Effects of variability in ambient vibration data on model updating and damage identification of a 10-story building

This study presents linear finite element (FE) model updating and damage identification of a ten-story reinforced concrete building using ambient vibration measurements. Structural damage was induced to the building by removing six perimeter infill walls. Ambient acceleration response of the structure was recorded before and after the induced damage which are referred to as the reference state and damaged state of the building, respectively. An operational modal analysis method is used to identify the natural frequencies, damping ratios, and mode shapes of the structure using different sets of ambient vibration measurements at the reference state and the damaged state of the building. An initial linear FE model of the structure is created based on in-situ geometry measurements and testing of material samples. The initial model is then updated to reference models using different sets of ambient vibration measurements at the reference state of the building. The updated model parameters reveal considerable variation despite the fact that the identified modal parameters exhibit a much lower level of variability. A subset of the updated reference models are subsequently employed to detect the location and extent of the induced damage by updating the equivalent stiffness of 12 wall substructures using the measured data at the damaged state. Although the identified damage is generally in good agreement with the induced structural damage, the results are found to be sensitive to the variation of the identified modal parameters.