Simulation of the E-Defense 2015 test on a 10-storey building using macro-models

Abstract

The capabilities of certain standard macro numerical models were evaluated by simulating a shaking table experiment that was performed on a full-scale ten-storey fixed-base building with a frame and dual structural system in two perpendicular directions (denoted as the frame and wall directions) at the largest shaking table in the E-Defense centre in Japan. The lumped plasticity model for columns and beams, the multiple-vertical-line-element model for walls and the scissors model for beam-column joints were evaluated. The results indicated that the experiment was simulated reasonably well. The most significant discrepancy was observed between the maximum drifts along the wall direction in the strongest cycle of the strongest test (calculated drift of 1.9% versus measured drift of 1.5%). In other cycles and tests, these differences were smaller. The calculated and measured maximum accelerations along the wall direction in the strongest test were 13.8 m/s2 and 13.5 m/s2, respectively. The discrepancy between the analysis and experiment results was smaller along the frame direction. The maximum calculated and measured drifts were 2.9% and 3.1%, respectively. The maximum calculated and measured accelerations were 15.8 m/s2 and 19.0 m/s2, respectively. In general, the standard input parameters were used in the evaluated models. However, some parameters required modifications, particularly when modelling weakly reinforced beam-column joints with substandard reinforcement that were considerably damaged. Their yielding rotation and near-collapse strength were, on average, reduced to 55% and 30% of the standard value, respectively. One of the most important parameters influencing the response was the effective width of the slabs, which was increased to the total span length for the highly loaded beams. The ratios of the strength, stiffness and amount of dissipated energy in the joints, beams and columns also significantly influenced the response. The adequate ratio of the dissipated energy was obtained by reducing the standard unloading stiffness in the beams and columns. The initial stiffness considerably influenced the response, particularly under weaker excitations. This stiffness was reduced threefold to account for various factors that typically reduce its value, which, among others, includes the influence of preceding tests on the same building with sliding foundations, as well as the assembly, transportation and handling of the specimen.