Learning Forum

[WolfgangNeil]

I'm trying to replicate a quarter symmetric model of a thin tube structure. The symmetry conditions have been assigned across symmetric surfaces using the ANSYS inbuilt "Symmetry" options using proper surfaces and symmetry normals. However, the results differ from the full-scale model. Any suggestions to resolve the mismatch in the results?

[peteroznewman]

Please post images of the quarter symmetric model. Show the loads and boundary conditions and the results. Also post the full model details and results so we can see the mismatch you are describing.

[WolfgangNeil]

I'm trying to replicate a similar model as described in the picture

[WolfgangNeil]

@peteroznewman
I've shared a picture.
Hope it helps to understand the problem

[peteroznewman]

Thanks for the picture.
The Bottom is showing a Fixed Boundary Condition.
The boundary condition: Ux = Uz = 0 and URx = URy = 0 is not a Symmetry Boundary Condition.
A Symmetry Boundary Condition has Uz = 0 and URx = URy = 0 for a face whose normal is parallel to the z axis.
If you create a Symmetry Region in Workbench, you have to be careful because it defaults to the normal parallel to the x axis, regardless of the face that you picked.

[WolfgangNeil]

@peteroznewman
Actually, Ux = Uz = 0 and URx = URy = 0 are applicable to both long edges on the top of the body. Please tell me know if I'm missing something.
Also, it would be very helpful if you can please share some relevant material to understand more about symmetry boundary conditions.
I've prepared a 1/4 model of the body and (referring to the coordinate system of the image) assigned end faces with z-axis as area normal of the model with Symmetry and normal as z-axis. Similarly, I've assigned end faces with x-axis as area normal of the model with Symmetry and normal as x-axis.
I'm using ANSYS 18.1 and the software provides an option to change the symmetry normal as per the geometry selection.

[peteroznewman]

You say Ux = Uz = 0 and URx = URy = 0 are applicable to both long edges on the top of the body, but that is not symmetry. That is symmetry on the plane normal to z and an additional Ux = 0. Where does that come from?

[WolfgangNeil]

@peteroznewman
Sorry, I didn't get your question.
Can I conclude that I will have to analyze this geometry as a full model?

[peteroznewman]

I'm saying that you can either use two Symmetry BCs on two orthogonal planes or you can apply the constraints shown in the image, but if you do that, don't call it symmetry.
Symmetry means the normal displacement to the plane and two rotations in the plane are set to 0.
The constraints shown in the image have symmetry and an extra displacement set to 0 which prevents nodes from moving in the plane of symmetry in one direction only! How is that achieved in reality? This constraint pattern is no longer what anyone would call a symmetry BC.

[WolfgangNeil]

@peteroznewman
Now I can understand the point you are trying to make.
Thank you for your kind cooperation in understanding the matter.