Metamaterial-based acoustic enhanced sensing for gearbox weak fault feature diagnosis

Abstract

In gear fault diagnosis of practical engineering, the gear fault signal is affected by external strong noise interference, the multipath superposition and interface loss of the internal fault excitation transmission, which leads to weak feature of incipient fault signals, making the gear fault diagnosis difficult. Inspired by the acoustic manipulation capabilities of acoustic metamaterials, this study proposes a method for weak gear fault feature diagnosis via gradient acoustic metamaterial (GAM), which utilises the acoustic rainbow capture and compression ability to reduce the difficulty of gear incipient fault diagnosis. According to the gear frequency modulation/amplitude modulation modulation principle, enhancement of fault feature frequencies can be achieved by collecting gear acoustic signals from the selected air gaps in GAM structure. In this study, the feasibility of GAM-based gear weak fault feature diagnosis is proved by experiments, which verify the multiscale feature denoising and frequency selective enhancement characteristics of GAM. The results show that the amplitude of the target signal is amplified more than eight times, the sideband component containing the fault signal is enhanced obviously, and the effect of denoising outside the target sideband is evident, making the fault feature can be easily identified from the weak fault signal after demodulation. By comparing with digital filtering in traditional signal processing, this method is more straightforward in extracting weak gear fault features. Additionally, this method adopts non-contact measurement method with micro-electroMechanical system (MEMS) microphone, which has advantages over acceleration sensors in overcoming the space limitation. All in all, the proposed method is effective and facilitates the identification of weak gear faults.