Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.
cell reynolds number
Posted Jul 12, 2010, 6:04 a.m. EDT 1 Reply
Please login with a confirmed email address before reporting spam
Hi there,
i'm using weakly compressible NS with heat transfer (transient). The flow is buoyancy-driven through heating.
I "experienced", that the temperature distribution, which i'm focussed on, depends on the mesh size in the domain where convection is active (+- 15K at 1000K. This is really annoying me. In this case refining the mesh causes higher temperatures. Further refinement has only little influence. So I don't know what is my criteria for knowing that the flow is modeled correctly.
I always look at the cell reynolds number, to be sure, that the flow is locally laminar. But I don't know if there exists a limit for the local Re-Number. cellRe<30? Using the GLS stabilization is needed for Re>2 for convergence. But is there any thumb rule for knowing: The boundary layer is correctly modeled within an error of about 1% (or 5% or whatever)? or things like that? What is your criteria for meshing flow domains? Your experience?
Thanks in advance!
i'm using weakly compressible NS with heat transfer (transient). The flow is buoyancy-driven through heating.
I "experienced", that the temperature distribution, which i'm focussed on, depends on the mesh size in the domain where convection is active (+- 15K at 1000K. This is really annoying me. In this case refining the mesh causes higher temperatures. Further refinement has only little influence. So I don't know what is my criteria for knowing that the flow is modeled correctly.
I always look at the cell reynolds number, to be sure, that the flow is locally laminar. But I don't know if there exists a limit for the local Re-Number. cellRe<30? Using the GLS stabilization is needed for Re>2 for convergence. But is there any thumb rule for knowing: The boundary layer is correctly modeled within an error of about 1% (or 5% or whatever)? or things like that? What is your criteria for meshing flow domains? Your experience?
Thanks in advance!
1 Reply Last Post Jul 12, 2010, 8:28 a.m. EDT
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
I'm no fluid expert, but as a general comment if a variable changes largely with the mesh density, for me, it mainly means that the gradient of the variable is not well resolved by the mesh, so once you refine to get some <10% changes it should be OK (for this particular variable) which does NOT mean that the mesh is sufficiently fine for another physics or variable your are also studying (plus any possible coupling efects).
That is why aI prefer to start solving my physics separately, as far as possible, before I mix them in, but theory is simple, in reality it does certainly not always perform like that.
Good luck
Ivar
I'm no fluid expert, but as a general comment if a variable changes largely with the mesh density, for me, it mainly means that the gradient of the variable is not well resolved by the mesh, so once you refine to get some <10% changes it should be OK (for this particular variable) which does NOT mean that the mesh is sufficiently fine for another physics or variable your are also studying (plus any possible coupling efects).
That is why aI prefer to start solving my physics separately, as far as possible, before I mix them in, but theory is simple, in reality it does certainly not always perform like that.
Good luck
Ivar