Damage detection on the Z24 bridge by a spectral-based dynamic identification technique

Abstract

The paper tackles the dynamic identification and the damage detection carried out by a spectral-based method on the well-known Z24 bridge, a three-span pre-stressed concrete bridge located in Switzerland. Before being destroyed, the bridge was progressively damaged and tested in the framework of the Brite Euram project SIMCES. Starting from this benchmark, the presented spectral-based identification technique is validated and the usefulness of this method as a non-destructive tool able to catch the dynamic behavior of a structure and locate the damage is widely discussed. Firstly, a FE model of the bridge was built and calibrated in order to analyze its response to different excitation types (free vibration, triangular pulse, swept sine, shaker and random vibrations) and several damage scenarios. Secondly, aiming at identifying both the modal parameters and the damage of the bridge, the spectral-based method is applied making use of the power spectral matrix decomposition. Finally, a proper index is defined and applied to this case-study in order to locate the damage.