CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   CFX (http://www.cfd-online.com/Forums/cfx/)
-   -   Different itterations i CFX 5.5.1 (http://www.cfd-online.com/Forums/cfx/19426-different-itterations-i-cfx-5-5-1-a.html)

Michael Rasmussen January 15, 2003 13:24

Different itterations i CFX 5.5.1
 
Hi

I have a question regarding the number of itteration pr. equation. In CFX 4.4 I remember there is an option where it can be defined that the hydrodynamical equation should be itterated f.ex. 5 times and the scalar should be itterated f.ex. 3 times.

This was very helpfull when solving transient transport of a scalar in stationary flow. First the stationary solution was found and the model restarted with the solution as initial guess. Then the hydrodynamical equations was not solved (no. itt =0), but the transport equations were (no itt = 3).

-It increased the overall computational speed.

Is it possible to do something similar in 5.5.1 ?

Regards Michael

Astrid January 15, 2003 13:42

Re: Different itterations i CFX 5.5.1
 
You have to set

Solve fluids = false

in the Expert Parameter Section. This will switch off u,v,w, and p. Similarily you can make adjustments for energy, turbulence equations and so on. It is not possible nor is it necessary to adjust the number of iterations. Except for the situations where Mach > 2.4, where you are advised to increase the number of continuity loops.

Astrid

Robin January 15, 2003 16:07

Re: Different itterations i CFX 5.5.1
 
Hi Michael,

For a transient simulation, one can specify how many inner iterations are performed per outer timestep and this may be done per equation class.

In general, what you are suggesting is necessary in CFX-4 as it uses a segregated solver. Since CFX-5 uses a coupled-multigrid approach, there is no need to specify additional inner loops since the solver is already doing the extra work.

If you are solving a scalar equation which does not effect the flow, then doing as Astrid suggests is OK. I would recommend converging the flow first, then solving the scalar equation with fluids off.

In the end, solving a scalar equation is relatively inexpensive, so I would just leave it on throughout and save myself the extra work. The progress of the scalar equation solution can be sped up by specifying a larger timestep for the additional transport equation.

Regards, Robin


Swapnil January 15, 2003 20:45

Re: Different itterations i CFX 5.5.1
 
Hi Robin,

Will you please explain the concept behind switching off particular equtions? What's the benefit in doing this (Will you please site some good examples)

Regards,

Swapnil

Robin January 16, 2003 21:16

Re: Different itterations i CFX 5.5.1
 
If you are tracking the path of an additional scalar and already have a solution of the hydrodynamics equation, you can speed things up significantly if you are not trying to further converge the hydrodynamic equations. Thus, in such a situation you may choose to turn these off and only solve for the additional scalar.

Robin


All times are GMT -4. The time now is 08:44.