Computation of Stretch for Ogden Material

[R2197]

Hi

I am trying to implement Ogden material for my model. I already have a neo-hookean model implemented and to test my implementation of Ogden material, I plan to simplify it for N=1 and a=2 which then becomes the NH model, and compare the stresses.

Now, I already have the required tensors F(3,3) and C(3,3). But I am stuck at computing stretches from these. I understand that stretches are the eigenvalues of C and hence I use 'call EIG3(C, L, r)' to compute them and use these to calculate the corresponding principal stresses, but the solution explodes. The other way I tried is getting the stretch tensor by taking the square root of C, and then its eigenvalue but the solution still explodes.

Can someone help me with this?

Thanks in advance
Raghav

[Prof. S. Govindjee]

What do you mean explodes?  Is eig3 throwing an error?

[R2197]

Hello Prof. Govindjee

I meant that the stretch values keep increasing from 1 to 100 to 1e13 and eventually the residual goes to Nan or Inf.

There is no particular error message that I get from eig3.

Thanks

[Prof. R.L. Taylor]

No way to tell from your information.  If you have equal roots then special care needed. Ogden’s book gives details.

[Prof. S. Govindjee]

The first thing to do is to impose a set of deformations and satisfy yourself that the stress and residual are correct. 

Then use FEAP's built-in material models to get a feel for the expected convergence behavior for your test problems.

Then you can try to use FEAP's numerical tangent options to see if your element will converge with numerical tangents (though probably slower than FEAP's models); see

Code: [Select]

TANGengt,NUME,1
Lastly you can use FEAP's feature to compute a numerical tangent of an individual element and compare it to your tangent; see

Code: [Select]

NTANgent,ELEM,<n1>