CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   CFX (http://www.cfd-online.com/Forums/cfx/)
-   -   why CFL is only in coupled solver? (http://www.cfd-online.com/Forums/cfx/23898-why-cfl-only-coupled-solver.html)

 john April 23, 2007 12:58

why CFL is only in coupled solver?

why CFL is only in coupled solver? and why cant we see it in segregated solver? thanks

 opaque April 23, 2007 14:45

Re: why CFL is only in coupled solver?

Dear John,

Would you mind elaborating on your question? It is not clear (at least to me) what you are referring to..

Opaque..

 Tim April 23, 2007 16:54

Re: why CFL is only in coupled solver?

The two solver types work differently. The CFL number defines the number of cells through which information passes in a single iteration/time step. The higher your CFL number the faster the case will converge, but it also makes it more unstable, so it's also more likely to diverge.

Segregated solvers do not have this concept.

 Glenn Horrocks April 23, 2007 18:12

Re: why CFL is only in coupled solver?

Hi,

CFL number is equally applicable to coupled and uncoupled solvers. The importance of the CFL number is in stability of explicit solution schemes where CFL=1 is the limiting time step size but implicit solution schemes have no CFL limit for stability. For implicit solvers (like CFX) time step size is limited by other factors such as numerical accuracy considerations.

CFX has no segregated solver option. For segregated solvers you must be comparing to the old CFX4, Fluent or another code. CFX does have a segregated option for homogeneous multiphase flow, but that only couples the volume fraction equation to the momentum & mass equations. The Momentum and mass equations are always coupled in CFX.

Glenn Horrocks

 noureddine April 25, 2007 05:46

Re: why CFL is only in coupled solver?

Hi,

CFL is used in time-marching (or density-based) method, when the governing equations are solved in conservative coupled approach more suitable for compressible flows.

when you choose an uncoupled solver (it means that you have choose a pressure-based method) which is solved implicitly (no CFL is required) and it is more suitable for incompressible flows.

Djeghri Noureddine

 HekLeR May 8, 2007 16:26

Re: why CFL is only in coupled solver?

Sorry, but this post is hogwash. CFL has nothing to do with solution variables or solve strategy.

A density based solver does not have to be coupled. I wrote an explicit uncoupled versions in grad school (FCT and ENO).

Perhaps what you mean to say is a density based solver which computes the surface fluxes using a reimann solver is 'coupled'.

Glenn posted the only correct answer.

 Djeghri Noureddine May 10, 2007 11:27

Re: why CFL is only in coupled solver?

Yes, you are right HekLeR, Since I did not express well my idea.

When I said "CFL is used in time marching method" I meant that explicit scheme are used specially in the time marching and not in the pressure-correction methods.

When I said that in explicit scheme the governing equations are solved in coupled manner, I meant that these equations are written in conservative form and you can solve all these equations in one boucle iteration for the vector of the conservative variables, but you can solve them also in different boucle iterations for each conservative variable as you have mentioned.

 All times are GMT -4. The time now is 02:14.