Residual Stress in an NLAD Forming Model

[nds88]

Hello, I am trying to experiment with nonlinear adaptive meshing and I made a quick test 2D plane stress model.

In the model, a rigid body 'tool' uses a prescribed displacement to contact a billet as a flexible body. The flexible body is fixed in the x-axis on it's sides and x/y on its base. The material is a copper with a bi-linear isotropic hardening plasticity model (a default ansys material). It's a simple model and it doesn't really represent any real process, but when I was looking at the results, I was wondering how do you interpret residual stress after gross-plasticity? Obviously yield is increased after the load is removed, but how is residual stress in this model is still way over yield? Initial yield is 280 MPa. Is this possibly because the data file does not contain a data point beyond 0.0127 strain? Maybe I am misunderstanding something about residual stress?

Thanks for your help!

 

[NRP99]

nds88

How billet takes shape of punch without undergoing large plastic deformation? So the stresses are expected to be beyond the yield to deform the billet to take final shape. Now, about residual stresses after any forming process, these are expected to be close to yield. As you know stress reliving is used to relieve these stresses after welding/forming process. The residual stresses are required to be in equilibrium with each other (tensile and compressive stress field co-exist). Otherwise there will be distortion/deformation in the direction of unbalance.

Check the directional stresses instead of von Mises to see exact magnitude of residual stresses.

After the last point of true stress-true (Plastic) strain curve input, the material behavior is considered to be perfectly plastic by any FEA Software. This is logical. Once ultimate stress reached, there will not be much strain hardening and test piece deforms without much load until it necks and breaks which is analogous to perfect plastic behavior.

 

[FEA way]

One thing is that the stresses displayed here are extrapolated from integration points and averaged. This distorts the values.

Did you check the value of plastic strain (PEEQ) ?

This material model is simplified, if you can get the data for the multilinear model (data point for the whole plastic part of the stress-strain curve) it should give you better results.

 

[nds88]

Thanks for your replies. I understand that bi-linear isotropic hardening model is a simplified model of plasticity and that a multi-linear kinematic model would be more true to life (this requires test data, which I don't have). I'm not actually interested in accuracy of the model, my question was more about the magnitude of residual stress after the load is removed. Is it expected that residual stresses after the load is removed are as much as 3x initial yield? I expected stress to return to a much lower value.

 

[NRP99]

I think residual stresses should not be more than yield.

-Did you Check the directional stresses instead of von Mises to see exact magnitude of residual stresses?
-Is this a random problem? Try with verification problem from Ansys verification manual or try a problem whose results are already known.
-Did you read the limitations of using nonlinear adaptive mesh? (Remote displacements can not be used with NLAD)
-Have you tried this problem with nonlinear static structural analysis without nonlinear adaptive mesh?
-In forming process, the strains are sometimes more than 1. Check with complete material model as FEA way suggested.

For results to make sense,

I'm not actually interested in accuracy of the model

this approach is not recommended. Try to achieve at most accuracy in your model to get meaningful results.