To measure cracks in reinforced concrete beam without prior detection, application to diffuse cracking evaluation
Year: 2025
Type: conference-paper
Abstract: In recent years, local Digital Image Correlation has become a standard method to monitor cracks in concrete samples during laboratory tests, aiding the understanding of fracture mechanisms.The significant challenge is to efficiently and accurately evaluate crack characteristics from the DIC displacement field.Most existing methods, when they do not employ machine learning techniques in the middle, start with crack extraction.This is usually based on DIC residuals, strain analysis or image processing.The discontinuity kinematic can then be evaluated using points on both sides of the crack and usually require prior crack local frame computation in addition.This paper applies a streamlined method for crack measurement to four-point bending tests on RC beams.The method bypasses the need for crack geometry extraction, local frame calculation and point selection, directly analyzing the full displacement gradient field.Taking advantage of the rather simple displacement field around cracks, the local crack kinematics-such as opening, slip, and orientation-are continuously calculated at all displacement points.The opening and slip measures are maximal along cracks, allowing easy crack extraction after crack kinematic evaluation.RC beams with varying shear reinforcement ratios are studied, demonstrating the method's capability to analyze complex crack patterns.Furthermore, this procedure can be easily adapted for 3D Digital Volume Correlation (DVC) displacement fields obtained from microtomography with minimal modification, offering broader applicability to both experimental and numerical displacement fields, an advantage over conventional methods.
DOI: 10.21012/fc12.1123
Authors: Guillaume Ewald, Yann Malecot
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: gold
OpenAlex ID: W4409434723