hss seismic connections

huangyhg

hss seismic connections
we've got a project that is a small roof (open building) - six columns holding up a perimeter beam that supports light gage steel trusses.  no walls at all.  way out in western us with high seismic.
the aisc seismic provisions for ordinary moment frames do not allow fillet welds for tensile forces and don't seem to deal with hss shapes at all....only wf shapes.  
the governing code is ibc 2000 which references the aisc seismic provisions with suppl. no. 1.  fema 350 doesn't deal with hss either it seems.
we want to use hss simply because this is an exterior condition and want to minimize irregular surfaces and avoid bird nests.  
so is there a prescribed or pre-qualified tube column to tube beam moment connection that can be used?  or do we have to use wf's?  
find a job or post a job opening
jae,
i do not know of any source as of yet that allows the use of hss moment connections.  personally, i think it should be allowed as long as the connection can take the stresses and be within drift limits and that appropriate testing has been performed.
this is a link to a discussion for whatever its worth...
jae,
the discussion will make sense if you read up from the first posting way at the bottom.
whyun - thanks a lot.
the narrative you referenced seems to imply that a lower r (=3) might be used and allow skipping the aisc provisions.  however, the ibc 2000 only specifically allows that for sdc a, b and c and not for the higher areas.
it seems like i've seen all sorts of tube-type frames in the magazines and such - just can't believe that they dont' use them out west at all.
for single story open structures, often the columns are designed as an inverted pendulum or a cantilevered column structure to avoid the moment connection issue.  it results in larger size columns but much less headache...
can you use flare-bevel welds? for "back-up bar", you can show a short piece of fitted hss just inside the main hss.
similarly, if you cut a long slot on all 4 faces, centered on the hss width, fillet welds to an inner fitted hss would be in shear, not tension.
for bevel welds, the backup bar must be removed "on the bottom flange" according to aisc...but granted, this is with the wf connection...at least it seems to be referring to it that way...and for a tube connection you cannot get to the back up bar as its inside the tube.
the long slots are a good idea, but doesn't this still beg the question as to whether the connections must be tested?
the cantilever option seems like a possibility as that then makes the beam to column connections not part of the lateral seismic resisting system and therefore we can just connect them as we wish as long as we meet deflection compatibility requirements.  with this in mind, we would analyze the system as cantilevered columns with pinned beams.  then after designing the columns and footings based on this, we could then fix the beams, deflect the frame by and amount of δ (derived from the pinned condition) and design the beam to column connection to simply resist the shears and moments that result from this specified frame deflection.  does this make sense?
if the post-beam ( short length of beam) connection is shop-welded, producing "half trees", and you provide link beams with end plate connections (the link beam is of length to put its end at the "dominant" moment inflection points), would that make sense? i did something like that in 1982 in the phoenix area, and it got a write-up in modern steel construction magazine. it wasn't a seismic issue - it was a constructibility issue. combined with fixed column base, and if your actual stresses are kept low, that could be justification.
jae, i agree with your approach on cantilever column design.  i would design a single column taking the worst case lateral load at the tip.  use the same for the other five.
assuming wide flange beams, i would rest the beams on a column cap plate (bolted to the bottom flange of wf beams).  this way, you don't even have to consider secondary moment in the beam to column joint...
one of the biggest preferences here (by the owner) is to use tube beams to avoid bird nesting ledges, etc.  so tubes are what we really would like to do.  i guess we could somehow connect the tube to column cap plates, allowing a looser connection of sorts.
sarebleu - that's a great idea - i really like it. but this project just isn't that big for that application and we're also trying to keep the top flange surface flat throughout to allow simple bearings for the metal roof trusses - just a covering for a fuel delivery pipe system.  ...also...you've got one great handle.  
jae - i was going to suggest the r=3 also, but i guess that's not an option!
i've used the cantilever design successfully in the past (interior stand alone mezzanines, for example).  the lateral loads are generally light for small structures like that, and you may find that your typical base plate, anchor rods, piers and footings are still ok for the moments.