Bottom chord buckling of a steel truss due to compression from wind uplift

[nivoo_boss]

Hey everyone!

This is sort of a different thread, since I'm not asking advice on any particular problem but would just like to hear some general opinions.

It's about steel roof trusses that develop compression in the bottom chord due to uplift from wind. The out-of-plane buckling length is pretty much the same as the span.

The thing is, I cannot really imagine how a truss would fail to this. Let's say the bottom chord buckles out-of-plane due to extreme wind for a second - what then? The wind is probably not constant, the chord would buckle for a second and fall back down and the chord would likely be fine.

Do you guys know of any real failures due to this same reason? How would you see a truss failing due to this?

 

[GeorgeTheCivilEngineer]

You'd probably reach serviceability failure, leading to misaligned and peeling cladding, before a ULS buckling failure. If the buckling is plastic then you've got permanent deformation causing changes int he load path and capacity of the sections (not least the truss connections).

I can't see any obvious failures in my disaster book(s). But there will be reserves of strength to prevent complete collapse but can still make the roof a failure.

 

[phamENG]

It will also depend on the type of buckling. If it remains entirely elastic, sure, it'll spring back. BUT - if it goes inelastic you'll have permanent deformations that will adversely impact the truss. And, as George mentioned, the connections are a big issue. As the bottom chord buckles and the top chord is held in place by pulins or sheathing, those connections (which are generally designed for pure axial or some in plane moment) will now suddenly experience a bunch of out of plane moment. If those fail, it won't matter if the truss springs back to normal.

 

[canwesteng]

Depending on how much it buckled and the proportions of the truss, the global stability of the truss could be impacted, and instead of snapping back to straight the truss could just roll over.

 

[jayrod12]

Truss/Joist suppliers have bottom chord bridging and bracing for specifically that reason, to shorten the bottom chord buckling length. Cheap and easy solution.

On large trusses, the chord are either chosen accordingly, or I believe the webs are often looked at as cantilevering bracing members.

 

[nivoo_boss]

Truss/Joist suppliers have bottom chord bridging and bracing for specifically that reason, to shorten the bottom chord buckling length. Cheap and easy solution.

On large trusses, the chord are either chosen accordingly, or I believe the webs are often looked at as cantilevering bracing members.

It's not like I haven't designed bracings for bottom chords, I know how to overcome the issue I explained. It's just the failure due to this that's a bit hard to imagine. But people here are making good points.

 

[jerseyshore]

There is a building where I grew up that has the first 3 or 4 bar joists near the exterior wall with buckled bottom chords. I'll try to see if I can find the photos. My first thought was that something hit them, but they are high up in a gym and lifts are rarely used (maybe just to change light bulbs) and it would seem a little difficult to smash into multiple joists like that. It's the only time I've ever seen anything like that in a bar joist framed building and my best guess for why they are distressed is some sort of wind uplift issue since it's right near an edge.

 

[JoshPlumSE]

The thing is, I cannot really imagine how a truss would fail to this. Let's say the bottom chord buckles out-of-plane due to extreme wind for a second - what then? The wind is probably not constant, the chord would buckle for a second and fall back down and the chord would likely be fine.

I was always told that this was pretty common for extreme wind events. Particularly with a warehouse roof. But, most of the failure pictures I have seen are not TRUSSES, but rather purlins or girts that are continuously braces by the roof deck at the top flange, but unbraced at the bottom flange. My guess is that you do have cases where the trusses are damaged from wind uplift. But, not the dramatic failures that you see with the cold formed Z, or channel beams in a roof.

 

[GC_Hopi]

You can find quite a few joist failure on older building. Sometimes it can be misleading because the joists will remain in place and the metal deck is missing. My understanding is the joist buckles, loses stiffness, then the deck has to span twice the length so the connections fail like a zipper. Bye bye to the deck but the joist can remain.
Joist failure

 

[GC_Hopi]

The photos are better in this one. Joist Failure

 

[nivoo_boss]

Yes, this one article I found myself as well, right after I posted here.

 

[SandwichEngine]

I don't know how much this will add to the discussion but I noticed a lot of what I assume are Victorian era structures in Paris have trusses that just have cables for the bottom chords so zero ability to resist compression. I'm thinking that if your main roof structure is heavy enough, you can ignore uplift compression forces on the bottom chord.

 

[phamENG]

SandwichEngine - I think the key to your pictures is that they can't resist compression even if they wanted to and, therefore, cannot buckle. That load path simply doesn't exist. Where it does exist and it is stiffer than the top chord acting as a beam on its own, the buckling concern is going to be valid.

 

[Tomfh]

I think the key to your pictures is that they can't resist compression even if they wanted to and, therefore, cannot buckle

Rods buckle just fine. They’re just as stiff in compression, but they buckle before any significant proportion of the compressive stiffness is mobilised.

 

[human909]

The second photo isn't a truss. It is an compression arch with ties to resolve the outward forces from gravity. Being curved negative with loads would be entirely different.

The first photo isn't clear entirely what is happening. If the diagonals are tensile members then the members you are referring to are for upwards forces! But these could just be ties to stop outward forces too. But I can't tell from the limited information.

 

[Tomfh]

The second photo isn't a truss. It is an compression arch with ties to resolve the outward forces from gravity. Being curved negative with loads would be entirely different.

It’s a similar scenario to trusses though. Transient compressive loads in unrestrained bottom part.

We’re dealing with one of these tied arches at present. 100+ years old. The rods are a pair of angles. The ties don’t work in theory against the uplift. It’s a tricky one. Do do the arches lift a bit, and then flop back down again, like those trusses in the photos above? Does it actually fail?

What the buffer between the bottom chord buckling and the structure actually failing?

 

[Denial]

A few years ago I wrote a spreadsheet to investigate the "lateral response of a sinusoidally-imperfect Euler column under the action of an axial compression load that is applied for a short (but not instantaneous) duration".

It uses the following general assumptions:
»[ ] There is no damping.
»[ ] Axial deformations can be neglected entirely.[ ] This includes neglecting the effects of any axial "shock waves" travelling down and up the column.
»[ ] Axial inertia is ignored.
»[ ] There are no localised effects in the vicinity of the point of application of the axial load.
»[ ] Material behaviour remains linear-elastic throughout.
»[ ] The cross-section remains undistorted at all times and in all places.

And the following specific assumptions:
»[ ] All the work done by the applied axial load as it forces the top of the column downwards manifests itself as bending strain energy in the column and lateral kinetic energy in the column.
»[ ] The initial "imperfect" shape of the column is a sinusoidal "half-wave".
»[ ] The deflected shape is also a sinusoidal half-wave, but with a larger amplitude.[ ] This is probably the most tenuous assumption, particularly for a column with only a small imperfection.[ ] An axially directed force does not "naturally" induce a lateral response, and any lateral response it does induce will probably include significant contributions from the higher frequency mode shapes.

This spreadsheet is available for downloading from my website (rmniall.com).[ ] It might or might not be applicable to your problem, which probably involves an "ultra-slender" member.


[sub][ ]—————————————————————————————————[/sub]
[sup]Engineering mathematician/analyst.[ ] See my profile for more details.[/sup]