Damage detection of a large-scale bridge model based on shaking table tests.

摘要

In the last twenty years, structural health monitoring (SHM) has become a major topic in civil engineering. The employment of acceleration sensors to monitor and assess the "health" condition of structures has proved to be one of the most successful techniques in SHM applications.;This thesis deals with a large-scale bridge model instrumented by accelerometers and subjected to shaking table tests. The study focuses on the development of a methodology to assess the specimen structural integrity condition based on the measured vibration response to seismic excitations. A stiffness-based damage detection procedure in which the damage assessment founds on the analysis of the linear part of the structural vibration response has been developed.;By means of continuous wavelet transform analysis, it was possible to identify the linear portion of the recorded signals. Structural identification was then performed processing such linear record segments by transfer functions analysis. Finally, damage detection have been conducted using a direct search algorithm that reconciles the modal characteristics predicted by eigenanalysis of a parametrized finite element model of the bridge and the measured vibration characteristics. It will be demonstrated that the results obtained are consistent with both the experimental observations and the results of other studies on the same bridge specimen.