Distribution Network Ultrasonic Live Detection Technology and Defect Classification and Identification Method
In the process of conducting live detection on the distribution network,conventional methods have poor processing effects on complex equipment such as switchgear.Therefore,the ultrasonic live detection technology and defect classification and identification method for the distribution network are proposed.Firstly,calculate the mean of ultrasonic live detection,further construct feature vectors at the cyclic discharge point in the audio,analyze the repetitive characteristics of the audio data,complete the processing of basic information data of distribution network defects,and extract data features.Then,based on this,establish a defect classification database,construct a simulation model,adjust training rules,and eliminate noise in the secondary circuit defect signal.Finally,defect classification and recognition can be achieved through the optimal hyperplane.The results show that the accuracy of this method is 98.7%,and its performance on all samples is relatively stable,which can improve the detection effect of distribution network ultrasonic technology.
ultrasoniclive detection technologydefect classification and identificationdistribution network
NETL
NSTL
万方数据
