Influence of Measurement and Model Parameters in Computational Model Updating using UHPC Sampless

The majority of the existing European transport infrastructure was built forty to sixty years ago. Due to the ascending transport loads and increasing traffic the associated costs to maintain our existing infrastructures is permanently increasing. A common solution for this problem is to extend infrastructure lifecycles together with reducing maintenance costs by selective measures. These objectives may be achieved by developing new sustainable materials and infrastructure components as well as innovative approaches to assess the structural condition in real-time. The method to include measurement data in the numeric model is called model updating. This may be achieved by adjusting model parameters manually or by a computational optimizing algorithm. The latter is therefore called computational model updating (CMU). The entire process from determination of test data, post-processing and inverse modeling involves numerous sources of errors and uncertainties. Some of these are caused by noise, incorrect sensor location, inexact equipment calibration, operation in a region of nonlinearity, and parameter settings in system identification algorithms. Some others are related to inappropriate theoretical assumptions, inaccurate estimation of material properties and boundary conditions, insufficient or incorrect modeling detail, computer input errors, and improper application of optimization algorithms. Although computational resources and the sensor sensitivity are increasing, the influences of the modeling parameters and the uncertainty propagation in the process have to be considered to obtain meaningful and interpretable results. This is fundamental if those methods are used to assess reliability and safety of structures in particular bridges.