Finite sliding Contact between poroelastic bodies

[krishnagoud]

Dear all,

First of all, I would like to appreciate the kinda response the people are giving in this forum. I learned a lot from this.

Description of my question:

I have modeled 3D complicated knee geometry where the poroelastic soft tissues are in contact with each other and also with its metal implant. In order to post my problem I have made an equivalent rough axisymmetric example which consists of
1.Implant----> Metal-isotropic elastic
2.CartF---> Poroelastic soft tissue attached to bone
3.CartT---> Poroelastic soft tissue attached to bone

Here is my problem.

In the Solis consolidation step, the model couldn't converge and it gives only warning NEGATIVE Eigenvalues.

Model is perfectly constrained and there exist no rigidbody motion.Then I understand that the negative eigenvalues are coming only from Interactions. But I could't figure it out what went wrong.The same thing happens in the complicated 3D Model also.....

I played with all contact controls option
(Automatic tolerances, Unsymmetric solver, stabilize.....)

Still I couldn't solve my problem............. I completely stuck here.

I deeply requesting you please look at my problem attached (.cae and .inp)

http://files.engineering.com/getfil...9d8a-4891-a331-369beb0891d4&file=x_finite.cae

http://files.engineering.com/getfil...7f74-499f-b059-bd298a55886c&file=x_finite.inp

I really appreciate your reply !

Thank you very much in advance !

/Krishna

 

[krishnagoud]

Hi,

I really stuck with this problem....
Please someone of you experts in Abaqus look at my problem.

Awaiting for your kind reply.

/Krishna

 

[rstupplebeen]

Krishana,

Attached is a picture outlining some of the issues.

You have an interference fit in your model and need to remove that some how. Depending on if you want prestresses in the model or not will determine if you wan to "Adjust only to remove overclosure" or "Interference Fit.

A second error in the contact definitions I believe is that the wrong face is in the CartF.bottom_side set.

I hope this helps.

Rob Stupplebeen

 

[krishnagoud]

Hi Rob,

Thank you very much for your kind reply !

In the model, I want to know the stresses/strains due to that interference and hence I used Interference fit option but not 'Adjust only to remove overclosure'.

Inorder to avoid 'falling of' the corner node, I used the face CartF.bottom_side and Also I expect surfaces CartT.top and CartF.side will come in contact during the course of the analysis.

Eventhough I use only CartF.bottom (instead of CartF.bottom_side) as a master surf in 'CartT-CartF' intercation, model didn't converge.

The same warnings as Negative Eigenvalues.

I don't know how to deal with this.

Thanking very much for considering my problem

Awaiting for your kind reply

Thanks !

/Krishna

 

[krishnagoud]

Any one???????
Someone could help me out please !!!!

/Krishna

 

[rstupplebeen]

How is this actually put together. If I have an interference fit that I really want to understand I will model the assembly. If that is not the real desire of the analysis then the Interference Fit option seems appropriate.

Alternatively you can shrink the part with thermal cooling then turn on contact and heat back up. This method does introduce an additional degree of freedom for each node though.

I hope this helps.


Rob Stupplebeen

 

[krishnagoud]

Hi,

Thanks !

The interference fit seems to be not causing any convergence difficulties.But I guess, the problem couldn't converge because of other contact in the model.

Eventhough i use 'Adjust only to remove overclosure' instead of interference fit, model didn't converge...

The same warnings appear again !

I would appreciate any other suggestion
Awaiting for ur reply.

/Krishna

 

[rstupplebeen]

It might be collapsing the element because your mesh size is small compared to the interference. For testing try a larger mesh of a different geometry. I would just assemble it if it were me though.

Rob Stupplebeen

 

[krishnagoud]

Thanks for your reply !

I would check with larger mesh size.
I am wondering what you mean by 'assembly' for interferencce fit.
I just modeled geometry of 2 parts sothat their instance will have some overlap as soon as they put them in assembly.

What you mean is different than this????

The interference is due to the wedge shape of the implant (this is one of the design variables in my parametric study).....

Thanks for your help again !

/krishna

 

[rstupplebeen]

What I meant is to start with the parts separated and lower on onto the other allowing the wedge of extra material to deform out of the way during the sliding contact analysis.

Rob Stupplebeen

 

[krishnagoud]

Thanks for your reply !

That was the first try of my simulations. But, since it is poroelastic there exist pore pressure continuity problems with Finite sliding application.

The actual model didn't converge due to complicated contact conditions.If I make the contact small sliding, then problem converges.

Inorder to make the contact sliding small, I used interference fit option instead of sliding the surfaces (Inserting the wedge shaped implant into CartF ).

If I use all small sliding contacts in the model, then it works absolutely fine but there exist some falling off nodes (slave penetrates into master) at the end of analysis.

Since these are small sliding contacts, abaqus detects intersection of the nodes at the initial step. But as the analysis progresses, there will be some sliding which results in penetration of slave into master(it behaves as if there is no contact master surface)

Finally, I decided to use finite sliding contact and make it converge. Thats the reason why i made an simple axisymmetric model and trying to ask the experts here.

I hope you understand what i mean to say !

Thanks again for your kind reply.

I don't know how this problem could be solved.