Title: Identification of Soil-Foundation Dynamic Stiffness from Seismic Response Signals
Author(s): E. Taciroglu and S.F. Ghahari
Publication: Symposium Paper
Appears on pages(s): 113-128
Keywords: blind modal identification; earthquake data; finite element model updating; soil-foundation dynamic stiffness; Timoshenko beam model.
Prediction of the seismic response of civil structures without considering the flexibility and damping provided by their supporting soil-foundation systems can be unrealistic, especially for stiff structures. In engineering practice, the substructure method is generally preferred for considering Soil-Structure Interaction (SSI) due to its computationally efficiency. In this method, soil is modeled using discrete spring elements that are attached to the superstructure; and the Foundation Input Motions (FIMs)—which are usually calculated through analytical transfer functions from recorded/anticipated free-field motions—are applied at the ends of these springs. Whereas the application of the substructure method itself is simple, the determination of FIMs and the soil-foundation systems’ dynamic stiffnesses are challenging. In the present study, we propose two new approaches to identify the dynamic stiffness of soil-foundation systems from response signals recorded during earthquakes. In these approaches, the superstructure is represented either by a numerical (finite element) or by an analytical (Timoshenko beam) model, and the soil is represented by discrete frequency-dependent springs. In both approaches, the superstructure and
soil-foundation stiffnesses are all identified through model updating. We present various forms for the second approach (involving the Timoshenko beam) and verify these through comparisons with the results from the first approach (involving the finite element model) obtained using earthquake data recorded at the Robert A. Millikan Library at the Caltech campus in Pasadena, CA.