Application of a Time-Domain Local Identification Methodology to Compact Analysis of Continuous and Complete Structural Response Data

There is need to develop and evaluate new strategies for structural health monitoring, focusing on efficient problem decomposition to rapidly and accurately determine the occurrence, location and level of small changes in the underlying structural characteristics of a densely and continuously monitored structure based on its response to known and unknown forces. Under certain conditions found in many mid- and high-rise buildings, one can model a structure using a chain-like MDOF model with story masses connected above and below by single springs. If complete set of response sensors are installed at every floor, one can then apply a robust time-domain identification technique for chain-like MDOF systems in which the identification of each link of the chain is performed independently to estimate inter-story restoring functions. These are the mass-normalized local stiffness and damping values, and can be linear or nonlinear. This study presents the methodology applied to a full-scale 17-story building located on the UCLA campus. The Factor Building is permanently instrumented by the USGS Advanced National Seismic System (ANSS) with a state-of-the-art dense array of 72 accelerometers with four accelerometers on each floor. High-resolution, 24-bit data are being continuously recorded and archived through ANSS. In this study, we considered 50 days of continuously-recorded ambient vibration data and also segments of elevator-induced vibration data. Chain system identification was successfully implemented using these output-only and input-output data. Probability distributions of the estimated coefficients of displacement and velocity terms in the inter-story restoring functions are presented. Variability of the estimated parameters due to environmental conditions is investigated for the 50 days' data to understand the underlying real-world statistical variability. Results show promise for the continuous monitoring and detection of structural changes due to earthquakes or other extreme loads if dense, high-resolution data are available.