CFD Online Discussion Forums

CFD Online Discussion Forums (
-   CFX (
-   -   Mesh Stiffness Option for Mesh Deformation (

Ste_Lakey January 9, 2006 09:00

Mesh Stiffness Option for Mesh Deformation
I am returning to a problem I encountered previously regarding the modelling of an oscillating aerofoil, and the mesh stiffness when using a prescribed sinusoidally varying oscillation.

In the manual it suggests the following for mesh stiffness:

1 [m^3 s^-1]/Wall Distance

Could anyone expand further as to what this term will do, particularly in relation to the term Wall Distance (does this require substituting with a value based on my mesh geometry?).

Thanks in advance

Mike January 10, 2006 10:41

Re: Mesh Stiffness Option for Mesh Deformation
The mesh stiffness controls which parts of the mesh will deform. The absolute value is not important - so a constant mesh stiffness of 1 or 1 million will give the same result. What is important is how the mesh stiffness varies throughout the domain. Areas with a lower stiffness will deform more that areas with a high stiffness. Generally speaking, you want the mesh near the walls to remain stiff so that your boundary layer mesh remains "nice"; therefore the volume mesh away from the walls will deform to any motion of the boundaries. Wall Distance is a variable equal to the distance from the nearest wall boundary. So a mesh stiffness = 1/Wall Distance will keep the boundary layer mesh stiff and cause deformations to be absorbed in the volume mesh. Another useful formula for mesh stiffness is: 1/Volume of Finite Volumes This will make small elements stiff - the boundary layer usually has small elements. Also, elements that have already deformed a lot and are close to folding will become stiff. M

Ste_Lakey January 10, 2006 11:19

Re: Mesh Stiffness Option for Mesh Deformation
Thanks Mike, will give these a go

Ersin January 19, 2006 17:33

Re: Mesh Stiffness Option for Mesh Deformation

depending on the geometry you are deforming it could be that there no difference between the formulations

<1/wall distance> or <1/volume of finite volumes>!

Perhaps you can also try the formulation

<1/wall distance^n> with n > 2

All times are GMT -4. The time now is 23:56.