Sarah Petry
In regions with low to moderate seismicity, unreinforced masonry (URM) is commonly used for the construction of low to mid-rise buildings. When these structures are subjected to seismic loading, the stiff URM walls attract considerable part of the lateral forces and need therefore to be considered in seismic design and assessment. Nevertheless, the response of URM walls subjected to lateral in-plane loading is not yet fully understood and given estimates for some of the crucial design parameters are unsatisfactory (the displacement capacity and the effective stiffness). Two series of URM walls tested under lateral in-plane loading are presented. First series was built at full-scale using a typical Swiss hollow clay brick and a commercially available standard cement-based mortar. Second series represented the same walls at half-scale. During all tests, the walls were subjected to quasi-static cycles of increasing drift demands, while controlling the boundary conditions (axial load and moment restraint at the top of the walls) such to simulate the typical loading of a ground floor wall in an URM building. From the experiments several new insights on the kinematics of URM walls are drawn. For instance, it is identified that once several diagonal cracks develop in the walls, the deformation capacity of the walls is governed by the separation of the rectangular wall into two triangles. From comparison between both test series at different scales, new recommendations for the correct scaling of hollow core masonry are derived and proposed. In order to compare our own results to other existing wall tests, an existing dataset of URM walls is extended and reanalyzed. Finally, a dataset of 64 quasi-static tests on modern URM walls of different heights and masonry types is presented. The evaluation of the dataset confirms the influence of the boundary conditions on the drift capacity. Moreover, it reveals the presence of a strong size effect and an empirical equation is proposed to estimate the ultimate drift capacity while accounting for these factors. Throughout the wall tests, an optical measurement system is used to track the displacement field of the walls. This measurement was synchronized with the force measurement such that global and local engineering demand parameters of the walls could be linked. This point was crucial for the following contributions of this thesis: (i) proposition of two sets of limit states (LSs) that link local damage states to characteristic points of the global force-displacement curve of URM walls; (ii) the study of different deformation parameters for the validation of mechanical and numerical models at local level. A new mechanical model is proposed which describes the nonlinear force-displacement response of flexural dominated URM walls. For this, first, an analytical part is derived based on the plane section hypothesis and a non-tension material with a linear-elastic constitutive material law in compression. It assumes that only the compressed part of the wall contributes to the wall resistance and accounts for a softening due to the reduction of the effective area. In a second step, new criteria are developed which predict the occurrence of the previous proposed local LSs, which are then incorporated in the analytical model. The new complete mechanical model is validated through the comparison with experimental evidence yielding to good estimates for the effective stiffness and the displacement capacity.
EPFL2015
