Fault Diagnosis of Electrical Equipment Based on Infrared Thermal Imaging
The fault diagnosis of electrical equipment in substations is very important for maintaining the safe operation of a power grid. This paper briefly introduced the principle of infrared image diagnosis of electrical equipment and the application of back-propagation neural network (BPNN) and convolutional neural network (CNN) in infrared image diagnosis of fault equipment. Then, a simulation experiment was carried out on the CNN-based infrared image diagnosis algorithm in MATLAB software and compared with SVM and BP neural network. The results demonstrated that the CNN-based diagnosis algorithm recognized the fault area in the image more accurately and hardly recognized the normal heating area as the fault area. Regarding accuracy, recall rate, and F-value, the CNN-based diagnosis algorithm performed the best, followed by the BPNN-based diagnosis algorithm, and the SVM-based diagnosis algorithm had the worst recognition performance.
Keywords: infrared thermal imaging, electrical equipment, fault diagnosis, convolutional neural network