Life prediction of composite insulator based on thermogravimetric test and support vector machine

Composite insulators are affected by ultraviolet rays, electric field, humidity and temperature difference, which inevitably aging and triggering insulation failure. Through the heat loss test, the heat loss curves of silicone rubber composite insulators with inner and outer layers, different heating rates and different operating years were obtained, the activation energy was calculated, and the Arrhenius life equation was established. The optimal penalty coefficient C and kernel function parameter g in the support vector machine are searched by genetic algorithm, and a life prediction model based on GA-SVM is proposed for training and testing. The results show that the content of silicone and other macromolecules in silicone rubber materials is significantly reduced. the prediction accuracy of the GA-SVM model is improved by 5%-12% compared with the thermal life equation and the BP neural network which is easy to fall into the local optimal solution. The prediction results are accurate and valid as verified by specimen appearance and SEM, which can provide a reference for the maintenance and operation and maintenance of composite insulators.