Digital magnetic encoder hardware and software integration for bearing fault detection in a modular condition monitoring system

This paper presents the integration of digital magnetic encoders into a modular IoT-based condition monitoring system. Encoder-based instantaneous angular speed data is a less explored, but potentially valuable resource for the condition monitoring of rotating machinery. This work explores the integration steps and challenges encountered in acquiring high-resolution data from encoders to detect bearing faults. The necessary hardware and software are examined, with a focus on error sources introduced by the encoder acquisition and methods for mitigating them. For initial validation, and as part of a larger study, the system was tested on a multi-gear stage test bench with a known faulty bearing. Vibration data from Integrated Circuit Piezoelectric accelerometers, collected by the same acquisition system, is compared to the instantaneous angular speed signals to assess the fault detection capabilities of the encoder-based approach. The results of this study indicate that once error sources are mitigated, the encoder can detect fault peaks in the amplitude spectra corresponding to bearing defects. However, these faults are only identifiable at rotational speeds between 50 and 500 RPM and appear less prominent than the fault peaks observed in the accelerometer data. Nevertheless, the encoder's ability to detect some fault peaks justifies further investigation. Future work should explore other fault types, sensor options, and mounting locations, as well as increase the counting clock frequency to minimize quantization noise at higher speeds.