Miami Pedestrian Bridge, Part XIV 78

[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

Part XIII
thread815-457935

 

[thebard3]

Is it possible to link to the source of that page,

I found it at Link

Brad Waybright

The more you know, the more you know you don't know.

 

[waross]

Thanks Brad.

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

 

[bridgebuster]

@waross,

One of my colleagues sent me the graphic from an article in New Civil Engineer.

Link

 

[hokie66]

thebard3,

The issue about "designing outside the area of engineering principles" is more appropriate to a bridge structure than to a very risky space exploration project. There was no need to experiment with this bridge without understanding its behavior beforehand. The issues with the design were that the form of the bridge as a concrete truss/frame of that shape and magnitude had never been done before, and then they relied on unproven shear friction theory in a critical joint. The design could have been proven or otherwise by load testing before erection, while the bridge frame was still in the casting yard.

 

[FortyYearsExperience]

Hi Hokie66
your comment "The issues with the design were that the form of the bridge as a concrete truss/frame of that shape and magnitude had never been done before, . . . " is not correct. The record is full of instances where concrete trusses were successfully used in bridges and other structures. However, I agree that in none of these instances was shear friction theory applied. Whoever came up with that theory should face a court martial.

 

[hokie66]

I think we had this argument before, and none of us could find many concrete truss bridges. Certainly not ones with central truss/frame web members with two wide, flat decks used as flanges. If you know of one, please point it out. I still contend that this was a novel design, and required more than just calculation to prove that it would work.

 

[wiktor]

Hi Hookie66 & FYE,
I 100% agree that there are few examples of RC or PRC trusses used as bridges, and definitely none with deck and canopy used as top and bottom chords. So, the bridge was definitely a novel structure, designed using wrong code, with the initial S.F. close to 1.3-1.35 at construction stage, a horrible detailing of the joints - the node in between the deck and #11 diagonal being classical example (positioning of the anchorage of post tensioning rods in the middle of the bottom joint causing likely delamination below the anchorage - and the sliding plane). The shear friction theory was discussed at length, so I will not comment on it. The simple approach is that the cracked concrete cannot carry any stresses, so the theory works only with uncracked concrete, and most likely in the lab condition.
The reports by OSHA and NTSB are both missing the crucial point - C/D ratios were computed using demand computed with load factor 1.3 for DL and for the bridge with DL/LL=9 this was suicidal. All RC trusses existing or proposed have overall SF of 2.5/3, and these are mostly redundant structures.

 

[dik]

I suspect strongly that the failure would have occurred in the finished product, too...

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

 

[RFreund]

Now that we are finally getting in "lessons learned", I think what has been pointed out is all good and fair. But when it comes down to it, it seems like there was a break down in the internal communication. For the sake of avoiding names you had the "analyst" the "designer" and the "manager". If you read the NTSB interviews for them, you will see they are quick to point out their "role". The designer used the forces provided by the analyst. The analyst only provided forces that were requested and the manager just was 'checking'. They used different models in different software (Larsa and Lusas), which yielded different joint forces. The designer 'requested' forces from the Lusas model which had substantially less demand. So they kinda mixed and matched models. It just seems like they never stepped back to look at the big picture or compare opinions.

Maybe there is a better explanation out there of why they chose the lower demand model, but I have not seen it.

What happened after the cracking started to occur is a different story.

EIT

http://www.HowToEngineer.com

 

[SFCharlie]

If they had cut out a profile of the "truss" and viewed it with polarized light, would they have noticed that there was a problem?

SF Charlie
Eng-Tips.com Forum Policies

 

[Vance Wiley]

I am not sure what kind of light bulb went off when this structure was conceived but not enough light was used in the design phase.
I have not worked with polarized light but it would seem it could reveal distortions from a true plane, but the model used would require great precision to start with. If the PL could show changes in density which could represent stresses the model would need to be loaded appropriately.
The problem developed at the joint of the truss diagonal to a 32 foot wide deck, so three dimensions would be required.

Perhaps the light of a setting sun (career) would have been most helpful.
So tragic.
Thanks,

 

[Alistair_Heaton]

I come from the FEA world of dos ANSYS 4.22 and running it on SGI with graphics cards which were more than my salary per year.

The issue with FEA is that even with my exposure to it which encompassed Dos to windows to integrated into pro engineer and autocad (I actually used to work for strucom who did the initial port to autocad) is that generating a FEA model became way to easy. Everyone just blasted it through and had no conception what they were doing with the mesh or the element types. It was pretty grim in the late 90's i would hate to think what its like now with colossal computing power and getting away with murder. We had to be tight and know what we were doing with 18 hours static load runs on a 486. I was mainly doing none linear which had its issues. But even with company's that should have know what they were doing you were presented with the issue of explaining that it was a nonlin geom problem and worse than that tet mesh was completely the wrong element type to use.

I suspect the model they were using was a static linear model which would give utter nonsense results especially if all they looked at was the von misses and didn't roll back to none linear geom and displacements.

 

[dik]

I'm not sure...

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

 

[ldeem]

Alistair Heaton really hits an important point, its so easy to use automated analysis tools and engineers just take the results as accurate. I work a lot in STAAD and on occasion have reviewed other peoples models, some have been downright scary.

Maybe load factors should be adjusted upwards with every automated analysis option used in a design. Of course on the other side the gross simplifications sometimes made in hand calculations can also lead to problems.

 

[SFCharlie]

To clarify my query: I was speaking of structural models made of clear photo-elastic plastic viewed through cross polarized sheets or glasses. Please see: photoelasticity

SF Charlie
Eng-Tips.com Forum Policies

 

[SFCharlie]

Yes, I'm old enough to be classified as ancient, I made structural models of the machina (cranes) for Stonehenge out of string and branches...

SF Charlie
Eng-Tips.com Forum Policies

 

[FacEngrPE]

I don't think photoelasticity would give a good representation of reinforced concrete, however that is not to say the visualization would not be useful to a designer skilled in the technique.
The over reliance on software solutions is a problem in all fields of engineering. I had the opportunity to introduce a newly minted engineer to the psychrometric chart today, just as good today as it was 50 years ago. For some tasks just as fast at the computer calculation.

Fred

 

[LittleInch]

This bridge was a disaster in the making all along.

An asymmetric concrete truss with different width flanges, made to look like a cable stayed bridge hence more forces concentrated at one end, The flanges themselves were not evenly tensioned.

Then the real killers - late extension to the length forces the two ends to become cantilevered during the rather fraught installation forcing extra tension tendons only for the transportation phase in members 2 and 11. Throw in the fact that at least some parties never analysed the part completed bridge, the joint area at 11/12 was changed to incorporate lots of ducts and tubes and the bridge really should have collapsed as soon as they removed the transport supports. IMHO.



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

 

[hokie66]

Yes, but if it had collapsed during load testing in the casting yard, no one would have been killed.

 

[dik]

I think it was a neat bridge, if it had have worked... quite attractive.

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