Classification And Thermal Fault Diagnosis Of Substation Equipment Based On Infrared Images

Substation equipment is the main component of the power grid,but also to ensure the national economic development and People's Daily life is one of the important infrastructure.How to ensure its normal operation and to detect the sudden failure has become a hot issue for researchers.With the development of science and technology,the method of fault troubleshooting for large-scale power failure has been transformed into a new method of thermal fault diagnosis.Including the use of hand-held infrared imager to collect images,and then fault screening,is now being promoted as a technology,but the current mainly visual image,a large demand for manual.In order to promote the automatic operation of power grid and reduce manpower input,this paper designs an automatic identification and thermal fault diagnosis system for substation equipment based on thermal infrared image.Aiming at the shortcomings of traditional segmentation algorithm which is sensitive to the clustering center of infrared thermal image of substation equipment,such as low clustering accuracy and details,an improved IFCM algorithm which is suitable for the infrared image of power equipment is proposed.The gaussian model is introduced into the global spatial distribution information of the image,and the membership function is optimized by using the spatial operator of the local spatial information to improve the IFCM algorithm and solve the problems of edge blur and uneven image intensity.Experimental results show that the relative regional error rate of the algorithm is about 10%,which is less affected by the change of the fuzzy factor m.It is proved that the algorithm can effectively suppress noise interference.In this paper,a SVM infrared image classifier is designed to effectively identify three types of common electric power equipment.Through HOG feature extraction of infrared image of power equipment,the classifier is combined with SVM multi-classification to improve the recognition accuracy.In the experiment,the classifier was used to identify three kinds of equipment,and the results showed that the comprehensive recognition accuracy reached over 95.3%,which was better than the traditional classification method and satisfied the demand of classification accuracy.Finally,the traditional method of relative temperature difference is improved by using the temperature data of infrared image.The experiment shows that the diagnosis system designed in this paper is able to grade the faults and give treatment Suggestions while judging whether there are thermal faults in three kinds of power equipment,which verifies the feasibility and effectiveness of the infrared diagnosis technology for substation designed in this paper.