[RJCHARLIE]

I know that laminar model is solving Navier-Stokes equations without any averaging. For 3D simulation, transient analysis of laminar model with time averaging and very fine mesh should give the result similar to that of turbulence modeling with relatively coarser mesh. Since turbulence is inherently 3D, then for a flow with high enough Reynolds number, does a 2D transient analysis of laminar model with very fine mesh and time averaging give the result similar to that obtained by turbulence modeling? Can the 2D laminar model give the correct turbulence velocity profile in the boundary layer?

[FMDenaro]

Originally Posted by RJCHARLIE

I know that laminar model is solving Navier-Stokes equations without any averaging. For 3D simulation, transient analysis of laminar model with time averaging and very fine mesh should give the result similar to that of turbulence modeling with relatively coarser mesh. Since turbulence is inherently 3D, then for a flow with high enough Reynolds number, does a 2D transient analysis of laminar model with very fine mesh and time averaging give the result similar to that obtained by turbulence modeling? Can the 2D laminar model give the correct turbulence velocity profile in the boundary layer?

No, in 2d model you have no stretching of voriticity. That is a very specific constraint that can be suitable only for geophysical flows.

[RJCHARLIE]

Originally Posted by FMDenaro

No, in 2d model you have no stretching of voriticity. That is a very specific constraint that can be suitable only for geophysical flows.

Let me clarify, if I use 2D laminar model for mid-to-high Reynolds flow problem, even though I get convergence and grid-independence, that still means its the wrong result and wouldn't exactly correspond to real world physics? If so, then for 2D model, is it based on user's experience to choose the right model? for example, the user should only use laminar model for very low Reynolds flow (Re<2000) in which based on user's experience for that particular flow problem has definitely no turbulence and if somehow there is some turbulence in the real world version of the same problem, then the results from the 2D laminar model are not correct? Does this mean that, for a 2D problem, I should always use turbulence model for the flow problem which I am not sure is completely laminar?

[FMDenaro]

Originally Posted by RJCHARLIE

Let me clarify, if I use 2D laminar model for mid-to-high Reynolds flow problem, even though I get convergence and grid-independence, that still means its the wrong result and wouldn't exactly correspond to real world physics? If so, then for 2D model, is it based on user's experience to choose the right model? for example, the user should only use laminar model for very low Reynolds flow (Re<2000) in which based on user's experience for that particular flow problem has definitely no turbulence and if somehow there is some turbulence in the real world version of the same problem, then the results from the 2D laminar model are not correct? Does this mean that, for a 2D problem, I should always use turbulence model for the flow problem which I am not sure is completely laminar?

2D laminar model only for:


- low Re flow

- LES/DNS of geophysical flows (with some care)


You can use 2D model for turbulence only for steady RANS.

[LuckyTran]

Note that 2D model even for laminar flows can still be wrong because of the same vortex stretching mechanism being missing. Laminar flows can also have vortexes.


Getting any solution, whether this be numerical, analytical, or even experimental, its merit is judged based not only on the tools and methods used but also the hypothesis it is trying to assess. Getting an exact solution to a Schrodinger equation for example doesn't help you do fluids.