Aging Assessment of Insulating Paper Based on Furfural Concentration Prediction in Transformer Oil

Furfural is a chemical marker that characterizes the aging state of paper insulation in power transformers. Current research lacks reports on the prediction of furfural concentration in insulating oils, which is not favorable for predicting the potential failure or aging state of transformer paper insulation. In light of this, this paper presents a method for predicting furfural concentrations utilizing the Multi-Strategy Particle Swarm Optimization (MSPSO) combined with the Backpropagation Neural Network (BPNN) algorithm (MSPSO-BPNN), and thus evaluates the aging state of the insulating paper. Firstly, oil-paper insulation samples are prepared under laboratory conditions, and concentrations of furfural in oil are measured at different aging periods; Secondly, the furfural concentration time-series data are predicted using the MSPSO-BPNN algorithm, and the RMSE of the predicted results reached 0.05; Ultimately, based on the predicted furfural concentration data and the furfural-polymerization degree (DP) assessment model considering the second-order aging kinetic theory, the aging assessment of the insulating paper has been achieved with an accuracy of 98.7%. Therefore, this work can provide technical and theoretical support in the assessment of the aging state of insulating paper through the application of furfural concentration prediction.