| dc.description.abstract | During seismic ground shaking, the motion of a soil–pile system is different from that of the free field due to seismic soil–structure interaction (SSI). Kinematic soil–pile interaction is known to cause significant filtering of the ground motion, resulting in the foundation input motion (FIM) differing from the free-field ground motion. In certain situations, where embedment of the pile cap is ensured, such as in the case of a piled raft (PR), previous numerical studies have shown that embedment effects can cause additional alterations to the FIM. In this work, we investigated the influence of an embedded pile cap on the seismic response of a pile group (PG) employing physical and numerical modeling. A shaking-table test program was designed to investigate the embedment effects on the seismic response of a scaled 2 × 2 PG in clay, in the absence of superstructure inertia. Two identical PG models––one embedded and the other free standing––were subjected to a series of harmonic and white-noise signals, following which, the responses were assessed in terms of the transfer functions and spectral ratios. The ratio of translational response amplitudes of the PR to the PG indicated that embedment effects can lead to significant filtering of the ground motion at higher excitation frequencies. Unique experimental evidence is presented showing that pile-cap embedment can result in additional filtering of the ground motion, even for a highly nonlinear soil response. The results from the experimental program were complemented by numerical analyses of a real-world bridge support system where the influence of embedment on the bridge deck response was studied for a set of seismic ground motion records of varying intensity. We confirmed that the embedment effect is a SSI problem, independent of the earthquake-induced soil nonlinearity, and loss of the soil–pile-cap contact can lead to higher energy being transmitted to the superstructure. | |
