You're using solid elements, why aren't you using beam elements ?
Meshing weldments with solid mesh is harder than with beams or shells.
If you want to have more success with solid mesh you should use curvature based mesher.
Results are not symmetrical because mesh is not.
Try using very fine uniform mesh, and you will see that results will be more symmetrical.
I personally wouldn't do that, I would rather:
- study half model
- put symmetry restraints
- use the option "display symmetric results" when post processing.
As Alessandro Tornincasa said, the results are asymmetric because the mesh is asymmetric. If you are going to use solid elements to study your model then you may need to have a smaller element size than this. There should be a minimum of 2 high quality elements across all your thicknesses in the model. You would need even more elements than that for accurate and symmetric results. The number of elements required can quickly become large and unmanageable. If the model and loading are truly symmetric then you would want to use a half model with symmetry fixtures and dial the element size down so you can get under the 2 element across thickness minimum.
Many thanks for your responses.
Ryan, what you said is what I usually do but I was just trying to save some time!!!
I normally split the weldments with a mid surface then remove them after splitting. The two (inner/outer) tubes then get bonded and this gives the two elements on the ends of the tubes.
Guess I just have to do it the long way to get the results I need!!!