Mesh stack orientation inconsistencies between hex and wedge cohesive elements

[justAnEngineer95]

Hi all,

I am trying to model a scarf joint using cohesive elements for the adhesive. The adhesive mesh is modelled using a mixture of wedge and hex elements to create a "structured" mesh.
Taking a look at the adhesive part, the way that this is accomplished is by assigning a structured mesh to the middle parts, and a hex-dominated sweep to the triangular parts.
(Do not mind the fact that there are two elements through the thickness of the adhesive, this is something I can control with a script.)

Problem?!

However, when taking a look at the mesh stack orientation of the adhesive (cohesive) elements, there seems to be some inconsistencies at the areas between hex and wedge elements, since the mesh stack is assigned using faces.
I want the mesh stack orientation to follow the Y-coordinate of the CSYS I have create denoted CSYS Bondline, which is depicted in the two first figures.
Is this possible or are there any workarounds?

One possible solution I have found is to remove the triangular parts of the adhesive completely - like the article "Application of a Direct/cohesive Zone method for the evaluation of scarf Adhesive joints (2018)".

 

[FEA way]

Have you tried using the Assign stack direction tool to obtain the desired orientation ?

There was a discussion regarding modeling of this type of joint in Abaqus some time ago, you may find it helpful:

https://www.eng-tips.com/viewthread.cfm?qid=488121

 

[justAnEngineer95]

Yes, using the Assign Stack Direction tool is what creates the result in picture. It works everywhere except for the transition between the wedge and hex elements.
In fact, from the previous post, I can see that the same problem is encountered in the figures.

Here is the picture from the post you mentioned. The OP has issues with his far field stress distribution through the thickness of the laminate, which I fortunately do not.

If you look at the end, the mesh stack direction is also not assigned correctly for the wedge element...

Assigning the material orientation is fine. There are no issues here. It is the mesh stack orientation which does display what I want.
I suspect it is a issue related to the wedge element and mesh stack orientation specifically. I don't know if anyone has experience with wedge type elements?

 

[FEA way]

Have you tried running this analysis already ? If so, do the results look correct ?

 

[justAnEngineer95]

I currently get the following error: The local-1 direction (after the additional rotation and permutation) of the material orientation , specified via *ORIENTATION, is normal to the SHELL, MEMBRANE, GASKET, COHESIVE, or SURFACE element. The local-1 direction must be projected onto the element surface. Check input data. The elements have been identified in element set ErrElemUserNormMatOrient.

The analysis works when I simply use solid elements in the adhesive, e.g. C3D8R and the linear wedge element. (Edit: Solid elements are also used for the adherends. Layup is [0/90]2s)
The analysis produces OK results - here are the bondline stresses from the str8 up linear elastic analysis: Validation case: Gunnion 2006 - Parametric Analysis of A Scarf joint.

However, when I switch and update my model to account for and use cohesive elements as the adhesive, I get the error above. This is what makes me suspect that something is going with the mesh stack orientation, which needs to be specified for the cohesive elements.
I found this video which talks about mesh stack orientation for CZ elements.

https://www.youtube.com/watch?v=jpMEVGN4c-Q&t=504s

However, I have trouble assigning this to the wedge elements.