Restraints for static stress of a cylinder

[kyong]

If I want make both end faces of a cylinder keep perpendicular to the axis while allowing free movement of one end face in axial direction, what would be the best way? I tried a few settings, getting not very satisfactory result. One of the settings I used is : 1) make a reference axis. 2) set axial movement of one edge at an end face to 0. 3) set circumferential movement of the face at the other end to 0.

 

[cbrn]

Hi,
unfortunately from your post it is not clear exactly what you want to do.
1- If you mean that sections perpendicular to the axis must remain perpendicular even when the axis bends, well, that depends on the overall model behaviour (loads position, restraints positions,...). Remember you don't have a Beam, where the "classical" formulation ensures this "a priori".
2- If you mean the faces will have to be restrained, please explain what is the physical behaviour you have to simulate. There are several ways you can play with "on planar face" restraints, associated or not to auxiliary coordinate systems or reference axes, and also with combinations of point / edge / surface restraints. They all have a different meaning and thus a different effect, though...

Regards

 

[kyong]

cbrn,
I think I missed one thing in my discription of the problem. The cylinder is hollow cylinder. The cylinder is receiving internal pressure/thermal gradient and so on.
The two end faces of the cylinder are intended to remain perpendicular to axis which is straight. That is my desired restraint. To model that restraint I have improvised by giving 0 axial movement relative to the axis on one end and calculated axial growth of the other end. Then in a second run, I gave the calculated value to axial movement of the other face. The result was OK. But it required two steps and bad thing is that locations of the two faces are fixed, which prevent any movement even though an additional affecting load is applied.

 

[cbrn]

Kyong,
if I understand well, your axis should be free to bend, but the end sections have to keep perpendicular to it though their centers follow the axis. If this is the case, unfortunately I don't see very well how to model it with CosmosWorks since it doesn't have "constraint equations" like in Ansys for example. But I think this wouldn't be a problem because the normal behaviour of the cylinder will achieve approximately this, unless external loads force the end sections to behave differently.
On the other side, if the axis will remain straight even with the loads applied (i.e. the loads have cylindrical symmetry), then "restraint -> on planar face" will give you good solutions, as you can refer also to a CSYS, which will allow you to control things like "no tilt / no twist",... and also with the reference axis the options seem to me satisfactory: If you want to investigate the stress rise due to thermal expansion, you have to constrain the axial displacement of both sections, so you wouldn't have to run twice the solution (first to determine free elongation, second to determine corresponding stress).

Regards