Based SVM Algorithm For County Intelligent Monitoring Information Statute

With the rapid development of the national economy and maturity of computer electronic technology,the power grid automation system which is the basis of power system management,and is also constantly developing.It mainly contains the following aspects:The substation is being a large-scale unattended integrated automation transformation,which makes the information of the substation increase rapidly.The remote signaling information of the substation changes significantly compared with the previous one,especially when protect the application of microcomputer.The position signal of the device,the protection signal of each device,the pressure plate information and even the unmanned station of the DC receive more than one hundred signal.Therefore,the monitoring requirements for the operation of the power system are also increasingly accurate,and the abnormal information is also increasing.The main alarm sources for grid equipment divided into two categories: telemetry and remote signaling.It is very important for a device to classify signals from a large number of alarm sources for one signal or multiple signals.Essentially this is a question of categorizing the original text.SVM supervised learning is considered to have a good effect on text classification.Therefore,the classification of historical alarm data of grid equipment by SVM is mainly studied.The classification of alarm text data using SVM include two processes of training and classification.Text training is to learn the relationship between categories and features through training text.Classification is the process of training the model to classify historical alarm texts.The main work of this paper includes:(1)This paper describes that the concept,characteristics,importance and development trend of power grid monitoring based on the State Grid Zhejiang Zhuji Power Supply Company.It points out the difficulties in developing intelligent monitoring of power grids.(2)Master the technology and development tools which will be used in system,and master the development technologies such as.NET,VISUAL STUDIO and SCADA.(3)The device historical alarm information is classified according to a predetermined theme,and a category is determined for each alarm information document.This paper focuses on the application of SVM in the text classification of historical alarm information of power gridequipment and its future development trend.The SVM supervised learning training model for the grid equipment historical alarm data is generated for predicting the category of the alarm content.(4)This paper design and develop the system which called county monitoring and monitoring information intelligence of the State-owned Zhejiang Zhuji Power Supply Company.The system changes the display mode of the alarm information in the previous text,and can display the alarm information in combination with the graphic classification.The system adopts C/S architeture,and the server mainly has functions of data information monitoring,data information analysis,data storage,configuration management,and authority control;the county-level monitoring information intelligent system client function module includes data exhibition,signal display,Information alarm,report management,system settings,alarm management,etc.