Hi everybody Can anyone send me or tell me a site that has the source code of a compressible unsteady flow using the pressure correction method? Thanks a lot in advance!

OpenFOAM:

http://www.openfoam.org

Beware, it's in C++ but it is well written, it's got a lot of functionality and it's just plain nice! The solver you want is called sonicFoam (laminar); there are several other variants, some including turbulence models.

Enjoy,

Hrv

is this sonicFoam a coupled solver or a segregated solver,

further, i wished to ask somethign more,

in my code i implemented the over-relaxed approach as you have mentioned in the thesis, and it works fine, but when i was reading further you have mentioned that for the test cases the number of corrector were , say 2 or three etc.

i wished to ask that how do you construct and solve the corrector equation,

i mean at the moment i take the contribution by k-vector in the descretisation, and hence my equation is complete as a form [ ap*phi + sigma(ai*phi_i) = su ]

so how do i solve the correction equation, does this mean that i construct the similar equation again and solve the matrix for two three times, before moving to next variable.

further, i am facing another problem that k-e or k-w equations are strongly coupled, and thus i find them very difficult to converge, (assumign bicgstab solver), do you have any suggestion for me what to do for treatment of these equations, or what approach openFOAM takes for them.

Thank you a lot for your quick answer, but I have one problem.I am not familiar with C++.Do you know any similar source code that is written in Fortran? Thank you!

Hi George,

Have a look at ISAAC, though it is a structured code,it is very good. Has lot of turbulence model and even LES.

Goto www.sourceforge.net and search for ISAAC. The link can also be found from cfd-onlines resource section.

Good Luck

A.S.

Yes, sonicFoam is a pressure-based segregated solver.

As for the non-orthogonal correction, try thinking about it this way: each Laplacian discretisation will produce two parts. The implicit part is fine because it is built into the matrix and "picks up" the latest field values. For the explicit part, we use a face-interpolated gradient, which is evaluated from the latest available solution, i.e. before the solver is called. The fact that it is explicit means it will not be updated within the solver loop.

Thus, every time the solution changes, its gradient changes and the explicit part of the Laplacian should be updated. In other words, the "non-orthogonal correction equation" is the same as the actual equation - all you do is update the non-orthogonal correction term after each solution.

Hope this was useful,

Hrv

P.S. Sorry, can't help you with a Fortran version, I've been doing only C++ for the last 12 years

i think i userstood it, thanks.

p.s.: its okey if you can't help with fortran, i also don't like it (or could never understand it), for me also c++ is the best (easy to use and goes with my way of thinking about prgramming).

dear sir could you please inform me how we can consider an incompressible flow as a compressible flow by acorrection flow

Try the precondition method developed by Turkel

I recently started to research the subject of the incompressible limit to compressible flow. I found a number of papers on asymptotic expansions in this limit. Do a google search on "Multiple Pressure Variables Method" and you will find a paper by Munz, Dumbser & Zucchini. That paper should be of interest as well as references therein. Also you will find some other similar papers. Follow the links.

I have not read these yet but I think that among them you will find the information you are looking for.

sir i'm working on model POM. In order to obtain a water flux in mediterranean sea the water must be compressible we will find the volume modified by heating. thank you for your intrest please send me paper attached as PDF file or .doc thank you again