Problems with self contact

[francoisgl]

Hi,

I'm tring to model a forging process with ABAQUS/Explicit, but I've got a problem with self contact (or maybe surface définition). one surface of the elastic-plastic material is schrivelling up during the simulation and after a certain elapsed time one of my boundary node is moving inside the matérial, making the simulation failing.

I defined my surface like this (I aslo tried without the internal surface, without succes, and my elements are CAX4R):
*Elset, elset=S1, instance=lopin, generate
1, 150, 1
*Elset, elset=S1_int, internal, instance=lopin, generate
1, 150, 1
*Surface, type=ELEMENT, name=S1
S1_int, S1

The contact is defined with:
*Surface Interaction, name=contact
*Friction
0.2
*contact Pair, interaction=contact
S1, S1

Any advice would be very appreciated, thanks in advance!

François.

 

[francoisgl]

Hi again,

While I was trying to solve my problem I ran the example problem "ale_sinusoid_forgingaxi" (contained in Abaqus example problem manual, section 1.3.9) to see how it was defined. Exactly the same problem happens at the end of the simulation if I divide de velocity by 10 and multiply the simulation time by 10.

Could someone explain me why some nodes/elements don't react correctly in this case?

I've made a lot of attempts to solve this, without succes, and my work is blocked...Thanks in advance!

François.

 

[bassmanjax]

Hi Francoisgl

Regarding your contact problem, you could look at using general contact rather than the contact pair algorithm. Look at *CONTACT and *CONTACT INCLUSIONS.

Good luck!

Martin

 

[bassmanjax]

Agh, my mistake - just re-read the original post and I see that you are using axisymmetric elements. I believe that the general contact method is only available for 3D elements, so it won't help you at all I'm afraid.

Martin

 

[francoisgl]

Ok, no problem, Martin...

Still trying to understand why the "ale_sinusoid_forgingaxi" failed, I just tried to divide the analyse in two steps (I was only using one step), specifying the option op=new for adaptive mesh, and I got a solution!

So maybe it will work for my work... I'll give news soon, it would maybe be helpfull for others.

Thanks anyway, Martin,

Francois.