Please follow the paragraph "I have problems when using the Interface (GAP) material on contacts" from our "Known problems" page http://www.cervenka.cz/products/atena/known-problems for the Interface Material settings.

However, if you have no information about the concrete-epoxy interface, I would recommend first running the analysis with "perfect connection" instead of GAP material, and only switch later if you see the results are too high due to not capturing the debonding.

Regarding the epoxy stregth, you could
a) keep the elastic model and only check in posptocessing that the stress does not exceed the strength
b) switch to Bilinear von Mises material with ideal plasticity (0 hardening modulus). However, if the epoxy behaviour is more brittle than plastic, you still may have to check in postprocessor
c) if you have more material data (e.g., direct compression and/or tension test on an epoxy sample), you may choose a more advanced material model, calibrated by modelling the material test.