No abstract

No abstract

Shaking tables are frequently used to determine the dynamic behavior of structures in the laboratory environment. In order to obtain realistic results in experimental studies, table response and performance should be consistent with the desired motion. In this multidisciplinary study, an application of a new method for determining and calibrating the mechanical response of a developed bi-axial displacement controlled shake table according to the desired motion data is presented. The bi-axial shake table's electro-mechanical components consist of stepper motors, ball screw sets, linear ball bearings, and linear potentiometers positioned on both axes for displacement measurements. For the control and data acquisition (DAQ) unit of the shake table, an open-source electronic prototyping platform Arduino was used. From several experimental results, it was seen that, with the presented calibration method, harmonic and earthquake simulations could be achieved with a relative root mean square error (relative RMS error) of less than 5% for desired displacement-time histories.

This study investigates the pounding phenomenon by shaking table experiments on two scaled building models. Representing the situation where the seismic gap is insufficient, two building models are adjacently positioned on the shaking table, and pounding was investigated for harmonic and strong ground motion excitations. Displacement and acceleration responses were obtained to observe the pounding effect experimentally from video and accelerometer recordings, respectively. The Kelvin-Voigt Model consisting of spring and damper was used for numerical pounding analysis. The most critical parameters of the Kelvin–Voigt model, which are the spring stiffness (ks) and the damping, are calculated according to the coefficient of restitution (r) and were investigated and compared with harmonic experimental results. The obtained parameters, compatible with the harmonic experiments, were used to examine structural behavior under earthquake effect for the case where the building models are positioned for an insufficient seismic gap. For comparison, numerous numerical simulations were realized using different spring stiffnesses and coefficients of restitution. The study shows that when the coefficient of restitution is taken as 0.2 or 0.4, and the ratio of spring stiffness to shear stiffness (ks/k) is 1 or 5, reasonable results in numerical earthquake simulations can be obtained.