Who says "we" don't consider self weight in mechanical designs? If its appropriate to my analysis, I include self weight. if its not, I dont. My guess is the rest of we do the same.
Ansys doesn't include self weight by default because it is a very general purpose program and geometry can be imported with random orientation. Im guessing that if it decided that global Y was up, and a part designed with global X as up was imported, an improper load would be applied that might not be obvious to newbie or casual users. I think their approach is they supply the model, but the user builds the model and runs the analysis. In any event, its easy to do with the ACEL command in MAPDL, and Workbench has a gravity load (which requires you to pick the up direction), so really its no big deal. I have no experience with Staadpro, but it looks like a dedicated civil engineering type structural package, so perhaps in that context it makes sense to include self weight by default. But the reality is that gravity is largely irrelevant to the analysis of all kinds of stuff. Is it more correct to include gravity? Yup. Is it always relevant? Nope. When you add in discretization error, variable material properties, etc, the effect of gravity gets lost in the noise.
Rick Fischer
Principal Engineer
Argonne National Laboratory