High-fidelity digital twin-based fault dynamics analysis of motorized spindle bearings under coupled mechanical and electromagnetic physics

Abstract Accurately capturing the dynamic behavior of spindle bearings is essential for reliable operation and intelligent maintenance of high-end CNC machines. Internal dynamic responses, however, are difficult to observe directly owing to high rotational speeds, compact structural layout, and complex electromagnetic interactions. This study develops a high-fidelity digital twin framework to analyze fault-related dynamic behavior under coupled mechanical–electromagnetic effects. A physics-informed dynamic model identifies electromagnetic disturbance sources, including magnetic-force asymmetry and torque fluctuations, and links them to vibration and fault development. Model accuracy is enhanced by incorporating experimental data through parameter updating and response matching. Validation experiments were conducted on a fault-free CNC form-grinding machine equipped with a multi-sensor measurement system, to acquire key mechanical and electromagnetic parameters for model calibration. The simulated fault-free responses align closely with the experimental measurements, confirming model fidelity under normal operating conditions. Vibration responses for inner- and outer-race faults are subsequently generated via simulation, and their spectral characteristics are consistent with theoretical fault-frequency signatures, demonstrating the framework’s ability to reproduce coupled multi-physics behavior and support condition-based monitoring and predictive maintenance strategies for motorized spindle systems.