An Evaluation Method of Insulator Aging Grade Based on Extreme Learning Machine

As an important part of the power grid, insulators play a role in isolating and protecting the power system. During the operation of the power grid, insulators on the line are prone to corona aging due to environmental factors such as electric field strength. Insulator aging will lead to grid failures, outages, and other problems. For the corona aging detection of insulators, the traditional methods are usually tedious, therefore, this paper proposes an approach for evaluating the aging grade of insulators, which involves the use of hyperspectral detection technology based on a limit learning machine. The primary objective of this approach is to address the corona aging problem associated with insulators. In this paper, firstly, corona aging of silicone rubber sheet was carried out from 0 to 300 hours, then the spectral data of silicone rubber sheet samples were obtained using a handheld hyperspectral camera, and finally, the limit learning machine model was established to realize the evaluation of different aging grades of the silicone rubber sheet with an accuracy of 94%. The hyperspectral testing technology can achieve non-contact and high-accuracy insulator performance assessment, which reduces the operational complexity and improves the economic efficiency of the power system compared with the traditional methods and helps to guarantee the normal and stable operation of the power system.