Image-Based Identification of Orthotropic Masonry Textures
Please login to view abstract download link
Analysis of existing masonry is strongly influenced by uncertainty in material properties and by texture heterogeneity. In engineering practice, mechanical properties are frequently assigned from visual/typological classifications and guideline tables [1], whereas the actual mortar–unit layout (order, orientation, and phase arrangement) may significantly affect the response. Despite image-based and homogenization-inspired studies on non-periodic masonry [2–4], a gap remains in providing a resolution-aware procedure that links segmented textures to directional moduli while systematically accounting for input variability. This contribution addresses this gap by proposing an end-to-end image-based workflow that provides direction-dependent in-plane stiffness measures together with uncertainty bands. Segmented binary textures are mapped onto a regular grid through a controlled coarsening ratio (CR); a CR sensitivity analysis is employed to select an operational scale exhibiting stable trends. The coarsened layout is then analyzed using a multi-material Q4 finite element model in plane stress, where the phases are idealized as linear-elastic isotropic. A small set of elementary boundary-value problems is solved to identify a homogenized in-plane (generally orthotropic) response. By rotating macroscopic quantities into an (n, t) reference frame, the workflow returns orientation-dependent moduli as functions of the loading angle. Variability in phase elastic parameters is explored through Latin Hypercube sampling, and the resulting ensemble yields uncertainty bands for the directional moduli. Illustrative applications on masonry textures are presented, and the influence of texture and its orientation on the directional response is discussed.
