Revisiting the de la Concorde overpass collapse

[GalileoG]

I was reading through the report issued by the Commission of Inquiry on the collapse, which clearly states that the failure mode is a shear failure of the thick slab (given how the thick slab did not possess any shear reinforcement.) What I do not understand, however, and is nowhere mentioned in the report (as far as I can tell,) is why would the bridge collapse by shear at a time when there were no vehicles on the deck? As far as I know, all of the injuries and deaths were from vehicles below the deck.

Thought perhaps this would induce an interesting discussion here.

 

[JAE]

Well, you're asking for us to speculate on aspects we don't know. But I'll try (grin). Here's a few possibilities:

1. Temperature got colder prior to collapse and the resulting shrinkage created enough of a straw to break the camel's back. Shear capacity in concrete doesn't like tension resulting from the shrinkage.

2. Long term degradation of the concrete simply resulted in a collapse that occurred when it occurred (under self-weight only).

3. Moisture got into some cracks and through freezing action expanded and initiated collapse (sort of like item 2 above).

4. A bird landed on a railing - just enough weight to push it over the edge.

 

[cooperDBM]

#2 & 4. Given their diet of left over bagels and smoked meat, Montreal pigeons have obviously been packing on the ounces.

The bridge collapsed under it's own weight. The failure crack started as a separation between the top long. steel and the lower bars that wrap down the face of the cantilever. The bulk of the concrete was below this crack and there were no stirrups to control the crack progression. The crack immediately extended into a diagonal crack as the shortest route to the base of the cantilever resulting in the sudden failure. The initial crack was due to deterioration and fatigue in the concrete around the top long. steel. There wasn't much to stop it from failing anytime during it's life.

The idea that minimum stirrups are needed in deep slabs has only been recognized in recent years. As recently as ACI318-05 (for buildings) minimum stirrups weren't required for slabs of any size. These old ACI shear rules were the accepted practice (buildings or bridges) when the bridge was designed.

 

[cooperDBM]

I see that ACI318-08 (the latest) still doesn't require minimum stirrups in any slabs.

Canadian codes now recognize this "size effect" and require minimum stirrups for thicker slabs thanks to the work of Michael Collins and friends at the University of Toronto.

 

[apsix]

cooperDBM
As I understand it the failure was due to poor detailing of rebar at a halved joint. This would seem to be a separate issue to your reporting of the Canadian requirement for minimum shear reinforcement for thicker slabs.
At what thickness is a slab required to have minimum stirrups?
It will be interesting to see if this idea will be picked up by other national codes around the world.

 

[cooperDBM]

The poor detailing (actually poor placement I think) allowed a horizontal plane to crack between the top long. steel and the face bars at the end of the cantilever which lay immediately below (bent horizontal below the main long. bars). There was no mechanical interlock between these layers of steel. The weight of the suspended span was carried by a bearing seat below this cracked plane, meaning this plane was in tension. This crack then extended quickly into the shear crack which caused the brittle failure. Shear stirrups would have hooked over the long. bars and taken the tension from the bearing seat (the original face bars were presumably supposed to do this). They would also have prevented propagation of the horizontal crack and shear crack. If a crack had formed it would either have been controlled or at least wouldn't have been a brittle failure.

I spoke too soon about the Canadian codes. The Canadian Bridge Design Code (S6) still doesn't require min. stirrups in slabs. To be fair many bridge engineers wouldn't consider a thick one-way full-width cantilever, such as the Concorde bridge, a slab. It's obviously been open to that interpretation in the past. The building code (A23.3) requires stirrups in a slab if the loading is high enough or if the depth is greater than 750 mm (30").

 

[apsix]

I know that stirrups would most likely have prevented the failure, but that points to the need for good rebar detailing, not to a blanket requirement for stirrups in all thick slabs.
There's no doubt that the cantilever 'slab' in question should have had shear reinforcement.

 

[hokie66]

A very interesting, if long winded report. There were a lot of factors including poor detailing, little inspection, botched repairs, and failure to recognize cracks on the face of the bridge as indicative of what was happening deeper in.

Truss analogy, or strut and tie design, was not standard practice at the time. If these joints were designed now, it would be routine to adequately develop the hanging steel at the corbels. Hooking these bars over the top bars and cogging the top bars would have prevented the collapse.

The worst thing about the bridge, though, was that the bearing area could not be inspected. One engineer, worried about the face cracking, wanted to lift the deck for inspection of the corbels and bearings, but he was overruled Even if this had been done, the impending failure may not have been discovered unless the deck was lifted far enough to discover the cracking, which was along the line of the top bars.