Influence of System Uncertainties on Structural Damage Identification through Ambient Vibrations of Steel Structures

Abstract

The practical difficulties presented by forced vibration testing of large steel structures, such as tall buildings, transmission lines or bridges, led to an increased interest in structural monitoring through ambient vibrations, which usually allows the proper identification of modal properties, natural frequencies, damping and modes of vibration. Changes in these modal properties constitute an indication of structural damage, which may then be assessed on the basis of experimental evidence. The authors proposed an approach to determine the so-called damage damping and stiffness matrices, which are essential to identify the location and intensity of damage. No restrictions were introduced on the damping matrix of the system. The approach requires ambient vibration data of all relevant coordinates used in the structural model, which are processed employing the SSI method. In practice, the identification method is seriously hampered by ambient factors such as temperature or humidity. In general those effects must be filtered out in other to obtain a reliable diagnosis of damage, approach that demands long term monitoring. In this paper, an alternative approach is explored, based on the introduction of error damping and stiffness matrices. Data on both matrices is generated on the basis of observed variations of structural member stiffness and damping caused by ambient factors. The influence of this uncertainty on the identified spectral properties is assessed by simulation.