Best practices about turbulent modeling#
Attention
In construction
Summary of turbulence model compatibility in TrioCFD#
Model |
VEF |
VDF |
PolyMAC |
|---|---|---|---|
\(k\)-\(\omega\) (Pb_Multi) |
✓ |
✓ |
|
\(k\)-\(\omega\) SST (Pb_hyd_turb) |
✓ |
✓ |
|
\(k\)-\(\varepsilon\) |
✓ |
✓ |
For the historical record, at first, there was only a \(k\)-\(\varepsilon\) model in TrioCFD. At some
point, the multiphase framework arrived with the PolyMAC discretisation. Someone needed turbulence,
so a first \(k\)-\(\omega\) model was made in PolyMAC. Then a need was expressed to have a \(k\)-\(\omega\)
SST model in the historical Pb_hydraulique_turbulent framework. A first implementation was started
in VDF but was quickly stopped to switch to the VEF discretisation. Then the VDF discretisation was
made available with Pb_Multiphase so the \(k\)-\(\omega\) model was extended to it in the
Pb_Multiphase framework. At some point someone needed to use the \(k\)-\(\omega\) model in
Pb_hydraulique_turbulent with the VDF discretisation. So rather than finishing the previous
implementation and validating it, they chose to minimize the validation work and instead allowed
using the \(k\)-\(\omega\) Pb_Multiphase version with Pb_hydraulique_turbulent. Later, new work was
carried on hybrid RANS/LES method. The main framework was the VEF discretisation (so
Pb_hydraulique_turbulent) but some validation was needed using VDF. As the code structure is
different between Pb_hydraulique_turbulent and Pb_Multiphase, it was decided to finish the VDF
version of the \(k\)-\(\omega\) SST in Pb_hydraulique_turbulent.
As of 2025, a complete redesign of the turbulence-related code is under consideration to reduce code duplication and to allow for an easier extension. But workforce is limited and the project is huge.
Recommandations#
The most used model in TrioCFD was historically the \(k\)-\(\varepsilon\) one with a wall function. This is robust and fast. Some low-Reynolds models are available but not yet fully validated. The \(k\)-\(\omega\) SST model being the most widely used one in the litterature, we recommand using it.