There are many buildings supported by composite foundation with piles and footings on inclined bedrock. In previous studies (e.g. Kakurai (2009)), seismic lateral forces of this kind of composite foundation structures has been pointed out that can be biased. On a previous report (Yamada et al. (1999)), they are biased to the pile side of the foundation. On another previous report (Seki et al. (2003)), they are biased to the footing side of the foundation. It is assumed that it depends on seismic input and system parameters that which is loaded bigger either on pile side or on footing side. In this paper, we studied the effects of torsional seismic response at this kind of composite foundation that affect seismic inertial force of a building by using Finite Element Method (FEM).
This FEM analysis was carried out for the purpose of investigating how seismic inertial forces of the building is distributed to the composite foundation. For this purpose, we created 2 models. One is the model of the whole system to calculate the whole seismic loads, the other is the model of the soil and the foundation (no building) to calculate the seismic loads not including the inertial force of the building. We separated the seismic inertial force distribution from the whole seismic loads by combining the results of these calculations.
According to the results of this FEM analysis, it is phase lag between torsional response of the structure and translational response of the structure that decide which side of the foundation is loaded bigger. Because torsional inertial force can be replaced by a force couple, on the one side of the structure, torsional inertial force and translational inertial force are overlapped each other, on the other side of the structure, they are canceled out. It depends on the phase lag between the torsional response and the translational response that which side is overlapped or canceled. So the phase lag decide which side is loaded bigger.
The phase lag is decided by the relationship of the various characteristic periods. They are the predominant period of the input ground motion, the natural periods of the surface soil, and the natural periods of the building. If the input ground motion is dominated by a single period, it is decided clearly which side is loaded bigger. But in general, it is difficult to assume the input ground motion as design input which is dominated by narrow-bandwidth signals, and it is normal to assume the design input which is dominated by broad-bandwidth signals. So it is necessary for seismic design of a building supported by composite foundation to consider both cases of that pile side is loaded bigger and that footing side is loaded bigger.
But on condition that the building is seismically isolated, there may be cases that the bias of seismic lateral force is not remarkable. In the case of seismically isolated building have natural periods that is much longer than natural periods of the surface soil, the bias of seismic lateral force can be not remarkable because the natural periods of the building and the predominant period of the seismic response of the surface soil (that is seismic input to the building) can be different from each other.
