Research On Ultrasonic Signal Characteristics And Pattern Recognition Of Partial Discharge In 10kV Distribution Network Switchgear

With the rapid development of the power system,switchgear as an electrical equipment has been widely used in the power distribution network system.The transmission line of the switchgear is responsible for cutting off overload and short-circuit current,and is a solid barrier to prevent large-scale power outages.Therefore,the stability and reliability of the switchgear is the key to ensure the safe and stable operation of the power grid.Partial discharge is one of the main reasons for the insulation failure of switchgear.Accurate and rapid monitoring of partial discharge in switchgear is of great significance to the elimination of equipment defects and the reliable operation of the power grid.In this paper,the ultrasonic method is used to detect the partial discharge in the switchgear of 10 k V distribution network.The causes of partial discharges are introduced by the equivalent circuit method,and the generation mechanism of ultrasonic signals is described in combination with the characteristics of ultrasonic waves appearing and propagating outwards when partial discharges occur,and then the propagation characteristics of ultrasonic signals in the medium are analyzed.The ultrasonic partial discharge detection system includes four typical discharge models,partial discharge simulation test platform and ultrasonic signal hardware detection device,and samples the four types of partial discharges respectively.Fast Fourier transform is performed on the collected ultrasonic signals.When partial discharge occurs,the ultrasonic signals will be concentrated in the range of 40 k Hz～100k Hz,and the intensity will increase significantly.By observing the spectrum of the ultrasonic signal in the frequency domain and time domain when no discharge occurs,it is concluded that the interference signal has the characteristics of low amplitude and wide frequency distribution.Therefore,the wavelet threshold denoising algorithm is selected to denoise the ultrasonic signal,and then the amplitude and phase information of the partial discharge signal are extracted,and the partial discharge phase distribution map is drawn.Combined with the distribution law of the four types of partial discharges in the map,the distribution characteristic parameters and time domain characteristic parameters that can characterize the partial discharge types are found,and the characteristic parameters are optimized.Finally,the support vector machine algorithm is used to realize the identification and diagnosis of four types of discharges by taking the collected 280 sets of experimental data as samples.By adjusting the kernel function and optimizing the support vector machine algorithm,the final recognition accuracy rate can reach more than 90%.