CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   Main CFD Forum (http://www.cfd-online.com/Forums/main/)
-   -   About Residual Distribution Schemes and Local Preconditioning for Unstationary NS Eqns. (http://www.cfd-online.com/Forums/main/77-about-residual-distribution-schemes-local-preconditioning-unstationary-ns-eqns.html)

D.A. Caraeni July 29, 1998 04:51

About Residual Distribution Schemes and Local Preconditioning for Unstationary NS Eqns.
 
Hi everybody in CFD!

I'm currently involved in testing my own code, Navier Stokes solver, FV, unstructured grid, using a Multidimensional Upwind Scheme (Matrix Distribution Schemes, like: Narrow, SUPG and LW). For turbulence modeling, I intend to use the LES concept. I will finish testing the laminar solver and implementing one or two SGS models inside my code before Dec. this year. Finally I intend to have a parallel version of my code before June 1999.

My first question is, if there is anybody having experience in this field willing to share these knowledge's with me. I know some Ph.D. students and Professor Deconinck at VKI and Professor Phil Roe and co. at Michigan University are doing almost the same things, but I have no contact with them.

For the unstationary problems (turbulent flow, means always unstationary!) I use a Jameson type scheme and the same kind of (pseudo) time marching scheme as for the stationary (stationary inviscid or laminar viscid flow) case. Therefore, I am also interested if there is anybody that has some experience in preconditioning methods for unstationary NS equation and for NS equation including source term (like centrifugal force and Corriolis force terms for instance).

(I know the work of Van Leer, D Lee, Phil Roe from Michigan, and the experience that they have in preconditioning for stationary NS equations!!)


Looking forward hearing from you,

Yours, D. Caraeni

Ps. "A nice day" everyone.

Farid Moussaoui July 29, 1998 08:54

Re: About Residual Distribution Schemes and Local Preconditioning for Unstationary NS Eqns.
 
Hi, You can found many publications on FS schemes on the web page of the CFD lab at michigan.

http://www.engin.umich.edu/research/...lications.html

Farid.

D.A. Caraeni July 29, 1998 09:42

Re: About Residual Distribution Schemes and Local Preconditioning for Unstationary NS Eqns.
 
Thank you Mr. Moussaoui for your kindness.

Yes, I know those papers (dissertations) of Lisa Messaros, Dohyung Lee, George Tomaich, and others. I was thinking to found some "real persons" that are now currently involved in research in the Residual Distribution Schemes or Local Preconditioning for Unstationary NS field, and to directly discuss with them. Which is the state-of-art, at the level of year 1998, in this field. There are any new publications(late 1997-1998)? Anything new about a new Matrix Distribution Schemes? Is there any solution to "fix" the convergence stability problems of the original matrix-PSI scheme? Any new solution for a high order scheme suitable to use also for NS problems (here I cannot use the HE decomposition!)?!

Sincerely yours, D. Caraeni

Farid Moussaoui July 30, 1998 14:23

Re: About Residual Distribution Schemes and Local Preconditioning for Unstationary NS Eqns.
 
There is an extensive research on FS schemes at VKI. You can see the VKI lectures series. VKI LS 1995-02 and VKI LS 1997-xx on CFD. You can also see the state-of-art in the volume booktitle = {Euler and Navier-Stokes Solvers Using

Multi-Dimensional Upwind Schemes and Multigrid

Acceleration}, editor = {H. Deconinck and B. Koren}, number = {57}, series = {Notes on Numerical Fluid Mechanics}, year = {1997}, publisher = {Vieweg Verlag},

I hope that this help. Farid.

D.A.Caraeni July 31, 1998 04:55

Re: About Residual Distribution Schemes and Local Preconditioning for Unstationary NS Eqns.
 
Dear Mr. Moussaoui,

Thank very much for your advice and help. These book, that you've indicate me, should be of real help for me. I will try as soon as possible to order it. Meanwhile I've found on internet some other references very interesting for me, but nothing about local preconditioning for unstationary NS. I have some results on that but not completely satisfactory and I wanted to see if there is also somebody else working on these topic, to exchange ideas. Thank you again for your help,

Sincerely yours,

Doru Caraeni

Farid Moussaoui July 31, 1998 05:09

Re: About Residual Distribution Schemes and Local Preconditioning for Unstationary NS Eqns.
 
Hi,

There is a web page on preconditioning maintained by Kleb at Nasa. I think it is very usefull. For the preconditioning of the Unsteady NS, you can see the work of Viozat at INRIA. She works with the Roe scheme, and all what she does is to precondition the artificial viscosity in the numerical fluxes without the special treatment for other methods of precondioning where accuracy in time is lost.

I hope that this help. Farid.

Oups, the preconditiong web page is

http://ab00.larc.nasa.gov/~kleb/precond/

D.A.Caraeni July 31, 1998 10:36

Re: About Residual Distribution Schemes and Local Preconditioning for Unstationary NS Eqns.
 
Dear Mr. Moussaoui,

The information's that you've give me were extremely valuable. In these two day I've found (due to your help) a lot of informations in the problems that I'm interested. Thank you very much for your help!

With esteem, Doru Caraeni

Ps. My web page is:

http://www.fm.vok.lth.se/FM/Doru/doru.html

and it contains some informations about me.

M.A.Olshanskii August 5, 1998 07:31

Re: About Residual Distribution Schemes and Local Preconditioning for Unstationary NS Eqns.
 
There is a number of recent results for the preconditioning of incompressible N-S flows, if you interested...

Best.

Aldo Bonfiglioli August 7, 1998 12:04

Re: About Residual Distribution Schemes and Local Preconditioning for Unstationary NS Eqns.
 
Dear Dr. Caraeni, as you were asking in a previous post about people involved in the study of residual distributions schemes, this is my home page:

http://www.unibas.it/utenti/bonfiglioli/www.html.

which is a little bit outdated, but still gives an overview of my activity.

Concerning preconditioning, (H-E splitting), to my knowledge this remains a problem whenever one encounters stagnation points. The only viscous computation I've seen using H-E involved a Mach=2. Re=106 flow over a NACA 0012, where the physical dissipation was probably enough to make things work. However, I have to admit I'm also out of date about preconditioning, so there might have been progresses I'm not aware of.

Using matrix distribution schemes, however, you should be able to solve the unsteady eqns. without H-E splitting (no preconditioning) which shouldn't be too much of a problem unless you're solving for very low Mach numbers.

Regards, Aldo Bonfiglioli


All times are GMT -4. The time now is 09:54.