please consider the following model.
The structure is under a thermal load. The temperatures are calculated by a prior sol 153 run. The following constraints should be applied:
Symmetric constraints on the lateral surfaces, no translation in z-direction on the front and the rear surface, no constraints on the top and the bottom surface.
I modelled this using Symmetric Constraints on the lateral surfaces and User Defined Constraints on the front on rear surface. Of course I received overlapping constraints which I tried to resolve via "Keep overlapping". This appeared to be logical to me because both constraints (symmetric and user defined) should be valid on the common edges.
The result was not correct in my opinion. Obviously there is a z-translation on the common edges inspite the User Defined Constraint which should suppress any z-translation. This takes place especially on the mid nodes of the elements.
I can't understand this behaviour. What's going wrong here?
I am grateful for any hint.
Thank you in advance
Solved! Go to Solution.
Using the "simplified" constraints (symmetry, slider, etc.) is very dangerous if what is going on behind the scenes is not understood. I avoid those constrain types at all costs
When you create a "simplified" constraint, the program is:
So, the program is honoring your instructions to enforce both constraints (symmetry and Z) and is writing out a SPC card with DOF 1 and 3 constrained. The problem is that the Z axis of the coordinate system created to enforce the Symmetry constraint is not parallel to the global Z axis, so an unintended direction is being constrained.
This "behind the scenes" creation of coordinate systems and the modification of the nodal displacement coordinate system is what makes these constraint types dangerous - the behavior of existing constraints can change when the displacement CSYS definition changes.
You need to carefully examine the nodal displacement csys definition and re-apply the "axial" constraint in terms of that coordinate system rather than global Z.
Thank you Jim,
may I add that the dangerosity is enhanced by the fact that I think the graphic representation of such "resolved' "simplified" constraints is wrong: what user sees on screen cannot be believed.
Which is quite not normal for such a graphic tool.