An analytical model for the out-of-plane response of URM walls to different lateral static loads

The analysis of unreinforced masonry (URM) walls to out-of-plane loads incorporating arching is rather complex. It incorporates large deflections and rather large strains; thus an appropriate model should account for both geometric and material nonlinearities. URM walls are characterized by a brittle failure mode and can sustain a relatively low magnitude out-of-plane load. Predicting the URM wall static and dynamic response is required to assess its possible bearing capacity and survivability under extreme actions such as different static loads, wind loads or blast loads. This article presents a new analytical model that advances the state-of-the-art in predicting the response of one-way arching URM walls, constructed with hollow concrete masonry units (CMUs), subjected to different types of out-of-plane static loads. The first part of the model analyses the initial cracking stages during loading and determines the cracks' locations. In addition, it provides insights for possible changes in the crack location during the wall response. The second part of the model presents a new calculation procedure to obtain the force-deflection curve of the arching URM wall. This procedure is based on the axial compression strains of the full set of masonry units and the in-between mortar joints, where the axial strains of the CMUs are ignored and all the axial deformations are attributed to the axial strains of the mortar joints. This procedure provides the relationship between the lateral load and the wall mid-height deflection and determines the magnitude of the thrust force during the loading process as well. The model takes into account the geometric nonlinearity and accounts for the variations of the thrust force and the restoring moment arm, resulting from the increasing large displacements. The analytical model predictions are compared with two experimental test series dealing with different types of lateral loading. In both cases, a very good agreement is obtained between the model predictions and the test results and insight of the tested specimens is gained.