If you have small deformation, did you set the deformation scale factor to 1 in the postprocessor?

Thank you for your answer Mr. Martin Baeker .

I set my the deformation scale to 1 in postprocessor but the distortion still remain.

Do you know some element type that reduce this deformation? I used reduced integration, second-order accuracy and hourglass to enhaced.

Do I need to use Distortion control in element type? Which length ratio use?

If abaqus thinks that this deformation is real, set NLGEOM on, different element types will not help you.

Mr. Martin Baeker

Thank you very much for your answers. I corrected and adjusted my mesh quality by increasing the mesh density around the crack tip.

Your solution for using NLGEOM on couldn't work because the J-values obtained were very low

Hi Marcus,

I came cross the similar problems before when using linear element. Refining the mesh can minimize the hour glass distortion, but may not able to yield an accurate result. Have you tried using quadratic element (C3D20R) to see if two element types generate similar J/CTOD and stress field?

Regards,

Longjie

Longjie Wang thank you for your answer. I have tried to use C3D20R and the distortion was better, but I couldn't to calculate the J-values?

Did you have similar problems?

Hi Marcus,

Did you request history output for J contours as you did for C3D8R ? Could you provide more details about how you obtain the J ?

Regards,

Longjie

Longjie Wang I did request history outputs for J coutours, but, when I try to do the Job, it shows a message that I cannot be provided J values, as it can be seen in the Figure below. I don't know how to manage with this information

Hi Marcus,

Sorry for late reply.

Thanks for your further details. Have you sorted it out ? I haven't come aross this kind of problems before. Would you mind sending me the .inp file so that I can take a look at all the settings of the model ?

Kind regards,

Longjie

Thank you for your reply Longjie Wang

I'm still struggling with it. I am not able to find a solution yet. I add the .inp file.

Thank you so much for your consideration.

I have run it in my machine. Two observations:

At the end of teh step, plastic strains are huge, maximum strain is about 3, i.e. 300%, so this is totally unrealistic in any case.

At the end, there are large distortions that seem to be some kind of hourglassing (due to reduced integration elements). You would probably fare better using quadratic elements at least in teh region of the crack tip.

Thank you so much Martin Baeker

I will use your recomendations about plastic strains and element type.

You should define mesh setting concerning to your simulated model manually in order to avoid strange mesh distribution. In addition you can refine the problematic parts of mesh distribution.

Hi Marcus,

I have had a look at your model. The settings looked fine to me. Did you have problems on the convergence of the solution ? Would you like to send me the model with C3D20R element and I can run it to see what the result is like ?

Kind regards,

Longjie

This is very usual problem in FE analysis for bend specimen with a crack. Basically, I guess the plastic deformation behaviour at the crack tip cannot be expressed by the plastic formulation in continuum mechanics. In order to avoid this problem, I have developed a selected division method. By my method, FE analysis result using Abaqus can express the plastic deformation profile at the crack tip of steel material which was measured by the silicone rubber casting method. However, this method can be applied only until ductile crack initiation. This limitation is ordinarily CTOD