Cohesive Zone Modeling 2

[meolo2002]

Hi to everybody,
i'm trying to do a simple test on cohesive materials.
The model is composed by a cantilevered beam made by two layer of steel glued together by a layer of glue.
Working units SI,mm
Plate dimension lenght 500mm width 100mm thickness: steel 10mm, glue 5mm, steel 10mm.
Type of mesh, hex8
Steel prop. E 210000, ni 0.3
Cohesive material prop.
cohesive energy 40
critical opening displ 0.01
max opening displ 0.1
ratios 1
energy dissip. 10E-5
i've applied a force at the end of the beam on the node of one of the steel layer, pulling the layer down.
I expected that, reached the maximum opening displacement,the layer of glue breaks... but this does not happens.
Tied to change the values of glue and value of the load but no difference in results.
Any idea?
Thanks

 

[hezhj2000]

When you say 'the layer of glue breaks', do you mean the glue elements breaks, or do you mean the glue elements have damage? In Nastran/MARC, the cohesive zone element will not break. You can check the damage of the element to see if it 'breaks' or not.

 

[meolo2002]

-__________- i mean helements break, like in VCCT (but without the calculation of direction of the crack)...
Anyway how i can determine if some elements have a damage?
Thanks

 

[hezhj2000]

If you post-process the results in the MASTER/DBALL file, you will see 'Damage Value'. It should be controlled by the 'NLSTRESS' case control entry.

 

[meolo2002]

The problem is that even if i change the Cohesive energy of the glue, the solution ( in particular the strain in the glue)doesn't change at all.


In addition i can not attach the MASTER result: the following error message is visualized:

$# An internal error occurred when retrieving Direct Results Access data in routine: "DraGetStateData".

 

[hezhj2000]

Did you request results in MASTER/DBALL? You need to add 'scr=post' option when you submit your Nastran job.

 

[meolo2002]

Yes, of course 've requested master/dball
No news good news... scr=post doesn't works... the problem remains

 

[hezhj2000]

You can try the attached Nastran bdf file. It is working on my machine.

 

[meolo2002]

sorry for the late... i had no time for reply...
Another question... why do you use prismatic elements for cohesive??

 

[hezhj2000]

Prismatic elements were used to test the cohesive mesh can be non-congruent with the structure mesh.

 

[Mechaero2006]

Hello,

I am interested in knowing if there is an analytical relation for the calculation of the critical opening of a cohesive material. Normally this is something that is calculated from experimental results, however from a given test (DCB) we have the following mechanical properties:
sigma_o = 55 N/mm2, tau_ο = 120 N/mm2 and Gc = 0.8 N/mm.

The constitutive model that will be used is the exponential. I suppose there has to do with a relation between Sigma_ο and Gc.

Any information (including additional info if for example a bilinear or linear exponential law is selected) will be useful.

Thank you in advance for your assistance.

 

[Mechaero2006]

I think I've found the answer myself:

MSC MARC 2012 Volume A: Theory and user Information pp. 517 to 525.

Let's hope that someone will find this useful.

PS: As to the model provided by hezhj2000 in a previous post, I confirm that it works on my machine too.

 

[meolo2002]

Sorry, ive seen this reply only now ..
thank you for your reference .... it's very useful!