Identification of multi-bolt head corrosion using linear and nonlinear shapelet-based acousto-ultrasonic methods

Abstract

Among different types of fasteners used in multiple industries, the bolted connection attracts the most attention due to its low costs and ease of operation. However, bolt failures may induce severe catastrophes if not timely detected. One of the most common bolt failures is bolt corrosion (especially at the bolt head), and bolt head corrosion can significantly affect the mechanical performance of the bolted connection and even the entire structure. So far, no relevant investigations have been conducted to detect multi-bolt head corrosion, and thus the author applies the linear and nonlinear acousto-ultrasonic methods to attempt to solve this problem for the first time. Notably, this paper’s main contribution is that two new shapelet-based acousto-ultrasonic methods, i.e. shapelet-based active sensing and shapelet-based vibro-acoustic modulation (VAM), are developed by utilizing the concept of time-series shapelets. Compared to existing approaches such as entropy-enhanced acousto-ultrasonic methods, the proposed shapelet-based active sensing and shapelet-based VAM can achieve better detection performance. Finally, by conducting a multi-bolt head corrosion test, the effectiveness of the proposed methods is verified. Overall, the proposed methods can provide a new direction for researching bolt corrosion identification, and they can also contribute to the development of structural health monitoring, e.g. the detection of other structural damages in future work.