Thrust Bearing Faults Diagnosis Using Multiphase Current Signals With Sub-Nyquist Sampling

In this article, we propose an integrated diagnostic method based on motor current signals approach, encompassing identification, enhancement, and analysis, for common fault diagnosis in thrust bearings under sub-Nyquist sampling conditions. A 3-D electromechanical coupling numerical model maps mechanical fault characteristics onto the current signals, achieving fault signal feature localization. The assumptions and limitations of the proposed model are thoroughly discussed. An original time-shifting method for mixing multiphase current signals is introduced to enhance the fault characteristics. The nonuniform fast Fourier transform is employed to analyze the mixing signals with nonuniform time distribution, and its kernel is modified to reduce computational complexity. This method addresses a critical research gap in the analysis of thrust bearing faults under conditions of low sampling rates and constrained computational resources. The effectiveness of the proposed method is validated through two case studies.