Study On High - Frequency Current And Ultrasonic Signal And SVR Estimation Model

Switchgear partial discharge is the important factor to cause switchgear insulation degradation and power accident. According to the on-line monitoring of switchgear partial discharge, we can detect the insulation defects which exist in the switchgear to avoid major accidents. This paper is based on the background of the above research and designs a set of on-line monitoring system by using high-frequency current sensors and ultrasonic sensors. And it makes a remote on-line monitoring to the changes in switchgear partial discharge of Chuxiong 220kV Waist Station substation. The estimated model based on support vector machine for regression is used to analyze and estimate the amount of switchgear partial discharge. My major works are as follows:1.According to the analysis of the partial discharge monitoring parameters, this paper selected high-frequency pulse current and ultrasonic as partial discharge monitoring parameters. Summarizes the wireless sensor network architecture and takes into account the limitations of the electrical equipment operating environment. Wireless sensor technology is used for real-time, online, remote, dynamic monitoring in partial discharge of electrical equipment.2.This paper establishes a set of sensor network based on wireless sensor network switchgear partial discharge monitoring. Sensor network uses IEEE 802.15.4 wireless communication protocol to achieve automatic network and could accommodate multiple sensor monitoring nodes. A fixed IP address is set for the server to implement the connection between server and peripheral internet, then the functions of real-time, online, remote, dynamic for switchgear partial discharge monitoring system can be implement.3.Through the analysis of the demand of the partial discharge monitoring system, this paper designs and develops a partial discharge monitoring system. Event-driven mode is used to deal with the interaction of serial ports, and multi-threading technology are adopted to write the communication protocol and analysis procedures, the ODBC database access technology are selected to access the database, then the analysis of data is displayed on the real-time monitoring interface.4.Through the sample data to train SVR model, the test sets are used to validate the accuracy of estimated model. The data of high frequency pulse current and ultrasonic are selected which measured by the switchgear partial discharge monitoring system, and put them into SVR estimated model which has been established, then simulate switchgear partial discharge value. It provides early warning support for switchgear while it emerge partial discharge.