Substation Equipment Classification And Its Application In Fault Diagnosis Based On Infrared Image

Infrared thermal imaging as a non-contact state detection technology has beenwidely applied in maintenance and fault diagnosis of substation equipment, whichcan find out faults of substation equipment without interrupting the production.However, the analysis and diagnosis of the infrared image mostly adopts manualway, whose efficiency and precision is low. Target recognition and classification ofsubstation equipment infrared image is researched in this thesis, and the researchconclusions are applied to thermal fault diagnosis. The main works are listed asfollows.(1) This thesis designs segmentation algorithm for substation equipment basedon the research of traditional methods of image segmentation. Firstly, thisalgorithm processes the infrared image through Mathematical Morphology, andthen the image is segmented by the method of combining Otsu algorithm andRegion Grow algorithm. The experiment result shows that the algorithm has goodeffect.(2) Shape feature is used to construct the feature vector for classificationaccording to the characteristic of substation equipment infrared image. Severalshape features such as Fourier Descriptors, Hu moments and Zernike moments arediscussed and compared in this thesis, and then a classifier is designed based on BPneural network. The experiment result verifies classification performance of shapefeature and classifier through practical substation equipment infrared image.(3) Based on above conclusions from theory research, this thesis designedfault diagnosis system combined with “Application Rules of Infrared Diagnosis forLive Electrical Equipment”. The system diagnoses infrared image according to thecategory of equipment and fault diagnosis rules, then provides the defect level andfault type and gives treatment suggestion.The practical application shows that the segmentation and classificationalgorithm designed in this thesis can resolve automatic analysis and fault diagnosisof substation equipment infrared image effectively.