Numerical and experimental study on scaled soil-structure model for small shaking table tests

The goal of this study is to determine an appropriate scaling coefficient for small-capacity shaking table tests representing the full-scaled field conditions rigorously for the soil-structure interaction (SSI) problem. The scaling approach discussed in this study includes not only geometric similarity but also kinematic and dynamic similarity with the real system. In order to analyze the soil-structure interaction effects on dynamic response of the mid-rise structures under different earthquake motions for both real system and scaled test model were directly accomplished by using 2D finite element method under plane-strain condition. Three different earthquake acceleration records as Chi-Chi (1999), Loma-Prieta (1989) and Kocaeli (1999) have been considered at the bedrock level of the soil-structure system for this study. The real system and the scaled soil-structure coupled model were investigated numerically for three different conditions under original and scaled earthquakes records, respectively: (a) fixed base model, (b) soil-structure model under assumption of linear elastic soil behavior, (c) soil-structure model accounting for plastic deformations of the underlying soil under Mohr–Coulomb failure criterion. To considerably simplify and realistically simulate the SSI analysis, viscous boundaries were used along the truncated interfaces of the model boundaries to avoid unexpected reflection of waves back into the soil region. In the first phase of this work, the representation capacity of the employed scaling factor of 1:45 are evaluated by comparing the structural lateral displacements obtained from fixed base and flexible base analysis for full-scaled field conditions with those for scaled test model to be experimentally examined in the small-capacity shaking table. The scaled geotechnical model is composed of the undisturbed soil specimen taken from real site and the laminar soil container designed to be formed of aluminum frames and rubber layers. Afterwards, the obtained results from the scaled numerical model of the soil-structure coupled system have been validated and verified by those of employing experimental shaking table tests. In addition, the kinematic interaction effect on the free-field ground motion and the local site effects on the earthquake ground motion were assessed numerically and experimentally in the scope of this study. According to the obtained numerical and laboratory test results, it is concluded that the carefully selected large geometric scale coefficients for soil-structure model used in the small scaled shaking table experiments can capture the response of the full-scaled soil-structure system with acceptable accuracy. It is clearly observed that the local soil properties have considerably amplified the earthquake response of the structures and the free field motions in comparison to the fixed base structure and bedrock excitations, respectively.