Computer vision-based stress estimation of concrete structures

Abstract

Losing tension forces of tendons is a critical issue in preand post-tensioned structures. The decreasing tension force can be considered as an initial step of structural damage, as it can further cause concrete cracks, reduced load carrying capacity, and even structural instability. Indeed, it is a
serious threat to structural soundness while difficult to identify. Several approaches for estimating current tension forces have been developed, including ultrasonic wavebased methods, vibration-based methods, and impedance of the piezoelectric material. Although these methods in the
literature have made certain progress in this field, practical use is still limited. Instead of measuring the tension force, this study presents a method that can directly measures the static stress level of concrete by combining the stress relaxation method (SRM) and digital image correlation (DIC). By drilling a small hole, a part of the current static stress can be released, inducing stress field change around the hole. DIC can identify the deformation due to the stress field change using two images taken before and after drilling the hold. This deformation is subsequently compared to one that is calculated using finite element model to finally estimate the current static stress level in concrete. The proposed strategy is validated using concrete specimen loaded by the universal testing machine.