Abstract:

In this study, uncertainty quantification was carried out on the damage detection results obtained through the sensitivity-based finite element model updating method for a two-story, single-span plane frame model, where damage was simulated as a regional stiffness reduction at the ends of columns and beams. The damage detection method employed is based on iteratively reducing the differences in modal parameters between damaged and undamaged conditions. The presence of noise in dynamic measurements introduces uncertainties in damage detection results by affecting the estimation of modal parameters. The reliability of the obtained results is possible by quantifying the amount of uncertainty they encompass. In the presented study, uncertainty quantification of damage detection results was carried out by adding noise data associated with different coefficient of variations to modal parameters related to the damaged condition. The model updating process was performed by using modal parameters with added noise data for the damaged condition and modal parameters for the undamaged condition, as well as by updating the element stiffness values of the finite element model representing the undamaged condition. Stiffness reductions of elements were calculated using the determined stiffness values before and after updating, thus enabling the detection of the location and extent of damage. Uncertainty quantification of the damage results obtained under different noise conditions was conducted using analysis of variance, meta-modeling, and sensitivity analysis methods. In this way, factors that introduce substantial uncertainty in damage detection results have been identified.

Keywords: