Forum: OpenFOAM Bugs
December 11, 2010, 20:54
|
|
Replies: 90
Views: 13,948
George, maybe changing the outlet BC is proper...
George, maybe changing the outlet BC is proper for the description of your particular physical problem, but in the case of the problem of this thread, using another outlet BC is solving a complete...
|
|
Forum: OpenFOAM Bugs
December 9, 2010, 13:35
|
|
Replies: 90
Views: 13,948
|
|
Forum: OpenFOAM Bugs
November 8, 2010, 15:16
|
|
Replies: 90
Views: 13,948
|
|
Forum: OpenFOAM Bugs
November 6, 2010, 21:30
|
|
Replies: 90
Views: 13,948
Eugene, I think things are not so polarized,...
Eugene, I think things are not so polarized, Versteeg way to assemble BC's is the same as FOAM, except for the case of downstream BC and upwind scheme. For this particular case I've found that FOAM...
|
|
Forum: OpenFOAM Bugs
November 6, 2010, 11:35
|
|
Replies: 90
Views: 13,948
Duncan, respect of:
this is I'm doing,...
Duncan, respect of:
this is I'm doing, like in Versteeg book, for the last cell in diffusive term, use the value given by the BC, because a centered scheme is used (CD), the upwind is used...
|
|
Forum: OpenFOAM Bugs
November 5, 2010, 23:19
|
|
Replies: 90
Views: 13,948
|
|
Forum: OpenFOAM Bugs
November 5, 2010, 10:37
|
|
Replies: 90
Views: 13,948
|
|
Forum: OpenFOAM Bugs
November 5, 2010, 09:37
|
|
Replies: 90
Views: 13,948
|
|
Forum: OpenFOAM Bugs
November 3, 2010, 21:50
|
|
Replies: 90
Views: 13,948
Hrv, I tested you case, try this:
-ddt(T)...
Hrv, I tested you case, try this:
-ddt(T) steadyState
-div(phi,T) Gauss Upwind
-DT={0.1, 0.01, 0.001}, so Pe_mesh={0.1, 1, 10}
Solutions are in the figure. Due the steady state, no tricks can...
|
|
Forum: OpenFOAM Bugs
November 3, 2010, 17:59
|
|
Replies: 90
Views: 13,948
Hrv, Alberto, problem was:
...
Hrv, Alberto, problem was:
V*dphi/dx-k*d^2phi/dx^2=0, phi(0)=1, phi(L)=0
exact solution,
phi(x)=(-e^(V/k*L)+e^(V/k*x))/(1-e^(V/k*L))
completely bounded between 0 and 1. Until I understand...
|
|
Forum: OpenFOAM Bugs
November 3, 2010, 15:32
|
|
Replies: 90
Views: 13,948
Hrv, this is we were talking about:
Test...
Hrv, this is we were talking about:
Test case explanation and exact solution is in #1 (FOAM case attached), examples of output in #2, correct formulation by Prof. Versteeg in #37.
BTW, What are...
|
|
Forum: OpenFOAM Bugs
October 22, 2010, 17:43
|
|
Replies: 90
Views: 13,948
|
|
Forum: OpenFOAM Bugs
October 20, 2010, 13:47
|
|
Replies: 90
Views: 13,948
|
|
Forum: OpenFOAM Bugs
October 4, 2010, 16:29
|
|
Replies: 90
Views: 13,948
Eugene, I think your solution is the most...
Eugene, I think your solution is the most elegant, I'll try to implement it. BTW I'll look for some FVM examples of this problem, because all my references are in FEM and FDM.
Duncan, I attached a...
|
|
Forum: OpenFOAM Bugs
September 30, 2010, 19:57
|
|
Replies: 90
Views: 13,948
Eugene,
if I understand well your...
Eugene,
if I understand well your proposal is to use the weights passed to valueInternalCoeffs and valueBoundaryCoeffs as a way to know what scheme is being used, so 1 is used for a face...
|
|
Forum: OpenFOAM Bugs
September 29, 2010, 17:45
|
|
Replies: 90
Views: 13,948
I was checking this today and for upwind the...
I was checking this today and for upwind the weights at boundaries are set in the same form that internal ones. 1 for downwind and 0 for upwind (they are set in 1 in the constructor but are changed...
|
|
Forum: OpenFOAM Bugs
September 29, 2010, 17:29
|
|
Replies: 90
Views: 13,948
|
|
Forum: OpenFOAM Bugs
September 29, 2010, 10:33
|
|
Replies: 90
Views: 13,948
Eugene's comment is a correct summation of my...
Eugene's comment is a correct summation of my proposal (really is not my proposal but what I learnt from my advisor). The value written in the solution is the same that user gave us, because it is...
|
|
Forum: OpenFOAM Bugs
September 27, 2010, 12:12
|
|
Replies: 90
Views: 13,948
Thanks, Alberto & Eugene, some comments,
I...
Thanks, Alberto & Eugene, some comments,
I was thinking a lot in Henry's answer and in writing to Hvr. about the matter, thanks for do it. Looking at the code is true that fvm::div is assembled in...
|
|
Forum: OpenFOAM Bugs
September 24, 2010, 20:52
|
|
Replies: 90
Views: 13,948
I'm downloading the -dev/-ext versión, it is...
I'm downloading the -dev/-ext versión, it is taking a very long while... I'll post the news once download and compiling are finished. Meanwhile, I'm studying Eugene suggestions and checking them with...
|
|
Forum: OpenFOAM Bugs
September 24, 2010, 09:42
|
|
Replies: 90
Views: 13,948
|
|
Forum: OpenFOAM Bugs
September 23, 2010, 22:23
|
|
Replies: 90
Views: 13,948
Eugene, your approach is interesting, few days...
Eugene, your approach is interesting, few days ago we (the group in which I'm working) were wondering how to circumvent the issue, and it is related in fact with gaussConvection scheme as you said....
|
|
Forum: OpenFOAM Bugs
September 23, 2010, 08:57
|
|
Replies: 90
Views: 13,948
Alberto, your observation is very clever
...
Alberto, your observation is very clever
I tried going down in Pe changing the diffusion and results were OK with central difference scheme (gauss linear), i.e. linear solution between inlet...
|
|
Forum: OpenFOAM Bugs
September 22, 2010, 11:42
|
|
Replies: 90
Views: 13,948
Similar problem was reported previously by...
Similar problem was reported previously by another independent person:
http://www.cfd-online.com/Forums/openfoam/76790-convection-diffusion-1d-wrong-solution-large-delta-x.html
Could you...
|
|
Forum: OpenFOAM Bugs
September 21, 2010, 10:35
|
|
Replies: 90
Views: 13,948
Alberto,
problem is simply steady...
Alberto,
problem is simply steady advection-diffusion so you can set two BCs. You have an advective velocitiy V, a fixed value at inlet and a fixed value at outlet. Due to diffusion both fixed...
|
