Real life truss failure! 8

[MIStructE_IRE]

I was called out to look at this truss which has failed. Its a Steel warren roof truss. The layout is shown below - but I’ve crossed out the vertical which doesn’t exist - which I think should have existed!

It looks to me as though the second last diagonal, the compression member buckled due to excessive compressive force, and as it buckled it pulled the last diagonal, which should be in tension, causing it to buckle also.

Any thoughts on the mechanism? There are hundreds of these trusses throughout the building and this one has failed.

Disclaimer - not my design!! I just got asked to review this failure!

 

[Tomfh]

Interesting failure. My guess is the additional load from the hangers has caused the final compression member to buckle, rotating the node and adjoining members.

Do you know the magnitude of the loads? If the truss has failed due to overload then are the other trusses safe?

 

[r13]

It wasn't caused by overload, but the heavy localized load may have contributed to the buckling - upward forces near the joist end meeting the confining force near the first hinger, resulting in excessive compression in the web members. The roof was confined by the heavy load, so it remains flat, evident by the unaffected top chord.

 

[BAretired]

It wasn't caused by overload, but the heavy localized load may have contributed to the buckling -

You may be right, but you do not know it wasn't caused by overload. Copper is a pretty heavy metal. It would be foolhardy to recommend a repair without determining the magnitude of the load from the "copper bar services".

BA

 

[MIStructE_IRE]

I will be doing analysis shortly, going into it with an open mind, and lets see how much compression was in that last compression diagonal.

 

[LittleInch]

The key for me is whether there is any deflection in that top beam.

If not then it got whacked from underneath at some point seems to be the consensus

If there is actually a visible or measurable dip or bend in the top T section then it could have been buckling under additional compressive force caused by the additional weight applied.

But the fact that the beam looks like it is surviving and hasn't apparently bent is very interesting. I can't see how the elements would bend that much without the top chord moving downwards and creating a hinge at the junction of the 2nd and third diagonal.

At the same time I would reinforce the attachment of those down tubes. They look like they are literally hanging on by their finger nails....

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[r13]

BA,

Whatever service was designed for, if overload, the top chord must deform too, and might not be just one place.

 

[kingnero]

Looks like uniform corrosion under that flaked paint.
If that is correct (can't really tell from those photos), it indicates that the damage has been done long time ago.
I have no idea, but would there be a way to determine when the damage occured using the corrosio as an indicator? I'm thinking wall thickness measurement at that point a e few inches further, under "good" paint?
Probably to small a difference to work with?

 

[IRstuff]

The thickness of the dust layer might also be a clue

TTFN (ta ta for now)
I can do absolutely anything. I'm an expert!

https://www.youtube.com/watch?v=BKorP55Aqvg

faq731-376 forum1529 Entire Forum list

http://www.eng-tips.com/forumlist.cfm

 

[oldestguy]

Back to what happened and fixing it. My take it is more likely to have been damaged before placement some how rather than a mistake with a man lift gong too far. Then for any repair with replacement, etc. Ya gotta lift it during repair so you will put a load on it for service after repaired. Jack it up against that roof planking and load? No way. Likely to just damage it elsewhere. It won't move. The least any repair will do is provide help against further loading. However cutting out the damaged parts, which now do carry some load, then you run into final repair leaving it more deflected than it is now. Needs some careful thought before doing anything. Might even be best to do nothing.

 

[rb1957]

who'd attach the out-of-plane stringer to a bent flange ?

At lest the fix is straight-forward … cut out the cancer, add new, paint, done.

another day in paradise, or is paradise one day closer ?

 

[oldestguy]

One possible fix that would take more loading, but will look pretty bad is as follows. Weld an angle at the truss end support and another at the lower chord, at the end. Leave all other parts as is. Between these two angles install (weld) two threaded rods, one on each angle, one left hand thread. On these threaded rods install a turnbuckle (before they are welded). Due to space limits, I'd make this turnbuckle from a short length of pipe and would weld nuts to each end, one of which is left hand thread. Turn to tighten and get a substanti0asl tension load. Tack weld so it can't turn or use a lock nut. Looks bad, but will work. A cheap fix. Might add some more metal to the buckled angle . If there is difficulty getting large enough threaded rod diameter, use two of them, side by side.

Edit: The so called easy replacement fix ain't so easy. Unload the bent diagonal to the support first? The fix won't be loaded so easy.

 

[LittleInch]

Well before this drops off the first page, please be sure and let us know what you find and what you intend to do to repair it.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[gtaw]

All the web members are single angles weld to one side of the top and bottom chords. Why couldn't new web members be welded to the opposite side of the damaged web members? Once the new web members are welded, the bent ones could be removed.

I would suggest checking the top chord member to determine if it is deflected downward from the roof load. It doesn't appear to be from the photographs, but you never know until you check it. Assuming the top chord is undamaged and there is no sag, you should be good to go.

It might be advisable to unload the open web joist to the extent possible, but it doesn't appear that the roof loads are that high considering the roof doesn’t appear to be deflected downward to the point where it is noticeable in the photograph.


Best regards - Al

 

[Joe591]

Was the building smaller at one point in time. It looks like the far wall might have been added in recently. The only times that I've ever seen damage like that in a factory has been when there was either a hyster loader or a backtipping truck involved. A backtipper truck can bend a thick I beam like paper if the truck driver is not paying attention to what he is doing. Hysters are very similar, but obviously not as powerful.

What other source of upward force could you possibly have. Maybe a skyjack or boomlift operator? Both can push upward, but how will you be able to tell for sure?

 

[KootK]

It looks to me as though the second last diagonal, the compression member buckled due to excessive compressive force, and as it buckled it pulled the last diagonal, which should be in tension, causing it to buckle also.

With every fiber of my being, I've wanted to believe that you did not, in fact, observe a unicorn. And by "unicorn" I mean a genuine, theoretically predicted, overload buckling event. I wanted it to be some pragmatic, pedestrian thing to do with forklifts or misbehaving construction workers. The usual. But, alas, I've come to the conclusion that your original explanation is my preferred explanation for what has actually taken place with this truss. Unicorn.

Below, I've attempted to tell a cohesive story of this truss failure as it might occur via overload. I haven't actually added anything to what has already been tabled above. What I have done, is attempt to put a bunch of it together into one cohesive story. I've also presented things as though they might have happened in sequence which is false. Obviously, all of this stuff would have happened, progressively, in unison. I just find it easier to conceptualize it this way.

The truss largely remain load supporting due to catenary action from the last tension angle and bottom cord.

I feel that aspect of things (yellow line below) is very important to the story of this failure and did not garner enough attention initially. I lean on it in what follows.

Everything below is also attached as a PDF in case others want to sketch over my stuff to show me how I'm wrong etc. Obviously, my theory is still just one theory of many. It's a pretty good story though... princesses and dragons.

 

[KootK]

It looks to me as though the second last diagonal, the compression member buckled due to excessive compressive force, and as it buckled it pulled the last diagonal, which should be in tension, causing it to buckle also.

I disagree with the bolded part a bit, at least semantically:

1) The last diagonal is in tension and, I suspect, always has been.

2) What you've described as the "buckling" of the last diagonal is really it having been rotated at the bottom in order to keep the second web from compression buckling to the side infinitely. So it is a member currently experiencing axial tension and end bending but not buckling.

3) The last diagonal, as a result of developing the canternary mechanism shown above, has likey had it's axial tension increased as it has effectively become a shallower tension web than it originally was. I'm not sure "caternary" is technically the right term here but human909 chose that so I rolled with it. I'm sure that we all get the gist of it: a thing that moved and resisted greater shear in having done so.

 

[BAretired]

Nicely presented, KootK. I think you have nailed it.

BA

 

[KootK]

Thanks for the vote of confidence BA. It'll be interesting to see what critiques inevitably come.

 

[3DDave]

My concern with this as an explanation is this - how much energy was absorbed in that amount of distortion and where did that energy come from? In typical buckling the energy comes from a change in vertical height of some amount of weight, but here no other part of the structure appears to have moved at all.

So how much energy was required and where did it come from?

This is why the energy supplied by a lift is more attractive an explanation.

 

[hokie66]

I agree with your analysis, KootK. I commented twice, once on each side of the argument, but hadn't focused at the time on the loading from those hangers.