Study On On-line Monitoring And Analysis System Of ZnO Arresters

With the fast development of economy in our country, more and more power energy is needed and the power system is more and more important. Power equipments are basic components of power system and have an important influence on the safety and energy quality. Since1960s, the traditional preventive maintenance mechanism shortcomings appear gradually:power cut, high maintenance cost, low reliability, long test cycle, etc.The development of the sensor technology, GPS and information processing technology makes the state maintenance mechanism which bases on on-line monitoring technology become the research focus. State maintenance can effectively improve the operation stability, reduce power outage time and save social resources, and will be the future developing trend in the field of the overhaul.To limit the lightning overvoltage and operating overvoltage and protect electrical equipments, ZnO arresters are widely used because of its good nonlinear characteristics. Theoretical analysis and the experience show that the change of the resistive component of leakage current can reflect the degree of dampness and aging.In this paper, the models of the ZnO arrester in normal and lightning situations are simulated. The leakage current is very small in normal operating situation, but very large and contains a lot of high harmonics in the situation of lightning strike. To measure the total leakage current accurately, the current sensors are studied and the transfer functions are analyzed in this paper. Due to the large high frequency components in lightning and operating overvoltage, traditional voltage sensors cannot meet the requirements, and a method of capacitance division is presented. External environments usually make the collected signals contain much noise, and the wavelet method to remove noise is used in this paper. On the basis of resistive currents calculated by improved capacitance compensation method, the rules change of the equipment are analyzed from the point of view of the statistics and the GM(1,1) is used to predict the future state. Finally, the interface of the online monitoring and evaluation system is designed, which makes the results intuitive.