Pre- and post-simulations of shake-table tests on a stone masonry building aggregate using finite-discrete elements

This work presents the simulations of the non-linear dynamic response of a three-dimensional finite-discrete element model. The model simulates a half-scale masonry building aggregate tested on a shake table by other Authors. The aggregate is made of two un-connected building units having different heights and slightly different wall thicknesses. The floors are made of timber beams and boards. The modelling approach accounts jointly for in-plane and out-of-plane responses, which can be expected given the high flexibility of the floors, and for the separation between the two building units. The simulations are related both to the blind predictions, according to a scheduled testing sequence, and to the post-dictions according to the actual testing sequence and some model calibrations. The prediction model overestimates displacements, underestimates base shear and fairly predicts the damage pattern of comparable experimental runs. The use of the recorded shake table motion improves the accuracy of the post-diction simulations, while still delivering beam unseating. A higher Young’s modulus of the blocks improves markedly the predictions. The strengthening intervention with steel angles connecting floors to walls is only approximately modelled and does not improve the outcomes of the simulations. In summary, the adopted modelling approach is capable of accounting for the pounding between the two building units, predicting the most significant damage as well as estimating approximate average of peak values of base shear and displacements, while individual time histories are less accurately estimated.