Research on fault type identification method for electrical equipment based on MFCC-CNN

Abstract

Abstract As a critical component of the power system, the secure and reliable functioning of electrical equipment is the key to ensuring the dependability of energy supply. In this article, a fault type identification method based on Mel-Frequency Cepstral Coefficients-based Convolutional Neural Networks (MFCC-CNN) for electrical equipment was proposed. Firstly, the reasons for the sound generated by electrical equipment during the operation were analyzed. Then the MFCC coefficient of sound features was extracted by collecting and processing the faulty sound of electrical equipment. Finally, a CNN model was established to train and recognize sound signals. This method combines the recognition and classification of sound with deep learning (DL), which can significantly improve the efficiency of fault diagnosis.