Colloquium: Javier Fatou Gomez (AWEP)

14 december 2018 09:30 - Locatie: Kronigzaal, Faculty of Applied Sciences, Lorentzweg 1, Delft.

Multi-Fidelity Co-Kriging Optimization using Hybrid Injected RANS and LES

Design optimization processes for complex turbulent flows is nowadays limited by the lack of accuracy of Reynolds-Averaged Navier-Stokes (RANS) and the infeasible cost of multiple Large Eddy Simulation (LES) evaluations in situations such as massively separated configurations. A novel method is presented, which injects LES information into a RANS model, constructing a Hybrid Injected RANS (HIRANS) formulation. It is validated against its injected LES data for the periodic hill test case, and its prediction capabilities are tested in a multi-fidelity Co-Kriging global optimization framework in combination with LES.  The width hill is chosen as the optimization parameter, with a combination of turbulent mixing and total pressure losses for the objective function. 

The HIRANS method satisfactory corrected mean velocity components and turbulent stresses within the 59 LES evaluations compared to the baseline RANS model. In contrast, it was outperformed by a single fidelity LES and multi-fidelity RANS-LES in the optimization process.