Re: [Getfem-users] bug in von mises stress calculation?

[Renard Yves]

Dear Andriy,

A priori, I found the reverse in the svn tree. the factor sqrt(3.0/2)  
has been added by Jean-Yves Heddebaut in r2953 (version 3.1) and is  
still present. Is this factor correct ?
If I follow the definition of the Von Mises stress in  
http://fr.wikipedia.org/wiki/Critère_de_plasticité
and after basic computation I found a factor of 3.0/sqrt(2.0). Am I wrong ?

Yves.



Andriy Andreykiv - 3ME <address@hidden> a écrit :

> Dear Yves and Getfem users,
>
> Want to report a bug. I was using von mises calculation in  
> getfem::mdbrick_nonlinear_elasticity<>.
> It was giving a wrong value even though all the other results of the  
> calculation were consistent with the benchmark.
> So I checked the code and found that in 3.0.1 version of getfem (the  
> version which documentation is on the website)
> the final expression for the value is
>
> VM[i] = sqrt(3.0/2)*gmm::mat_euclidean_norm(sigma); (which is  
> correct, I believe)
>
> while in 3.1 version the multiplier sqrt(3.0/2) is absent.
>
> Best regards,
>                  Andriy