Article révisé par les pairs
Résumé : This article describes the calibration of the numerical model of a bowstring-arch railway bridge based on modal parameters. An ambient vibration test allowed the identification of the natural frequencies, mode shapes and damping coefficients of several global and local modes of vibration of the bridge by the application of an output-only technique based on the enhanced frequency domain decomposition method. The calibration was performed using a genetic algorithm that allowed obtaining the optimal values of fifteen parameters of the numerical model. For the mode pairing, a new technique based on the calculation of the modal strain energy was used. The stability of a significant number of parameters, considering different initial populations, proved the robustness of the adopted algorithm in the scope of the optimization of the numerical model. The updated numerical model was validated based on an experimental test for the characterization of the modulus of deformability of the concrete and a dynamic test under railway traffic. The results showed an excellent agreement between numerical and experimental results. © 2012 Elsevier Ltd.