Research On Fault Diagnosis Technology Of Power Cable Accessories Based On HHT

With the development of society and the improvement of people's living standards, the national demand for electric energy is increasing, and higher requirements for the quality of the electricity operating system have also been put forward. Our urban power grid line is currently walking on the road of comprehensive high-voltage power cable. Once a fault occurs, it will lead to serious power outages, which will put a serious threat to the lives and property of the people. Most of the faults on the electrical line operating system are caused by the defects on the power cable accessories. So it is necessary for us to make a study on fault diagnosis technology on the power cable accessories. With the help of this research, we can achieve accurate fault diagnosis of the cable accessories, and take effective measures to avoid the occurrence of insulation accidents after that.The fault diagnosis technology of power cable accessories mainly includes partial discharge signal analysis and identification of defect types. With a large number of study, domestic and foreign experts and scholars found that the most effective method for fault detecting on the power cable accessories is partial discharge signal measurement, and the analysis of partial discharge plays an important role in the preprocessing before the application of the identification of defect types. In this paper, the HHT mode mixing suppression method is applied to conduct the decomposition and analysis of partial discharge signals, which are collected from the partial discharge test. On the basis of signal analysis and the denoised signal, the support vector machine and extreme learning machine are used to do the defects recognition, respectively. With the comparison of recognition accuracy of the two algorithm, the optimal recognition method is determined. The main research contents are as follows:(1) After the deep learning about the positions of power cable accessories fault defect prone and the defect types and other issues, the analysis of the causes of the defects of the power cable accessories has been done. (2) The partial discharge test of power cable accessories has been carried out. With the application of HHT method for partial discharge analysis, it is found that there is a large deviation between the results of the analysis and the actual.To solve this problem, the HHT mode mixing suppression method combined wavelet packet transform and HHT is put forward, and it is applied in the analysis of partial discharge signal to complete the verification; (3) The two theoretical methods support vector machine and the extreme learning machine are introduced. And with the application in the classification of iris flower, the recognition accuracy of the two methods is compared and the advantages and disadvantages of the two methods are expounded; (4) The HHT mode mix suppression method, support vector machine and the extreme learning machine described above are all applied in the fault diagnosis test of power cable accessories. Some comparative analysis of corresponding methods has been done in the link of partial discharge analysis and defect type identification, and the HHT mode mix suppression method is validated, besides, the best recognition method is sifted.The test results show that the application of the HHT mode mix suppression method mentioned in this paper make the analysis results expressed more clearly. And filtering and denoising of the signal has been done in this process, and the denoising effect is good. The support vector machine and extreme learning machine are used to recognize the defect types, and the comparison of the accuracy of the two methods shows that the extreme learning machine has a higher accuracy, therefore it is more suitable for cable fault diagnosis.