Hello,
this question is not as trivial as it may seem.
The Laminar model in Fluent/CFX is pretty stable even when the physics are not laminar at all (internal flow, Re=20k, bifurcations/pipe bends that turbulise the flow even further...)
If it were a simple direct solver, undamped turbulence (combined with a typical RANS grid and a large timestep) should cause divergence, as it is the case with icoFoam.
If you know what the "Laminar" model is based on or you have any references, please let me know.
I think that your statement is not 100% true. As it is known from theory of numerical methods for computational fluid dynamics, the usage of certain discretization schemes like 1st or 2nd order upwind schemes is equivalent to a certain amount of "numerical diffusion", i.e. the coarser the mesh than the larger is the contribution of these schemes to amplification of the diffusion coefficient as a consequence of the "numerical diffusion". This has an (intended) stabilizing effect on the numerical solution. On the other hand, if you apply a pure second order central differencing scheme, then I would bet that you observe the expected divergence of the solver at such high Re number immediately.
Regards,
Dr. Th. Frank.
I completely agree with the opinion of Thomas! The answer to your request strongly depends on the numerical scheme you wish to use.
there is no meaning your question, laminar model consists in the use of the original set of NSE without averaging and turbulence models.
- Badges
- Science topic