Hello everyone,

I am still trying to model a unit cell with a fibre and an interlayer. Since I am having convergence issues I was taking a step back and took a look at the interface again.

I've found two type of approaches in the literature: Use a small interface (e.g. 0.1µm) or a very small interface (0.0001µm) - Assuming the composite has to be modeled as one part and not be connected by a cohesive layer.

So I decided to compare the effect of the thickness and to my mind it should not make a significant difference.

I calculated the compliance as: K = E/t and chose a E = 3 GPa (the same as my matrix material).

Parameters:

Version1 (0.1µm):

Thickness: 0.1µm, MaxsDamage(0.06,0.06,0.06), Elastic(30,30,30), Displacement at failure: 1

Version2 (0.0001µm):

Thickness: 0.0001µm, MaxsDamage(0.06,0.06,0.06), Elastic(30000,30000,30000), Displacement at failure: 1

(The units I use are µm, ms and GPa)

In the section settings I set: Initial thickness = 0.1/0.0001

All other parameters were choosen to be identical. The left side was fixed, the right side with a displacement of 1 over 30000ms as Implicit Dynamic (quasi static).

Here are the results

0.1µm (not finished)

0.0001µm Thickness Finished:And the deformed shapes:

As you can see, with the thin interlayer the part fails at a significantly lower force. Also there seems to be a stress concentration at the crack tip.

Does anybody know where the difference comes from?

And do you have tips for me on how to improve the simulation time? Despite having only one element in thickness direction, the simulation takes some time to finish due to the very small time steps (E-9).

Regards

Mike