Research on High Voltage Disconnector Based on MoCo Comparative Learning

In order to improve the accuracy and generalization ability of fault feature extraction, and reduce the cost and time of data annotation, this paper proposes a fault diagnosis and identification method of disconnector based on MoCo comparative learning to reduce the dependence of labeled data. Firstly, the signal decomposition reconstruction of the original vibration signal of the disconnector is carried out by wavelet packet transformation, and then the MoCo momentum comparison method is used to carry out semi-supervised learning to extract the vibration signal characteristics of the disconnector, and finally the deep neural network is used as the basic model to simulate the fault diagnosis results. The simulation results show that the fault classification effect of the model is obvious, and the overall accuracy rate can reach 98.84%, which has certain practical value and can be widely used in the field of fault diagnosis.