CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   Main CFD Forum (http://www.cfd-online.com/Forums/main/)
-   -   SIMPLEC Method (http://www.cfd-online.com/Forums/main/69872-simplec-method.html)

meshkati November 7, 2009 09:13

SIMPLEC Method
 
Hey everyone,
I try to develop a CFD code by solving 3D Navier-Stocks equations in curvilinear coordinates. But I have a problem with coupling of pressure and velocity (pressure correction equation - pí) that I use SIMPLEC method for it.
In initial iterates (that no pressure gradients exists in domain) the momentum equations have no source term and then: ap=sigma(a_neighbors).
Therefore the parameter of [ap - sigma(a_neighbors)] that appears in the denominator of SIMPLEC method equals to zero and then the divides by zero error was occurred.
What can I do?
Thanks for advance.
Hadi
(Some of my code specifications are: Finite volume method, non -staggered grid, incompressible fluid, steady calculation, no free surface calculation, Rhie and Chow interpolation method, laminar or k-e turbulence model, SIMPLEC pressure correction method and the test case is straight channel with uniform velocity boundary condition in its inlet and if any another comments was needed just tell me)

ztdep November 7, 2009 21:22

Ap is the value after under-relaxation. for example , Ap/0.6 Ap/0.6-Sigam (A neighbour) >0

hadian November 8, 2009 07:03

also, it is recommended to add Delta_F (Fe-Fw+Fn-Fs+Ft-Fb) to Ap.

meshkati November 14, 2009 07:48

Quote:

Originally Posted by ztdep (Post 235446)
Ap is the value after under-relaxation. for example , Ap/0.6 Ap/0.6-Sigam (A neighbour) >0

That's right, but in SIMPLEC method we can assume all under relaxation factors equals to one (for faster convergence). Then ap equals to ap/urf.

ztdep November 15, 2009 03:47

Quote:

Originally Posted by meshkati (Post 236229)
That's right, but in SIMPLEC method we can assume all under relaxation factors equals to one (for faster convergence). Then ap equals to ap/urf.

" in SIMPLEC method we can assume all under relaxation factors equals to one "
i do not agree with you. if a 1 is used. the simplec will destinated to divergence

universez November 18, 2009 22:58

Quote:

Originally Posted by hadian (Post 235475)
also, it is recommended to add Delta_F (Fe-Fw+Fn-Fs+Ft-Fb) to Ap.

what is meaning of Delta_F (Fe-Fw+Fn-Fs+Ft-Fb) ?

hadian November 19, 2009 01:48

Delta_F=Fe-Fw+Fn-Fs+Ft-Fb and these are convective fluxes passes control volume faces. actually this is the continuity equation and should be zero but it is not zero during the iterations. have a look on the book by Versteeg and Malalaskera.
also, note that in unsteady form of equation you have a A0 which added to Ap. so that Ap-SUM(Anb) will not be zero. moreover a relaxation factor is recommended.


All times are GMT -4. The time now is 03:46.