Soil–structure interaction (SSI) is one of the key factors governing structural response during earthquakes. Among its components, kinematic interaction induces differences between free-field ground motions and the motions actually transmitted to th...
Soil–structure interaction (SSI) is one of the key factors governing structural response during earthquakes. Among its components, kinematic interaction induces differences between free-field ground motions and the motions actually transmitted to the foundation (foundation input motion, FIM), thereby directly affecting the estimation of design seismic loads. However, studies that comprehensively evaluate, from the perspectives of two representative approaches—the Ratio of Response Spectrum (RRS) and the Transfer Function (TF)—the effects of major design parameters (building width, embedment depth, and soil shear-wave velocity) on kinematic SSI remain limited. In addition, the theoretical basis and applicability of the SSI-related procedure presented in the Korean design standard have not been sufficiently validated.
In this study, a two-dimensional model considering only the underground structure was developed to isolate kinematic SSI by minimizing the influence of inertial interaction from the superstructure. Dynamic analyses were conducted using the finite-element analysis software, OpenSees. A parametric investigation was performed by varying building width (16–40 m), embedment depth (3–12 m), and soil shear-wave velocity (185–480 m/s), and by applying six ground-motion records with different spectral characteristics.
From the numerical simulations, the response spectrum ratio (RRS) and transfer function (TF) between the free-field motion and the foundation motion were derived and compared with the RRS-based method in ASCE 41, the TF-based method in NIST document, and the input-motion reduction procedure specified in Korean design standard. The results show that the TF-based method most consistently reproduces the frequency-dependent attenuation of input motion due to kinematic SSI across variations in building width, embedment depth, and shear-wave velocity, and thus exhibits the best overall agreement with the numerical results. In contrast, the RRS-based method shows good agreement with the numerical results in low-to intermediate frequency ranges as embedment depth increases, whereas discrepancies grow beyond a threshold frequency (approximately 4–6 Hz). These discrepancies become particularly pronounced for stiffer soil conditions, indicating the need for further examination of an upper applicability limit. The KDS procedure, which implicitly assumes that free-field displacement is identical to the soil displacement adjacent to the structure, fails to properly
capture kinematic SSI and becomes inapplicable under certain conditions.
Moreover, the input ground motions reduced by the three methods exhibited distinct time-domain characteristics, and these differences were directly reflected in structural responses obtained from time-history analyses using a substructure SSI model. Specifically, depending on the kinematic SSI procedure adopted, peak floor accelerations varied by up to 35.2% and maximum interstory drift ratios varied by up to 14.1%. Overall, under the conditions investigated, the TF-based method is considered the most rational method for deriving FIM in practical seismic SSI analyses. The RRS-based method requires quantitative criteria to define its applicability under high shear wave velocity and high-frequency conditions, and the KDS procedure warrants further theoretical refinement and validation of its applicable range.