I’m working through an online course, using PrePoMax.
One example is a plate in tension, modelled with shells. The course provider has modelled the plate within the software (FEMAP), and meshed with S4 elements. I’ve modelled the surface with Onshape, imported it to PPM, then meshed with quad dominated 1. order elements. The results are similar, and good.
The example shows a few scenarios, iterating through different levels of accuracy with different ways of constraining. One of the scenarios wants to allow for transverse shrinking at the boundary condition (initially fixed, i.e. not allowing for it), and then goes on to show what happens if the transverse constraint is removed from the entire edge, and only applied to the middle node. This is where my problems occur.
I can’t pick a single node in the BC menu, so i tried to create a node set with just the 1, although i can’t seem to do this either. I can only pick the edge, and get a node set with all 11 nodes. I’ve then tried creating a reference point from the middle node and constraining it, but the ref point is ignored and rigid body movements occur. I then figured perhaps adding a rigid body constraint between the ref point and the middle node would work, but then i’m back at the beginning - i can’t pick the node!
I wonder, does this have to do with the way Calculix handles the expansion of shell elements compared to true shell elements?
I’m expecting a stress distribution similar to this:
I added a node set with the middle node, constrained it in transverse (Z), but my stress distribution is still strange, and not what i expected, although i realise that the stress is quite uniform.
What you get is correct for the applied BCs - basically uniform stress distribution with just some small variation after the comma due to numerical reasons. In Abaqus you would get results like this:
I finally worked out the issue. Apparently FEMAP overwrites a nodes DOF with the newest defined BC, rather that adding it. That’s why in the FEMAP model the middle node suddenly started translating in the force direction, causing the strange (well, not so strange now that i know why) distribution compared to mine. He mentions this briefly, but i didnt catch it until a few times around. I believe i was finally able to somewhat replicate it by creating a node set for the middle one, and one node set for the rest of the line.
Stresses and displacements should be perfectly symmetric in this problem. I recommend you increase the deformation scale enough to detect asymmetries just in case.
Ask or read about the 3-2-1 method to apply BC. It’s a simple and powerfull idea to start with.
You should be able to get exacly the same result as Abaqus with ccx in this case.
Try to figure it out by yourself first. Each failed option will be a discard that will make you more efficient. If you don’t get it in three days, I’ll send it to you but I’m pretty sure you can solve it.
I’ve studied the 3-2-1 method, and tried applying it, but it seems no matter what BC’s i set up i’m not able to get completely uniform stresses in my shell model. I still get stresses within 0.7 MPa like in my earlier comment.
If you don’t mind i would love to see how you solved it @ANYS
Sure.
Look at the reaction forces when applying boundary conditions.
It can help you to know if the selected points are the right one.
Ideally, the supporting points preventing from rigid body motion in directions without load component should only generate marginal reactions.
Thanks a lot @ANYS
Your model helped me realise a few things. I was trying to use the 3-2-1 method to constrain individual nodes in the model, since every article i could come across mentioned this approach. By restraining the entire edge in force-normal translation, i got the stable result.
The traditional 3-2-1 method involves constraining only 3 nodes but nothing prevents you from experimenting with other variations like combining this method with constraints applied to edges.
If you are interested, I suggest you try to apply it to the gas cylinder. It has real world application to any system under internal or external pressure like vessels , pipes ,….
NOTE: Don’t feel bad if you don’t get the uniform Stress in the sphere. It is challenging.
