Research on Evaluating the Ageing State of Transformer Insulation Paper Based on Variable Weight Theory

Current methods for multi‐characteristic jointly evaluating the ageing state of insulation paper mostly have the problems of single evaluation indicator (based solely on content) and fixed weight factors. Fixed weight factors cannot accurately reflect how the relative importance of each characteristic varies with ageing. Therefore, the applicability of the evaluation methods needs to be improved. This paper proposed a multi‐characteristic joint evaluation method on the ageing state of transformer insulation paper based on multi‐indicator. Methanol (MeOH), furfural (2‐FAL), carbon monoxide (CO), and acidity were selected as evaluation characteristics. Content, increase rate, and evaluation error were selected as evaluation indicators. A weight model was constructed based on the CRITIC method, and the Bayesian Optimisation Algorithm (BOA) was further utilised to optimise the weight factors, thereby establishing a multi‐characteristic joint evaluation model. The optimal weight values for each characteristic were determined through an accelerated thermal ageing test conducted at 393.15 K. The experimental results showed that the relative evaluation errors of the proposed method were within 6.81%, which was significantly lower than those of the single‐characteristic evaluation methods and demonstrated better applicability. These results verified the effectiveness of the proposed method.