Study on Vibration and Temperature Rise Characteristics of Power Transformers Based on Multiphysics Data Fusion Method

Temperature variations during the routine operation of power transformers can affect the material properties of structural components, alter the viscosity of insulating oil, and induce thermal expansion, all of which influence the dynamic behavior of the transformer. This study investigates the temperature-dependent vibration characteristics of the transformer core using a combination of simulation and empirical data. Given the limited availability of fault samples under real operating conditions, we integrate onsite vibration measurements with simulated data obtained through COMSOL Multiphysics. A detailed 3D simulation model is constructed, incorporating electromagnetic, structural, and thermal parameters to perform multiphysics coupling analysis. The electromagnetic force distribution is first calculated through magnetic field simulations, and then transferred to the structural module to simulate mechanical responses. A thermal module is applied to account for temperature variations and their effect on the core’s vibration behavior. Several mathematical models—including linear, exponential, logarithmic, and polynomial functions—are used to fit the resulting vibration data. Among them, polynomial fitting demonstrates the highest accuracy and lowest fitting error. Based on the best-fit model, a temperature-vibration compensation curve is established, enabling correction of vibration amplitudes under different thermal conditions. This curve provides a reference for assessing core performance more accurately. The proposed approach facilitates intelligent monitoring and predictive maintenance by fusing temperature and vibration information, offering a practical tool for transformer health diagnostics and operational reliability.