We all know that at sharp inside corners (e.g in welded parts) high stresses can be seen due to stress singularities and also at the supports due to rapid change in stiffness. These two cases always converge to an infinite stress. My question is how we can interpret the results. How far from the support or corner are the stresses valid for judgment?
I tried following example:
Simple cantilever (hollow circular cross section) with solid mesh and appropriate mesh refinement at the support, resulting bending stress is very high at the support which we expect to be. This stress is substantially larger than resulting stress from a beam mesh or manual calculations.
The same cantilever with shell mesh and immovable fixture at the support gives results close to theoretical values. However if you restrain the rotations as well resulting stresses are high.
In one of the Solidworks companions we learn some techniques how to size the weld at the bonded areas but when it comes to member capacity itself, what is the appropriate way to interpret the results?