Have you considered changing the fracture initiation criterion at the surface? The default one is usually stress-based, thus it won't work. But it is possible to modify it, there are a few already implemented. Among them, there's a strain-based one. As displacements (and thus strains) are well defined in your model, the program will be able to correctly determine fracture initiation.

You should also check the fracture propagation criterion. It is usually energy-based (i.e. energy release rate vs fracture toughness). You should verify that energy release rates' calculation is strain-(i.e. displacement-) based and not stress based. The formulation does exist and I think it is an option already available in Abaqus, but not the default one.

Hope this helps.

Best,

Luca

 Hi Luca,

First of all, thanks for your reply.

I indeed changed the fracture initiation criterion to use the displacement-based one (maximum separation described in the Abaqus Analysis User's Guide). 

I don't have any idea of the order of magnitude of the displacement, but I'll try. 

My problem also may be linked with the stiffness definition considering the bending-compressive behavior of the interface depends on it. I'm working with MPa and mm and assumed my stiffness was 10^3 higher than my modulus. 

Regarding the fracture propagation criterion, I chose energy-based. But as far as I know, I can't verify or change the type of calculation and no information is given in the .inp file or in the Abaqus documentation.

Thanks for your help

Fabien

I too was thinking about the same. Why have you chosen plane stress in an impact problem? I'm not sure the plane stress state's assumptions are fully satisfied in such scenario. Is there any reason except reduction in computational complexity? 

Hi Majid,

I agree with you. Solid elements are more relevant for an impact application. Nevertheless, the reason I chose continuum shell elements is that a VUMAT in-built user Subroutine is available with plane stress formulation. This subroutine permits to model in-plane damage and shear response.

Using 3D elements would force me to re-write a user subroutine and I'm not experienced enough in Fortran programming to do so. 

I have just tried to give my shell section material properties such as through the thickness modulus, Poisson's ratio and transverse shear stiffness but I still don't have any result on my output cohesive damage variable.

Thank you both for your replies

Dear Majid, 

Thanks for your remarks, I take it into account.

I succeeded to model fiber failure and delamination with continuum shell elements. The reason why it wasn't working was that my contact was not well defined.

I'm aware the accuracy would be better with solid elements, but considering the computational effort, and the re-writing of a user subroutine, continuum shell elements seem to be a good alternative.

Regards

Fabien

Dear Fabien,

I am experiencing a very similar problem and I am wondering if you can tell me how you solved it. Do you have an input or cae file of your example?

I am also modeling a continuum shell body with cohesive elements in places where the cracks are expected. In my case the reason for using this type of elements is in Hashin damage criterion, which is only used in combination with plane stress elements.

How did you determine the nominal strain in damage criterion?

Regard

Marko

Hello Fabian

I am also working on a similar problem, and was wondering if you were using damage evolution with Hashin damage or just the damage initiation for ply damage. If yes, how did you determine the longitudinal and transverse fracture energies required for damage evolution.

Also, I want to learn to write VUMAT for this project. I have no experience in this regard. Where should I start?

Thanks

Hello Nooruddin,

Hashin damage criterion is used to determine damage initiation only. In my case, modelling a woven fabric, I used maximum stress criterion but it works the same. Once the criterion is reached, starts damage evolution, which in my case was ruled by a power law, depending, as you said, on the fracture energies. Those fracture energies can be determined experimentaly (I don't remember the standards exactly), or you can find some values in the litterature.

I can't help you much about VUMAT subroutines as I used a built-in subroutine.

Good luck!

Fabien

Thanks Fabian

One other confusion i have is that can the built-in hashin damage initiation criteria be used for woven fabric composites. The matrix damage (compression mode) in the warp and fill direction is not the same (as should be in a plain woven fabric composite). Can the given damage criteria be used for woven fabrics without modification

Hashin criterion has been developped to model UD. As soon as you want to model a woven fabric, you should use another initiation criterion as the maximum stress or displacement. As you probably know, you could as well create your own criterion (by creating, I mean use any criterion of the litterature), by coding it and implement it as a subroutine.

Good luck!

Fabien

Thanks Fabien

I have started using a built-in subroutine (VUMAT) for woven fabric composites provided in abaqus. The model is based on this paper 

https://drive.google.com/open?id=0BynEgKdijyaHZnFLYVdDcXZjNHM

Will let you know how this works out.

Regards

Nooruddin

Hi Fabien,

I am also trying to model fiber fraction and delamination with continuum shell elements. I created one part and partitioned it using datum planes, so normally I don't need to specify any contact anymore. I gave a shell, homogeneous section to the fibers and a cohesive section to the matrix. but still it will not work. Do you have any idea how to fix my problem?

Thanks! 

Frede

Hi Frede,

I think you have to define the contact properties of the surfaces you've been partitioning. To model delamination, you have to define the stiffness of the contact, and the stress/displacement cirterion ruling the initiation and propagation of the delamination. I'm not sure I understand how your model works. Did you model any single fibre AND matrix ?

Hope it helps

Fabien