CFD Online Discussion Forums

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   Main CFD Forum (https://www.cfd-online.com/Forums/main/)
-   -   moving grid with domain size change (https://www.cfd-online.com/Forums/main/7535-moving-grid-domain-size-change.html)

Hu May 12, 2004 05:24

moving grid with domain size change
 
Hi, Everyone, I want to use CFX 4 to solve the moving problem. The moving grid techniques are used in the case but the problem is as the grids move, the domain volume will increase or decrease. Should I add the mass source or sink to account for the volume increase or decrease in order to meet the continuity equation? Many thanks

Praveen May 12, 2004 05:57

Re: moving grid with domain size change
 
Yes, something like that is required. Search for geometric conservation law.

John Luo May 12, 2004 06:54

Re: moving grid with domain size change
 
No, you should not do anything if the moving mesh effects are already taken into account in the Navier-Stokes equations. Say you have engine cylinder with no inlets and outlets, then when piston moves up and down, the mass will stay the same, only pressure and density will change to satify mass conservation.

Hu May 12, 2004 11:06

Re: moving grid with domain size change
 
John, Thanks for your reply. I agree with you about the piston case. But for my case, the mass did increase or decrease. It does not stay the same. How to keep mass conservation without adding mass ? Hu

John Luo May 12, 2004 15:04

Re: moving grid with domain size change
 
In this case, either CFX does not do a good job in moving mesh implementation (I heard someone mentioned this before), or you didn't set your case up properly. You could try to describe your case in more detail. In any case, the mass has to conserve.

Chris Sideroff May 13, 2004 21:14

Re: moving grid with domain size change
 
I agree with John Luo lasts reply - CFX is not implementing the moving grid properly. If the numerical discretization if conversative in time as well as in space then continuity will be satisfied.

A previous colleague of mine wrote a CFD code with a moving grid and by constructioning his temporal discritization conservatively autmatically satisfied continuity.

The geometric convservation law (or GCL) is a product of improperly constructing the time discretization and used to be an ad hoc way of satisfying continuity. It is essentially a tranport equation for the transformation of the Jacobian matrix.


All times are GMT -4. The time now is 10:36.