Thank you for the feedback. I don't recall discussions so far about how interference in the assemblies are handled....so this is an interesting topic to explore.
In a few models that I have tried with interference and without interference....I don't see a behavior that looks like 'twice' the material is there. I get the same results with and without interference. I've also validated this in Ansys Discovery Live.
So possibly the difference in the models has to do with the performance of the graphics card fidelity used to solve in those corners of the models. Did you look at the 'deformation'? Does the deformation result look like twice the material is there? The sharp corners that you indicate will be troublesome for FEA programs in general due to stress concentrations or singularities in those areas.
I would like to investigate further if the deformation results show a significant difference between interference and no interference. Can you please let me know the differences for that measure?
Thank you for your time,
PTC CAD Product Manager
I've performed some more tests with a 100% interference component : main component is a plate with a hole and the second one is a small cylinder completely interfering the plate close to the hole (in the high stress area).
What i've found out is really strange : if both components have the same material, the system seems to do as if the cylinder was not existing (nothing can be seen on the area compared to the symmetrical area). But if i put different materials, then the comportment is really different and we can clearly see the stress concentration around the cylinder border.
I'm really curious to know exactly how the system is handling those cases so we can anticipate it in the model preparation.
I will look further into finding out what should happen when two part geometries occupy the same space and they have different material properties.
I've gotten a little more information to share. The overall assumption is that the interference is small or should be small, so the simplest/fastest solving approach is used. The interference doesn't try to do any blending of materials, it takes the material properties from the part model that is highest in the model tree currently and uses that for the interference zone.
I hope that helps answer your question.