Research And Platform Development Of Switch Cabinet Abnormal Detection Technology Based On Infrared Thermal Imaging

In daily production,most electrical equipment needs to work continuously for a long time.This kind of operating strength will cause the internal heating phenomenon of the equipment.Because the overall structure of electrical equipment has the characteristics of large current,high voltage,strong magnetic field,and slow heat dissipation,the heating phenomenon will lead to electrical equipment failure,affecting the service life of electrical equipment and the stable operation of the power system.In order to ensure the safe,stable and normal operation of electrical equipment,the study of fault diagnosis and early warning of electrical equipment during operation has very important theoretical value and practical engineering significance.This paper takes the detection of electrical switch cabinet faults as the research object,and realizes fault monitoring and early warning through infrared thermal imaging technology.First of all,it summarizes the methods of infrared detection faults commonly used at home and abroad,analyzes their advantages and disadvantages in detail,and introduces the basic principles of infrared fault diagnosis,the technical standards of infrared technology detection,the classification of electrical equipment heating defects and defects definition of grade.Secondly,based on the experience of using infrared imagers in the field,several common infrared diagnosis methods for detecting faults are introduced in detail,and the photos of abnormal infrared heating of switchgear collected on the spot are established and saved.Then,according to the infrared abnormal heating fault photos of the existing electrical switchgear,based on the Gaussian-YOLO v3 target detection algorithm,K-means algorithm,new parameter optimizer,data enhancement method and non-maximum suppression method are used to improve The detection effect of Gaussian-YOLO v3 algorithm on infrared abnormal heating faults,the detection accuracy of its detection algorithm on the self-made temperature abnormal infrared data set reached 93.45%,and the detection speed reached 49 frames per second.Finally,in order to monitor the temperature status of multiple electrical switch cabinets on site,a temperature alarm detection platform was developed.The platform can not only remotely monitor the operating temperature of multiple electrical switch cabinets,but also record the abnormal heating conditions of the electrical cabinets,facilitating maintenance personnel to check and repair electrical equipment in time.This thesis collects images based on the on-site thermal imager,uses the optimized Gaussian-YOLO v3 abnormal heating detection algorithm and real-time monitoring of the temperature alarm detection platform of the electrical switchgear.It is not only suitable for electrical switchgear fault detection,but also can be applied to various other electrical equipment.