Bearing Fault Detection Under Low-Speed and Variable-Speed Conditions: An Adaptive Multiperiod Cumulative Cross-Reference Correlation Coefficient Signal Reconstruction

The optical encoder-based instantaneous angular speed (IAS) signal offers advantages such as direct correlation with torque, and a non-time-varying transmission path, making it widely studied in fault diagnosis applications. Nonetheless, under low-speed and variable-speed conditions, the IAS fluctuations caused by the faulty rolling element bearings (REBs) remain weak since REBs do not transmit torque. This study discusses the elapsed time-based IAS estimation method and analyses factors influencing the magnitude of IAS fluctuations, including operational speed, fault severity, radial load and random slip. Furthermore, phenomenological models are established for both inner and outer race faults in REBs, incorporating the encoder installation position, random slip and other relevant factors. Finally, to enhance the weak characteristics of the faulty REBs, a multi-period cumulative cross-reference correlation coefficient (MPCRCC) signal reconstruction scheme is proposed, leveraging the IAS fluctuation characteristics, multi-period accumulation, and cross-reference correlation. Simulation and experimental results demonstrate that the MPCRCC algorithm effectively enhances fault-related features of REBs under low-speed and variable-speed conditions.