unexpexted results with a mesh test with hexahedral elements.

[briandebont]

I have analyzed a cantilever beam with the aid of Abaqus. My goal was to determine which hexahedral element can be used best with a natural frequency analysis. It went very well accept for the linear C3D8 element.

It gives good results with a very fine mesh, as expected. But i imagined that coarsening the mesh would make the element stiffer due to shear locking. as a consequence the natural frequency should increase. This is however not the case, the natural frequency decreases.

Anyone knows what is going on here?

 

[testing]

The continuum elements do not have rotational degrees of freedom, so for a mesh that is too coarse, I'd expect you'd have much less stiffness as a result (since you have no bending resistance at the nodes), which would mean a lower frequency (~K/M)

 

[briandebont]

Thank you very much for your answer.