Research On Partial Discharge Detection Technology Of Switchgear Based On Ultrasonic Method

The switchgear is an important equipment for relay protection and power flow control in the power system,among which 10 k V and 35 k V high voltage switchgear are widely used,and its safe and stable operation is crucial to guarantee the power supply reliability of the power system.Partial discharge(referred to as local discharge)is a form of discharge before the formation of penetrating discharge in the insulating medium.It is not only the characterization of the insulation deterioration of the switchgear,but also the factor of accelerating the insulation deterioration.The detection of switchgear release information can provide the reference basis for the operation and maintenance personnel and users of insulation deterioration,so as to timely find out the potential safety hazards,prevent switchgear accidents,and ensure the safety and reliability of the power system.According to the characteristics of ultrasonic signal caused by partial discharge in the switchgear,this paper designs the ultrasonic detector of switchgear local discharge,including the hardware acquisition system of local discharge ultrasonic signal and the software system of local discharge data analysis on PC.The article studies the mechanism and characteristics of partial discharge switch cabinet,through the simulation to establish the mathematical model of ultrasonic partial discharge,put blew a model based on the research of ultrasonic signal de-noising algorithm and pattern recognition algorithm,and design the ultrasonic board based on ARM kernel put acquisition circuit and applied to the PC board put PD＿Check ultrasonic data analysis system.At first,this paper introduces the design and manufacture,transportation,installation,operation environment and discharge of four trigger switchgear insulation aging factors,analyzes the creeping discharge in the switch cabinet,air gap discharge,the discharge signal point discharge three different characteristics,introduce poisoned with ultrasonic signal attenuation changes in the process of generation,transmission and communication;Then,the mathematical model of local discharge ultrasonic wave is established by using data fitting method.The mathematical model of local discharge source based on IEC60270 standard is introduced.Based on this analysis,the new multi-order Gaussian mathematical model is selected as the fitting function,and the parameters of the local discharge fitting function are determined by using MATLAB data fitting toolbox cftool.Combined with the transmission attenuation function of ultrasonic wave in switchgear and the field noise simulation function,the ultrasonic observation signal model of local amplifier is established.Secondly,this paper proposes the self-adaptive LMS-wavelet soft threshold denoising algorithm and studies the feature extraction and pattern recognition algorithm of local amplifier ultrasonic signal.After denoising,the local amplifier ultrasonic signal extracts the signal’s time-domain waveform feature and wavelet packet coefficient feature as the feature quantity,and uses the kernel principal component analysis(KPCA)to reduce the feature quantity dimension.Probabilistic neural network(PNN)is used to realize the pattern recognition of local release,and the recognition effect is good.At last,the hardware circuit based on ARM kernel and PD＿CHECK system of ultrasonic signal analysis are designed.The hardware circuit includes the design of ARM core development board,the selection of ultrasonic sensor and operational amplifier,and the design of wireless data transmission module.Through the hardware circuit to achieve the local amplifier ultrasonic signal acquisition,and the use of wireless data transmission module to achieve data transmission and control display;PD＿CHECK system includes ultrasonic signal noise reduction and macro feature analysis algorithm,to achieve the local ultrasonic signal noise reduction and feature analysis.