Miami Pedestrian Bridge, Part XIII 81

[JAE]

A continuation of our discussion of this failure. Best to read the other threads first to avoid rehashing things already discussed.

Part I
thread815-436595

Part II
thread815-436699

Part III
thread815-436802

Part IV
thread815-436924

Part V
thread815-437029

Part VI
thread815-438451

Part VII
thread815-438966

Part VIII
thread815-440072

Part IX
thread815-451175

Part X
thread815-454618

Part XI
thread815-454998

Part XII
thread815-455746


Check out Eng-Tips Forum's Policies here:
faq731-376

 

[Vance Wiley]

HA HA HA - no, Fred - to be twice stout. (I hope).

 

[Earth314159]

I think the main difference with Lemessurier was that he informed the emergency services who then put into place an evacuation plan of the surrounding blocks and monitored the wind forecasts. The engineers also kept the city and emergency services informed of the degree of retro fit and the increased wind load capacities.

 

[Earth314159]

Cables vs. pipes

I think pipes were not such a great idea. The long term upward cambre from the PT of the truss puts more load on the truss as the pipes go into compression. As you mentioned, temperature can also put the pipes into compression. This is where redundancy and indeterminacy can harm you. You better hope the trusses tries to sag rather than cambre up.

If vibration was the purpose of the tube stays, I think a viscous damper would have been a better choice rather than a solid tube.

 

[PA PE]

EarthPi, you're right, I do remember reading something about vibration with respect to the pipe stays, and I think it was to dampen vibration from the live loads. I remember thinking "how are pedestrians going to make that much vibration?" Maybe if they were all marching double time, or doing a "wave" across the bridge. Could you input that into a computer model?

 

[Tomfh]

I think the main difference with Lemessurier was that he informed the emergency services who then put into place an evacuation

I’m sure that would have been very comforting to the unaware people inside the building, especially had it fallen over.

There’s a good reason they kept it all secret for 20 years.

 

[jrs_87]

Questions for structural engineers:

Was the deck suspended from the canopy and if so, what percentage? What was the intent for the canopy curve? (It is reminiscent of a retractable metal tape measure.) Why did the deck not have any similar feature to increase stiffness? Is it fair to refer to deck as a ribbon? How much capacity was in the moment of the deck between the pylon pier and 10/11 node? Did FIGG depend on that moment? Why was steel avoided on span for the sake of ease of maintenance but lots of it was used for the pipe faux stays? Why was the sorter span designed exactly like longer span? I.E. same deck thickness, same diaphragm sizes, same canopy, same number of truss members. The long span was about 910 lbs per inch. What about the short span?

 

[Earth314159]

Tomfh

There was actually a news release at the time. Reporters were going to follow up on it but then there was a media strike. The emergency plan was tracking the wind forecast and not only planned for the evacuation of the building, there were plans for the evacuation for blocks around the building. I think people behaved more ethically in that case than the FIU bridge. It was the New Yorker that picked up on the story 20 years later and contacted LeMessurier. It wasn't actively hidden.

You also can't really evacuated a chunk of New York city for a few months until the work is done.

 

[Earth314159]

Pa Pe,

It is difficult to model accurately. The response of long term creep is educated guess work. You can model with a lower E value (the long term creep)and keeping the temperature on the PT (you can use temperature on the PT to model the pretensioning forces) the same as the short term model. But a structure doesn't necessarily follow what the computer says. I have a non-PT two way slab spanning about 40 feet (in both directions) and it was suppose to be creeping down by now (we cambered some of the dead load out of the slab) and it is still arched up. We did an FEA models for both short and long term and included cracking (we altered the modulous for a more flexible cracked slab).

If the truss starts to arch up instead of sag, with compression pipes, you are in trouble again.

I also thought they could have used something like a take up device that tightens a cable option if there is any slack (a torsion spring on a nut) but that is complicated. On other designs I have added springs to keep exposed cables in an architectural design taught if there was even a "compression" load combination.

 

[Tomfh]

There was actually a news release at the time.

The public statements were false. They said there was no problem, and the rectification was merely an additional safeguard. Same line FIGG was using

LeMessurier maintains that the. . .tower has well over the structural support it requires to withstand anticipated wind loads and that the purpose of the extra bracing is simply to supplement it

As with FIGG, they suggested all was well, that there was no real risk, but some extra strapping was required just as an additional safety measure.

The key difference is not their behaviours, but that Lemessiur got lucky and pulled it off, and FIGG didn't.

You also can't really evacuated a chunk of New York city

Saying you can't evacuate around the building is precisely the same logic as we can't shut down the freeway.

 

[Earth314159]

Tomfh,

Yes, at the least, the developers put a positive spin on the news release. It is still a false equivalence. I don't see these as the same at all. Closing down a highway for a day to install cribbing and shutting down part of New York city for months are not the same. Once cribbing is in place, you can even reopen some of the lanes until transporters can be arranged to remove the bridge.

There was also no release to the public at all with the bridge. The engineer did not come clean to the authorities.

Statistical analysis was done with Citi Corp building to determine a point at which an evacuation was needed and public safety plans were in place.

It was only by chance that the media didn't track down more details from Citi Corp, the city or LeMessurier.

The Citi Corp release was also public and not just in a meeting with those responsible for the build.

 

[Earth314159]

Tomfh,

Just out of curiosity, what would have done differently from LeMessurier? I can definitely say what I would differently from Figg but I am not as certain about LeMessurier.

 

[Earth314159]

Questions for structural engineers:

Was the deck suspended from the canopy and if so, what percentage? What was the intent for the canopy curve? (It is reminiscent of a retractable metal tape measure.) Why did the deck not have any similar feature to increase stiffness? Is it fair to refer to deck as a ribbon? How much capacity was in the moment of the deck between the pylon pier and 10/11 node? Did FIGG depend on that moment? Why was steel avoided on span for the sake of ease of maintenance but lots of it was used for the pipe faux stays? Why was the sorter span designed exactly like longer span? I.E. same deck thickness, same diaphragm sizes, same canopy, same number of truss members. The long span was about 910 lbs per inch. What about the short span?

1) the deck is not really suspended from the canopy. The canopy is in compression and the deck in tension (but pre-compressed so the deck concrete is in compression but the PT is in more tension for a net tension force). It is all part of a truss system. Just like when you were a kid bending an eraser until the bottom split in tension and the top was in compression (then your teacher gets made at you for wrecking your eraser).

2) The deck did not have the same features primarily for functionality. The canopy also needs more curve and depth to minimize buckling since it is the element that has a compression member force. The curve also allows for a longer spanning length. You can say the concrete in the deck is also in compression but PT can't actually buckle the concrete that it compresses (unless you have external PT).

3) It is a ribbon in the sense that it was a tension member (although pre-compressed so the PT is more like a ribbon). It also had a significant local bending force to transfer the loads to the diagonals of the truss. Seemingly contrary to my comment above there is some load that hangs but it is not really "hanging from the canopy".

4) The moment on the deck was only relied upon for the local loads and not the bridge as whole. The deck could not span 175' without the truss action.

5)The faux pipes could be allowed to corrode to some degree. Maintenance could be slack. The coverage on the pipes was relatively small but I have to say, access to the pipes would not be easy.

6)A different design could have been done for the shorter span. However, the back span (the shorter span) helped the longer span out. With the canopy and deck lining up, you get a negative or hogging moment over the support that reduced deflection. It also helped reduce the shear demand at the underside of #11. But it also increases the loads on the diagonals which were also under designed. The truss shape was to allow clearance for the roadway while not having to raise the deck too high. You would need a taller elevator, more stairs, a longer walk for people, etc. It is also not as clunky looking as a concrete girder. The shorter answer is that it looked better to the designer and client.

 

[jrs_87]

Earth314159, thank you for response. It goes to show contemplating bridge design is more complicated than it looks. Hogging is an interesting concept. As a kid I tried to make longest possible metal tape measure "bridge" before it would buckle. I remember adding hog or at least tilting tape body upward would extend the distance.

It seems more and more to me FIGG knew what they were doing and simply took risks that ended up going the wrong way and they did not give the project the attention it needed.

 

[jrs_87]

It's off subject, but it is a bridge in Florida under construction. The lift-bridge looks interesting.

https://www.google.com/maps/@26.675...4!1s8w-BhPsweSubfAuXJrdcHw!2e0!7i16384!8i8192

 

[Vance Wiley]

Some points from NTSB Meeting Miami and further down, comments about loads and capacities at failure.

2:08 clock on video
Question : “What was different at south end? There was a stairs there”. Staff said the south end was supported in a block out with a back form and that provided resistance which the north end did not have.
They forgot to mention the 1-1/2” expansion joint. That is more than twice the size of measured movement at north end just before collapse. There is nothing to help at the south end before a collapse has started.

2:17 Chairman “Joint finish is a red herring because it still would have failed”. Roses to MCM.
Staff described joint as “not finished so it was rough”. Roses to MCM.
“Rough would not have changed the outcome”.

1:24 Joints: Staff comments that FIGG disregarded the cohesion factor and that was a conservative move. But apparently staff gave them no credit for that decision.

Re: Moving/Transport: Board accepted as “documented there was no increase in cracking during the move”. And condition “after move was nearly identical to its initial state”.

2:23 Redundancy. Figg considered it redundant, used 1.0. If correctly considered non-redundant, should have used 1.05 but that would have made no difference.

My comments about capacity of node 11/12 under shear friction at failure:
Given known conditions at time of collapse as follows:
1. Self weight + maybe 3 kips construction load on node 10/11
2. Concrete tested 9300 psi
3. Steel tested 62000 psi
4. Cross plane reinf of 4 - #7 hoops and 2 - #11 from member 12 = 4.8 in^2 + 3 = 7.8 in^2.
5. Use AASHTO design and include cohesion.

At collapse, load factor is 1.0.
At collapse materials are operating at 100% so phi = 1.0

FIGG intended joints roughened to 1/4” amplitude.

Truss reaction is about 810 kips. That is 940 kips minus trib of canopy and deck which is supported directly by pylon which is about 130 kips and adding 3 kips construction load.
[pre]Cohesion is 0.28 ksi X 21 X 40 = 245 kips
Steel = 7.8 in^2 X 62000 = 490 kips
Pc = 813 kips X 1.0 = 813 kips[/pre]
Expected Capacity at failure on March 15 (if undamaged) is 1548 kips.
Actual demand at time of failure is 813/tan 31.8 deg = 1311 kips

Conclusion? Should not have failed.

OK - only 3 - #7 hoops worked - so lets count the 3rd #11 in member 11.
OK - there were 4 - 4” sleeves in a bad place. But those had nothing to do with the capacity of the shear plane UNDER member 11 at the deck surface.
OK - diaphragm 2 was cracked to hell. That is not in the shear plane.
OK - there was an 8” pipe 20 inches below the top of the deck. Not in the shear plane.

What did happen? The joints were not prepared - at all. They were as poured. So that loses 40% of 1303 kips = 521 kips lost, leaving a capacity of 1027 kips, much less than the unfactored demand of 1311 kips. It fails.
It cracked at the joints either before or during moving/transporting, and that negated any cohesion. That loses 245 kips and leaves a capacity of 1548 - 245 = 1303 kips = 8 kips less than demand in this set of calcs. Flip a coin.
So if we lose the cohesion and consider non-roughened joints the capacity becomes 1027 kips - 245 kips = 782 kips, sure failure.
The real question is how did it stand for 5 days?

Then there is the thing about restoring the PT in member 11.
EDIT ADD
560 kips PT causes 560Xcos31.8deg = 476 kips slide
560 kips X sin31.8 = 295 kips clamping
Joint load becomes 1311 kips + 476 = 1787 kips
Joint capacity = 1548 kips + 295 = 1843 kips

An undamaged joint should not have failed. ( Close, though)
END EDIT
Your mileage may vary.

 

[3DDave]

Did you include the resistance to tearout of member 12? I think that's what kept it up.

Also, 11 was also shedding the outer sheath of concrete as well as the toe that held the 4th hoop, so the section area resisting shear was reduced; more like 20x36.

 

[jrs_87]

"Feds figured out why FIU bridge fell. Now prosecutors must decide if deaths were a crime"

https://www.miamiherald.com/news/local/crime/article236597488.html

 

[The Mad Spaniard]

jrs_87 (Mechanical)1 Nov 19 13:42
"Feds figured out why FIU bridge fell. Now prosecutors must decide if deaths were a crime"

https://www.miamiherald.com/news/local/crime/artic...

This is where money comes into place to avoid justice. With South Florida politics, with "some people" that are heavy connected with the Florida GOP, it is very possible that the "Justice"department in Florida will find ways not to prosecute. I will be surprise if they will. Remember Epstein...

 

[samwise753]

I am still perplexed by the final continuity condition between the front span and the back span. Going back through the RFC plans, Tendons C1 and C4 are called out to be stressed after the closure pours between the front and back spans have cured to 6000 psi. There were 2 - C1 tendons and 2 - C4 tendons, both sets were 0.6" diameter. The stressing occurs, of course, after both the front and back spans have been set with all the temporary supports/falsework removed. This means both spans have already under gone their response to simple span, dead load deflection. That is locked in. So, at best, this structure was continuous for pedestrian live load only. The only way you achieve continuity for the dead load of both spans (i.e. continuous beam behavior) is if you can take back that simple span response with the post-tensioning in the canopy or you avoid the simple span dead load condition altogether by keeping the spans temporarily supported until the closure pour and C1/C4 post tensioning is complete. I need to run the numbers, but how in the world are 4 - 0.6" strands going to relieve the simple span dead load condition for the 950T front span and the 520T back span at the same time? Was this even the intention? Was this even realistic?

 

[jrs_87]

samwise753, it's simple, just look at the precursor of this bridge design.