Miami Pedestrian Bridge, Part IX 33

[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


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

 

[3DDave]

Neither bridge failed for political reasons - they failed for fundamental engineering mistakes. I expect it is rare for any project to be touted as "spending extra money for no reason" so any claims about costs are universal and indicate zilch. The more common complaint is cost overruns. Likewise since most bridges are publicly funded for public use there are almost always political overtones. Unless one fails in a spectacular fashion neither the money nor the politics will get any note at all. Most people are so relieved that the construction is done they don't care.

 

[waross]

The Saskatchewan bridge as been discussed here.
See:
thread815-444241


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

 

[TheGreenLama]

For what it's worth, here's a member load summary extracted from the LARSA model. The issue with the calculations, however, is that they do not include any results for the construction stage at the time of collapse. As far as I can tell, the LARSA construction check done in the calcs only reflects the span during the move. As a result, the closest state I could find is the "end of construction" (EOC) conditions, which is NOT the load state on the span at the time of collapse.

So, a couple of things to keep in mind:
1) EOC state includes continuity PT, pylon, and faux stays.
2) It's presumed the PT case below excludes the permanent diagonal PT; see note at end.
3) Only two load cases are shown below. Others, such as LL, creep, etc., can be found in calcs.
4) A maximum load has been extracted for the LARSA member numbers shown below. These member numbers constitute the bulk of the length for each of the diagonals, top, and bottom chord members. As a result these loads are approximations only.
5) Loads shown are axial only (kips). Bending moments exist, but are not included in this summary. Axial loads are calculated by multiplying the "P/A @ centroid" by member area. If I'm interpreting LARSA output incorrectly please let me know.
6) These are uncombined loads w/o load factors.
7) Negative is compression.


LARSA Members:

Area (canopy) = 16.46 sq ft
Area (deck) = 45.67 sq ft
Area (diag 2) = 5.25 sq ft
Area (diag 3-11) = 3.5 sq ft


Self Weight Forces (kips):

Chord loads found on pp.579-583, diagonals on p.807 of calcs.
Diagonal 2 components (for 23.88 deg):
Horiz = 1653 k
Vert = 732 k
Diagonal 11 components (for 31.62 deg):
Horiz = 1155 k
Vert = 711 k
Total vertical = 711 + 732 = 1443 k

Rough hand calc of self weight of span based on input areas (excludes blisters and some incidental end zone area of each of the diagonals):
Clear span = 174.83' - 2.83' - 3' = 169' (this will exclude weight from Members 1 & 12 and end diaphragms, which are supported on the bearings)
Self weight of deck + canopy = (45.67 + 16.46) x 169' x 0.15 k/cu ft = 1575 k
Self weight of diagonals:
L2 = 29.1'
L3-11 = 162.8'
Wt = [(29.1 x 5.25) + (162.6 x 3.5)] x 0.15 k/cu ft = 108 k
Total self weight to be carried by truss = 1575 + 108 = 1683 k


Post-tensioning Forces (kips):

Chord loads found on pp.603-7, diagonals on p.811 of calcs.
Compare with PT stressing forces from design dwg B-69:
Total deck = (835 k x 10) + (527 x 2) = 9404 k
Total canopy (w/o cont.) = (531 + 534) x 2 = 2130 k
Total canopy (incl cont.) = 2130 + (556 x 4) = 4354 k
>>

Note that the two self weight loads, 1443 k for the LARSA model end diagonal forces, and 1683 k for the rough concrete volume, conflict with the 950 x 2 = 1900 k that has been quoted from the first post in this thread. Not sure if it's a result of my approximations, or if it's something else? The 950 tons was reported as the "lift" weight, which would include Members 1 & 12 and end diaphragms and additional 5.83' of span, so that may be where the difference lies.

Also, if you look at say Member 10, PT shows a positive value. This implies that diagonal PT is not included in the LARSA run for PT case shown above. Diagonal PT is sized on p.842 of calcs based on service load combinations, among others, for the critical axial and bending stresses. After having determined this diagonal PT, it's unclear if the model was then rerun to check long term stresses.

And finally, looking at Member 11 end loads. As Gwideman has pointed out earlier, the values used for calculating the shear reinforcement on p.1283 of calcs (Horiz = 589, Vert = 1233), which were supposedly pulled from the FEA, wildly contradict the loads found in their own complimentary LARSA analysis.

 

[Vance Wiley]

This is a first post for me. I have read from the first line to end. Not familiar with tools so have to just do it.
FYE 18 May 19 22:49
AASHTO and ACI allow the general formula used by FIGG to design reinforcing steel across an assumed and /or defined plane which transfers shear. This is called "shear Friction" - I did not coin the term. It is particularly useful in transferring shear from a deck slab to girders and designing corbels on columns. In the use of welded studs - "Nelson Studs" to the top flange of steel girders, the studs are inert and do not provide an active clamping force, but a 3/4" dia headed stud is allowed over 10 kips shear in developing composite action between the slab and steel girder. The concept of "shear friction" should work in the case of member 11 at the deck if correctly applied. I do have a concern about the amount of fractured aggregate I see in the photos, and whether a "mu" of 1 is perhaps too high for this concrete in this project. Also I have seen no confirmation of the preparation of the "cold joint" surface at the top of the deck.
In the use of Equation 5.85.1-3 of AASHTO 2016 (perhaps there is a newer one?) I do agree with FIGG that the "cohesion" should be disregarded, particularly in the case at hand.
" gwideman" has well described the intent of the influence of reinforcing across a "shear plane" (cold joint in this case) as used in the "shear friction" design concept.
Since the horizontal force at the base on member 11 is only present concurrently with a vertical component creating compression across the shear plane, it seems appropriate to allow use of the vertical component (unfactored) as a clamping force.
If the FIGG design complies with with the appropriate formulas for shear friction given in the applicable code (which allows reinforcing as across the joint a clamping force)and is properly applied to the particular joint , FIGG is off the hook for this item and codes need revision.
FYE "The way I understand the shear capacity of reinforced concrete is to ignore the concrete when analyzing the steel." Correct for designing a normally reinforced beam - when the shear stress exceeds a defined value, a crack is assumed and stirrups provided to support the shear at that section. Prestressed concrete design allows a contribution to shear resistance at sections which are under compression from the prestressing. The compression closes any cracking and maintains aggregate interlock. If one does not accept this concept, crossing a prestressed concrete bridge should be avoided.
The "10 minute calculation" (I use a lot of those - some not so quick) appears to develop the horizontal force of 725 tons in direct tension and tie that force back into the deck using longitudinal reinforcing of 27 sq in at 53 ksi. In this case the 725 ton force never made it across the cold joint.
I'm interested - what do you find the tension across the "cold joint" to be using Mohr's Circle? Member 11 is in compression, the deck is in compression from both longitudinal and transverse post tensioning. The existence of the cold joint defines the orientation of the tension developed by Mohr's - perhaps that analysis is more appropriate for a condition like this.
Regarding the development of the reinforcing across the cold joint as questioned by TheGreenLama, the dimensions for vertical placement should have been provided on the drawings. Perhaps they are on the rebar shop drawings.
Great forum - great questions - great responses - great format and tools.
Thank you.

 

[Vance Wiley]

Reinforcing across cold joints at deck surface
In reviewing the details of reinforcing which connects the diagonals to the deck slab, I see some interesting things.
First, a bit of background - in a simple Warren Truss, the diagonals transmit loads through a shear connection to the top and bottom flanges, creating compression in the top chord (canopy in this case) and tension in the bottom chord (deck). From the center (where shear is zero if the truss is uniformly loaded and symmetrical in geometry), each panel point (node) adds loads to the canopy and deck. The sum of these loads will add to become the horizontal shear at the bottom of the end diagonals in the case at hand. When viewing sheet B-61 titled DECK REINFORCEMENT & P T -MAIN SPAN (2 OF 2) and focusing on the #7 ties which resist the shear across the "cold Joint"s at the top of the deck, we see:
Truss Member 1 & 2 has 5 - ties "7S01"
Truss member 3 & 4 has 9 - ties "7S01"
Truss member 5 & 6 has 6 - ties "7S01"
Truss member 7 & 8 has 6 - ties "7S01"
Truss member 9 & 10 has 6 - ties "7S01"
Truss member 11 & 12 has 4 - ties "7S01" <<<<< the shear load is the sum of shears from nodes 7&8 plus 9&10 (plus node 10&11) but node 11&12 has only 4 ties instead of the 6 that nodes with less loads have. In effect the deck connection the north diagonal 11 is resisting loads generated by 12 ties plus node 10&11 and has only 4 ties to do so. Of course the vertical component of #11 reduces the steel requirement but ------- apparently not enough, judging by the results.
Without doing any calcs, it can be said "That's not right". Particularly at the end where breakout is a concern.
When watching the news on the west coast about 1:30 PM on March 15, I could see the bridge failed at the north end. Soon after it was learned that it was a concrete truss. Knowing that heel joints in trusses are critical, that became the suspect joint. Here we are.
I am saddened at the loss, and those lost and injured and their families have my sincere condolences.

 

[human909]

Whie I agree with HotRod that her license status is not really relevant to her owning a compnay, it does sound really off for her to say "it's all in compression" when it clearly is and was not. She's representing an engineering firm that specializes in high-level bridge design, so she really needs to know better what she's talking about.

Back to the collapse though. I did a Google search to see if a beam with an open truss-frame web was ever looked at before, and I came across some research done on a beam of this type to be used for a roof. I haven't delved into too deeply yet, but it seems they had a lot of trouble dealing with the force concentrations where the struts met the flanges. They're section was also symmetrical about the midline, and they used consistent strut angles through out. They also had full prismatic sections for the 1st and last quarter (roughly) of the beam. The most telling (and obvious) statement of the paper was:

"The fact that concrete does not work efficiently in resisting tensile stress makes it very
difficult to design a truss made of solely reinforced concrete."

https://digitalcommons.unl.edu/cgi/...redir=1&article=1013&context=constructiondiss

Great article. Though I would say that quote barely scratches the surface of the good and relevant material in it. Worth a read or even a skim look at the pictures for anybody following this thread. The examples shown show clearly that failure similar the Miami bridge are of key concern hence solid webs at the ends of the trusses. The ultimate failure of the test truss also bares similarities with the Miami bridge though at a much more reasonable load.

 

[jrs_87]

Upon re-reading HWY18MH009 NTSB report, I found a single mention of The Corradino Group, Inc. According to the report, this firm was involved in the FIU project. Does anyone here have any insight as to the probable nature of this firm's involvement?

 

[SFCharlie]

I can't prove it, but they do transportation planning, and may have been involved in the bus stop planning...

SF Charlie
Eng-Tips.com Forum Policies

 

[BridgeSmith]

Re: The Corradino Group, Inc. - Did they perhaps do the traffic control planning for the move?

 

[Vance Wiley]

LOADS ON SHIMS, FIGG POWERPOINT PG 4 OF 44, 15 MARCH, 2018
The last paragraph states:
"The bending moments that develop in the continuous structure, when the falsework of the CIP Back Span is removed, will reduce the load on the shims from their current values."
I do not agree - in fact, I see exactly the opposite happening. Creating continuity over the interior pier and subsequently releasing the vertical load of the back span will leave a negative moment over the pier which did not exist at the time that statement was presented, and that negative moment will draw shear from the south span onto the intermediate pier, increasing the load in member 11 and adding load onto the bearing points.
Perhaps it was intended to grout fully under the north diaphragm when casting the filler concrete over the intermediate pier and that grout would support any increase in load at the pier? That does not sound like it will "reduce the load on the shims from their current values". It could share any subsequent loads which develop (after any shrinkage in the grout has been accommodated).

In a somewhat related issue, has anyone seen in the FIGG calculations an analysis of the vertical movements of the long span under changing temperatures or night/day differences in expansion or contraction of the upper flange (canopy) with respect to the bottom deck? It seems to me the canopy would cool much faster after sundown than the deck, causing a downward movement and subsequent strains in the web members and joints.

And one more, if I may. Was vortex shedding considered in the Figg calculations for the 16" pipe "suspension strands"? It seems to me to be a very real possibility. Imagine all those pipes vibrating at different frequencies. They will also expand and contract with changes in temperature and some are quite long.
Thank you.

 

[Wetlander]

Upon re-reading HWY18MH009 NTSB report, I found a single mention of The Corradino Group, Inc. According to the report, this firm was involved in the FIU project. Does anyone here have any insight as to the probable nature of this firm's involvement?

According to the contract between FIU and Bolton Perez, The Corradino Group, Inc. was a subconsultant with the task of Structural Engineering/ Bridge Inspection. All the contracts (and other files of interest) can be found here: Link

 

[jrs_87]

It's worth repeating this link:

http://facilities.fiu.edu/projects/BT-904-PRR.htm

Especially 17. Time lapse videos (Caution File is large 16GB)

Digging around here you will find high-resolution time lapse videos that you may have seen before, but perhaps not at this resolution. I was able to see someone was indeed down at 11/12 spotting while PT work was being done at blister right before collapse. It would be very difficult to dispute tension was being added to get some type of mechanical result at 11/12. It is also easy to detect uneven lifting and rocking of span (not to say it was undue amount). And you can observe how many times in the two days before collapse, workers visited 11/12.

http://facilities.fiu.edu/projects/BT-904-PRR.htm

14. Recordings of Design Build team selection

This link has direct source of Linda Figg's pitch and other pitch from Facchina. Finally UCPP Deliberations 11.5.15.wav directly reveals FIU/Sweetwater efforts to get the fancy bridge they wanted at federal taxpayer expense.

 

[TheGreenLama]

I'd like to say that clearly, by re-stressing the temporary PT in Member 11, it was FIGG's intention to both 1) increase the vertical clamping force across the shear plane, and 2) reengage the PT rods in a wild attempt to quickly "add" steel across a failing joint that was suddenly found to be inadequately reinforced. But, sadly, I don't know FIGG's intention. No one does, yet. Prior to this reckless operation, the only things holding Member 11 to the deck were the Member 12/end diaphragm connection, and the unstressed PT rods laying loosely in the ducts of Member 11.

A question: If the PT rods were being engaged in shear at the time of re-stressing, would that have been apparent to the jacking crew?

Disclaimer: I just want to say that in no way do I approve of the above course of action.

 

[jrs_87]

Apparent actual superstructure designer of FIU bridge is one Eddy D Leon. He is fully qualified.

https://www.linkedin.com/in/eddy-leon-p-e-b524b3a6

Quote from his calculations document:

"The nodal zones of the superstructure were designed for interface shear transfer between the interface of the diagonals with the deck and canopy. AASHTO LRFD 5.8.4 was used. Shear forces at the nodal regions were [highlight #FCE94F]extracted[/highlight] from the LUSAS finite element model. The required size of the interface shear reinforcement was calculated for the critical region and provided conservatively to all regions." Highlight added by me.

The AASHTO LRFD Bridge Design Specifications: Section 5 Reorganization
by R. Kent Montgomery, [highlight #FCE94F]FIGG[/highlight], Dr. Shri Bhide, Bentley Systems Inc., and Gregg Freeby, Texas Department of Transportation

http://www.aspirebridge.com/magazine/2017Winter/AASHTO-LRFD.pdf

 

[Sym P. le]

GreenLama, I can't answer your question but I've often wondered what type of feedback the work crew was getting by means of their equipment. I wrote off thinking about it for lack of info on the procedure or monitoring of the structure. If the PT cables were pinched, it was at the far end so tension could still develop over the greater length of the cable and seem normal. I was thinking more along the line of the damaged #11 deflecting and not allowing tension to develop fully, i.e. just pulling the 10/11/canopy node down. Nothing would move much but the crew could be left scratching their head wondering why the tension wasn't coming up all the while pumping more energy into the structure. The reality, of course, was that this whole approach was just a pipe dream.

Some other thoughts:

From the Investigative Update, it is not unreasonable to suggest that #11 has shifted longitudinally w.r.t.

the deck or that #11 is no longer in contact with the deck for the purpose of engaging friction. My next question was what deformation is allowing these gaps? Is the deck sagging or rather how much? One thing is obvious, #11 is tortured and cannot be reliably modelled.

From the FIGG Presentation, it's acknowledged by FIGG that the photos do not even do justice for the gravity of the situation.

Is FIGG saying that they were up on the deck again just prior to walking into the meeting? This being almost two days after the photo above was taken? I really can't imagine what they were thinking. The presentation w.r.t. interim remedies is just bonkers, even the idea of reconsidering their analysis of an ideal structure in the face of obvious shortcomings defies rational thought when it comes to considering interim remedies. They just didn't want to lose face by admitting that a more obtrusive intervention was required and bullied everyone else with a hurry up offence while throwing caution to the wind.

But that's just my opinion.

 

[TheGreenLama]

Sym P. Ie, clearly the movement is longitudinal. From those other photos it's apparent that end tear out has already begun. What would be an interesting project for someone who's tech savvy in this area would be to generate a 3D model from those crack photos. Something similar to this:

https://www.youtube.com/watch?v=BLxJGFzX0Qc.

Maybe it's not possible with the few photos available, but if they could somehow be stitched together maybe we all could get that same "in person" sensation that FIGG noted in their presentation.

 

[SFCharlie]

"No. 3 - Those involved in the deadly FIU bridge collapse case will head to court today. A judge is ordering the company who designed the bridge to testify over why they aren't able to provide investigators the text messages they sent or received about the bridge in the days leading up to the collapse." That's the whole item.

SF Charlie
Eng-Tips.com Forum Policies

 

[jrs_87]

One thing that has concerned me about this bridge from when I first heard about it: Holding the casting form diaphragm base to be exactly relatively coplanar to the permanent pier landings would be absolutely required, and yet difficult to achieve. What if the form supports settled during casting or were not built with this in mind to begin with? The entire surface area of span was supported during casting, so I admit settlement would not be expected, but it should still be kept in mind.

Underside of north side diaphragm shown here. Note what appear to be four through holes for the 190 109 foot pylon.

 

[Sym P. le]

You've touched on one of the biggest unassuming issues with the placement and it screams from this perspective. The primary vertical load vector passes directly through the pad beneath #12 and yet there was no shim at this location. Instead the forces had to be carried through the slab, around those four holes to the laterally placed shims. This was apparently addressed early after the placement but I have not seen details of how this was carried out.

That was an oops they thought they got away with!

 

[Vance Wiley]

The lack of vertical support directly under node 11/12 was the main focus of the FIGG PowerPoint Presentation of March 15 meeting. They could not explain the cracking at the top of the slab so they proposed to place more shims directly under the node ( between the spaced bearing pads).
I think I recall a 1/2 degree limit for handling/moving the bridge. That is about 1-7/8" across the top of the pier. They would set forms much more level than that. Then there is the chance to level the tops of the bearing pads.
As to the outcome of the meeting, I am reminded of the Kenny Rogers song.
You got to know when to hold them
Know when to fold them
Know when to walk away
Know when to RUN!