AMBIENT VIBRATION AND EARTHQUAKE-DATA ANALYSES OF A 62-STORY BUILDING USING SYSTEM IDENTIFICATION AND SEISMIC INTERFEROMETRY

Abstract:

A 62-story residential tower in San Francisco—the tallest all-residential building in California—was recently instrumented by the USGS’s National Strong Motion Project in collaboration with the Strong Motion Instrumentation Program of the California Geological Survey to monitor the motion of a tall building built with specifically engineered features (including buckling-restrained braces, outrigger columns and a tuned liquid damper) to reduce its sway from seismic and wind loads. This 641-ft tower has been outfitted with 72 uni-axial accelerometers, spanning through 26 different levels of the building. For damage detection and localization through structural health monitoring, we use local micro-earthquake and ambient monitoring (background noises) to define linear-elastic (undamaged) dynamic properties of the superstructure including its modal parameters (fundamental frequencies, mode shapes and modal damping values) and shear-wave propagation profile and wave attenuation inside the building, which need to be determined in advance of strong shaking. In order to estimate the baseline modal parameters, we applied a frequency domain decomposition method. Using this method, the first three bending modes in the reference east-west direction, the first two bending modes in the reference north-south direction, and the first two torsional modes were identified. The shear-wave propagation and wave attenuation inside the building were computed using deconvolution interferometry. The data used for analyses are from ambient vibrations having 20 minutes duration, and earthquake data from a local M4.5 event located just north east of Geyserville, California. We show that application of deconvolution interferometry to data recorded inside a building is a powerful technique for monitoring structural parameters, such as velocities of traveling waves, frequencies of normal modes, and intrinsic attenuation (i.e., damping). The simplicity and similarity of the deconvolved waveforms from ambient vibrations and a small magnitude event also suggest that a one-dimensional shear velocity model is sufficiently accurate to represent the wave propagation charactersistics inside the building.