Simulation and experimental verification of vibration signals for bearing faults in the framework of digital twin

Purpose This study aims to construct a thermo-mechanical-lubrication multiphysics coupling dynamic model for aero-engine bearings under the Digital Twin (DT) framework. It aims to realize high-fidelity simulation of bearing fault vibration characteristics, overcome the spatiotemporal constraints and measurement blind spots of physical tests, solve the low simulation accuracy of pure mechanical models under extreme conditions, and optimize bearing DT modeling in industrial lubrication and tribology. Design/methodology/approach A dedicated test rig matching actual bearing service conditions is built for data acquisition and model validation; inner ring eccentricity is integrated into dynamic modeling, with a thermo-mechanical bidirectional coupling model and virtual-real closed-loop data link established, and friction heat generation and oil film damping fully characterized. The modular model adapts to various bearings via parameter adjustment, and a steady-state temperature field running-in test verifies its heat generation characteristics. Findings Simulation results agree well with tests: relative errors of inner/outer race fault damage cycles are 4.67 and 3.33%, fault characteristic frequency errors are only 0.0704 and 0.243%, and the outer race temperature MAE is 1.034°C with a 0.9922 Pearson correlation coefficient. Originality/value This work upgrades bearing test digitization, breaks traditional test limits, cuts test costs and provides a new method for bearing fault diagnosis and condition assessment with important theoretical significance and engineering application prospects. Peer review The peer review history for this article is available at: Link to the website of publons