Overhead sign crushes car on freeway in Melbourne Australia. 3

[MDEAus]

This happened a few days ago, but the article has just been updated with dash-cam footage from the car immediately behind the one involved.
very quickly it can be seen that the failure point is at the welded joint for the bolted connection. the grainy footage makes things harder to see, the base off the standoff looks weird, why would you close the end off the SHS to be welded to the main beam? Galvanising requires drainage points and a minimum area open between closed sections. Unless that is a solid block off steel? I'm not sure why they would have used a solid block off steel though.

https://www.abc.net.au/news/2019-01-08/tullamarine-freeway-sign-falls-and-crushes-car/10700470

 

[EdStainless]

I think that hokie hit the nail on the head.
One party designed the gantry, another the sign (though it looks like a very standard sign and mount).
No one designed the integration of them.
At the least there should have been internal reinforcements in the gantry where the stools attached.
Perhaps both ribs and a heavier plate section.
But that would take planning ahead......

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, Plymouth Tube

 

[BridgeSmith]

The holes in the gantry beam would seem to indicate a failure of the base metal at the toe of the welds around the structural tube base. For a sign in that position, fatigue from truck-induced wind gusts could easily be the culprit. I think it would be classified as Category E or E' fatigue detail. Depending on how it was welded, at best, it would be adequate for about 7ksi with the typical number of cycles for truck-induced gusts. At worst, it could fail at anything over 2.6ksi under what we call natural wind gust, which is the gust pressure it would see several times per day.

 

[MDEAus]

The gantry does look like it was made up from plate sections.
At this point we know one thing for certain the joint connecting the stool/flanged stub post to the gantry member.
The videos and still frames shows a few things.
1) The flanged stub posts remain bolted to the sign at all times.
2) There are two holes left in the gantry while that material stays intact with the flanged stub posts.
3) Cracks around the area that the stub posts where welded to the gantry member.

Number 2 can rule out that the welded joint wasn't completed (the weld bead was stronger than the base material). There are a few scenarios to fit the video evidence which suggests a fatigue failure. Below is what I think happened that lead to the failure.

1) An undercut weld at the stub post/gantry connection on the gantry could have caused a stress riser.
2) Insufficient material strength for the gantry(lack of reinforcement, wrong material used or spec'd).

In my opinion the design safety factor for gantry material should have been enough to cover some welding defect like an undercut. However as others have noted this is likely a case of the designer of the gantry not knowing that the sign is going to be attached there at all and as such not being able to provide adequate strength.

 

[jhardy1]

It has been reported that the sign is a new sign, installed after some significant roadworks - I wonder if the gantry is an older structure, and the new sign was attached to an existing gantry as part of the recent roadworks project?

I would have expected that for a new-build design, you would have some internal stiffening under the top flange at the locations of the stubs, to transfer the base moment into the webs of the main gantry beam. However, the apparent failures of the top flange suggest that there was no internal stiffening, so the base moments are being resisted solely by flexure of the top flange plate. It is not hard to imagine fatigue of the fillet weld at the base of the square hollow section stubs if the sign is subject to fluttering / buffeting.

It is also possible that there might have been some sort of resonance between the torsional mode of the main gantry box beam and the sway mode of the cantilevered sign, which might amplify the moment at the base of the stubs (and increase susceptibility to fatigue of the fillet welds around the base of the stubs).

http://julianh72.blogspot.com

 

[MDEAus]

It has been reported that the sign is a new sign

I've seen it reported that the sign and gantry were installed in late 2017 at the same time(as part of a new build after roadworks). So to me there absolutely should be internal stiffening of the gantry; which admittedly could be obscured by shadows, but evidence points to it's absence.

 

[human909]

Agree with most of what has been said by others. A thin gantry material combined with what was likely poor welding. Just poor detailing. All that was needed would be a piece of flat plate welded onto the gantry. That would give it a larger weld circumference, reduced load on the weld and parent metals and an all round stiffer connection.

There has been suggestions (unsure if confirmed) that it was fabricated overseas. Which has much more variable and usually poorer quality work than local fabrication. (We have had some good stuff done, but the last load of 'food grade' stainless steel work we had done needed weeks of re-work.)

 

[IRstuff]

Hokie, I went by a previous news report that stated that the sign was bolted and welded. What you said makes more sense, but that was a mickey-mouse design, even to my eyes. The plate should have been welded directly to the gantry, and there should have been studs where the bolts went. That's how street light columns are installed, and those seem to work pretty well.

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

 

[hokie66]

No studs on street light columns where I am. They use bolts out of the footing.

 

[bugbus]

I've done proof engineering for similar sign gantries in the past. It's worth mentioning that the fatigue requirements in AS5100-2017 for highway gantries refers to the 2015 AASHTO code, which doesn't explicitly disallow square/rectangular hollow sections for these structures, but requires some pretty onerous FE modelling of the weld details to demonstrate that the connections are suitable (the FE model - which must be solid by the way - has to be so refined as to provide several finite elements just along the toe of an individual weld - fractions of a mm. Simple welded connections may require 10s of thousands of solid finite elements to model). AASHTO heavily favours CHSs for these structures. In practice, it seems that the rigorous FE analysis required by AASHTO is not performed by designers, and equivalent fatigue categories for circular sections are simply adopted for square/rectangular sections.

My understanding is that RHSs tend to attract stress to the corners of the sections near connections, where they are relatively stiff and do not warp in-plane to the same extent as at the middle of the walls. This is particularly the case at connections with stiff plates. Additionally, where the end of one RHS (member A) is welded to the face of another (member B), especially where they are the same size, the relatively low out-of-plane stiffness of the face of member B allows member B to warp locally. The high in-plane stiffness of the RHS walls tend to attract a large share of the load instead and you end up with some fairly non-uniform distributions of stress. For a CHS, this isn't as big an issue. Based on my own FE modelling of similar RHS connections, the stresses in the walls may vary by a factor of as much as 3-4 from the corners of the section to the centre of the walls.

I can't say whether or not this would have contributed to this particular failure, but it's likely that not enough consideration was paid to the fairly complex flow of stresses through the RHS to RHS connection. My first thought is that a simple bracket clamped around the perimeter of the SHS gantry would have provided a simple weld-free connection that is better able to distribute the loads from the sign into the gantry (not so concentrated to a single wall of the section).

 

[charliealphabravo]

It looks like a short HSS upright with a bolted flange on top was welded directly to the top of the gantry. Due to the low height of the HSS and the width of the bolted flange they had to fillet welded it from inside the HSS further reducing the moment couple. This would also help explain the clean tear of the gantry material since the outer edge of the HSS acted like a brake.

I don't see how that configuration could have worked without a taller HSS upright and stiffeners to the gantry. The thought of a 16 ft high sign fatigue bending the wide plates of the gantry is a bit terrifying.

 

[hokie66]

charlie,
Just how did they weld from the inside of the SHS with that plate on top? Limited access for welding may have played a part, but your explanation is not logical.

 

[charliealphabravo]

I think the plate is just a bolting flange welded to the top of the SHS with a full size opening or hand hole. That would have allowed them to weld from inside the SHS stand prior to bolting on the sign. The flanges look to be at least 4 inches wide and the height of the SHS stand is less than the flange width. I don't see how it could have been welded from the outside.

 

[hokie66]

OK, hadn't thought of an access hole for welding because that would be a poor detail, as it would decrease the strength and encourage corrosion. But if the stools are not high enough for welding, that is a poor detail as well. Something went wrong, either in the design or execution.

 

[waross]

As I read this I keep wondering;
"Was this the only similar sign installed in 2017?"

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[Settingsun]

I've done proof engineering for similar sign gantries in the past. It's worth mentioning that the fatigue requirements in AS5100-2017 for highway gantries refers to the 2015 AASHTO code, which doesn't explicitly disallow square/rectangular hollow sections for these structures, but requires some pretty onerous FE modelling of the weld details to demonstrate that the connections are suitable (the FE model - which must be solid by the way - has to be so refined as to provide several finite elements just along the toe of an individual weld - fractions of a mm. Simple welded connections may require 10s of thousands of solid finite elements to model). AASHTO heavily favours CHSs for these structures. In practice, it seems that the rigorous FE analysis required by AASHTO is not performed by designers, and equivalent fatigue categories for circular sections are simply adopted for square/rectangular sections.

AASHTO is poorly written IMO. Clause 5.6.2 does appear to outlaw SHS/RHS but then rules are provided for them, however there is nothing provided for them in terms of fatigue (aside from the impractical FEM requirements). I wouldn't be surprised if designers are getting it wrong.

[Edit: Or are the RHS/SHS rules only for assessing existing structures? I'd have a hard time justifying them in a new structure given what 5.6.2 says.]

This failure is intriguing because it seems as though there wasn't much wind which would tend to suggest fatigue failure, but it's a young structure. Could be a fairly diabolical oversight in design or construction, or material supply.

 

[Retrograde]

"Two more freeway signs removed after Tulla collapse":

https://www.brisbanetimes.com.au/na...ved-after-tulla-collapse-20190114-p50r7y.html

 

[waross]

Retrograde; Thanks for answering my question.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[charliealphabravo]

This failure is intriguing because it seems as though there wasn't much wind which would tend to suggest fatigue failure, but it's a young structure. Could be a fairly diabolical oversight in design or construction, or material supply.

The lack of wind actually suggests that the failure was caused by fatigue. From the photos you can see that the gantry metal broke away cleanly in two squares, with only a small tear at one corner of one upright. This means that the gantry metal had been almost completely compromised by previous wind or gust effects and finally laid over in near calm conditions. If there had been substantial capacity in the connection then the gantry plates would have distorted and torn and it would have required a significant wind load.

 

[Settingsun]

I agree small load at time of failure makes fatigue the prime suspect but, at less than two years age, exactly how bad would the detail have to be? Standard sign structures with no particular fatigue consideration have lasted decades. I don't know when it was designed but I expect the criteria included infinite life to AASHTO.

 

[BridgeSmith]

As it was posited earlier, fatigue of the base metal of the built-up box section that spans the roadway, may have never been considered - one engineer designs the box section to span the roadway while another engineer designs the sign supports, and each assumes the other (or someone else) checked the connection between them. Even in only a couple of years, it may have been subjected to tens of thousands, maybe hundreds of thousands, of cycles of truck-induced wind gusts.