Simulation with 2 disconnected parts

[vazaar]

Hi all,

I am trying to do a simulation where a hemispherical cup is placed into some sort of mold (see image). The cup is a little bit larger than the mold, and the idea is that the cup is forced in and deforms (in the plasticity range). Then, by friction the cup should stay in place. My goal is similair to this paper (but then in 3D) :Link

For this I use 4 repeating forces that 'hammer' the cup into the mold. However, my problem is that there is an initial distance (a gap) between the cup and the mold. Because of this, the cup only receives the hammering forces on one side and there is no equilibrium, and I get no covergence.

I was wondering what would be the best approach to solve this. If i model some elements that connect the parts it is not mimicing the reality anymore. As a boundary condition I use friction with a penaly formulation.

Any advice is much appreciated,
Kind regards

 

[Erik Panos Kostson]

As you say it is not in contact when you apply the load thus in static analysis it will not converge.

What you need to do is to just move the top part (cup), down so it touches with the bottom one. In that way the contacts will activate as soon as you apply the force.

I would also consider applying quarter symmetry making it faster to solve and the model more stable (e.g., preventing cup from spinning around).

 

[FEA way]

You can even move the cup so close to the mold that these 2 parts overlap a little bit. Then if you check "Adjust only to remove overclosure" in the contact settings, Abaqus should adjust these surfaces and convergence will be more likely.

Anyway, remember that contact with initial gap is always difficult to handle. Rigid body motions may occur and even "Automatic stabilization" option won't help in most cases.

 

[karachun]

Also you can use enforced displacements instead of forces and check reactions.

 

[Mustaine3]

Just position the cup in a way, that it directly contacts the mold. There is a "Translate to" option in the Assembly-Module.
Also activate the adjust option in contact with a small value to overcome small numerical differences in the initial state.

 

[vazaar]

Thank you all for the suggestions. I did move the cup towards the mold and now I get a solution.

The only problem now is that the cup just won't move downwards into the shape of the mold. I use relatively large forces (4000N) and the materials that I use have a very low E-modulus and low yield strength. As a result I get that only some elements are deformed, but the cup is still in the same spot.

When I decrease the time steps (which would result in a faster increase of the force) I still obtain convergence errors.
I find it strange that it doesnt deform that much and doesnt move downwards.

Is it because of some mistakes in my model?

Thanks in advance