Integrating experimental and analytical data for validating finite element models

Finite element (FE) analysis is a powerful technique for structural analysis. However, the analytical results by FE method are often not consistent with the experimental results due to uncertainties or assumptions in modeling structures, which can lead to a lack of confidence in reliability of FE models. Therefore, it is extremely important to validate an FE model so that the structure can obtain accurate prediction for linear response under unusual loads. This paper describes a methodology for validating FE models by integrating experimental and analytical data to correct uncertainties in modeling structures, through systematic comparison with dynamic response measurements. As an example an exercise on a full-scale reinforced concrete bridge was used to investigate the technique. An FE model was set up and uncertain parameters such as Young's modulus and mass density of concrete and boundary conditions were updated to provide an accurate structural representation.