CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   OpenFOAM Running, Solving & CFD (http://www.cfd-online.com/Forums/openfoam-solving/)
-   -   Which Solver should I choose? (http://www.cfd-online.com/Forums/openfoam-solving/89106-solver-should-i-choose.html)

mikemech June 4, 2011 13:51

Which Solver should I choose?
 
1 Attachment(s)
Hello,
I would like to simulate incompressible laminar flow over a semi-infinite flat plate at zero incidence, with constant density throughout the flowfield.

The thing is that I want to compare the results that I will get with OF, with a homemade algorithm that I have built in FORTRAN, that simply solves the Continuity and X-Momentum equations, like listed in the attachment.

I am thinking of choosing icoFoam or simpleFoam with turbulence turned off, but I can't understand how they are different (Transient vs Steady State solver) and what will the impact on my obtained solution be.

If anybody has a remark, I would be grateful :)

Mike

ramhari June 4, 2011 14:04

If you are interested in steady state solution, u can opt for simpleFoam.

mikemech June 4, 2011 14:13

So, if I choose simpleFoam, what would be a reasonable value for timeStep in ControlDict? Should I set timeStep= 1 sec and run until convergence?

alberto June 4, 2011 16:38

Yes, and if you want 1D case, create a mesh with only one cell in the Y and Z direction.

Best,

mikemech June 6, 2011 17:09

2 Attachment(s)
Well, I want to increase Reynolds (~10^7) so that I have turbulent flow. I used k-epsilon model. I just introduced 2 files (k and epsilon) under 0 folder and the necessary fvschemes and fvsolution modifications. Is this correct? Or do I need other changes and files under 0 folder, like nutilda ?

And after that, how can I plot the u-velocity in gnuplot or paraview (if possible), towards timesteps ?

Thanks!

PS: i added 2 images. The first is the laminar fully developed boundary layer and the second is the turbulent fully developed boundary layer.

rvmedina20 December 13, 2011 14:38

mikemech, Alberto....

why would one need to give a time step to this type of problem, when you are doing steady-state simulation??? if you have steady-state the time dependence would disappear wouldn't it??? Thanks


-Ricardo

Quote:

Originally Posted by alberto (Post 310523)
Yes, and if you want 1D case, create a mesh with only one cell in the Y and Z direction.

Best,


alberto December 14, 2011 00:47

Quote:

Originally Posted by rvmedina20 (Post 335715)
mikemech, Alberto....

why would one need to give a time step to this type of problem, when you are doing steady-state simulation??? if you have steady-state the time dependence would disappear wouldn't it??? Thanks


-Ricardo

In steady-state solvers, the time-step is just a counter for the number of iterations. Simply use 1.


All times are GMT -4. The time now is 08:24.