Designing a system for detection and analysis of transformer electrical insulation contamination (ICAS) based on radiographic images processing

In the process of transformer insulation production, unwanted occurrence of contaminations in the insulation will cause the electrical discharge and major damage. Therefore, in order to detect impurities in the insulation, radiographic test is used in the quality control unit of transformer insulation production. In the available radiographic test, the radiographic operator observes the output image of the clear shiny spots in the field of images as impurities, but some factors such as the noise of radiographic images and the small size of the existing contaminations, reduce the accuracy and speed. On the other hand, the existing method does not detect the type of contaminations. Therefore, if contaminations are identified in the insulation, it will be impossible to identify the source of contamination and clean the insulation production line. In this research, a system for analyzing the contaminations of electrical insulation of transformers called ICAS is proposed in which by using image processing and machine vision methods, while improving the quality of radiographic images, contaminations can be automatically found in Identified the electrical insulation components of the transformer and obtained information about the occurrence of impurities in the insulation production line by identifying the impurities of iron and aluminum.