stent crimping of nitinol quarter model abaqus

[Josh1313998]

Hi i'm modelling possible frames for anchorage in valve prostheses ( larger stents) made out of nitinol and im exploiting symmetry by modelling only a quarter of both the cylindrical crimptool and the stent. Howwever, when i do so by imposing radial BC's on a reference point mapped to the crimptool nodes via an equation constraint the ends of the stent fold in and im pretty sure ive got suitable BC's (also the Reaction force output is high and noisy)
can anyone help me out?

 

[FEA way]

Are equations constraints necessary in your case ? Can't you use symmetry boundary conditions or cyclic symmetry ? It would be much easier and less prone to errors if applicable to your model.

 

[Josh1313998]

Hi, im using cyclic BCs to mimic symmetry and the equation constraint is only to map all nodes of the crimptool to one point for easy RF extraction. does that make sense?

 

[FEA way]

I would try without that equation to avoid overconstraint - using just the cyclic symmetry interaction.

 

[Josh1313998]

Thanks so much for your time and help. unfortunately this doesnt seem to solve my issue.
could you think of any other reasons? would sharing an inp help?

 

[FEA way]

Yes, it will be easier to help if you share the .cae or .inp file.

 

[Josh1313998]

[link file:///C:/Josh_Abaqus/RF-crimp-models/12-copy-shell/continuos-strut-design/NoEQNconstraintandbcsaresymmbcs.inp]Link[/url]
[link file:///C:/Josh_Abaqus/RF-crimp-models/12-copy-shell/continuos-strut-design/thickness030newmesh.inp]Link[/url]

The first link shows the newer design inp which i created to reduce regional stress concentrations but doesnt seem to be doing so and the second one is an older design where it ran pretty ok. the problem is it is quite noisy, has high stress concentrations and most importantly the RF values im getting are too high especially for the new design.

 

[Josh1313998]

"C:\Josh_Abaqus\RF-crimp-models\12-copy-shell\continuos-strut-design\NoEQNconstraintandbcsaresymmbcs.inp"
"C:\Josh_Abaqus\RF-crimp-models\12-copy-shell\continuos-strut-design\thickness030newmesh.inp"
I cant seem to attach the links so that you can open them . is there a way to do so on this?

 

[FEA way]

There is an Attachment bar at the bottom of the reply window. There use the "Click Here to upload your file to ENGINEERING.com" option and upload the file from your computer.

 

[Josh1313998]

here is the 1st one where the only probelmes are folding in of parts of the stent, high regional von mises and noisy/high reaction forces

 

[Josh1313998]

Here is the newer design using fillets that i am trying to use

 

[Josh1313998]

this picture shows the folding in of the ends of the struts i have referenced as well as the high von mises stress values.

 

[FEA way]

Here are a few remarks after looking at your model:
- you are using fixed incrementation within a dynamic explicit step - is there a reason for it ? The use of automatic incrementation is advised in the majority of cases.
- your general contact domain is limited to one selected surface pair - it's not wrong but make sure that no other contact (like self contact) can occur in this case
- did you check the global energies in the model ? Do they all look correct ? Also, I've noticed that you limited the energy output to just two energies. It would be good to include other types as well - for verification purposes. In Abaqus/Explicit analyses it's really important.
- if you encounter some noise in results, consider increasing output frequency or using built-in filters
- prescribed velocity (with smooth step amplitude) might be better than prescribed displacement in the case of Abaqus/Explicit analysis

 

[Josh1313998]

Thanks a million! i have taken these into account but still no joy unfortunately. The only problem im really having now is the high reaction forces.is there anything else that could induce this?

 

[FEA way]

What are the expected reaction forces and how did you estimate them ? Also, what's the quantitative difference between them and the values from the analysis ?

 

[Josh1313998]

so from literature similar crimps would generate 30-80 N of radial reaction force which im getting from my one quarter model so i must multiply the 40 n im getting by four to get the real values of the radial reaction force of the whole stent.. im extracting this by using an equation constraint to map a reference point to all nodes on the crimptool and plotting the reaction force of the reference point in the field output.