calculating rho prime for long-term deflections

huangyhg

calculating rho prime for long-term deflections
this might sound like a silly question, but i have to ask.  the question is which b to use when calc'ing rho prime in order to get the long term multiplier for long term concrete deflections.
let's say you have a concrete t-beam with a flange width of 120" and a stem width of 36".  you will get a significantly different rho prime depending which b you use for as/(bd).
i have always used the flange width, but have never had such a big disparity before and needed so much compression steel to cut down the long term deflections so this is the first time i've thought this much about it.
check out our whitepaper library.
i would go with flange width as well.  b it typicaly defined as the width of the compression face, which will primarily be the flange of the t in this case.   
i have to think that if you used 120" to compute short term deflections, then you need to use 120" for ρ'.  however, if using b=36" for short and long term deflections results in a lower long term deflection, then i'd use 36" for both.  are you considering the longitudinal slab steel in your as'?   
ρ' is based on bd where b is the compression flange and this is taken at the midspan (by definitions from aci 318).
so you'd use the 120" and the area of compression steel at midspan.