designing raduised steel beams

huangyhg

designing raduised steel beams
what are some thoughts on designing radiused steel beams? are there any rules of thumb that experienced engineers use? i have a condition where i have a radiused beam over the entrance to a four story bldg spanning about 20' between columns. it is about 12" eccentric from the cl of columns.
just to clarify, is it a curve in plan or a curve in elevation (i.e an arch)? how is it connected in the ends?
assuming it's curved in plan, and since it appears to be a single span, treat it as a simply supported beam (to determine the torsional moment only) and fasten the ends for this moment.  
for a 20' span, although the amount of curvature appears to be small, you might find that an hss or ts (depending on the local practice) is more efficient.  w sections are very poor for torsional resistance in general and the size increases substantially.  although the curvature is small, both sections will distort a bit.
dik
yes..it is a horizontal curve. i am planning on having full penetration connections at the columns.
with the oversized sections required for torsion, you may not require a full penetration weld... try for fillet welded end plates and hs bolts...
dik
when u say simply supported beam to determine torsional moment, r u suggesting wl^2/8 just like a straight beam? would u use the cl to cl dimension between the columns or use the arc length of horizontal beam in the equation?
i'd use the simply supported beam to obtain the reaction that the slab is causing and then calculate the eccentricity of the load to determine the torsional moment.  if you consider an axis through the reactions and determine the moment of the reaction about this axis.
dik