CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   OpenFOAM (http://www.cfd-online.com/Forums/openfoam/)
-   -   Turbulence SIG - OpenFOAM Workshop (http://www.cfd-online.com/Forums/openfoam/64415-turbulence-sig-openfoam-workshop.html)

grtabor May 11, 2009 06:33

Turbulence SIG - OpenFOAM Workshop
 
Hi. As those of you who have applied will know there will be a Turbulence SIG meeting on the last day of the OpenFOAM Workshop in Montreal next month. I put my hand up to organise this, and I was wondering what topics we ought to discuss? The intention of this thread is to start getting people thinking about what we want to talk about and what the functions of the SIG should be; I will try and create a summary of posts to this thread to use as a starting point on the day, but contributions on the day would also be very welcome. Additionally, contributions from people with an interest in Turbulence modelling who are unable to make it to Montreal would also be welcome - feel free to raise suggestions for us to talk about.

Gavin

djbungee May 14, 2009 04:41

Hi Gavin,

there is one point that is always discussed when it comes to applying OF in our company. That is the universal boundary condition for turbulence models. In the existing RAS models you have either a wall function or a low-Re formulation. Therefore, you have to have a good idea about the solution in advance or you have to put more effort in generating your grid according to the model you use (and, consequently, your choice of models is limited).

I have some experience in implementing that, but this experience was gained at university when teaching about turbulence modeling a long time ago and the code is implemented using FORTRAN. Moreover, the code structure was somewhat more flexible regarding turbulence modeling and wall/boundary treatment as far as I can see it in OF with my rather limited experience in C++ and object-oriented programming.

I think that this is an important point. It is pretty much standard in commercial codes. I am not aware how this idea has been developed in the OF-extend or whether this is already a major topic.

Best regards, Ulf.

grtabor May 14, 2009 05:16

Hi Ulf,

I wasn't quite sure what your query was about OF - sorry! The existing OF RAS models currently either implement wall models (high Re models) or are low-Re. I know that Fluent (for instance) implements a wall modeling method which switches between wall models according to local y+; which could be implemented in OF fairly easily I would think, if one is prepared to rewrite the individual turbulence models. More significant would be to rewrite the whole turbulence model structure to allow run-time selection of different wall models.

We should certainly discuss this in the SIG - will you be there??

Gavin

djbungee May 14, 2009 05:33

Hi Gavin,

I am sorry but I cannot attend and personally contribute to the SIG as I will not be in Montreal.

Yes, one way would be to manipulate each single model (adding new models). However, you can implement universal wall boundary conditions for turbulence models based on a Taylor series, e.g. refer to http://www.cfd.tu-berlin.de/research...brid_wall.html in a much more flexible way.

The basic advantage is that you keep your models and just manipulate the (wall) boundary conditions for the corresponding quantities as described in the link cited above.

Best regards, Ulf.

braennstroem May 20, 2009 05:41

Hi Gavin, hi Ulf,

there were some discussions about this before, but I would like to mention it again. Similar to the turbo group we should provide a base for validation and further improvement. 'Base' in this case means, that there is a collection of e.g. Ercoftac test cases available to quickly evaluate own developments. These cases should provide the needed scripts for pre and post to compare the results with others and experimental work.

Obviously, something like this (
http://openfoamwiki.net/index.php/Si...vaned_diffuser
) but even a quite simple script for model comparison would be enough. The new model or wall treatment could be tested on a range of 'simple' test cases!?

Regards!
@Ulf: What do you think about an joint-open-project for the development of the mentioned wall treatment... this could be a xing-group contribution.

djbungee June 10, 2009 08:00

Hi Fabian,
I like the idea about the XING contribution, however, this has to be coordinated somehow.
And as far as I see the code right now, especially regarding turbulence modeling, this will not be easy.
Best regards, Ulf.

waynezw0618 June 13, 2009 12:14

Hi

i am interested in the rotating turbulence.
there is lots of works about that. and recently i have try to adding some other`s model to kEpslion and SST in openFOAM.
the test will be taken on a water pump impeller. but i don`t know if it would be on the time for your workshop
wayne

waynezw0618 June 14, 2009 00:44

BTW the modification is based on the Bradshaw Richardson number. i am interested in how to directly add the rotating speed to the turbulence model now.

can any one tell me how ?

thanks


Quote:

Originally Posted by waynezw0618 (Post 219204)
Hi

i am interested in the rotating turbulence.
there is lots of works about that. and recently i have try to adding some other`s model to kEpslion and SST in openFOAM.
the test will be taken on a water pump impeller. but i don`t know if it would be on the time for your workshop
wayne


grtabor June 15, 2009 09:34

Update
 
Hi. I thought I should give a followup on the Turbulence SIG meeting in Montreal. The SIG meetings were scheduled for the morning session on the last day (Thursday) and although a few people had already gone, we had about a dozen people come along to the Turbulence meeting, making it (I believe) the 2nd most popular meeting after the Turbomachinery SIG. I had volunteered to `facilitate' the meeting, and after we had gone around and introduced ourselves I attempted to keep up with the discussion scribbling on the blackboard. (Someone photographed this as a record but I havn't received the pictures yet). I introduced the OpenFOAM Wiki pages for the Turbulence SIG and encouraged people to sign up to the mailing list.

From memory; the following points were raised;

1. It was suggested that a general list of resources related to turbulence modeling would be useful. I've added a page to the SIG Wiki and put on a few entries from my own bookshelf - please add to this list! (Amazon.com links not obligatory!)

2. There was a general request for more tutorials and information about the turbulence models and their implementation in OpenFOAM. This would be useful for people wanting to use the turbulence models, but also for people wanting to implement their own models into the code.

3. There was some discussion about validation of the code and comparison with other (principally commercial) codes. There was a suggestion that `best practice' guidelines and better documentation would be useful in encouraging people to adopt OpenFOAM.

4. Also discussed were ideas for collaboration. Individuals implementing new modelling into the code should be encouraged to document it and make it available through the repository; people with similar interests (eg. DES) should collaborate on this. The SIG could provide a useful `matchmaking' facility for this.

I think that sums things up; if anyone attending has different recollections please correct me! I have edited the Wiki pages somewhat; to take account of these suggestions and some of my own ideas (I've added a section to list all the turbulence-related OpenFOAM papers - please add your own or any you know of to this list!). I'm also aiming to upload some of my teaching materials for OpenFOAM, including some tutorials on the code. I would suggest that improving the SIG Wiki would be a valuable community activity, so please make an effort.

I am also going to post this to the SIG mailing list to see what happens there!

Gavin


All times are GMT -4. The time now is 10:25.