I agree with Nige and others, just take off the sharp edges and leave it alone. Welding creates a stress concentration at the toe of the weld and "heat treats' the metal, so you can easily make it worse. It is not near any welds or joints so it should be fine. If it was a critical location they wouldn't have put a light bracket there.
"So it's the tops and bottoms (or in a corner radius?) were all the stresses can make a gouge crack out?"
I work on structures along with designers...and lets just say it can get complicated. All metals deform under load and can fatigue when subjected to pure tension, pure compression, or a combination of both. Any change in cross sectional area along a load path creates a stress riser. The question is that stress riser large enough under the loading to fatigue the material and start a crack? Depends on the loading per mm², base metal, severity of section change, angle relative to load path, aggravating factors such as weld beads and bolt holes, random overload events, etc. Then for good measure throw in multiple simultaneous loads.
(That's why computer modeling has had such a dramatic effect on structures over the last 25yrs.)
The idea of slowly changing the cross section is why you see radii in place of sharp corners, why doubling plates are often fish-mouth or diamond shaped. On the negative side blending costs money because they are harder to manufacture. So it comes down to a balance, what do I need to do to make this machine live X hours for most of our customers. (cant design for worst case customer, it would cost too much and machines wouldn't sell) On the mining equipment I work with we have some critical welds that must be hand ground and blended - a huge amount of labor but a very significant increase in life. Little sensors 5-10mm long can be attached to the parts to measure the actual deformation while in use to confirm the computer analysis.
In this application the most critical part is almost certainly the top plate in tension, especially between the cylinder anchors on the corners. Hard to see from the pictures exactly how that joint is constructed but it appears to be a relatively large weld bead. My guess is they were trying to get the weld toe farther down away from the corner.
This just scratches the surface. Hundreds of people have written very large books on this subject and many more make a career out of this. (Airplanes get real interesting - stressed skin, over 1 million rivets with corresponding holes, ~400 lives, big lawsuits potential)
ISZ
