Learning Forum

[jinglun]

Hello,
I am trying to simulate the compression of a hemispherical shell, using Ansys Academic 2020 R1. The shell is a composite layup defined in ACP. The compression setup is as shown below:

[BenjaminStarling]

Split your model in half and use a Symmetry Boundary Condition. Otherwise restrain one (ore a few) nodes on the crosshead in the X direction.n

[jinglun]

thanks for your suggestion! Sorry but I forgot to mention that the hemisphere is not symmetrical. So symmetry would not be applicable in my case.Other than that, did you mean restrain the nodes on the hemisphere, instead of the crosshead?n

[BenjaminStarling]

nHow is a Hemisphere not symmetrical?nBut if that is the case, then yes, apply the x constraint to a node on the hemisphere. Just one node is probably preferred if you are restraining the hemisphere.nI originally looked at your image and interpreted the hemisphere and crosshead as sharing mesh, but paying closer attention it appears you have a contact between the two? nAnother option is to use frictional contact. Even a frictional contact between the hemisphere and crosshead may permit applying the x constraint to the crosshead. n

[jinglun]

ArrayMy bad, should not have used hemisphere here. The curvatures at the 2 ends on the long axis are different, so the shell is not symmetrical. nYes, you are right about the contact between the shell and crosshead. But the contact area is quite small as the shell has a high curvature. I specified a small vertical displacement to the crosshead in the first load step, hoping to get more elements into conact before ramping up to the full displacement. But the pivot warnings appeared even before that was done.nThanks for suggesting frictional contacts, will give it a shot!nBtw, also curious if it would make a difference if I specify contacts in the following 2 ways (all frictionless / frictional):nshell curved surface + rim to 3 wall faces , as 1 contactnShell curved surface to bottom + back wall as 1 contact, shell rim to front wall as another contactn

[BenjaminStarling]

nThe definition of the contacts is not important. Defining more contacts (splitting up larger definitions) just allows for more granularity and control i.e. each contact definition can have different settings. Particularly for postprocessing where it allows you to view results on smaller areas at a time.nAlso just checking, now that I know the crosshead is separated, is that set to fixed displacement in X and Z? It should be. You can also use the adjust to touch option under the contact definition for the cross head to hemisphere contact. However small initial displacement increments is still good practice.n