Stuck? 
support@mecway.com
support@mecway.com
Mecway
FEA
Node rotation BC as a function of position
Hi Victor
I’m working on that old semi-sphere problem which I resume in a picture.
When solving the whole hemisphere with shells, there is no issue and results agree well with analytical solution.
I’m trying next to solve it using symmetry BC which requires apart from imposing some constrains in the displacements to constrain some rotations on the Symmetry faces.
Mecway throws that known warning about rotations not aligned with the global axis.
To sort it I was thinking about defining a local coordinate spherical system for each node by means of multiple Transform cards. Mecway can do that when constraining displacements.
If I’m not wrong that would mean to write a node set for each node, a transform for each node set and a boundary constraining the rotations around two of the new axes on each symmetry plane.
My question is , does that have any sense?.
I’m pretty sure you have been thinking about this before. Is there a reason why the “node rotation” window doesn’t allow to input the direction as a function of position in the same way “displacement” window does?
Regards
I’m working on that old semi-sphere problem which I resume in a picture.
When solving the whole hemisphere with shells, there is no issue and results agree well with analytical solution.
I’m trying next to solve it using symmetry BC which requires apart from imposing some constrains in the displacements to constrain some rotations on the Symmetry faces.
Mecway throws that known warning about rotations not aligned with the global axis.
To sort it I was thinking about defining a local coordinate spherical system for each node by means of multiple Transform cards. Mecway can do that when constraining displacements.
If I’m not wrong that would mean to write a node set for each node, a transform for each node set and a boundary constraining the rotations around two of the new axes on each symmetry plane.
My question is , does that have any sense?.
I’m pretty sure you have been thinking about this before. Is there a reason why the “node rotation” window doesn’t allow to input the direction as a function of position in the same way “displacement” window does?
Regards
Howdy, Stranger!
It looks like you're new here. If you want to get involved, click one of these buttons!
Comments
*EDIT: *TRANSFORM allows a cylindrical coordinate system which looks like it might work for this.
I don't think there's any particular reason for not having position dependent rotation constraints. It might have just not been a useful as displacement so I didn't spend the time on it.
I completely miss that. Now I see you have managed to build a local coordinate system aligned along the symmetry plane and following the shell face. Impressive.
I finally managed to solve 1/4 hemisphere problem. This problem has shown ridiculously complex to me for a long time, and it still is.
Things found in the process:
-I had to generate knots along the shell BC face. +0.0001mm makes the difference.
-Transform also impose some limitations on the load’s positions. Directly on top of the Symm BC face throws:
"Error generating inp file: Node 2 has a coordinate transform (*TRANSFORM) together with more than one *AMPLITUDE block. This is not allowed by CCX."
-I have need one additional refinement to get good results. That should be considered before using Symmetry BC.
-Keep the symmetry planes aligned with the main axis.
Expected:
Result:
This new tool makes a big difference. Really nice Victor