Fixing thickness direction in cohesive element

[DhirajKumarSingh]

I want to model crack propagation in CT specimen using 3D cohesive elements (COH3D8). Exploiting symmetry I have modelled only quarter of the specimen. The bottom face of the specimen has COH3D8 elements. As the cohesive elements open up under displacement loading, the normal and shear separations at integration points are calculated w.r.t. a coordinate system lying at the midplane of the cohesive element. Now this coordinate system keeps on changing with the opening of cohesive element.
Is it possible to specify a coordinate system at the bottom face such that ABAQUS determines normal and shear separations w.r.t. that coordinate system?

Regards,
Dhiraj Kumar Singh
PhD student, IIT Bombay