Wave method for structural system identification and health monitoring of buildings based on layered shear beam model
Year: 2014
Type: dissertation
Abstract: Monitoring the integrity of a building based on instrumental data and detecting damage during or soon after an earthquake or some other natural or manmade disaster may significantly reduce loss of life and injuries caused by potential collapse of a weakened structure during shaking from aftershocks and facilitate emergency response in large cities. Likewise, it would help prevent or reduce loss of function of critical facilities and monetary losses caused by needless evacuation of a structure that is safe. This dissertation presents a wave method for structural system identification and structural health monitoring (SHM) of high?rise buildings. The method is robust when applied to real structures and large amplitude response, and is not sensitive to the properties of the underlying soil and its changes. It is intended for use in seismic alert systems as well as for general condition monitoring. ? The method uses data from an array of accelerometers, identifying the velocities of waves propagating vertically through the structure, and detects changes in these velocities, possibly caused by damage. The identification is based on fitting a layered shear beam model of the buildings, such that the layers correspond, in general, to a group of floors. The fitting involves matching propagating pulses in impulse response functions computed at different levels in the building. Three identification algorithms are presented: 1) the direct (ray) algorithm, which is based on reading the pulse arrival time, 2) the time shift matching algorithm, which is an iterative version of the direct algorithm, and 3) the waveform inversion algorithm, which involves nonlinear least squares (LSQ) fit of the pulses as waveforms over a selected time interval. The latter is extended to an automated moving window analysis, i.e. time-velocity analysis, for detecting changes during an earthquake. ? Detailed applications of the method are presented for three full?scale buildings: (1) a typically instrumented tall steel building during six earthquakes over a period of 19 years, none of which caused damage (Los Angeles 54?story Office Building), (2) a densely instrumented high?rise reinforced concrete (RC) building during one smaller local earthquake (Millikan library in Pasadena), and (3) a damaged high?rise RC building (Sherman Oaks 12?story Office Building, during the San Fernando earthquake of 1971). Also, analyses of models are presented, which demonstrate the insensitivity of the method to the effects of soil?structure interaction, the effects of the slabs on the wave propagation, and the modeling error when the method is applied to buildings with significant bending deformation (e.g. buildings with shear walls). ? It is concluded that the presented method is robust, accurate and sensitive to damage, being able to detect considerable changes in an RC building that has been only lightly damaged, and permanent changes over time in a tall steel building that has shown no signs of damage. With careful consideration of the frequency bands for the fit, it can be applied to many high rise buildings typical of metropolitan areas, such as, e.g. Los Angeles and San Francisco in California, and urban areas in general. Directions for research on further development of the method are identified.
Author: Rahmani, Mohammadtaghi (author)
Language: English
Subfield: Civil and Structural Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W2518859746