Adaptive spectral amplitude modulation: An optimized processing method for rolling bearing fault IAS signals

Abstract

Fault diagnosis of rolling bearings is of great significance for the monitoring and maintenance of mechanical systems. Spectral amplitude modulation is an automated nonlinear filtering method known for its good robustness. However, under different magnitude order (MO) weights, the amplitudes of interference components may be amplified, which can affect the accuracy of fault diagnosis. Therefore, selecting an optimal range for MO is crucial. To address the above issue, an adaptive spectral amplitude modulation (ASAM) for rolling bearing fault feature extraction based on encoder signals is proposed. Firstly, the encoder signal is processed using the forward difference method to obtain the instantaneous angular speed signal, followed by the Fourier Transform to derive its amplitude and phase. Subsequently, while maintaining the phase unchanged, different weights are assigned to the amplitude, with the MO range set from −0.5 to 10. An inverse Fourier Transform is then performed to obtain the modified signal. The normalized squared envelope spectrum of the modified signal is obtained through the Hilbert Transform. Next, a Normalized Hoyer Envelope Spectrum Energy Ratio index and an adaptive threshold are established to characterize and differentiate the richness of fault information in the modified signals, enabling adaptive selection of the MO range. Finally, a normalized squared envelope spectrum with an adaptively selected MO range is generated to extract the fault features of rolling bearings. The effectiveness of the ASAM is validated through simulation and experiments.