Ambient vibration studies for system identification of tall buildings
Earthquake Engineering & Structural Dynamics
Copyright © 2002 John Wiley & Sons, Ltd.
The performance of a building under wind and seismic loads depends on stiffness and mass distribution, and may be estimated using finite element codes. Experience has, however, shown that such finite element models often fail to predict accurately the fundamental natural frequencies. Usually the frequencies will be underestimated, that is the building will turn out to be stiffer than anticipated, meaning the design would usually be conservative. On the other hand, effects like torsional eccentricity and foundation compliance may not be correctly modelled, which could be less desirable. A full understanding of linear performance under lateral loads can be obtained through experimental evaluation of the vibration modes. Traditionally only a limited range of modal analysis procedures and software has been applied to civil applications and the ‘special case’ where no input forces can be measured has been the usual situation for large civil structures. Recent developments in system identification, which is the set of procedures to build mathematical models of the dynamic structural systems based on measured data, have added significantly to the potential of ambient vibration or ‘output only’ testing. The aim of the research reported here has been to apply and evaluate the procedures on typical buildings. The procedures are briefly explained and two experimental programmes are then described; a long-term tremor monitoring exercise on a 280m office tower and an ambient vibration survey of a smaller office block. The different forms of response data are examined to study the performance of the analysis procedures and expose benefits and limitations in their use. There is a growing interest in output-only modal analysis procedures in civil engineering. The experience reported in this paper has shown that quick and reliable estimation of mode shapes and frequencies can be obtained, even with small amounts of data. Judgement of modal participation and damping ratios requires more detailed study yet the results are at least as convincing as existing and relatively limited frequency domain methods.
This is the peer reviewed version of the article, which has been published in final form at DOI 10.1002/eqe.215. This article may be used for non-commercial purposes in accordance with Wiley Terms and Conditions for Self-Archiving.
Earthquake Engineering & Structural Dynamics, 2003, Vol. 32, Issue 1, pp. 71 - 95