Prediction of Aging Degree of Oil-paper Insulation Based on Raman Spectroscopy and Fuzzy Neural Network

In this paper, a method based on Raman spectroscopy to diagnose the aging degree of oil-paper insulation is discussed. According to IEEE guidelines, oil-paper insulation samples were prepared in the laboratory, and the degree of polymerization values of the samples were measured to obtain the exact aging states. Raman detections of the samples were carried out based on the Raman experimental platform. T-S fuzzy neural network was used to analyze the Raman spectrum data of insulating oil. By using the experimental data to train the diagnosis model, constantly modifying the membership function, and mining the internal mathematical relationship between the Raman spectrum characteristics of oil paper insulation and the degree of polymerization of insulating paper, this paper managed to use the Raman spectrum data of insulating oil to predict the aging degree of oil paper insulation. Finally, the deviations between the predicted and the measured degree of polymerization values of the five test samples in this experiment are both less than 150.