Fully automated peak-picking method for an autonomous stay-cable monitoring system in cable-stayed bridges

Abstract Stay cables are the most important components of cable-stayed bridges, in which bridge loads are directly carried. Because the stay cables dominate the alignment of the bridges as well as their internal force distribution, cable tension forces are crucial to both construction and maintenance. In this regard, a vibration-based monitoring system is considered as a cost-effective method to monitor the cables, and they require the extraction of modal frequencies via peak-picking in the frequency domain. Manual peak-picking is susceptible to human errors, making it challenging for real-time and continuous monitoring. For autonomous monitoring, this study develops a fully automated peak-picking method for stay cables. The proposed method combines information derived from two sources without any prerequisite implementations. The proposed method is extensively evaluated with three methods using real field data. The experimental results demonstrate that the proposed method outperforms the other methods in terms of accuracy and computational efficiency.