Hyperelasticity + Viscoelasticity 1

[MechanicPW]

I intend to use a model where the instantanous response is nonlinear elastic (hyperelastic - Neo Hooke potential) and is followed with a viscoleastic response (linear model - Prony series).

It doesn't calculate a single increment and I get information about "negative eigenvalues" and heavily distorted elements.

I've tried with a finer mesh (Hex). I've tried various element types. I've lowered the inintial increment value and trid with various values of strain tolerance (for viscoelasticity). Still error.

Step type is visco.

Thanks in advance.

 

[vumat721]

Have you tried your model with only the hyperelastic portion of the material model in a static step? If not, you can use CAE's "evaluate" function to test that the hyperelastic constants work properly.

Also, I would start with a one element model and pull it in tension or compression at a reasonably slow rate during a visco step. Make sure the stress/strain response seems correct. If no errors, then move on to using this material in your big model.

 

[MechanicPW]

Thanks for reply. I'll try this. I thought that I must have test data in order to use the "evaluate" option.

there might be something wrong with those hyperelestic constans as D is relatively high, which means that the Poissons ratio should be close to zero (and I get a warning message that the Poissons ratio is negative or close to zero.

I've tried with a purely hyperelastic model in a static step and it's alright. It completes the analysis.

 

[MechanicPW]

I've tried some compression tests with a single Hex element and that's the resolut:

http://img18.imageshack.us/img18/6562/creepn.gif

I think that probably it's alright but the creeping has been seriously linearized (I don't know - maybe it's due to the plot scale etc).

I've evaluated both hyperelastic and viscoelastic models and Abaqus says that they are stable for all cases.

I get a warning about a negative eigenvalue (1). One negative eigenvalue for one-element analysis. For a more complicated part there is much more of them and that's the reason why the analysis crashes. Still, I don't know what causes those negative eigenvalues.

 

[vumat721]

Abaqus won't evaluate the viscoelastic part of the model when you use the CAE Evaluate button. It ignores it and just evaluates the hyperelastic portion.

It looks like your creep time constant is really really small, which is why you see the initial spike and then the flattening out. Try to increase your time constant, or decrease your time scale.