Plastic behaviour. 2

[psantunes]

Hi!

I'm running a model that cointais plastic data in the definitionf of the steel mechanical behavior. The plastic region is defined using a hardening modulus (classical plasticity theory), but I obtain several messages of excessive distortion followed by negative eigenvalue messages when the strees reach the yield values. I know that the negative eigenvalues messages are due to the loss of stifness or geometrical instabilities in the structure, but this loss is ruled by the plastic flow of the material and I can´t change this, unless I change the hardening modulus of the material.

I´m using first order, reduced integration hybrid elements to account the volumetric behaviour (incompressibility) in the plastic region.

I try to use the Riks method in order to account the geometrical instabilities in the model, that may occur due to the geometry of my model, but this approach don't work..

Do I have some problems in the definition of the best element formulation for plastic deformations???

Thanks for your help!

P.S. This models runs very well when no plasticity is defined!

 

[mrgoldthorpe]

I assume the analysis eventually converged and reached the final target load?.
I have ofen had these messages in elastic-plastic analyses. I think they arise during particular displacement estimates within the iterations for some increments. The important point is that the analyses eventually converge at each increment.

 

[psantunes]

Hi mrgoldthorpe!

Thanks for your answer!

No! I'm not able to complete the analysis! I know that the problem is located in the rings that are placed above the bolt healds. These rings are made from steel and they are not flat rings. They have a small concavity and because of that they change rapidily tehe geometry due to the large inelastic deformations (the forve per bolt is equal to 10000N. At this moment I'm refining the mesh for the rings, in order to understand if this problem can be solved by this way.

Thanks!!

 

[harry123456]

Yes, refining the mesh is a good idea. You may also want to try *Static,stabilize. This puts on some viscous damping and might be useful here.
harry

 

[psantunes]

Yes harry!

You are rigth!! I know that I can use the *Static, stabilize, to improve the behaviour of instable models (loss of stifness for instance). If this approach (refinement) don't work I will use the static stabilization procedure!

Thanks