Insulator Contamination Diagnosis Method Based on Deep Learning Convolutional Neural Network

Insulators are one of the most used electrical equipment in the power system. Long-term exposure to the air environment can easily cause pollution flashovers and have a serious impact on the safe operation of the power transmission system. In this paper, after obtaining the discharge image of the insulator taken by the ultraviolet imager in the artificial pollution test, the image is preprocessed, and the convolutional neural network (CNN) algorithm is used to evaluate the pollution state of the insulator after reducing the image dimension. The results show that the pollution discharge status of insulators is positively correlated with their surface pollution and humidity. The average recognition accuracy of insulator pollution status obtained by the CNN algorithm can reach eighty-five percent, which can meet the requirements of engineering insulator pollution recognition.