CFD Online Logo CFD Online URL
www.cfd-online.com
[Sponsors]
Home > Forums

Search Results

Register Blogs Community New Posts Updated Threads Search

Showing results 1 to 17 of 17
Search took 0.00 seconds. Search: Posts Made By: eugene
Forum: OpenFOAM Bugs December 8, 2010, 18:07
Replies: 90
Views: 13,948
Posted By eugene
Thanks Alberto. Here are the images in question...

Thanks Alberto. Here are the images in question (I hope!).

The one with the high velocity streak along the bottom (outlet) and left wall is the normal convection assembly formulation. The upwind...
Forum: OpenFOAM Bugs December 6, 2010, 11:19
Replies: 90
Views: 13,948
Posted By eugene
Just a quick update on a practical application...

Just a quick update on a practical application that benefits from the approach discussed here. I have a boundary similar to Fluent's porous jump outlet. The normal velocity is zero gradient, but the...
Forum: OpenFOAM Bugs November 9, 2010, 12:29
Replies: 90
Views: 13,948
Posted By eugene
Well I think we are pretty much in agreement...

Well I think we are pretty much in agreement then. FOAM imposes boundaries in the strong sense, most other approaches do so weakly to ensure stability for upwind and Pe > 1. This doesn't mean the...
Forum: OpenFOAM Bugs November 7, 2010, 07:32
Replies: 90
Views: 13,948
Posted By eugene
Santiago, if you look in the code, Versteeg's way...

Santiago, if you look in the code, Versteeg's way and the OpenFOAM way are not equivalent. They might produce equivalent results most of the time, but there is no consideration of the convection...
Forum: OpenFOAM Bugs November 6, 2010, 20:24
Replies: 90
Views: 13,948
Posted By eugene
It seems clear (as mud) to me that there is no...

It seems clear (as mud) to me that there is no right or wrong here, just two different ways of approximating the same thing. I am not aware of any physical systems that are even conceptually similar...
Forum: OpenFOAM Bugs November 5, 2010, 08:03
Replies: 90
Views: 13,948
Posted By eugene
Lets not get all excited. Hrv, I can second...

Lets not get all excited.

Hrv, I can second Santiago's latest analysis. This is a fundamental issue, but it is only apparent if you use upwind or an upwind-biased scheme (more than 50% upwind) in...
Forum: OpenFOAM Bugs October 25, 2010, 03:42
Replies: 90
Views: 13,948
Posted By eugene
The correct formulation turned out to be...

The correct formulation turned out to be unnecessary for me as there was no flux over the new boundaries I had to introduce.
Forum: OpenFOAM Bugs October 21, 2010, 04:55
Replies: 90
Views: 13,948
Posted By eugene
Well that seals it in my opinion. Good timing...

Well that seals it in my opinion.
Good timing too, I was just wondering about how you were getting along as I have run into another problem where this div thing at the outlet is an issue. I will...
Forum: OpenFOAM Bugs October 1, 2010, 10:50
Replies: 90
Views: 13,948
Posted By eugene
Yes. The formulation I suggested will result in...

Yes. The formulation I suggested will result in the correct proportional combination of upwind and the boundary value for any scheme that is a combination of upwind and downwind provided the input...
Forum: OpenFOAM Bugs September 29, 2010, 17:43
Replies: 90
Views: 13,948
Posted By eugene
The thing is Alberto, there are a wide range of...

The thing is Alberto, there are a wide range of (admittedly marginal) problems that will be better served by, lets call it a scheme dependent boundary. In every case, you could go and write a...
Forum: OpenFOAM Bugs September 29, 2010, 17:20
Replies: 90
Views: 13,948
Posted By eugene
Here is my suggestion for a simple weighting...

Here is my suggestion for a simple weighting scheme on the boundary:
pure upwind = 1
pure downwind = 0
central/linear = 0.5

Since we don't know anything concrete about what is outside the ...
Forum: OpenFOAM Bugs September 29, 2010, 12:27
Replies: 90
Views: 13,948
Posted By eugene
Santiago, well it seems to me that a simple...

Santiago, well it seems to me that a simple switch would be half a job. If your div scheme uses 50/50 upwind+linear blending, then your face value should consist of 50% internal value + 50% face...
Forum: OpenFOAM Bugs September 28, 2010, 08:55
Replies: 90
Views: 13,948
Posted By eugene
There is one fundamental problem with your...

There is one fundamental problem with your proposal, it does not distinguish between the gaussConvection and gaussLaplacian operators. The really fundamental aspect of what Santiago is proposing is...
Forum: OpenFOAM Bugs September 27, 2010, 08:12
Replies: 90
Views: 13,948
Posted By eugene
I just had a more in-depth look at how the matrix...

I just had a more in-depth look at how the matrix is constructed from the boundary conditions and the values of the weighting field (for upwind at least).

Some things to note:
1. gaussConvection...
Forum: OpenFOAM Bugs September 27, 2010, 06:55
Replies: 90
Views: 13,948
Posted By eugene
This was my point. It isn't 100% clear to me that...

This was my point. It isn't 100% clear to me that there is a right or wrong in this case. I wrote a long reply about it on Saturday evening, but then lost it due to some browser hiccup. Grrr.
...
Forum: OpenFOAM Bugs September 24, 2010, 06:18
Replies: 90
Views: 13,948
Posted By eugene
I think the light has gone on. Its not a problem...

I think the light has gone on. Its not a problem with convection in particular, although this is where you are most likely to encounter the issue, because it is the only operator which commonly uses...
Forum: OpenFOAM Bugs September 23, 2010, 19:38
Replies: 90
Views: 13,948
Posted By eugene
I have have run into this issue a few times...

I have have run into this issue a few times myself without knowing what the source of the problem was (other than "knowing" that advecting something into a fixed value outlet is bad). Some of my...
Showing results 1 to 17 of 17

 
Forum Jump

All times are GMT -4. The time now is 20:02.

Contact Us - CFD Online - Privacy Statement - Top