Dear Abaqus users,
I am using cohesive surface-based interaction property to model the debonding of rigid polymer from CFRP specimen. Different materials so different CTE's. 

I've created two static general steps:
Step 1: Thermal loading (cooling from room temp. to colder temp.).
Step 2: Mechanical loading. (Displacement BC is applied on the CFRP at one end and the other end is fixed.  Temperature from step 1 is propagated in step 2)

I've defined a surface-to-surface contact interaction with small sliding, using a cohesive surface-based interaction property. Damage initiation and evolution criteria are also defined. 

Damage Initiation: Strength values of polymer
Damage Evolution: Fracture values of polymer (under the assumption of failure in polymer as it is the weaker part)


I am facing a challenge specifically how the damage variable propagates across the step 2. When cooled from room temp. to colder temp., the rigid polymer contracts, and due to this deformation, the damage initiation index CSQUADSCRT gets fulfilled during the cooling step (Step 1), and then the damage evolution index CSDMG variable starts increasing monotonically. However, when the simulation transitions to the mechanical loading step (Step 2), the CSDMG remains constant and does not change further.

Interestingly, if I disable the first step (thermal loading) and only run the mechanical loading step (Step 2), the damage initiation and evolution criteria are fulfilled correctly, but of course, I had to reduce the strength and fracture values up to 90%.

I can play with strength values (Damage Initiation) to delay or initiate the initiation criteria but it can get fulfilled only in the cooling step and not in the loading step. When I somehow get initiation criteria fulfilled in the cooling step then my Damage evolution (csdmg) starts but it also does not increase when it goes to the mechanical loading step. If I reduce the fracture values, I can see the larger value of CSDMG but only during the cooling step, it does not change in the entire loading step.

I should also mention that the rigid polymer does not experience direct mechanical loading, so shear lag would be the dominant failure mode. My loading condition is very similar to the picture attached (Reference: Davila, Camanho and de Moura)


Any help would be greatly appreciated.

Thank you