A Probabilistic Neural Network Assessment Method for Insulator Pollution Levels Based on Infrared Images

The surface pollution of transmission line insulators is still an important factor affecting the safety and stability of the power grid. Regular cleaning has become the most widely used anti-pollution flashover measure in the power system. However, because it cannot obtain the current pollution state of the insulator surface without contact, easily and quickly, or it is susceptible to being interfered with by the ambient noise and the accuracy is not high enough, the cleaning cycle can only be determined in advance. Therefore, it will cause insufficient or excessive cleaning. In this article, infrared images of insulators under different pollution levels, ambient temperature, and relative humidity are collected by artificial pollution test, and based on infrared image characteristic parameters, probabilistic neural network (PNN) assessment method is established and verified. The results show that the infrared image is denoised by bilateral filtering, and the half-shed surface of the insulator is selected as the research area by an improved watershed algorithm, it is effective in extracting the characteristic parameters of the infrared image of the insulator. The higher the pollution level and the higher the relative humidity, the more obvious the heating phenomenon of the insulator, and the higher the temperature rise and the dispersion of the temperature distribution. The ambient temperature has no obvious effect on the insulator heating phenomenon. The identification precision ratio of pollution levels I, II, III, and IV were 93.5%, 83.9%, 89.8%, and 91.8%, respectively. The research has some references for insulator pollution level assessment and pollution flashover prevention.