This study proposes predictive equations for estimating the peak-experienced drift ratio of unreinforced masonry walls based on the visual characteristic of the damages. In this regard, a comprehensive database comprised of 190 images associated with 30 unreinforced masonry walls at different drift ratios between 0.0 and 1.1 percent is collected, and the visual features of the progressive damages are extracted. Various image processing filters are implemented to the images to quantify the cracking length and crushing areas. The filters are capable of distinguishing different crack patterns, such as joint cracking and block cracking. In the following, four scenarios are introduced based on the accessible parameters of the damaged walls to perform the post-earthquake damage assessment. Predictive equations are proposed based on the apparent characteristics of cracking and crushing as well as the mechanical property of the walls and loading conditions. The models provide nonlinear regression equations for estimating the drift ratio with an R factor up to 90 percent. The peak drift ratio predicted by the proposed models of this paper is also attributed to the damage states using existing fragility models. The proposed framework of this study plays a crucial role in post-earthquake structural damage assessment to determine the updated damage state and performance level of the damaged unreinforced masonry walls.
展开▼
