Power Cable Fault Diagnosis Based on Chaotic System and Discrete Wavelet Transform Convolutional Neural Network
Aiming at the problems of information redundancy and inaccurate fault model diagnosis in traditional power cable feature extraction methods,this paper proposes a fault diagnosis algorithm of convolutional neural network based on chaotic system and discrete wavelet transform,that is,using discrete wavelet transform to filter the collected partial discharge signal,and using Lorentz chaotic system to establish dynamic error scatter diagram extraction.The fault features are finally identified through a convolutional neural network(CNN).Combining four typical power cable insulation faults and test platforms for verification,the results show that the proposed algorithm can quickly and accurately identify the fault status of power cables,and the recognition accuracy rate reaches 97.5%,which proves the feasibility and effectiveness of the proposed algorithm.It provides a certain reference value for the fault diagnosis of power cables.
NETL
NSTL
万方数据
