Uncertainty Modeling and Quantification
The goal of the iWG "Uncertainty Modeling and Quantification" is the adequate representation of data uncertainty in the context of the development of the Digital Twin Road. In the context of the SFB/TRR 339, a large amount of measurement data is acquired, for example for the determination of material parameters or for the determination of highly accurate position data. In the numerical subprojects, this data is used to introduce and validate simulation models for different aspects of the road. The uncertainty present in this data cannot be ignored.
The uncertainty is due to several aspects: inherent variability of material and geometry parameters as well as inaccuracies of measurements and missing data. Based on the origin of the uncertainty, different modeling methods (e.g., probabilistic and possibilistic modeling approaches) are applied. For this reason, an exchange between the subprojects generating the data and the simulation subprojects is necessary (Interface I).
In order to be able to consider the uncertainty of the input parameters in a simulation and to perform an uncertainty analysis, several requirements for the numerical model need to be considered (Interface II). On the one hand, a functionality for the automatic change of the input parameters is necessary. Furthermore, a numerically efficient model is essential. Strategies to increase the efficiency are investigated, for example, the parallelization of simulations and the application of surrogate models.