A new iterative near-field coherent subspace method for rub-impact fault localization using AE technique

Acoustic emission (AE) localization is an important method to detect defects in bearing of rotatory machine for faults maintenance. However, only the faults near the sensor array can be detected due to severe attenuation in the recorded AE signals. Therefore, we propose a highly reliable new Iterative near-field Coherent subspace method (IN-CSM) for multiple rub-impact faults localization. The proposed approach contains four improved processes: Modal plate wave theory (MPWT) analysis for the multi-modes decomposition and group velocity revision; Discrete wavelet transform (DWT) for the useful narrow band extraction; Near field Multiple signal classification (N-MUSIC) method for the preliminary position estimations; the IN-CSM algorithm for the multiple coherent sources separation and the precise localizations. The simulations based on N-MUSIC and IN-CSM methods were compared by rubbing teston the test rig of rotation machinery. The results indicate that the proposed method can effectively locate multiple coherent rubbing faults at once. Thus, it is an effective analysis tool for rub-impact fault detection.