Learning Forum

[Sebastien ARRIVET]

Hello,

I'm having  issues with convergence for a large model, and while trying to debug it I made a small pressure vessel using SHELL181 with KEYOPT(1) = 1 (Ux, Uy and Uz dof only)

The vessel is loaded with pressure using SFE, and constrained with what I beleive makes the model isostatic:

4 co planar nodes constrained in Z (the plane is normal to Z axis)

2 co linear nodes constrained in X (the line is perpendiculaire to X axis in the Z plane)

1 node constrained in Y

Yet I get the following error:

I dont understand how it can be unconstrained given the boundary limitation i've set.

You can find the model here: https://sendanywhe.re/TW7Q74O3

*use,main.inp will build the model and start the analysis. If you have any questions about the model I'll be happy to answer them.

Thanks in advance for your help

 

[Dave Looman]

A mebrane-only shell like you defined with only ux/uy/uz dof is unstable unless it is stiffened by prestress.  It's tricky even then because the elements  have to be supported laterally while you are applying the prestressing load.  It's almost never a good idea to just use ux/uy/uz dof in a shell element.

[Sebastien ARRIVET]

Okay, so I tried to go with KEYOPT(1) = 0, which allow for rotx, roty and rotz and indeed it worked. 

Thanks for your help.