XFEM analysis of thermal fatigue

[gianlucaelle]

Hello, I am quite new to ABAQUS and I am trying to analyze the effects of termal fatigue on the crack initiation and propagation in a 2D unite cell of composite material.

I made a program which calculates the termal stresses induced by the temperature cycles, using temp/disp. coupled elements and another program which calculates the crack front using XFEM technique (mechanical fatigue, plane stress elements). How can I link those two jobs? My idea was to eliminate the load in the second program and insert the temperature in the predefined fields from the first program, and then run again the job; this is not working. There is a more functional way to do it? Is it not posible to use coupled temp/disp. elements for XFEM analysis?

Thanks a lot for your answers.

 

[rb1957]

if you know the stress in the intact model along the intended crack face, you can apply these stresses to the structural model (or use a handbook solution ... a uniform tension stress along the crack (ie pressure within the crack) is equivalent to a crack in a sheet under remote tension.

Quando Omni Flunkus Moritati

 

[gianlucaelle]

I don't get how can I apply what you said...

The various parts of the model are expanding or contracting according to their coefficient of termal expansión. I would like to:

1) analyze the fatigue issue due to these cyclic thermal stresses directly using the same ABAQUS program

OR

2) run the programme with the thermal cycles, recover the stress data and implement them in a new "mechanical fatigue" program, which uses stress elements, and not coupled temp/disp. elements.


Which is the best solution? I think the first one woould be better, but it gives me some issues about the time incrementation of the step (I put the cycles of temperatura under "predefined fields" as a table).

 

[rb1957]

if 1) means you can get it all done with one model, why not?

i take 2) to mean get thermal displacements from one model and apply them to a structural model, that should work too.

Quando Omni Flunkus Moritati

 

[gianlucaelle]

I got the model, but results are quite unrealistic since the cracks propagates from the matrix inside the fiber section. Do you know how to describe the interface between fibers and matrix in a 2D unit cell?