminal,which makes the management of the device more flexible.Secondly,an intelligent fault diagnosis algorithm(CSRF)based Random Forest is proposed according to the data characteristics of manufacturing enterprise network.The algorithm improves the Random Forest from two aspects:sample sampling and model combination.The former uses the classified sampling technology to generate the training samples separately for each basic classifier,which alleviates the problems caused by sampling bias and data imbalance.The latter takes into account both the number of votes and the confidence of the basic classifier,which improves the accuracy of diagnosis.Thirdly,a muti-neural network fusion health assessment algorithm(MNN)is proposed based on the existing problem of the current network health assessment technology.The algorithm takes full consideration of single point features and link features,and it uses convolution network and BP network to model the features of different latitudes of network so as to assess the health status of network.Fourthly,the validity of the proposed algorithm is verified through the experimental design and the result analysis.