Very early loosening detection of bolted structures using acoustic micro-perturbation signatures via a piezoelectric active sensing method

Bolts are widely used to connect individual structural components across various fields, and monitoring their loosening is crucial for ensuring structural safety. To overcome the limitations of saturation observed during bolt loosening detection, a novel Acoustic Micro-Perturbation Signatures (AMPS), derived from the first-order derivative of the signal, is proposed for the detection of very early bolt loosening within the 80% to 100% range of the rated torque via a piezoelectric active sensing method. Simulated signals are firstly generated to evaluate the high sensitivity of the proposed AMPS-based index in identifying the subtle variations of different signals. Both simulation and experimental results consistently reveal that the proposed AMPS-based index demonstrates a linear decreasing trend as the torque increases, with correlation coefficients exceeding 0.99. This strong linear relationship can be employed to detect the very early bolt loosening. Furthermore, experimental results indicate that neither the transducer placement nor the bolt type adversely affects the detection performance, thereby validating the strong robustness and broad applicability of the proposed method for reliably identifying very early loosening states across different bolt configurations. In addition, the tightening torques of M10 bolt are successfully predicted with errors within 4.5%, indicating that the proposed AMPS-based method can accurately and effectively predict very early bolt loosening states. Compared to traditional methods, the proposed approach exhibits a stronger linear relationship between the AMPS-based index and tightening torque without saturation or fluctuation. This study proposes a reliable method for the health monitoring of bolted structures at very early stage.