•  
  •  
 

Keywords

multiphysics simulation, vibration propagation characteristics, inversion of the point of maximum vibration, random forest, oil-immersed transformer

Abstract

Accurately obtaining data at the point of maximum vibration in a transformer core is one of the challenges in power equipment health management.To address this problem，a new method based on multiphysics modeling and inversion techniques was proposed for detecting the point of maximum vibration acceleration in a transformer.First，a multiphysics load-transfer coupling model incorporating electromagnetic，structural，and fluid fields was used to analyze the vibration mechanism under an alternating and nonuniformly distributed magnetic field.The point of maximum i nternal vibration was then determined through harmonic response analysis.Second，considering data correlations，vibration measurement points on the outer wall of the oil tank that were strongly correlated with the point of maximum vibration in the core were selected.A random forest algorithm optimized through K-fold cross-validation and grid search was used to establish an inversion framework for the point of maximum vibration，enabling accurate and timely detection of the internal vibration state.Finally，the proposed method was validated using a simulation model of a 35 kV three-phase oil-immersed transformer.The results show that the multiphysics simulation samples generated under various operating conditions are valid and reliable.The root mean square error （RMSE ） of the inversion model for the point of maximum vibration is 1.2%，and the coefficient of determination is 0.92，thus accurately reflecting the internal vibration state.

DOI

10.19781/j.issn.1673-9140.2026.04.026

First Page

316

Last Page

326

Download

Share

COinS