problem with abaqus standard Convergence after normal loading of half sphere shell (buckling)

[brainsama]

Hello Ings I'm getting a problem reaching convergence with a half spherical empty solid ball when applying a normal load on its external surface leading to divergence when the shape deforms to reach a certain point of displacement ; this drives me crazy , I've read that this could be buckling problem but I couldnt manage to make a linearization of the model during post buckling .
Any help will be appreciated

 

[Erik Panos Kostson]

If it is a buckling instability the solver needs to sub-step (e.g., using arc length, load control cannot capture this), in order to get through this point. You could first run a linear buckling analysis to see at what load it buckles, that will give an indication whether and at what load the nonlinear solver will experience some difficulties.

If that is the case that it buckles, you could use the Riks nonlinear method to get you through say a snap through or buckling instability where the stiffness of the structure changes a lot.

Also feel free to send a model or CAD, with some load restraint and material information and we can have a look.

-------------------------------
Student see here: Link & Link

 

[brainsama]

thank you I 've tried the buckli,g step technique but when the simulation completes it gives me a non sens buckling results it gives it on the opposite direction. I'll attach the CAE model with the buckle step and everything . I tried to overpass the buckling by adding an other material in the inside to add some resistance.
salutation

 

[Erik Panos Kostson]

That is OK, I will have a look for you.

Talk to you later

 

[Erik Panos Kostson]

I have not looked at it, but if it is similar to snap through buckling of a spherical shell (see how that looks: Link)

then the linear buckling solver would not be able to capture that. We possibly need to run it using arc length sub steps, in order to capture that phenomena.

 

[brainsama]

it's like this, exactly like this .
how to do so and what is the arc size I donc know about these procedures

 

[Mustaine3]

Use displacement control or the Riks step (arc length) or a dynamic procedure (quasi-static in A/Std or Xpl).

 

[Erik Panos Kostson]

Ok I understand.

Then we need to use as mentioned a sub stepping technique and in this case arc length (called Riks method in Abaqus/ choose Static Riks when you define the step).

Instead of applying a force, we can use a displacement at the top, forcing the shell down. Then the arc-length method will go in and help when snap through buckling occurs.

Below is a small example of a spherical shell where all this has been used (enforced displacements at the top, and arc-length). The force reaction versus displacement curve is seen at the top (snap through occurs at about 250 N), while different stages of the snap through are shown in the Strand7 GUI, below.

 

[brainsama]

I need to apply a pressure nor a displacement , my friend my goad is to find the displacement

 

[Erik Panos Kostson]

That is OK, you can apply a force there. That is fine.

 

[brainsama]

I attached the model just run it on bucle step to find the first eigne value and you will see what I am talkin about

it's like you apply the load on one direction and you get bucling on the opposite direction

 

[Erik Panos Kostson]

Unfortunately I can not open it up. No worries though because linear buckling should not be able to predict this nonlinear (geometric NL) behaviour. As we said use Static Riks (+geometric nonlinearities) in Abaqus, and apply a high load that will be enough to cause it to snap through.

This link shows how to set the Static Riks method:Link

 

[IceBreakerSours]

Can you explain the application a little? If you have access to experimental data, then why can you not apply a displacement (iteratively) and find out the appropriate force?

In any case, please go through the documentation and run a few example problems that use arc length methods. It will be slow going for a little while but the hard work will be well worth the investment.

*********************************************************
Are you new to this forum? If so, please read these FAQs:

http://www.eng-tips.com/faqs.cfm?fid=376

http://www.eng-tips.com/faqs.cfm?fid=1083