Problems with tetrahedral elements in non-linear elastic analysis

[aekiep]

Hi,

I've written an FEM code based on hierarchical, isoparametric tetrahedral elements. Now, I'm having some serious problems getting the 3D version of my code to produce sensible results in a simple static simulation of the stretching of a bar consisting of neo-Hookean material. The test setup is conceivably simple and works perfectly with the 2D version of the code, which is based on isoparametric, hiearchical triangular elements and shares great portions of the source code with the 3D version, including the Newton-Raphson solver used for this problem. Also, since I've used this 3D code in a number of other applications, and tested it thoroughly with unit tests, I'm pretty sure the code is correct, hence my question: Is the type of elements, that I use, somehow intrinsically not suitable for this kind of simulation?

Thank you in advance for your answers.

 

[kellnerp]

What do the results look like and what do you expect them to look like?

TOP
CSWP
BSSE

http://www.engtran.com

http://www.niswug.org

"Node news is good news."

 

[aekiep]

Thanks for the reply.
Basically, since the load was quite low, like 10% of the Young's modulus used in the computation of the Lame coefficients, I'd expect something very close to the results of linear elasticity. However, the code produced wildly varying results, i.e. higher loads would sometimes yield less elongation and the like, or did sometimes not converge at all.
In the time, since I posted this question, I've managed to produce realistic results by increasing the order of interpolation to 3rd order. However, there are still some stability woes, and I can't find any reasonable explanation as for why quadratic interpolation would not work, so I'd still be glad for some input.