incompatibility between elements

[MatthieuM]

Hello,
I'm using the "coupled temperature-displacement" procedure for creep with two kinds of elements :

-C3D10MT

-C3D4T (where the Von Mises stress is low) to reduce the computing time.

My problem is that Abaqus gives me this error message :"the selected geometry order will introduce an incompatibility between the selected region and the neighboring region", which appears in the Warning too.

What are the consequences of this incompatibility on my results ?
Generally speaking, are the incompatibilities between elements a real problem ??

Thank you !

 

[corus]

If there's incompatability then Abaqus will tie the two regions together. This will give some inconsistencies in the results along the bounding surface between the two regions as the displacements are interpolated between nodes. If this is well away from regions of interest then it shouldn't be of much concern but generally it's better to have a consistent mesh throughout. Personally I'd avoid using tetrahedral elements at all unless there was absolutely no way of being able to mesh the region with brick type elements in a reasonable amount of time.

 

[MatthieuM]

Thanks for the answer,

I know that linear tetrahedron are really too stiff, however with quadratics like C3D10M I get good results (but the computing time is high). Besides, i read that for plasticity and contact, the modified tetrahedral (and triangular) elements are better to avoid shear and volumetric locking.