Impedance-based Induction Motor Bearing Failure Detection Applied to X-Ray Tube

Modern bearing condition monitoring systems for induction machines typically rely on spectral analysis performed on mechanical or electrical signals measured at high sampling rate to detect vibrations at signature fault frequencies. However, in some applications, these methods may be infeasible or prohibitively complex or expensive. For such applications, we propose an alternative low-complexity strategy for detecting bearing health issues by monitoring the induction motor equivalent circuit impedance. As the equivalent circuit impedance depends on the slip of the machine, abrupt drops of the impedance are indicative of sudden onset of friction in the bearings, which is a precursor of bearing failure. We quantify the sensitivity of this approach and demonstrate the robustness to other variables such as winding resistances. We illustrate the predictive power of impedance-based methods for detecting ball-bearing failures in X-ray tube rotors by evaluating the algorithm on data collected from run-to-failure experiments performed on five different tubes. Our results indicate that the method is valuable for short-term failure prediction of induction machines in critical applications.