Research On Power Grid Integrated Datanetwork Operation Situation Assessment And Forecast

Situation awareness, can be subdivided into situation assessment and situation prediction, is the overall status of the network and possible trends composed by the structure and configuration of network, the use of network resources, the performance and business factors of network showing. By evaluating the current situation and predicting the future situation trends of network, can provide a way to understand the behavior of network and network user deeply for network administrators.Power integrated data network, worked as support network for the informatization of the power grid, which bears the transmission of all the information services like internal high-speed data, voice and multimedia services of the power system, is presenting some new features: large network scale, complex network topology, variety of business, rich network protocol. In order to ensure the normal and efficient operation of power integrated data network, grasp the status and the future trend of the network operation situation accurately is the most important concern for the network administrator. Thus, it is an effective way establishing a system for situation assessment and prediction of the power integrated data network.In this thesis, through gathering the SNMP protocol information of the backbone network of the power integrated data network, extracting the network situation awareness factors, and classifying the network awareness factor vector into different assessment and forecasting the time series of factors, to achieve the purpose of assess the entire network current operating situation and predict the future situation. The specific work is as follow:1. Research the network situation awareness system model, present a situation assessment and prediction model based on backbone router operating status,which take SNMP protocol information of the backbone network routers as the data sources, create a network awareness factor index system, and finally implement a network operation assessment and prediction system for power integrated data network.2. Research the method of situation assessment, present a method to assess the operation situation based on K-means clustering pre-label and support vector machine for power integrated data network backbone network, which combined the expertise and K-means clustering to label the situation awareness factors. This method is to solve the problem of a large number of situation factors samples without status label, while using the supervised classification to assess situation.3. Research the method of situation prediction, present a method to predict the operation situation based on the accumulated error correction and support vector machine to overcome the disadvantages of time series forecasting. In this method, do time series forecasting of situation factors and prediction error values after accumulation and normalization synchronously to improve the prediction accuracy. The error prediction values are used to modify the preliminary situation prediction values and reduce the errors caused by the lag of support vector machine prediction and the sample mutation fluctuations.Based on the existing research on situational awareness, this thesis presents a operation situation assessment and prediction model for power integrated data network. In this model, a method to assess the operation situation based on K-means clustering pre-label and support vector machine is applied to make assessment, and a method to predict the operation situation based on the accumulated error correction and support vector machine is applied to make prediction. Through experimental analysis and system implementation, the model and method in this thesis has a certain practical value.