For my graduation project I’m working on a model to describe cracking behavior. For this model I want to model a fibre-epoxy interface, see the attached figure 1 (due to symmetry only a quarter is modeled). The fibre part is modeled with lineair elastic material. The epoxy part is modeled with CPS4 elements, with elastic behavior and MAXPS damage (damaged evolution included). The connection between the fibre and epoxy interface can be modeled with both a cohesive surface and cohesive elements. However cohesive surface will give inaccurate results due to peak stresses (see attached image 2).
For the attached image 2 I tested single cohesive elements, and two lineair elastic elements (very high E-module, thus these approach infinite stiffnes) connected with a cohesive surface.
Therefor I want to model the fibre-epoxy interface with cohesive elements instead of cohesive surfaces. I previously used a cohesive surface to model the interaction (before I found out about the inaccurate results). The cracking in the epoxy interface was modeled with XFEM. The results of the model where reasonable, however not completely correct. The normalized traction separation diagram is added in figure 3. However this normalized traction separation diagram should look as attached figure 4 as found by Alfaro and Suiker (2010).
Alfaro, M. Suiker, A. (2010) Transverse Failure Behavior of Fibre-epoxy systems, Journal of Composite Materials, Vol. 44, No. 12/2010.
So I tried to adjust the model, and change the fibre-epoxy interface from a cohesive surface to cohesive elements. When these cohesive elements are modeled the behavior of the model is however completely incorrect.
The input (.inp) files for both the cohesive surface and the cohesive element fibre-epoxy interaction are attached.
Why does Abaqus give completely different answers to the model when cohesive elements are used to model a fibre-epoxy interface.
And how can the model be adjusted to behave correctly.
Thank you in advance,
Any help would be very helpful!
Steven
For the attached image 2 I tested single cohesive elements, and two lineair elastic elements (very high E-module, thus these approach infinite stiffnes) connected with a cohesive surface.
Therefor I want to model the fibre-epoxy interface with cohesive elements instead of cohesive surfaces. I previously used a cohesive surface to model the interaction (before I found out about the inaccurate results). The cracking in the epoxy interface was modeled with XFEM. The results of the model where reasonable, however not completely correct. The normalized traction separation diagram is added in figure 3. However this normalized traction separation diagram should look as attached figure 4 as found by Alfaro and Suiker (2010).
Alfaro, M. Suiker, A. (2010) Transverse Failure Behavior of Fibre-epoxy systems, Journal of Composite Materials, Vol. 44, No. 12/2010.
So I tried to adjust the model, and change the fibre-epoxy interface from a cohesive surface to cohesive elements. When these cohesive elements are modeled the behavior of the model is however completely incorrect.
The input (.inp) files for both the cohesive surface and the cohesive element fibre-epoxy interaction are attached.
Why does Abaqus give completely different answers to the model when cohesive elements are used to model a fibre-epoxy interface.
And how can the model be adjusted to behave correctly.
Thank you in advance,
Any help would be very helpful!
Steven
