Discussion Forum

Hi

Well if you take the definition of the Reynolds number, i.e. from your CFD books or just from

www.cfd-online.com/Wiki/Reynolds_number

You will see that the "relative importance of inertial and viscous forces in a flow" is iven by Re but it contains a "characteristic length" scale value "L" in [m]. COMSOL cannot say what this value is, often it's related to the pipe diameter, but what if your fluid domain is no pipe ? How can COMSOl know?
However, the mesh average size "h" is known and that also gives you an indication how well the mesh resolves the physical inertial and viscous forces which for me is the reason you have the "Cell Reynolds" number in COMSOL.

I treat it that way, and consider it as an indication of the Reynolds number of my flow, and how I can resolve any turbulence versus laminar flow. But there are better CFD specialist out here that have probably more elaborated explanations, and some practical absolute values to use as scale for Cell Reynolds number considerations

--
Good luck
Ivar

Hi,

I also find the cell Reynolds number little confusing. When I plot the volume-plot cell Reynolds number in COMSOL, the maximum value is almost 10 time lower than what I compute analytically. Is it OK to have such discrepancy?

Thanks,
Kapil

Hi

Check the definition of the "cell Reaynolds" number its based on "h" the average mesh size and not on the geometry size, so it is not to confuse with a standard pipe flow Reynolds number. It is coherent for a regular mesh, but requires some adaption for elongated mesh shapes

--
Good luck
Ivar

Hi，

So how can we use cell Reynolds number if it varies with how fine the mesh is.

Thanks,
Yilin

Hi Ivar,

I do not think the difference is due to the characeristic length. In my opinion, the cell Reynolds number defined in COMSOL is based on velocity increment (delta_u or relative velocity) in each cell while not the absolute velocity used in the common definition of Reynolds number. The velocity increment (delta_u) is much smaller than the absolute velocity. That's why the spf.cellRe calculated in COMSOL is so small.
I am not an expert in CFD either. This is just my thought. Thanks.

Hi I agree, I'm missing a good explanation from COMSOL (i.e. in the doc/help) how the Cell Reynolds number can be related to the classical Engineering one. And how really to use it as a quality measure, with some good examples. Perhaps they are tuned in to give us some hnts here ?

"My way" has always to ensure Cell Reynolds number "reasonably" uniform within a domain, and more or less comparable to the classical one for the different flow types. Even for my clients, I mostly reject to push my models into the turbulent regime, estimating it's a domain where you need far more specific experience than what I have, and can manage to get in the active time left. Anyhow I know a few specialists to whom I'm happy to subcontract turbulent flows to.

-------------------
Have fun COMSOLing,
Sincerely,
Ivar