Research On The Evaluation Method Of Composite Insulator Aging Degree Based On Hyperpectral Technology

Due to the vast territory and complex terrain of our country,the terrain and environment of transmission lines across regions are changeable.Silicone rubber composite insulators are important supporting equipment for power transmission lines,and the convenient detection of their performance status has become an important issue for the safe and stable operation of the power system.Due to the long-term operation of insulators in outdoor conditions,which is harsh,inaccessible,and heavily polluted,the insulator is influenced by environmental factors such as ultraviolet radiation,temperature,acid-base erosion,and electric field.With the continuous increase of its operating time,its surface is prone to different degrees of aging or deterioration,resulting in a decrease in the insulation performance of the insulator,and a decrease in the flashover voltage,which is likely to cause grid failures and cause great economic losses.Accurate and rapid acquisition of the aging degree of the insulator surface is an effective measure to prevent external insulation accidents.In view of the fact that there is no convenient and fast method for the existing insulation aging detection,in this paper,a non-contact and fast non-destructive testing method for surface aging of composite insulators based on hyperspectral technology is proposed.First,four typical artificial aging test platforms were set up,and four typical artificial accelerated aging tests of UV aging,corona aging,thermal aging and acid-base aging were performed on the silicone rubber composite insulating sheet,and aging samples were prepared and characteristic tests were performed.Through the hydrophobicity test,Fourier infrared spectroscopy test and flashover voltage test of samples with different aging degrees under typical aging methods,the effects of sample aging on its hydrophobic properties,group changes and electrical properties are analyzed;Secondly,the hyperspectral experimental platform was built.On the basis of explaining the aging mechanism of composite insulators,the principle of applying hyperspectral technology to distinguish the aging degree of insulators is described.The hyperspectral information of the aging sample is obtained.In response to the influence of noise interference and light scattering,the spectral data is corrected for black and white image and multiple spectral line scattering correction,and then the spectral line is de-enveloped,and finally the spectral characteristic parameters are Extraction includes absorption position,left and right shoulders of absorption valley,absorption depth,absorption width,total area and symmetry,which makes the spectral line characteristic information of insulators with different aging degrees under typical aging methods more intuitive and obvious;Furthermore,the acquired sample characteristic test data,hyperspectral line characteristic parameters and sample appearance and topography information are comprehensively analyzed.According to the quality and performance of the insulators,there are 5 levels of excellent,good,medium,poor,and reject.The aging degree of the samples is calibrated to 5 levels from 1 to 5 respectively.Finally,based on the deep extreme learning machine,the insulator aging degree evaluation model was established,and the aging degree prediction of the measured data was carried out.The accuracy rate of classification reached 98.00%,and the accurate classification of the insulator aging degree was realized.and the accurate grading of the aging degree of insulators was realized.And compared with the BP neural network and support vector machine algorithm model,the model training and testing time and the accuracy of the detection results show that the model used in this paper can have both speed and accuracy.The hyperspectral detection method for the aging degree of composite insulators on transmission lines described in this article is conducive to visually detecting the aging degree of insulators,effectively formulating targeted insulator replacement plans,and providing technical reference and theoretical support for on-site detection of insulator aging degree.