A new formalism for improving predictions with RANS eddy viscosity models
by
OHSA/E13
Abstract:
RANS (Reynolds-Averaged Navier-Stokes) eddy viscosity models to approximate the mean flow behaviour are the preferred solution in engineering applications, yet their inability to capture turbulent fluctuations and their difficulties with laminar and transitional flows — due to their tendency to overpredict stresses in adverse pressure gradients — represent a hefty trade-off for their low cost.
We stand on a theoretical framework interpreting the effect of the turbulent stress introduced by RANS eddy viscosity models as that of an implicit temporal filter, to derive a dynamic procedure for the definition of the cμ coefficient in the calculation of the eddy viscosity. The procedure is fully automatic and requires no empirical coefficients or calibration, and it seeks to allow the model to resolve turbulent fluctuations supported by the spatial resolution and to adapt to laminar and transitional flows, preventing the overprediction of stresses. Results on the Benchmark on the Aerodynamics of a Rectangular Cylinder and the ERCOFTAC T3L suggest that both objectives are achieved.
Bio:
Yves earned his Master’s degree in Advanced Automotive Engineering – Racing Car Design from the University of Modena and Reggio Emilia in 2022, and followed it with a research project on the aerodynamic instabilities that lead to proposing in modern F1 cars. Still at UniMoRe, he enrolled in the "Enzo Ferrari" PhD Course in Mechanical and Vehicle Engineering at the end of 2023, resuming his work on the Dynamic Tensorial Eddy Viscosity Model (TVM). The development of turbulence modelling solutions has been the main focus of his research, with the temporal filtering formalism for RANS eddy viscosity models currently at the forefront of his work. In January, he will shift the focus of his studies to racecar aerodynamics for the final year of his PhD programme.
The Laboratory for Simulation and Modeling
Computational Fluid Dynamics Group