massive hss truss design

huangyhg

massive hss truss design
i am currently doing a very preliminary sizing of an hss roof truss with the following info:
span = 136 ft
trib. width = 40 to 60 feet (we can space the trusses out according to whatever works for us in that range)
service dead load = 15 psf
service live load = 50 psf
just going through some very rough numbers i am getting into a 12-15 ft. deep hss truss, with 12"x12"x1/2" sections for the shallower depths.
now, assuming the webs are the same size as the chords, is it correct to assume that the joints in the truss act like pins, or would i need to assume that they take moments?(since the sizes of the members seem so large, my guess is that there are secondary effects that can't be neglected)
any guidance is much appreciated!
check out our whitepaper library.
the major assumption of truss analysis and design is that all member connections are pinned, (no moment is transferred). however, you will need to look at each   
actually, in this case i will have steel joists at 10 ft o/c bearing on the top chord of the truss, but i will also have a vertical web member at each joist bearing location to avoid bending in the top chord.  therefore i'm not too worried about failure of the chords in flexure.
my biggest concern is what is actually happening at the joints in terms of potential local failure, due to the large sizes of the   
the connections as you have described them certainly deserve a closer look. the internal moments and stresses caused by rotation can be signifcant, especially for members with a larger cross section. i had a similar truss design a while back and found knife plate connections to be the best suited for the tube   
hi
just wanted to add some other thoughts.  first, you said that your web members are the same size as the top/bottom chord.  why?  if you are doing a direct welded connection, it would be easier if the webs are smaller, so that you can do fillet welds without any additional edge preparation.  of course, don't go too small so that prying of the tube face is a problem.  but for 12" chord, i would use 8" or 10" webs.
second, unless you really want the look of direct welded connections, you might consider an alternative like gusset plates, as motorcity mentioned.  but with the gusset plate connections, you have to consider teh prying and tearing of the tube face at the plate connection.  the benefit of gusset plate is that the tubes can be square cut at ends, and fit up is not so critical.  you may want to check with some local steel fabricators to see what they prefer.  some don't like dealing with knife plates and tubes, while others love it.  also, if going with direct weld connection, you may want to go with eccentric connection so they don't have to make double miter cuts on the diagonal web members.  saves a bunch of fit up time and hassle.
third, just some comments regarding the moment at the joints.  if you are modeling this truss by computer, it would be very easy to model it both ways (pinned & fixed) and see for yourself the difference.  direct welded would definitely be close to fixed, particularly if the webs are same size as chords.  even gusset plate is more fixed than pinned.
fourth, how do you plan to brace the bottom chord for compression during uplift?  if that is a concern.  i have designed several long span (up to 250') truss roofs, and uplift compression of bottom chord almost always governs the bottom chord selection.  of course, these have been in florida, so my wind loads are higher, and i don't have snow loads.  if you don't brace the bottom chord, you will find that you need a much larger bottom chord   
hi
you might find this book helpful:
"hollow structural section connections and trusses
- a design guide"
  - j.a. packer & j.e. henderson
- canadian institute of steel construction - 1997.
design procedures for truss connections that i am aware of assume that bending moment is negligible. as long as the connections are ductile any moments that can't be sustained will be relieved by rotation.
richard c