Stress change with mesh density.

[FC2008]

Hi, I'm trying to become a bit familiar with ansys workbench before my employer sends me off to a course. (a slow process)

I have a straight pipe section that I simply want to pressurize from the inside.

I use a cylindrical support on the outer surface with Radial set too free and Axial and Tangential set to Fixed.

Algorithm is patch conforming, method: tetrahedrons.

When I set the mesh from Course to Fine, the Equivalent Stress readings go up pretty drastically. I think this is telling me that there is something wrong..

Are there some simple things I need to do so this doesn't happen?

Thank you for your patience on this boring question

 

[FC2008]

- I also add fixed supports to where the flanges would have been..

 

[Wecos]

Are the large stress and drastic increase located at your fixed support? If so you could be getting a singularity effect from the constraints and no matter how much you refine the mesh the stress will incress infinitly.

The mesh density does effect the final results. In general, a finer mesh results in more accurate stresses. But its gets to the point that refining the mesh further would significantly increase the analysis solve time and not produce a meaningful difference in the final result.

 

[Transient1]

I would add to Wecos's post, that refining the mesh to the point where the stress does not change is referred to as a Mesh Convergence Analysis.

 

[cbrn]

... and to add to Transient1's post, if you refine the mesh in the zone where you do have the singularity, then the results will diverge, never converge... This is one case of what is generally called "mesh paradox".