3D crack growth in abaqus 1

[satyasrinivas]

Hi,

Is it possible to simulate 3D crack growth in ABAQUS.
Earlier, I developed my 3d models :they consist of two solids. I defined the contact between the solids as an interface.

Now, I would like to try to model the above interface as a 'crack' and also simulate it's growth.

The initial crack geometry is a 3D curviplanar surface.

Is it possible to do this in Abaqus. Any suggestions/help ? Thanks
Rajesh

 

[amubashar]

Hi,

One of the possibilities is the use of cohesive zone elements at the interface. Abaqus documentation provides details about that.

Regards

Aamir

 

[elkhorn]

Hi:

From my own experience, there is no easy way to simulate a crack growth using ABAQUS. In particular, if your model is a quite complicated one.

Maybe it is possible to do it with a simplified geometry. My model was quite complicated.

Good luck.

Elkhorn.

 

[satyasrinivas]

Thanks Amir and Elkhorn.

I am attaching an image file and a video file. The image file shows 'faults' (shaded portions in the block view) - these are what I would like to develop as shear fractures. Upon extending the block in the direction of the thick yellow arrows, I want to see
1) how slip occurs on the interfaces and
2) how do the interfaces themselves grow

I should have attached these before. Please take a look at these and let me know if I can still go ahead and do this in Abaqus (standard or explicit).

Thanks
Satyasrinivas

 

[satyasrinivas]

Sorry Amir and Elkhorn.. the image file did not come through. I am attaching it here again

Thanks
Satyasrinivas

 

[ccheric]

You can easily simulate crack growth given the way you set up your problem - you have predefined crack path (interface), so you don't need predict crack growth directions.

As amubashar mentioned, you can use cohesive elements and corresponding traction laws to model the "bond" between the solids. I'm not sure if ABAQUS has automatic cohesive element placement, you should check. Watch out for mesh dependence.

Another way available in ABAQUS 6.8 is VCCT, which is a fracture energy method. You can also use COD and other criteria like that. You can do fatigue on it too.

Also, the video you posted seem to have a lot of plastic deformation. I suppose to have plans to set that up.

Once again, since you have predefined crack surface, it's pretty easy to simulate growth on it.

I hope this helps.

** XFEM can simulate growth in arbitrary directions depending on loading. You can put singularity into the formulation. It's the sweet deal of the future.

 

[satyasrinivas]

Thanks Ccheric.

As you pointed out right, I am predefining the initial crack surface (shaded region in picture). But once the deformation begins, I would like to see how the surface propagates - in other words, I will not dictate the trajectory of the 'new' surface and would like to see how it grows (beyond the initial geometry) in Abaqus. Is this do-able in Abaqus ?

Also, I am pretty new to ABAQUS and am not an engineer. So I am still grappling to understand the ABAQUS package and it terminology. The procedure you and Ambushar mentioned -namely using cohesive zone method - is it doable in Standard or do i need explicit for that ? Please let me know. Unfortunately I have limited access to Abaqus 6.7-4 only. So I will not be able to try the VCCT approach you suggested. Thanks so much for the information though.

Could you suggest any good online tutorials for downloading and using XFEM. Is it something I can couple with ABAQUS ?

Thanks
Satyasrinivas

 

[ccheric]

What you are looking for (simulate trajectory of the new fracture surface with predefined interface) is exactly what cohesive elements and VCCT supposed to do! (XFEM is fancier, with no predefined interface; I don't know where to get it)

Cohesive is available in both Standard and Explicit; VCCT is only for Standard. Typically, (like the video you've shown), standard should be more appropriate.

You can find enough info on cohesive elements in your Abaqus manual. You can make something simple to start with. Make sure you allow many increments and small time steps, as they are probably needed for convergence.

Good luck!

 

[satyasrinivas]

Thanks Ccheric. That was quite informative.