GIS State Intelligent Diagnosis Research Based On Multiple Physical Quantities

The GIS(Gas-Insulated metal-enclosed Switchgear)is widely applied in power systems for its smaller size,less maintenance,higher reliability and longer lifetime.Although the reliability of its operation is high,but once the fault occurs,the consequences are very serious,for instance,the maintenance work is very complicated,the power outage time is long,the blackout range will involve non fault components.Therefore,it is necessary to strengthen the intelligent and information construction of GIS,to realize the intelligent diagnosis of equipment status,and finally to develop the corresponding maintenance strategy.The intelligent and information technology of GIS,mainly refers to use sensitive means of monitoring,reliable evaluation method to determine the operating status of the GIS,and perform failure analysis when the device is in a state of abnormal,finally make a preliminary judgment.At present,most of the charged detection and on-line monitoring equipment for detecting GIS at home and abroad use single sensor to detect and diagnose the single physical quantity,the comprehensive detection using multi-parameter to study the developing trend and law of insulation state of GIS during operation is used less,such as Ultrasonic and UHF(ultrahigh frequency)etc.It will be an important research direction that how to organically use various types of sensors,data acquisition device and signal transmission system.This paper focused on the research of the methods and key technologies of GIS condition monitoring and intelligent diagnosis based on acoustic and electrical signal sensor arrays.First of all,the complexity of signal detection and localization was analyzed,according to the propagation characteristics of UHF and Ultrasonic signals in GIS and the impact caused by the insulator,L type and T type corner etc.In the third chapter,the signal delay estimation algorithm based on high order cumulant and bispectrum estimation was studied,verified the accuracy of the algorithm using double exponential decay function to analog signal.Finally,the delay time was calculated by the time delay algorithm,which was applied to the partial discharge location of the GIS device.The spatial location of partial discharge source would be predicted.In the fourth chapter,the design of GIS state diagnosis technology platform was studied,studying works were as follows,Ultrasonic and UHF sensor as well as their optimal array,monitoring IED device architecture design,multi-channel synchronous acquisition device,and the method of PD(partial discharge)type identification considering the time domain pulse and statistical maps.The principal component analysis method,the information gain method and the support vector regression method were used to reduce the dimension of feature space,which could improve the efficiency of pattern recognition,finally,the state information and the fault type pre display of GIS were realized through the visual monitoring platform.In the end,this paper designed the GIS simulation experiment platform,and five kinds of typical insulation defect models,with them the physical experiments of acoustic and electrical integrated PD location method was carried out,and the main fault types of fingerprint and typical defect envelope were obtained.