View Issue Details
| ID | Project | Category | View Status | Date Submitted | Last Update |
|---|---|---|---|---|---|
| 0000751 | OpenFOAM | Bug | public | 2013-02-19 12:02 | 2013-03-14 10:30 |
| Reporter | Assigned To | henry | |||
| Priority | normal | Severity | minor | Reproducibility | always |
| Status | resolved | Resolution | fixed | ||
| Platform | Linux | OS | Ubuntu | OS Version | 12.04 |
| Summary | 0000751: rhoCentralFoam: diffusion coefficients in the two last terms of energy corrector equation | ||||
| Description | In my opinion, the thermal diffusion coefficients in the two last terms of the diffusive energy corrector equation in rhoCentralFoam should both be effective. If I understood correctly, the diffusion would be in principle fully taken into account by the first diffusion term. The last two terms are there to drive consistency between energy and temperature, so that e --> CvT + 0.5magSqr(U), and the two terms would cancel each other out in a fully converged solution ( e=Cv*T 0.5magSqr(U) ). Am I correct in this? If so, shouldn't the second diffusion term have a diffusion coefficient alphaEff (turbulent + laminar) instead of alpha (only laminar)? At the moment the thermal conductivity coefficient (k) is turbulent + laminar. Here's the code: volScalarField k("k", thermo.Cp()*muEff/Pr); // laminar + turbulent thermal conductivity solve ( fvm::ddt(rho, e) - fvc::ddt(rho, e) - fvm::laplacian(turbulence->alphaEff(), e) // laminar + turbulent diffusion + fvc::laplacian(turbulence->alpha(), e) // laminar diffusion, should be alphaEff? - fvc::laplacian(k, T) // laminar + turbulent diffusion ); If I replace the alpha with alphaEff (or comment out the two last terms) in my modified rhoCentralFoam with specie transport for a system I'm considering, I get temperatures consistent with a reactingFoam solution. Am I missing something, or is there a bug? Tatu | ||||
| Tags | No tags attached. | ||||
|
|
There is certainly an issue with the code, either the k should be laminar only and the turbulent part handled by e or if the k is actuall kEff then the alpha sholud be alphaEff to cancel. So the improntant question is what energy/temperature does the turbulence "diffuse"? |
|
|
I would like to note: current calculation of k is erroneous because Pr number for turbulent part of k should be Pr_turb ~ 0.9 not Pr_laminar (0.7 for air) P.S. Apropos: when Cv is constant it is indifferent what to use for turbulent diffuse (e or T) since e=Cv*T |
|
|
As it happens the handling of thermodynamics and transport in rhoCentralFoam has been updated for the next release which resolves the issue being discussed here. However, I am happy to correct the version in OpenFOAM-2.1.x once we have agreement on the form of energy/temperature "diffused" by turbulence. Clearly if Cv is constant it does not matter if T or e is diffused. |
|
|
Resolved in OpenFOAM-2.2.0 and OpenFOAM-2.2.x |
| Date Modified | Username | Field | Change |
|---|---|---|---|
| 2013-02-19 12:02 |
|
New Issue | |
| 2013-02-19 12:11 | henry | Note Added: 0001929 | |
| 2013-02-20 08:56 | Sahas | Note Added: 0001933 | |
| 2013-02-20 09:57 | Sahas | Note Edited: 0001933 | |
| 2013-02-20 10:02 | henry | Note Added: 0001934 | |
| 2013-03-14 10:30 | henry | Note Added: 0002004 | |
| 2013-03-14 10:30 | henry | Status | new => resolved |
| 2013-03-14 10:30 | henry | Resolution | open => fixed |
| 2013-03-14 10:30 | henry | Assigned To | => henry |
