Mexico City subway overpass collapses, killing at least 23 and injuring dozens 2

[dauwerda]

According to the story, this section of the line was built in 2012.

https://www.cnn.com/2021/05/04/americas/mexico-subway-collapse-intl-hnk/index.html

The very beginning of this video shows the collapse via security camera footage.

https://www.cnn.com/videos/world/20...hnk-vpx.cnn/video/playlists/around-the-world/

 

[hokie66]

That concept may be simple, but it would require well planned and supervised execution. I've never seen it done, but that means little. Anybody else used that system?

 

[Tomfh]

This article says the line was closed for structural repairs in 2014, including the collapsed span, and that there was cracking after 2017 earthquake.

https://www.bbc.com/news/world-latin-america-56985164

 

[Tomfh]

So, it would be either fatigue crack at the tension flange, or buckling of the entire beam.

Beams seem well aligned after collapse, so presumably the former?

 

[dik]

To help aggrivate matters, from the photo, it appears that the spandrel was discontinuous in the vicinity of the failure to accommodate a light standard of some kind.

Rather than think climate change and the corona virus as science, think of it as the wrath of God. Feel any better?

-Dik

 

[hokie66]

I think that is just a non structural fascia, not a spandrel beam as such.

 

[dik]

Quite likely, hokie... just added dead load... it looked substantial enough that it would provide some stiffness and a marked discontinuity. At about 0.6 span...

Rather than think climate change and the corona virus as science, think of it as the wrath of God. Feel any better?

-Dik

 

[LittleInch]

That connection of the divergent beam at the mid span I don't think is coincidental.

It looks quite strange, the main beam is the same size, it's a point load in the middle of the span and appears to rest on the bottom flange only.

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

 

[bugbus]

What is striking to me is how cleanly both of the girders are fractured, suggesting fatigue failures of both girders? Is it possible that both girder bottom flanges were substantially cracked through without being noticed at any point?

If it were only one girder which had a crack through the bottom flange, I would have expected one of the girders to maybe fail in fracture, and the other to fail in a more ductile manner. But this does not appear to have happened.

I wonder how often these bridges were inspected?

Agree with the previous poster that the precast decks were probably intended to be made composite by grouting a shear stud into a recess and also grouting between the deck units. This would explain the narrower top flanges. But there appears to be cracks between the adjacent panels, which might suggest that there was less (or no) composite action than anticipated and the bottom flange was working much harder than designed for (leading to fatigue cracks through the flanges at the butt welds).

 

[dik]

Is there a chance the steel may not have been weldable? or maybe poor low hydrogen welds?

Rather than think climate change and the corona virus as science, think of it as the wrath of God. Feel any better?

-Dik

 

[FacEngrPE]

Since hydrogen contamination of welds only requires storing the rods in humid conditions and not in a rod oven, that could explain both the poor appearance of the welds, and cause hydrogen embrittlement.

This is easy to prove either way with relatively simple forensics.

Photos published in the Washington Post

As often occurs this disaster is already impacting[sup][/sup] the politics of Mexico.

https://www.nytimes.com/2021/05/04/world/americas/mexico-city-subway-accident.html

 

[hokie66]

Yes, a material failure, either in the base metal or the weld metal, could be at issue.

But I keep going back to the small top flange. While the concept of developing composite action sounds attractive, the reality of introducing enough shear connection between a narrow flange and those precast deck units is questionable. That would require a lot of connectors, I think in a single line, so the deck units over the beam would have to be like Swiss cheese. And then the joints between the panels have to be grouted, leaving no gaps, at least over the beams.

And as gusmurr said, if composite action is not achieved, the bottom flanges would have to work too hard.

 

[LittleInch]

Same picture as tomfh, but there's something odd about that truss. It's the only one which is not in line with the lines of the slabs above. It actually looks slightly bent but I thought that was google street view being not precise.

Did it get "shifted" so they could repair the gap in the concrete above it?

The column set up also changes at this point to move over for a three track layout vs 2 track leading to a bit of instability in the column forces.



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

 

[Burloak]

That flared beam coming in does appear suspicious. It would add a mid-span reaction to the exterior girder, and also the load is eccentric to the web centre-line, adding a torsion to the main girder as well. At the same time, the same simple cross frame is located on the back-side as exists along the span.

The "butt" weld is likely (hopefully) a complete joint penetration weld (CJP), also called Full-Penetration weld. In the highway design codes, if properly tested and ground in direction of stress, the stress category is "B", which is pretty good and better than the weld of the vertical stiffeners to the bottom flange ("C'").

My guess is that with 1 day of looking at the drawings, they already know the causes - but have to properly document it before it is released to the public.

 

[waross]

This may be indicative of the workmanship or lack thereof.

Bill
--------------------
Ohm's law
Not just a good idea;
It's the LAW!

 

[dik]

sometimes composite members have smaller top flanges, in particular in areas where material costs are high and labour costs are low... I have no idea of how it was designed or misdesigned.

Rather than think climate change and the corona virus as science, think of it as the wrath of God. Feel any better?

-Dik

 

[TugboatEng]

Waross, your derailment first idea does have a bit of truth to it. When the story first broke the first reports mentioned the operators complaining about overweight trains. They complained that the overweight trains were wearing out the tracks for over a year. A worn out track could lead to a derailment. Operators usually aren't keen on observing structural condition so track likely does relate to the rails, sleepers, or ballast.

 

[hokie66]

Yes, rail wear, or perhaps a switch problem, could have led to derailment and an impact caused collapse. In some way, that switch location is probably involved.

Dik, yes composite members can have smaller top flanges, but the question is how composite this steel/precast connection was in practice.

 

[NorthCivil]

I had the good fortune of visiting Mexico city in 2017 right after the Puebla earthquake. there was a lot of odd damage around mexico city, mostly settlement related. The ground that the city is built upon is so soft, all the buildings are leaning one way or another.

Mexico city is such a lovely place to visit. beautiful thing about their metro train system, its only 5 pesos (~25 cents US) to ride that train anywhere along the vast array of lines.

 

[waross]

One possibility may be a split switch.
That is, the switch hangs up part way through its travel.
A split switch will cause a derailment.

Bill
--------------------
Ohm's law
Not just a good idea;
It's the LAW!

 

[LionelHutz]

Except it sure looks like the train was on the other track, the one that doesn't split.