Characterizing Soil-Foundation-Structure Interaction Using Experimental Data From A Test Structure

Abstract

Most of what is currently known about soil-structure interaction (SSI) is based on theoretical models. Since SSI can lead to seismically unsafe designs, as demonstrated by the devastating damage caused by soil amplification during the 1985 Mexico City Earthquake and the 1989 Loma Prieta Earthquake, it is important that these theoretical models be validated by field data. The primary objective of this research is to use experimental data from a well instrumented test structure to characterize the effects of SSI. The soil-foundation-structure-interaction (SFSI) test structure at the Garner Valley field site, operated by the Network for Earthquake Engineering Simulation (NEES) at the University of California Santa Barbara (UCSB), is designed to specifically study SSI. Using data collected from three years of nightly forced vibration tests, the effects of SSI were evaluated using parametric system identification procedures. An average period lengthening and foundation damping of 1.28 and 1.57% were found, respectively. These results are higher than predicted by the theoretical model of Veletsos and Nair (1975). Non-parametric system identification procedures were also used to confirm the fundamental mode of the structure. Both the parametric and non-parametric analyses showed a first mode frequency of about 5.7 Hz for the SFSI test structure. The effects of environmental conditions on structural response were also investigated. This is important for the structural health monitoring since environmental effects could mask damage and interfere with detection. A definite positive correlation between temperature and the first mode of the structure was found. A less definite but observable relationship was found between the water level in the soil and the first mode. In general, as the water level approaches the surface, the frequency decreases.