Technical Paper

Best Practice: Generalized k-⍵ Two-Equation Turbulence Model in Ansys CFD (GEKO)D

Two-equation turbulence models have matured to a point where a consolidation seems desirable. The main models used in industrial CFD codes today are k-e [14] (standard and realizable/RKE), k-w [18,33,34] (SST, BSL, Wilcox) and to a lesser degree V2f [12] (different versions). The differences between the models are not fundamental, but can nevertheless have a strong impact on results. For boundary layers, the models differ mostly in their ‘aggressiveness’ to predict separation onset. Furthermore, in the very near wall region, models can predict vastly different results, especially for heat transfer simulations, due to their differences in wall-treatment. There are also noticeable differences for free shear flows, where each model tends to favor certain flows over others. Finally, different models feature different limiters, which typically do not affect the baseline flows, but can have substantial influence in complex applications.

Ansys developed a new turbulence model family called Generalize k-w (GEKO) model with the goal of turbulence model consolidation. GEKO is a two-equation model, based on the k-w model formulation, but with the flexibility to tune the model over a wide range of flow scenarios. The key to such a strategy is the provision of free parameters which the user can adjust for specific types of applications without negative impact on the basic calibration of the model. In other words, instead of providing users flexibility through a multitude of different models, the current approach aims at providing one framework, using different coefficients to cover different application sectors.

The generic idea behind the model will be discussed. Not all details can be provided as the model is at present unpublished. However, the variability of the model will be demonstrated for a variety of generic flows and Best Practice Guidelines for optimal usage will be provided.