Vibration Based Bayesian Inference for Finite Element Model Parameters Estimation and Damage Detection

Abstract

Structural identification is a very important task especially in all those countries characterized by significant architectural and historical heritage and strongly vulnerable infrastructures subjected to degradation with time and to natural hazards, e.g. seismic and wind loads. Structural response of existing constructions is usually estimated using suitable numerical models that are driven by a set of geometrical and/or mechanical parameters that are usually unknown and/or affected by different levels of uncertainties. Some of these information can be obtained by experimental tests but it is practically impossible to obtain all the required data for reliable response estimations. For these reasons it is current practice to calibrate some of the significant unknown and/or uncertain geometrical and mechanical parameters using measurements of the actual response and solving an inverse structural problem. The identification of structural dynamic characteristics (e.g. natural frequencies, mode shapes and damping ratios) is usually carried out by means of monitoring data provided by Ambient Vibration Tests (AVT). Ambient vibration sources include wind, seismic micro tremors, pedestrian and traffic, which are not deterministic and can be only described by random processes. In the Operational Modal Analysis (OMA) technique this input excitation is not measured and it is assumed to be a stationary white noise. Nevertheless these ambient vibrations can be very far to be stationary: therefore, tools are needed to take into account all these sources of uncertainties. In this paper a new robust framework to be used in structural identification is proposed in order to have a reliable numerical model that can be used both for random response estimation and for structural health monitoring (SHM). First a tentative FE model of the existing structural system is developed and updated using probabilistic Bayesian framework. Second, virtual samples of the structural response affected by random loads are evaluated. Third, these virtual samples are used as “virtual experimental response” in order to assess the effect of the signal sampling parameters on the estimation of structural modal features. Finally, the information given by the measurement uncertainties are used to assess the capability of vibration based damage identification methods using Bayesian approach. The obtained results will be crucial to follow the structural performance and reliability in-time (SHM) and to develop suitable damage detection procedures to be used in a early warning framework.