channel-braced flange

huangyhg

channel-braced flange?
say for instance you have 2 channels, running side by side 2 feet apart. the channels are connected at 3 feet on center with ts members welded across the top of each flange, only grating is sitting on top of the ts members, (like on a catwalk).
i would not consider the ts memebers as brace points for lateral torsional buckling since i think that both channels could buckle in the same mode.  
however i have found disagreement among colleagues and am wondering what your opinion is.
check out our whitepaper library.
haynewp - that is a good question. i consider these applications to be braced for the following reason:
assume that any meaningful load that the channels must carry is applied to the grating. then friction (along with any minor anchorage) between the the grating and the ts   
i would consider the top flange braced at the weld points. especially considering the rigidity of the grating. is the grating welded down too? or mechanically fastened? if the stresses are based on the channel support alone rather than on the composite section, the structure should perform better than predicted.
is the catwalk simply supported, or do you have continuity over supports? you might have an issue with the bottom flange buckling.
the channels are oriented in opposite directions, like this  ) (
grating is mechanically fastened, resting above the ts members not in contact with the channels. only welds at the channel flanges are for the horizontal ts   
i guess i don't agree.  grating has very uncertain amounts of diaphragm stiffness.  also, the twin channels can bow/translate sideways with nothing to prevent this.  so your unbraced length would be the full span length.
but you can use the combination of the two channel's properties to get the capacity.  not as a combined section property, but as the sum of the two parts.
so if you are in asd, your allowable stress would be based on the summation of the rt's or whatever goes into the allowable stress formulae.  so you combined rt would be based on
sqrt((i1 + i2)/(a1 + a2))
where i1 and i2 are each channel's moments of inertia and the a1 and a2 values represent each channel's area.
the grating is usually a square or rectangular arrangement and i would be hesitent to use it as a diaphram bracing these   
jae,
this is interesting as i now see there is disagreement here.  when my peers and i design bridges for the construction stage with no composite deck or rebar, there is no way we could use the full span as you mention.  we place lateral bracing against compression flange in the compression zones and thus use an unsupported length equal to the spacing of these braces.
regards
vod
i have another case where the grating is screwed into to the top flange of the channels and there are no ts members or welds. the grating is segmented. i can see where this would kind of act like as a wood floor diaphragm, the 2-d panels resisting deformation along the length of the members.  

i originally didn't have the grating sitting above the ts members in mind, maybe this justifies the ts points as being brace points.
however another case,
looking at the horizontals that are placed for joist bottom chord bracing for wind uplift for instance, i think there are usually enough members tied together that the chance of them all buckling together in the same mode is very slim.
what if there were no grating and only horizontal ts   
i would agree - braces need to have a load path just like everything else.  in your case of two channels running parallel with ts braces at intervals between, all the ts's are doing is tying the two channels together.  so my comments above still apply - that the unbraced length is the full span but you can use the sum of the properties of the two channels in determining the capacity, vs. using each channel by themselves.
jae i don't agree. consider the behavior of a verendiel truss. consider how much force must be developed to provide the restraint. you are considering that the connections are theoretically pinned. and the out of plane stiffness of the bracing elements is zero.
i've seen attibutable assumptions being peculiar to the client. ie, power plant work getting the most conservative assumptions. i which case, diagonal angle bracing was added to the traverse bracing below the level of grating within the depth of the main   
perhaps the differing views on the capability of grating comes from the many types of grating available. for example when grating is mentioned, i think of "industrial strength" such as 1.25" x 3/16" 19-w-4, good for a udl of 716 psf for a 3 ft clear span. this size is quit rigid. others engineers will certainly have their own frame of reference, just as valid.
considering the possibility of friction providing lateral restraint, how does this sound?
the coefficient of friction could likely be anywhere between say 0.05 to 0.35 depending on the following:
1. grating (say plain, painted, or galvanized)
my preference is never to rely on friction to provide structural ties. i practiced in california, so i was accustomed to justification for this. grating is associated to me with industrial applications and vibrations which may tend to undermine a friction connection.
even nominal tack welding of grating seems unreliable with temperature affects and vibrations naturally working to undermine that connection.
so i specifically engineer it and detail the welds or connections.
i could understand where someone might doubt the rigidity of swaged grating. but only in comparison to the welded types. then the camparison becomes a matter of relative rigidity. so in all fairness, consider the rigidity of metal decking, or plywood sheathing, even gypsum wall board is considered effective to stabilize the flanges of steel