Export antisymmetric tensor about underworld2 HOT 3 CLOSED

Most of the constitutive models assume the tensor is symmetric and this is probably assumed in most parts of the code. Perhaps the function you are using is not compatible with being fed a symmetric tensor ?

There is also the issue of which solver setup to choose if the problem does not have the expected symmetry. We may need to discuss this somewhere other than the issue tracker.

from underworld2.

Due to the equilibrium of momentum, the stress tensor is always symmetric, so its asymmetric part should always be zero. However, the velocity gradient usually incorporates both the symmetric (strain tensor) and asymmetric (rotation) parts.

from underworld2.

Thanks both @lmoresi and @tingyang2004 . I'll asume that tensor is symmetric.

from underworld2.