Modal Identification and Model Updating: A Case Study on Historical Iron-Truss Bridges
摘要
Operational Modal Analysis (OMA) is widely used in the frame of Structural Health Monitoring (SHM) to assess the dynamic properties of civil structures from continuous measurements. The results, given as natural vibration frequencies, damping ratios and mode shapes, can be compared to the theoretical results of a Finite Elements Model (FEM) of the structure. In case of differences, some parameters might be tuned to update the model and make it fit to the observed reality of the structure. Moreover, changes in the modal properties can be used to detect anomalies, and precisely assess them by a new update of the model. This method has been recently applied on two historical iron truss bridges spanning the river Loire in France, in the frame of the “GeRICO” research project, as part of the “Programme National Ponts”. This contribution presents the results and feedback of this activity. The monitoring system is described first: one of the two bridges was equipped with a comprehensive set of accelerometers, to distinguish modes of bending, torsion, and distortion of the iron truss box girder. In addition, long-basis strain sensors were included in a comparative analysis to assess their ability to record bending vibration modes. The second bridge had a partial monitoring, on two spans out of eleven only, with a reduced number of accelerometers, to optimize the quantity of sensors. The datasets used for the OMA are discussed: an interesting result is the possibility to use concatenated, short trigger-based records, instead of continuous hour-long ambient noise records, without significant loss of information. Secondly, the link with the FEM is established by the means of advanced correlation analysis between measured and theoretical modal properties. The way to use this correlation to update the FEM is investigated.