Contact Analysis Solution Diverging

CMZA

Hi All,

I recently installed Calculix and have been playing around with using contact between surfaces. I don't have much of a mathematical background in FEA so wonder if anyone could give me some pointers.

I have a simple two part assembly; a portion of a pin inserted into a matching bore. A load is applied to the end of the pin setting up a cantilevered effect. I have defined contact surfaces; being the outer surface and one end of the pin as well as the bore that the pin is inserted into.

At a low loads it seems to run correctly and the results make sense. What I have found though is that as I increase the load, at a certain point the solver gives me an error to the effect that the solution is diverging. Does this mean that I have set the problem up incorrectly or does it point to something else?

Regards,
Christopher

Sergio

Can you please upload the model so we can play with that?

Victor

Yes, uploading a file will be helpful. Contact often has strange ways of failing and needs playing around with to diagnose. One way is to use small time steps and ramp up the load. Then you might see that the final solution before it fails reveals what's going wrong, like the pin penetrating the surrounding material or sliding out of the hole.

CMZA

Hi Sergio, Victor.

Thanks for your response. Attached is the model.

Two observations from playing a bit with it today.

1) At a 20kN load it solved fine, at 50 kN load it diverged, then at 100 kN it was fine again.
2) At the point of contact I'd expect the stress in the shaft to be a bit higher.

It's my first time at playing with contact, so I'm simply trying to get a feel for how it works, and the correct way to approach such problems. I have followed my understanding of the manual as best I could. But I'm sure I must be doing something wrong.

thanks for spending your time looking at this.

Regards,
Christopher

Victor

I suspect the problem is that the pin has no forces or constraints to prevent it either rotating or sliding out of the hole. In real life, you'd expect it to just pop out as soon as it starts to bend because there's no friction and the reaction force from the bottom edge of the hole has a small outward component. I found it sliding out when I reduced the contact stiffness, ramped up the load from zero and added a 0.5 friction coefficient. See picture.

So I added an axial constraint and it converges. Now it's a different problem though because the free end is no longer free to rotate. You might have to rethink what physical situation you're really modelling. Is the end of the pin really free to rotate? Is the applied force really perfectly perpendicular to the pin's axis and not slightly towards the hole?

There's also a lot of penetration now so you'd need to slowly increase the "Slope of pressure-overclosure curve" to reduce that, but going too high causes it to not converge.

CMZA

Thanks Victor,

That helps my understanding of the problem a bit more.

Kind regards,
Christopher