Looseness monitoring of multi-bolt connection using acoustic emission

A novel approach based on low sampling rate data is proposed to detect early-stage bolt looseness in structural joints. This study investigates how bolt loosening affects the acoustic emission signal in a multi-bolt connection using low sampling rates. Utilizing a low sampling rate sensor enables continuous and cost-effective structural health monitoring. To validate the method, an experimental set-up was conducted on carbon steel plates fastened with M8 bolts. The proposed technique consists of two main stages. First, the effect of bolt loosening in a single-bolt joint on acoustic signals is analyzed. Second, various bolt loosening configurations are examined in a linear three-bolt setup. The influence of different permutations of bolt looseness in the linear arrangement on the final results is also discussed. The results indicate that even in the presence of fully tightened bolts capable of transmitting stress waves, the initiation of loosening can be successfully detected using time-frequency domain features combined with support vector machine (SVM) classification. Experimental results demonstrate that the proposed method achieves an accuracy of 97.53 % in detecting early-stage bolt looseness. The findings highlight the method's potential as a practical and scalable solution for improving the safety and reliability of bolted connections in industrial applications.