You may have to sketch out your sphere first in AutoCAD to see how your nozzles are coordinated, and then figure out where the sphere itself is supported. Have a rigid element going up from the support to the height of your nozzle, and out to the nozzle-shell interface. You will have some funky angles, but if you put nozzle flexes in, and simulate the thermal growth of the spherical vessel itself with rigids, then I would think you would have a pretty good model.