Aging Estimation of Transformer Insulating Paper in Hotspot Region Under Nonuniform Thermal Field Based on Furfural Markers

The insulating paper at the hotspot region of the transformer winding is the weakest part; thus, it is essential to grasp its aging condition. Furfural is employed as a chemical marker to characterize the aging condition of insulating paper. However, the existing studies focused on the aging assessment of insulating paper at uniform aging conditions based on furfural, and there is a lack of reports on the insulating paper aging estimation in the hotspot region. Given this, an improved beluga whale optimization algorithm optimized multilayer perceptron (IBWO-MLP) model is proposed to predict the degree of polymerization (DP) value in the hotspot region. First, nonuniform aging experiments are conducted on oil-paper insulations. Furthermore, the DP values of insulating paper in the hotspot region were predicted by the IBWO-MLP model, and the accuracy of the prediction models was compared. Moreover, based on the enhanced data after the prediction of the IBWO-MLP model, a DP estimate model for insulating paper in the hotspot region applicable to field conditions was established. Eventually, the reliability of the constructed model was verified by data from the field transformer. In this regard, this study aims to contribute a novel approach to the aging estimation of the insulating paper in the hotspot region, which is supportive of the diagnosis of insulation failures.