contact between two poroud mediums

[krishnagoud]

Dear friends,

Greetings !

I am Krishnagoud Manda, pursuing doctoral studies at KTH Mechanincs, Stockholm.I have been using Abaqus FE software for my research,especially in modeling of porous medium.

I would greatly appreciate your help if you consider my question.

I briefly explain about my problems which I keep encountering in ABAQUS for last 4 months.I have been working on 'Finite element modeling of human knee joint', which in turn useful for development of cartilage implant. Basically, human knee consists of two bones in contact through the thin materials called cartilages.

Cartilage is highly porous medium and it contains water inside, hence it can be modeled as biphasic(poro elastic) medium using Abaqus (pore pressure elements and soil analysis)

Unfortunately, i could not be succeed in the modeling contact of porous articular cartilages in ABAQUS. I have been keep encountering problems with this. I am completely lost at this point. Unless I get some help from you i can not move further.

In order to verify and to understand behaviour of porous medium, I modeled a SIMPLE axisymmetric geometry(two cylindrical parts are in contact). But the contact between porous mediums is getting diverges and solution was terminating after certain load. Here are details...

1.Axisymmetric two cylindrical parts in contact, meshed with CAX4P elements
2.Part.1 is bottom cylinder of 3.1mm thick and 16.5mm radius
3.Part.2 is top cylinder of same size
4.left side of both parts are XSYMM
5.Part.1 bottom nodes are fixed in all directions (Encastre)
6 Contact established by moving top nodes of top part(part.2) in step 1 (*Soils,consolidation)
7.Top nodes of part.2 are tied together to move only vertical diretion(*equation,2)
8 load is applied on top surf 2MPA equivalent to 3500N axial load (in step 2, *soils, consolidation), using amplitude (increasing from 0 to 2MPA n 2 sec and the constant for 98 sec)
9. permeability is strain dependent
10. initial void ratio and pore pressure are defined in input file

I am wondering how ABAQUS handles the fluid behaviuor during transient consolidation.
It has not been understood why the solution is getting diverged. I am wondering whether I am missing some tricky portion of the problem or contact between porous mediums in ABAQUS requires any special specific criteria ?

I thought there could be some problems with continuity of pore fluid,in order to avoid fluid pore pressure problems at contact surfaces, I used the membrane between them, but still problem persists. details are...
1. Membrane is tied to part.1 (bottom part)
2. Surface to surf contact between membrane and bottom surface of top part.

But still the solution was terminating ! I don't understand why this is happening ?

Would you please help me whether I am missing something or ABAQUS can not handle this type of problems ?

Problem has been solved(with or without membrane) for low loads, but for high loads solution is terminating.

Human knee cartilage generally subjected to morethan 4000N (2.4MPA) of load. My model should work for atleast 2MPA pressure load.

Here i am attaching two input files.

1. porous_cantact.inp (without membrane)

http://files.engineering.com/getfil...cd9-af69-3aed0c79187f&file=porous_contact.inp

2. porous_contact_memb.inp (with membrane)

http://files.engineering.com/getfil...bad-8c1138ab0daa&file=porous_contact_memb.inp

Would you please look at my problem.

Please let me know how ABAQUS handles these type of problems(contact between porous media) ?

Thanking you in advance !

Awaiting for your kind reply !

Regards,
Krishna

 

[geotechniqa]

-Go to message file and see details on the error which you are getting and report it to us.
I suggest that you first consider full bonding (cancel the interface initially) and see if the model converges.
Modelling a porous medium usually requires special attention.
Some general guidelines may include:
-Initial time step in consolidation analysis is important to be calculated properly otherwise the model will diverge at the beginning,
- Make sure that the pore pressure tolerance you have inputted is high enough to avoid divergence,
- Make sure that effective stress does not go to zero at many points. In other words, undrained case should be avoided. This case may be a result of one or combination of : improper drainage BCs , very low permeability-materials, Poisson ratio close to 0.5 should be avoided, fast load. Avoiding this problems may be done by using elements biquadratic in displacement but bilinear in pore pressure,

- If large deformations are expected, then put NLEG on,
- Try to introduce dissipative energy ratio to dissipate/suppress the instability/rigid motion that could be produced at some points deforming excessively due, for example, to high permeability, large force in comparison with the modulus of elasticity,


Unfortunately , I do not have access to ABAQUS any more, otherwise I could look at your problem which seems interesting and challenging.

Pls update me with what happenes

 

[krishnagoud]

Dear geotechniqa,

Thank you very much for your reply and your guidelines.

1. Initial time step should more than critical value and pore pressure tolerance(UTOL) should be large enough- abaqus documentation

2.Effective stress is not going to zero at any point. But I did not give any drainage BC s. As I understood, cartilage is saturated medium, there is no drainage, but one of the surfaces is given zero pore pressure.

would you please explain "undrained case should be avoided"?

3. I expected large deformations, I put NLGEOM is on

Permeability values are taken from literature (void ratio dependent)

What happens if poisson ratio close to 0.5 ? (any way in my model it was 0.018)

would you please elaborate "Try to introduce dissipative energy ratio to dissipate/suppress the instability/rigid motion"

is this damping factor? i think there is no rigid body motion exists.


As you said, I even used TIED contact, but still the problem persists !
here I am attaching my .msg and dat file.

http://files.engineering.com/getfil...2ad-bbb3-8c124182f454&file=porous_contact.msg

http://files.engineering.com/getfil...89a-a8f2-7c027a781f5b&file=porous_contact.dat

http://files.engineering.com/getfil...850-a25c-0e23ef43c70c&file=porous_contact.sta

In these files i found only two things...

1. Zero Vol flux

2.Negative Eigenvalues (excessive distortion) in step 2, increment 5, attempt 1. But if you see in viewer, there is no excessive distortion.i pictue right top corner 6 elements

http://files.engineering.com/getfil...-4100-b27b-81b8bd3813b8&file=Screenshot-1.png

3.Finally *Error:Too may attempts for this increment or Min time increment required is less than specified.


I am completely lost with this problem........


unless i dont understand what is happening inside, and why it is not taking load of 2MPA(4000N), I can not move further with my actual complicated knee model.


Do i need define fluid flow (seepage coeff and sink, velocity....) in user subroutine. If it is the case, would you please help me with this.

I think zero vol. flux is coming, because I am not defining any fluid flow through porous medium. As i understood, zero pressure indicates Fluid goes in and out freely.

Saturation is always showing as 1, does it mean that there is no fluid comes inside and goes out?

Material parameters are
*Material, name=Material-1
*Elastic
0.69e6, 0.018
*Permeability, specific=1.
9.6E-19, 0
3.072E-17, 1
2.3328E-16, 2
9.8304E-16, 3
3e-15, 4
7.46496E-15, 5
1.613472E-14, 6
3.145728E-14, 7



Is my solution terminating because of solid phase ? Since E is 0.69MPA and I am applying 2MPA.( for elastic body never sustains this load in compression).


As I know, fluid takes around 90% of the load based on time history.


I completely stuck at this point........Dont know what to do...


please help me.........

Any help would be greatly appreciated

Thanking you

Regards
krishna

 

[rstupplebeen]

I usually set up a model with as few nonlinearities as possible. I usually start with contact and linear elastic material properties. Then add nonlinearities as I go. I hope this helps.

Rob Stupplebeen

 

[geotechniqa]

-Undrained case means that there is no flow of fluid within the porous medium (insignificant consolidation/volumetric strain at a point). This happens, for example, when the permeability is low. As you may know also when Poission ratio is 0.5 (or close to 0.5), one will also have no volumetric strain.

-By the way what is the unit of this permeability? It is very even if it is m/s. If this is the realistic permeability, we have to find solution to overcome numerical problem induced by this poor drainage causing volumetric locking.

-Try to use consistent units , for example: Pa ( pressure) , S (time) , m/s (permeability)
Is the unit of the Specific Weight of your fluid in your permeability inputs which you give 1 value. I do not know what type of fluid you have. If u use Pa for the pressure ( as I see), then SPECIFIC Wight of the water is 9810 N per cu m.

- When you apply the load, flow should occur (fluid leave the system from the drainage surface which is defined by the BC of zero pore pressure and thus consolidation will occur) if the load is not applied very fast compared to ability of the system to drain the fluid (for example if the permeability is not very low).

- If you define the porous medium as saturated (S=1) this means that all the pores of the medium are filled with one fluid ( say water) and no air or another fluid exists in the medium. Ideally the medium would will remain saturated (S=1) after application of the load as water will be squeezed out and volumetric strain will occur ( materials will compress) with pores size becoming less but still filled fully with water (S=1).

 

[krishnagoud]

Thanks for your concern about my problem.

My permeability units are m^4/Ns. Since sp. weight of liquid is 1, the permeability is same as in m/s

I am always using SI units in this analysis.

Since articular cartilage is biphasic medium, it is quite reasonable to have saturation=1.

as you said, because of zero pore pressure BC, fluid squeezes out from the surface. If it happens, then fluid does not take any external load. Is it true?? please elaborate this ! will amount of fluid present in porous medium gets reduced ? if yes how can i know this in ABAQUS

I requested MFL and MFLT in field output, but it was saying those are not relevant.

If i dont give any Zero Pore pressure BC, it does mean that fluid stays inside the porous medium and no squeezing occur. Then, since fluid is incompressible, it shoud take any amount of external load on porous medium and it should not deform !

But in the analysis with low external pressure load on porous medium(0.2MPa, problem gets solved with this load), still there exist deformation, and I noticed there is no much change in pore pressure with or without zero pore press BC.


Would you please explain, physics of the problem

1.If zero pore pressure BC, fluid goes out, means there exist mass flow rate

2. If no zero pp, then no sqeezing. can fluid take any amount of compressive load?

3. How the external load will be shared between solid matrix and fluid phase with or without zero pp

4. How the pore pressure changes with external load ? (simple cylindrical, surface pressure load on top surf, bottom is fixed)

5. How can i know load sharing between fluid and solid phases in ABAQUs.(is it effective equal to von mises stress, pore pressure indicates load shared by fluid ?)
As i know total stress is effective+pore pressure

6. does Zero Vol flux indicate there is no fluid coming in out ????


Would you please explain above !

Would you please give your contact number and free time, sothat i can contact you if need something else ? if it is ok with you !


Regards
krishna

 

[geotechniqa]

"as you said, because of zero pore pressure BC, fluid squeezes out from the surface. If it happens, then fluid does not take any external load. Is it true?? "

-Yes but only at the surface on which pore pressure is zero

For example, if you you have a layer with thickness H allowed to drain from the top only and you apply total constant preeuse on the top, the load at each point within the medium will be taken by pore pressure and the skeleton. However, you have to look at the space and time : Regarding space, the load taken by the skeleton will increase by moving to the bottom of the layer (far from the drainge surface).Regarding time, the laoad taken by the skeleton will increase with progress of time while that taken by pore pressure will decresae until reaching equlibrium : full load is taken by skeleton. If

-If the material is highly permeable the equlibrium will be reached fast meaning the water will drain quickly upon load application and after short time the whole load will be taken by the skeleton of the material. Note that points far from drainge surface will reach equlibrium later than those close to it

- If the material has low permeability then the equlibrium will take very long time ( in 10 m clayey soil , it could take 30 years to reach equlibrium at each point in the layer)

- plot the fulid flux along the surface at specific time and you will see that there is flux indicating moving of fluid out of the surface. It will be very small, though, because you have very low permeability.

" If i dont give any Zero Pore pressure BC, it does mean that fluid stays inside the porous medium and no squeezing occur. Then, since fluid is incompressible, it shoud take any amount of external load on porous medium and it should not deform !"

-The porous media should not have VOLUMETRIC DEFORMATION if
the skeleton is also incompressible. But the shear deformation will exist. You should be clear with the fact that the stress at a point is physically divided into Volumetric or confining or consolidation P = (sigma1 + sigma2 + sigma3) / 3 and distortion stress which is expressed by Mises stress q. The same thing is applied to strain or deformation. If no squeezing of pore pressure occur , no volumetric strain will occur but there will be deformation that is shear/distortional deformation.
To know how much load portion is carried by soil skeleton and how much carried by pore pressure during the time of loading application:
Plot the the time history of the pore water pressure vs effective confining pressure P' .

 

[geotechniqa]

Sorry . I want to say :
Regarding space, the load taken by the skeleton will DECREASE by moving to the bottom of the layer (far from the drainage surface). I attach a graph demonstrating what I am talking about.

 

[krishnagoud]

Thank you very much for spending your valuable time to answer my question.

I understood so much info from these scraps !

I have been feeling very happy since we entered into contact

I am very glad to stay connect with you

Thanking you very much once again !

Regards
krishna