seismic embed plate design for gravity beams

huangyhg

seismic embed plate design for gravity beams
occupancy category iv, ss=0.89, s1=0.31, 2006 ibc, aci 318-05, pci dhb 6e, v=145mph, puerto rico, f'c= 5 ksi, 8" cip walls
i'm having some trouble meeting the ductility requirement at an edge condition for embed plates in a concrete wall supporting steel gravity beams.  in a field condition, no problem.  even using the 2.5 times the load exception, a few connections are still giving me problems while others are then ok.  i'm checking both the aci appendix d method and the pci method (fun).
i'm thinking instead to use tail bars, as shown in an example in the pci dhb 6e page 6-21.  my question is in doing so, since the bars are supporting the reaction in tension, would you consider this to be a ductile failure mode so that i don't need to design the bars for 2.5 times the beam reaction?  how do others usually handle the ductility requirement vs the increased load?
find a job or post a job opening
if you add the tail bars, doesnt the steel fail before the concrete, thus a ductile failure. if you just use headed studs, and its a brittle failure (i.e. concrete breakout) can you add some rebar for shear friction to beef up the connection and then have the headed stud steel strength control. i have done this.
make sure if you are designing from appendix d, to follow d3.3.3 if need be, thats another 25% reduction, then also making sure you are ductile. the other thing to do is to reduce the number of studs, which brings down your steel strength of the plate, and maybe then will control. (this worked alot in 2003 ibc where you didnt have the 2.5 factor and call it good)
let me know if my thinking is way off
i don't understand the additional reduction on the steel strength.  i know it's in there, but can someone explain it?  if you reduce the steel strength artificially (the additional phi factor), isn't there a greater chance that the actual steel capacity would exceed the concrete capacity, thus creating a non-ductile condition?
i understand that the concrete is being bumped up by 2.5,and 1/0.75 is only 1.33, but it still doesn't make intuitive sense to me.
the 0.75 is just like another phi factor. its 0.75(phi)nn or 0.75(phi)vn
ucfse - yes i would consider the tail bar yielding as a ductile failure mode so they can be designed for the actual load.  precast industry uses this method a good bit...
i've thought about the shear friction bars also.  i should be able to develop them to either side with a hook.  yes, i have the additional 0.75 factor in there.  i think it makes sense to say the rebar is designed for 100% of the load, so the rebar then provides a ductile failure mode.  i guess i'm looking for a reality check. between the 2.5 increase in load and 25% reduction in in strength, i'm having issues with a few beams.  i'm planning to size the plate as normal, sizing the studs for the reaction, but also use the rebar with 100% load in order to provide a ductile connection where i can't manage the requirements near an edge.
seit, i haven't found an explanation for the factor either, after checking aci, pca notes and the pci.  i took it as additional safety for high seismic and important facilities but that's an assumption.  in my case for an essential facility, i think it makes sense, but i don't think it would kill them to provide some explanation of where the seemingly arbitrary number comes from.  when comparing failure modes, i compared concrete and steel strength without any reduction factors.  i think though if you multiply the numbers by the same reduction, then one that was bigger before reduction will still be bigger after.
the 0.75 is intended to reduce the steel strength for low-cycle fatigue under reversed cyclic loading (seismic events).  
lots of tests have been performed with various results on this...for example see:
the additional phi factor applies only to non-ductile, concrete failure modes (i.e. concrete break-out, etc.), it was a mistake on aci's part to include the factor for steel failure modes they have corrected the error in aci 318-08, as you pointed out multiplying the phi by both strengths cancels the intent of the phi factor.  using supplementary reinforcement kicks you out of appendix d, which is what i do very often.  also the idea behind the phi factor is to ensure ductile failure of the steel.
ucfse - this may be a stupid question but with the tail bar, i could see the ductile failure of the bar itself being ductile.  but what about the connection of the bar to the plate?  (am i seeing this right? - i don't have the pci example in front of me).
for a tail bar that is welded to a plate, wouldn't the bar-to-plate connection be non-ductile?
just askin'.

jae - bar to plate connection should be sufficient to develop the ultimate tensile strength of the bar (just like a stud weld) so the failure mode would be pushed into bar yielding (ductile).   
jae, willis beat me to it.  i sized the weld for the ultimate strength of the bar instead of the yield to push the bar to yield before the weld fails.