A Genetic Optimization Model for Aging Analysis of Paper Insulation in High-Temperature Region via Frequency Domain Spectroscopy

Existing methods based on frequency domain spectroscopy (FDS) cannot address the nonuniform aging of transformer insulation induced by the temperature gradient. To solve this issue, the genetic algorithm (GA) and a finite-element method (FEM) are combined to obtain the aging information of insulation at the transformer high-temperature region via the analysis of the FDS data. In this work, the FEM model is constructed to provide a mapping relationship between the aging state distribution and the FDS data of the oil–paper insulation system. Furthermore, the GA is designed to solve the parameter distribution of the FEM model based on the collected FDS of the whole insulation. Finally, the verification results indicate that the extracted aging information can characterize the actual aging states of the high-temperature region. The obtained relative error is less than 7% in laboratory conditions, showing comparable accuracy with existing methods on the uniform-aged sample. In that regard, the contribution is in exploration to propose a novel method for analyzing the aging state of insulation in transformer high-temperature region.