Determination of open boundary conditions for computational fluid dynamics (CFD) from interior observations

An optimization approach for the determination of open boundary conditions for Computational Fluid Dynamics is introduced, whereas the error between the solution σ and interior observations ω is minimized. The numerical weather prediction (NWP) model ALADIN–Austria provides data of wind speed and wind direction at virtual weather stations within the area of interest. Also, data from real weather stations and other sources can be incorporated into the model, respectively. In this work, the optimization method is applied to the constant density Navier–Stokes Equations. Thereby, for stabilizing the ill-posed pseudo inverse problem several regularization methods are reviewed. Further, numerical studies are carried out to identify the supreme regularization method for the presented application. Finally, the algorithm is applied to the micro- and meso-scale flow over the Grimming mountain, Austria. The results are compared with real weather station data and show suitable correlation with the measurements.