ponding, length of a cantilevered member for Cp calc

[tkhartle]

I have an existing building with a spliced beam system supporting joists and metal deck. The bays are 40'x40'. The roof slope is 3/16" per foot. I am checking for ponding instability. The formula for Cp in AISC is Cp=32xLsxLp^4/10^7xIe. The splices are 5'6". For the beams that cantilever 5'-6", can the length of the beam (Lp) be taken as the distance between the inflection points of the moment instead of the full length of the beam? This would reduce Lp and help out greatly. Does anyone know of any documentation on this subject?

 

[haynewp]

There has been discussion on taking inflection points as braced locations here before, and it is not recommended. Also see research by Yura.