Research On Network Associated Error Warning Information Analysis System

The task of Telecommunication Networks Management (TNM) mainly focuses on the real-time monitoring of the network status, which ensures the network to be stable, efficient, and reliable during daily maintenance. Meanwhile, as the mobile communication technology and the mobile services develop rapidly, the scale and the frequency of alarm rapidly increase accordingly, which cause it more complex to locate the accurate fault device.According the issue of mobile communications network, this thesis focus on the study of following aspects:Firstly, a novel multi-level fault alarm model is presented by studying the distributions of fault alarms and the model of the distribution. Gaussian perception network is used so that the alarm data artificial neural network to deal can be reduced and the efficiency can be raised.Secondly, BP neural network is chose to classify the fault alarm data by comparing various training methods of artificial neural network. BP neural network have adaptive ability and pattern recognition capability, so that the fault of GSM network is monitored and quickly located in the multi-level algorithm. Various methods such as momentum method, adaptive learning rate and Levenberg-Marquardt method are used to train the BP neural network.Finally, the fault alarm correlation analysis system of mobile communication network is designed and implemented. The system is based on Client/Server architecture, and a distributed alarm collection schema is used to monitor all the alarms produced by radio network, switch network and transfer network in the mobile network and collect them by the device types and their alarm acquisition mechanism. And, the coupling between modules can be reduced and the scalability and stability of overall system can be improved using hierarchical solution and Chain of Responsibility design pattern.