Transferring vision-based data to discontinuum analysis for the assessment of URM walls

Abstract

This research presents a new workflow to assess unreinforced masonry (URM) walls based on vision-based data. The morphological features (i.e. brick pattern) of URM walls are obtained through high-resolution orthogonal images and then imported into a three-dimensional discrete element code (3DEC) to perform structural analysis via a semi-automatic procedure. URM walls are represented using rigid blocks that can mechanically interact along their boundaries. The mechanical interaction is simulated based on the point-contact hypothesis considering spring frictional elements. Once the numerical approach is validated using the results available in the literature, the proposed data-driven modelling strategy is applied to different URM wall cross-sections, with or without openings. The results show that the adopted semi-automatic workflow can be a fast and robust solution to simulate the structural behaviour of existing URM buildings by considering their construction techniques efficiently.