Abstract: The comprehensive information perception of power equipment is essential for the high-quality development of new electricity system. A panoramic perception framework for transformer vibration states is proposed in this paper, which is based on digital twin technology. Upon structural fields, the transformer twin model is employed to model the vibration states of cores and windings under the coupled electric fields, magnetic fields, and vibration fields, so as to build the three-dimension vibration twin for the research of the transformer vibration mechanism. With the increase in service time of transformer insulation materials, carrier migration will occur, and leakage flux will increase, leading to additional vibrations. To ensure the fidelity of the transformer twin model, the aging characteristics of insulating paper and oil are considered. The model is updated based on the time-dependent characteristics of the dielectric loss factor tanδ. An S11-M-2 000 kVA oil-immersed transformer is studied as the experimental subject. The tests show the error between the twin data and the measured vibration acceleration is reduced by 6.1% by updating the aging parameters, achieving accurate panoramic mapping of the virtual twin model. Furthermore, the virtual-real error as an indicator is used to identify the abnormal health state, offering a new conception for the digital management of power equipment.