a new one on me

huangyhg

a new one on me
have a new project - 47 individual three story townhouse  structures in a complex - with three to six townhouses per structure.  
for fire separation reasons, the floor diaphragm between the units must be discontinuous, but the roof diaphragm is continuous.  this will leave a 1/2" to 1" full height separation in the party walls from the foundation to the underside of the roof diaphragm.
properly applying the wind forces in this condition i have no problem with.  my problem comes in properly applying the seismic forces considering the link of the roof diaphragm.  it seems to me that due to that link, more force than normal will be thrown to the roof diaphragm than just analyzing each townhouse separately, because they could deflect differentially, causing hammering without the roof diaphragm link.  this implies a need for stiffening at the roof diaphragm level to prevent hammering.
my question is if anyone has run into this condition before, and how did you quantify and detail the seismic link problem at the roof?   
mike mccann
mmc engineering

perhaps accept that there will be some local damage to the roof during an earthquake and detail so the rest of the structure isn't affected.
we had a similar narrow 3-story townhouse design. it was seismic design cat b and we were having a difficulty even with the wind loads on the the front and rear shear walls (19' wide, lots of windows.) we got a variance from the code official allowing us to connect all the diaphragms for lateral load only. we used tnemec paint on flat steel straps to penetrate the fire separation walls. although 6 units were built together, we only connected them in pairs.   
irhg:  what is the significance of tnemec paint?  i didn't know it was a fire retardant (intumescent) paint, i've only seen it specified as a primer paint for structural steel.
this is interesting.
i have not designed this type of structure in the us, but
the australian codes specifically allow for the floors of these types of structures to be connected together as long as there is solid timber between the fire rated walls above and below.
if you could justify that the charring width of a solid piece of timber gave equivalent fire resistance to the walls above and below, could you then apply for a variance based on that?
of course there may be associated sound transmission issues.
csd:
it's interesting to me too in that it seems like the structural integrity of the building is being trumped by the fire requirements.  
i've asked for the specific code section requiring this disconnect to see if it is interpreted properly, but, in lieu of that, i'm wondering how to analyze it.
could use the computer, i guess, but was looking for a hand method to keep the analysis method simple.   
mike mccann
mmc engineering
do you know approximately the distance required to prevent pounding? (in the latest asce, i believe now you have to add the deflections of each building directly after magnifying each by cd.) is this wood or steel?
  
can you link the floor diaphragms at the front and back walls?  
i would think the intention of not linking the floors is to prevent progressive collapse due to thermal/fire failure in any one unit from spreading to adjacent units (or to the party wall).
but, if you make a connection at the front and rear walls, the corners can be self braced enough to prevent this progressive collapse.
if this works, the diaphragm span from front to rear wall should have a small amplified deflection that might be less than the 1" provided.
be careful with steel straps from unit to unit.  there was a project recently where those straps were blamed for carrying the sound from one unit to the next.  apparently it was as if you could here the people next door as if they were standing next to you.
for the strap thru the wall, the arch speced the paint to meet the fire rating requirements. i am not sure what type of tnemec paint met the requirement.
we modelled the variance request from code language in oregon:
haynewp:
apparently, according to the local jurisdiction, i will have to live with the floor diaphragms being cut.  they are all wood structures, with steel w shape frames at the lower garage door end.  i guess this is a requirement for condos now.  
most structural engineers are just ignoring the continuity of the roof diaphragm from what i understand.  i guess there will be a lot of similar struuctures out there with roof damage after our next big one here.  not my idea of good engineering.
i do not know the deflection yet at the roof if it is cut too, but i am leaning that way, essentially pressing for a flexible seismic joint at that location and increasing the separation up to 2 or 3", whatever it calcs out to be.  i will be forced to design each townhouse individually for wind and seismic.  the architect is going to love to read this flexible joint going up the roofline.  oh well...  what do i know...   
thanks for the comments guys.   
mike mccann
mmc engineering