I have been using simulation to model the stress on a welded round tube frame. I have created a surface model of my frame with the surface diameter corresponding to the average of my outer and inner diameter, meaning I set my shell thickness as the wall thickness of my tubing. To get the elements to line up at the joints, I have inserted split faces. On most of the joints, there are no stress singularities. However on some of the joints there a single elements reporting extremely high stresses. Finer mesh often just makes this stress singularity worse, is there anything I can do to improve my model for these welded joints?
I have also attached my model, I have been using the simulation called "test point 3" The tubing is all 1.25" od with 0.083" wall, the three large joints on the center are separate machined components. In an effort to simulate the weld around these joints. I modeled a slight chamfer around the faces that interact with the these joints. I've been playing around with the model, and the version I uploaded shows a lower max stress, however it still exhibits the singularities around some of the joints, particularity where the max stress is notated.