Full life-cycle monitoring and earlier warning for bolt joint loosening using modified vibro-acoustic modulation

摘要

Structural health monitoring (SHM) has played a crucial role in continuously assuring adequate integrity across equipment, especially early loosening estimation of bolt joints. Recently, the vibro-acoustic modulation (VAM) technique for bolt early loosening monitoring has received much attention due to its high sensitivity to contact-type defects. However, the current early loosening monitoring mostly focuses on the bolt head and interconnecting componen-t loosening while ignoring the earlier loosening stage. The central contribution of this work is to develop full life-cycle monitoring for bolt loosening and consequently achieve even earlier warning in the stage of thread loosening. Firstly, regarding the bolt joints in the sand spreading device of high-speed trains, a modified VAM method, in which the designed sensor layout is more applicable for engineering and a swept signal is applied as a probing wave to overcome the dependence of detection accuracy on the exact resonance frequency excitation, is conducted for the full life-cycle monitoring. Secondly, the mechanisms, performances and dependency on the applied torques of the loosening-induced nonlinearity are explored with the proposed nonlinear indexes, theoretically and experimentally. The results demonstrate that both the nonlinear elastic mechanism and the amplitude-dependent dissipation effect contribute to such nonlinearity. Further analysis suggests two valuable and not yet mentioned findings: 1) there is a thread loosening stage that exhibits sufficiently sensitive nonlinear modulation index variations to predict bolt early looseness before the bolt head and interconnecting component loosening; 2) such nonlinear indexes are non-monotonically related to the torque levels in both the above two loosening stages, consistent with theory. Further, such full life-cycle monitoring is achieved by associating such nonlinear indexes and a linear transmitted energy index. All of the obtained results will aid in understanding the nonlinear symptoms under various bolt tightening levels and further improving the reliability of the VAM-based bolt early loosening monitoring.