EngrRC:
If you really want to transmit less moment to the concrete support structure, you will have to increase the stiffness of your RHS steel beam, so that it tends to induce less rotation at it end connections. And, you do not do that by trying to fiddle with the end connection welding details. In the extreme, substitute a 36WF for your light RHS. You will have essentially zero beam deflection and end rotation, so you just won’t be transmitting much potential fixed end moment, just gravity load shear and torsion which are fairly uniform You are either going to pull the top bolts out of the conc.; or the end pls. might be light enough (flexible enough) to bend enough btwn. the bolt line and the top (tips) of the side welds; or you are going to start to unzip the side welds at their top tips. The stresses in the last few inches of those welds will be very high when you add any moment component to the unit length average stresses from gravity load shear, torsion, and axial beam tension. The shear and torsion will cause fairly uniform stresses over the lengths of those welds, but the beam axial tension will cause some end pl. bending and prying on the welds and particularly the weld tips. Then you add the fixed end moment component potential and some end pl. bending and prying on the welds, again, particularly the weld tips, and you have a very high combined stress at the weld ends (tips). Then, add to that, that the weld ends with their starts and stops are the most critical (QC sensitive) areas on the whole weld, and that leads to a really nasty weld condition in your detail. As a means of thinking about and studying those kinds of details, look at how the various components tend to flex, rotate, deflect, etc., and wherever you get the max. interactions of these movements is also likely a very high stress area, and a likely failure initiation location.