ABSTRACT
The high-strength bolts are widely used in the infrastructures over the last decades due to the convenience and reliability. However, after years of repeated vehicle or wind loads, the bolt looses and its pre-tight force is gradually wore down, which may cause it invalid. Thus, the monitoring and in-time maintenance for the bolts are significant in engineering. Nowadays some methods have been applied for the management of the bolts. Mostly the manual marker and torque wrench are used to check and re-tight the bolts, which is time-consuming and low efficiency. The stress and vision based non destructive methods are widely studied, but the long term and high precision monitoring can still be a challenge. In this study, a machine vision-based system is proposed. We design a coded marker to precisely describe the rotation status of a bolt. Then we develop a device and an embedded algorithm to monitor the bolt in a long time. The system has been normally operating in a wind power tower for 3 months. The looseness process of several bolts were recorded in high inspection frequency. Thus the stress loss of the bolts were easily obtained. The system is not only easy to deploy in infrastructures with high efficiency for bolts looseness monitoring, but also meaningful for the researches on the mechanism of bolt loosening.
