Steel Truss

[txeng91]

I’m looking at doing steel girder roof trusses to frame a 70’x200’ structure. The trusses will be spanning the 70’ dimension, probably 5’ or 6’ to 7’ in depth, and will be spaced at about 20’ on center. The configuration resembles a typical bar joist but with 2 HSS 12x6 top and bottom chords sandwhiching 5” or 6” HSS web members. There are ~10’ top chord cantilever extensions, and with a 1:12 roof slope it makes the truss a little under 100’ in total length. The trusses will be exposed so the plan is to have no bottom chord bracing. I am planning on putting enough concrete on the roof to negate any uplift forces. My 2 concerns are:

1) Are there any concerns with running a truss this long without any bottom chord bracing, even if the bottom chord never goes into compression?


2) Should I anticipate having to provide a splice detail due to shipping limitations? If so, should making the cantilevers separate pieces to limit the total truss length to about 76’ be considered?

 

[Lomarandil]

Yes, especially when cable sag is taken into consideration... I once calculated the effective stiffness of an inclined cable including the behavior of pulling sag out under load. It's very, very low.

On that topic.. how much "buckling movement" before a brace kicks in is too much (without getting into FEA analysis)? At some point you'll run into obviously problematic second-order effects... but is that the only limit?

----
The name is a long story -- just call me Lo.

 

[KootK]

The 2nd order effects were all that I had in mind. Once you draw out any slack and start reengaging the full stiffness of the cable, I suspect that stiffness gets in line in a hurry. I'm not sure that I know how to calculate that without making a thesis of it though. And the stakes are pretty high.

Interestingly, camber actually makes this system less stable as it raises the load above the potential point of LTB-ish rotation. Another way to see it is that it reduces the lever arm available to the bottom chord to use its self weight to stabilize.


I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.

 

[txeng91]

Hmm I see your point. I would say that items 1 & 4 could be resolved with proper analysis and detailing, assuming the numbers work out, while 2 & 3 rely on proper and diligent erection of the trusses. It’s design/bid/build and I know the types of contractors that get the specific job I am working on and I probably shouldn’t rely on them carrying out “advanced” construction techniques for the stability of the structure. That said, assuming I could extend the bottom chord to the columns and strength and stiffness calculations were satisfied, I feel like I could just as well design a cable bracing system that should be much stiffer and stronger than the bottom chords spanning, although I could be completely wrong, still need to verify with calculations. If elongation and some construction tolerance is considered and adequate pre-tension is verified before and after the dead load is applied is there really an issue if the components are adequately sized and detailed correctly for all loading conditions? My only question to you is would be if you think the same issues would arise with using another form of tension-only bracing such as rods or other slender rolled members.

Obviously, I need to quit assuming what’s architecturally acceltable and just sit down with the architect and workshop some ideas. I’m really just trying to get as many options as I can into my back pocket for when I do. When I’m starting from a blank slate like this I tend to get too caught up in what I think the strucuture should look like instead pushing what’s best structurally and letting the architect do their job.

Really appreciate the time and effort you put into your responses Koot. I can’t tell you how many of these threads pop up on google searches and you’re in the replies with some good insight and literature. Maybe you have to play the role of devils advocate/Debbie downer sometimes but it’s worth it if it saves some poor engineer in the future getting the call asking to clean up someone else’s mess.

 

[txeng91]

Sorry, last 2 replies came in while I was drafting mine, and I see your point that it might not be possible to justify cable bracing with analysis. While the initial stiffness can be calculated and designed per AISC guidelines, when you start considering 2nd order effects it gets hazy and would require some assumptions and the right software and quite a bit of time resesrching how to accurately model and analyze the system. I’ll plan on staying away from the cables for now but I’m still curious on anyone’s opinions with other forms of tension-only bracing.

 

[Lomarandil]

Other means of tension-only bracing (or at least the only other I can think of, rods) are fine. #1 and #3 of Koot's comments still apply, and #2 and #4 as well, just to a lesser extent.

Going back to my cable sag exercise once more, I don't recall the exact deflections I got before cable stiffness began to kick in (besides, absolute numbers wouldn't mean much without more context), but it was roughly an order of magnitude above the 1/8" or so that I could justify by inspection. In my case, the 2nd order effects quickly ruled that sort of movement out.

----
The name is a long story -- just call me Lo.

 

[KootK]

Thanks for the kind words dnlv. For what it's worth, I suspect that many engineers would be fine with the cable bracing. I might even get there myself were it my project and my client for whom I was striving to provide value. I really think that your path here may well be justifying the no-bracing option. Based on the direction that you seem to be taking with member profiles, I think that you'll have two things working in your favor here:

1) It sounds as though your top chord will have a high torsional stiffness. That helps as your failure mode is effectively LTB about the top chord.

2) It sounds as though your bottom chord will have a high lateral stiffness. That probably means that you can span unbraced between supporting columns or, at worst, bridging/bracing lines at the the far ends of the truss.

For the no bracing option, you're essentially counting on bottom chord weight to brace against LTB. A little off the wall but, if it doesn't check out, I wonder if one might concrete fill the bottom chords to improve matters. That's right where the weight would be most effective.

I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.

 

[XR250]

For the no bracing option, you're essentially counting on bottom chord weight to brace against LTB

Koot,

Wouldn't you have to get a fair amount of truss rotation before the weight of the bottom chord would provide meaningful restraint?

 

[KootK]

I like to think of these things in energy terms as I find it easier to make good decisions that way. Buckling in an LTB-ish mode doesn't happen unless:

1) Truss rotation sets up a condition where truss weak axis bending is involved in downwards deflection.

2) #1 results in the applied loads moving closer to the earth than they otherwise would, thus reducing potential energy (mathematical instability).

Wouldn't you have to get a fair amount of truss rotation before the weight of the bottom chord would provide meaningful restraint?

You would. But, by the same token, you would also need a fair bit of rotation before the the applied loads would start moving closer to the earth in a meaningful way. So I think that you have balance in that regard. By the time that you need the resistance, you probably have it. It's certainly worthy of some numerical exploration of course, starting from an appropriate, presumed out of plumbness.

I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.

 

[XR250]

You would. But, by the same token, you would also need a fair bit of rotation before the the applied loads would start moving closer to the earth in a meaningful way. So I think that you have balance in that regard. By the time that you need the resistance, you probably have it. It's certainly worthy of some numerical exploration of course, starting from an appropriate, presumed out of plumbness.

Makes sense. So based on this discussion, should we assume a no-heel gable (triangular) truss with is always stable under gravity loads as long as the top chord is sufficiently braced?

 

[KootK]

So based on this discussion, should we assume a no-heel gable (triangular) truss with is always stable under gravity loads as long as the top chord is sufficiently braced

That makes for an interesting example. And the answer isn't at all obvious to me. My take:

1) In theoretical terms, I don't believe that a gable truss would be perfectly stable in this respect by default. The compression webs still need to be kept from kicking out at the bottom chord.

2) You've got a bottom chord that extends to the bearings which is great for restraining the compression webs from kicking out.

3) You've got a situation where kicking out the compression webs and bowing the bottom chord would actually raise the peaks of the truss and thus move load further from the earth initially. This is obviously a pretty great situation from a stability perspective. It might be that the stability graph here would be one that has multiple peaks and valleys. While you're initially raising the load, you may eventually crest and descend into an unstable region beyond. I'd think that this would take pretty extreme amount of movement though.

4) My conception of this actually differs a bit from the presentation given in paper that I referenced. They deal pretty much exclusively with the kicking out of the bottom chord. I see that phenomenon as really a symptom of a form of truss LTB which the paper doesn't really speak to. Obviously, a gable truss with the top chord restrained is in pretty good shape with regard to anything that resembles LTB. By comparison, the classic "bad truss" in these discussions is always an upside down king post truss (it is in the paper). That's effectively your gable truss flipped upside down which reverses all the good stability stuff listed above. And I do recognize that, in my disagreeing with the paper, I'm putting myself up against Fisher, Yura, and Galambos. I acknowledge that my odds of being "righter" than that gang are pretty low. That said, this is democratic space and even dummies are allowed to express their wayward opinions.

5) For practical purposes, I do consider the top chord restrained gable truss to be inherently stable to the point that I normally wouldn't waste real design time looking into it. I know from experience that the wood truss guys certainly don't.

I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.

 

[XR250]

For practical purposes, I do consider the top chord restrained gable truss to be inherently stable to the point that I normally wouldn't waste real design time looking into it. I know from experience that the wood truss guys certainly don't.

I figured as much which is why i asked you in particular. Thanks for the comprehensive insight.

 

[TLHS]

Don't have a copy right here with me, but I'm pretty sure the truss section of CSA S16 (Canadian steel code) has a specific requirement for tension chord bracing to keep it below a certain slenderness limit.

 

[BAretired]

They do but it isn't mandatory. Their suggested limit is 300 for KL/r of a tension member.

BA

 

[TLHS]

There's a separate clause specifically for trusses. Just checked it:

15.2.7 Maximum slenderness ratio of tension chords
The maximum slenderness ratio shall be limited to 240, except when other means are provided to control flexibility, sag, vibration and slack in a manner commensurate with the service conditions of the structures.

 

[BAretired]

Okay, that is good to know. I wasn't aware of that.

BA