Multifractal analysis of two-dimensional images for damage assessment of reinforced concrete structures

The most common assessment technique for reinforced concrete shear walls (RCSW) is Visual Inspection (VI). The current practice suffers from subjective and labor intensive nature as it highly relies on judgment and expertise of the inspectors. In post-earthquake events where urgent and objective decisions are crucial, failure of the conventional VI could be catastrophic. Conventional VI is mainly based on width of residual cracks. Given that cracks could close partially (e.g., due to weight of the structure, behavior of adjacent elastic members, earthquake displacement spectrum, etc.), methods based on crack width may lead to underestimating the state of damage and eventually an erroneous decision. This paper proposes a novel method to circumvent the aforementioned limitations by utilizing the information hidden in crack patterns. Crack patterns from images of the surface cracks on RCSW are extracted automatically, and Multifractal Analysis (MFA) are applied on them. Images were taken from two large scale low aspect ratio RCSW under quasi-static cyclic loading, and MFA showed clear correlation with tri-linear shear controlled behavior of walls which was observed in their backbone curves.