I am attempting to model a joint between two tunnel sections.
The joint's compressive behaviour can be approximated as a tri-linear elastic. I have created (and tested) a UDSM with tri-linear properties and that seems to works.
However, because the joint is very thin in comparison to the tunnel elements, when I tried modelling this as a soil block it wouldn't mesh well even on the finest mesh setting. Therefore I am attempting to model the joint as an interface by approximating the stress-dependent stiffness behaviour using equation 356 in the Plaxis Materials Model manual.
I am trialling this model on a very simple model comprising two concrete blocks, the first fixed at the model boundary, the other linked to the first with an interface set to this UDSM material. I am then applying a pure horizontal load (Compressive & Tensile) to get a force-displacement graph to compare to my known material behaviour. It's not behaving quite as expected which I suspect is to do with the interface thickness being undefined. So I have the following questions:
1. Equation 356 in Plaxis manual doesn't seem to have matching units (Left side kPa^2, right side kPa) - Why?
2. Is using an interface the best approach to modelling a thin, but very important, element between concrete/rigid elements?
3. If so, how can you set the real thickness of the interface?
4. If not, what alternative approach could I take?
5. Is there a way of adding a tensile strength cut-off for a UDSM interface? (Like in Mohr Coulomb model)