J-Integral for non-sharp growing crack

[Gabriel Coelho]

Hello, everyone

I am trying to model a C(T) specimen to derive the resistance curve from it. As it can be seen from the figure above, I have modeled it considering a non-sharp crack with crack front defined by the row of elements ahead of the pre-crack tip. The growth is simulated by element deletion considering a damage evolution law (GTN model, in abaqus known as porous metal plasticity).

I want to obtain the J-Integral values for each time step so I can derive the J-Δa curve. How could that be done in Abaqus? What woyld be the most easy and straightforward way to do it?

Thank you in advance

 

[FEA way]

In Abaqus you can request J-integral and stress intensity factor output for analyses with stationary cracks (using either regular FEM or XFEM). But you have to define crack first (Special —> Crack —> Create).

 

[Gabriel Coelho]

Hi FEA way

That would work only for a seam crack, i.e., a sharp crack with front defined as a line or point. As I explained, I am trying to make a more "real" crack front considering it a non-sharp (blunt) crack. For that, I do not know how to proceed.

If anyone could give me some help, I'd appreciate it. Thanks

 

[FEA way]

It's possible to analyze blunt cracks this way too. In such case you should select the blunted notch surface as crack front.

 

[Gabriel Coelho]

I tried that, FEA way, but it only allows me to select vertices and points when it comes to define the crack tip region (see print attached). What am I doing wrong?

 

[FEA way]

For crack tip you have to select nodes at the actual tip of the crack. Here's how it should look like for rounded crack (modeled using symmetry):

 

[Gabriel Coelho]

I tried and got this error message:

***ERROR: THIS KEYWORD IS NOT AVAILABLE IN Abaqus/Explicit
LINE IMAGE: *contourintegral, crackname=H-OUTPUT-3_CRACK-1, frequency=0,
normal, contours=6, cracktipnodes
***NOTE: DUE TO AN INPUT ERROR THE ANALYSIS PRE-PROCESSOR HAS BEEN UNABLE TO
INTERPRET SOME DATA. SUBSEQUENT ERRORS MAY BE CAUSED BY THIS OMISSION

***ERROR: ANALYSIS WILL BE TERMINATED

It seems that contour integral does not work with Explicit. Any ideas on how to work this around?

 

[FEA way]

Yes, you have to use Standard (implicit) for that. There’s no option to evaluate J-integral and SIF’s in dynamic explicit analysis. So you would have to change solver.

 

[Gabriel Coelho]

Would there be a way to calculate J-Integral by using the History Output results?

If so, how could I proceed that?

 

[FEA way]

I'm afraid that the only way would be to export field outputs (stress and displacement components, strain energy density) from Abaqus and postprocess these results in some external software.

 

[Gabriel Coelho]

Any ideas on how to perform that which is not manually?