How to get the logarithmic strain E=lnU?

[Jluc]

Hello everybody

I am writing a VUMAT for thermoplastics. This model is based on Boyce model with a spring + dashpot in parrallel with an hyperelastic spring. I have some trouble when I try to do computations in compression for a problem with several elements. For the elastic part of the model, I use the classical expression 2mu*Ee+(K-2/3mu)tr(Ee)I...I get Ee for lnUe and Ue using Franca algorithm (from Simo and Hughes book) to get the eigenvalues of Ce. Is this algorithm very robust for large deformation and different stress state...Is there other way to compute the eigenvalues of Ce to get Ue and then Ee.

Thanks for your help

Jean-Luc

[Jorgen]

There are many different algorithms that can be used to determine the eigenvalues. Since Ce is a small matrix (3x3), for example something as simple as Jacobi rotations works well.

-Jorgen

[Jluc]

Thanks Jörgen! I am going to continue to debug my VUMAT :-(

Jean-Luc