striped line
Department of Civil Engineering
FIU Bridge Collapse—Hard Lessons
set cell vertical height

Future of Bridges | Florida bridge collapse has some hard lessons


BY MARK HANSFORD - 16 DEC, 2019

A litany of failures ultimately combined to cause the collapse of the Florida International University (FIU) Bridge onto a live highway, according to the official investigation into the accident.

Key companies and organizations involved in the project were criticized in the final report into the collapse published by the US National Transportation Safety Board (NTSB) (New Civil Engineer last month).
Calculation errors
It concludes that design calculation errors made by Figg Bridge Engineers were ultimately to blame.

But failures by the independent design checker, client, contractor and on-site construction supervisor also contributed to the disaster, which killed six people (see box). All had the chance to act when serious cracks appeared in the structure in the two days before the collapse, but none did.

So this month we ask: what can we learn from this failure?


Tony Gee & Partners group director of structures Akram Malik, has spent 20 years examining why structures collapse – particularly as a tool for learning. So who better for New Civil Engineer to ask to review the FIU collapse?


 

The structure included a 53m long prefabricated main span made up of a 5.5m deep concrete truss of 12 diagonal members. The truss supported the 10m wide deck and a 5m wide canopy running along its length. The main span connected to a side span around a central pier onto which a 33m tall central pylon was to be built. This was an architectural feature with 10 diagonal steel “pipes” hanging from it to provide additional stiffness but no other structural function.


It was at the connection between the two spans and the pier – known as 11/12 nodal region – that the failure occurred.


The client for the bridge was Florida International University. It hired MCM as design and build contractor, and MCM hired Figg as designer and to serve as engineer of record. Figg then contracted consultant Louis Berger to conduct the required independent peer review of the structural design. FIU also hired Bolton Perez as construction engineering and inspection contractor to monitor and inspect the bridge as it was constructed. Florida Department for Transportation (FDOT) is the highway authority responsible for SW 8th Street, the road onto which the bridge collapsed.

 

Independent checker role


Malik believes there are key lessons to be learned, the clearest concerning the role of the independent checker.
“Independent checking when required should include every critical stage of construction,” he states.

“Here, we’ve got Louis Berger who was engaged by Figg, and Figg only asked it to look at the final design, not the interim construction stages – and both failed to see that the intermediate construction stages were actually critical in the failure”.

 

At face value the independent peer review process was good. The structure was deemed a “category 2” structure, for which the required checks are similar to the UK Highways England’s “category 3” process. “So a proper check was required,” notes Malik, although he adds that the exact requirements were unclear from documents he has seen.

Computer image of Florida International University Bridge

 

What is clear, from the NTSB’s report, is that Louis Berger dropped its price from $110,000 (£84,000) under pressure from Figg whose budget for this work was £46,500. Emails obtained by investigators show Louis Berger warning Figg that in seeking lower bids, FIU should be aware that “a lesser fee may be associated with less effort/value”.


More emails show that on contract award the original scope of work remained unchanged, but that Louis Berger had reduced the time it was prepared to spend on checks from 10 weeks to seven.

 

Peer reviewer requirements not met


Investigators also found that Louis Berger failed to meet FDOT requirements for a peer reviewer to employ at least three registered professional engineers, each with a minimum of five years’ experience in designing complex concrete bridges.

Louis Berger performed the independent peer review of Figg’s design using Adina, a finite element analysis software program, in accordance with FDOT standards. Post-collapse, Louis Berger confirmed to NTSB investigators that it analyzed the design as one structure in its completed state. It only analyzed the design for the completed structure and not for its various construction phases.

“Doing construction sequence staging analysis was not part of our scope. Doing such an analysis requires much more time than what we agreed about [with Figg],” it told investigators.

The significance of failing to consider the structure in its construction configurations became clear when investigators began examining the construction sequence, explains Malik.
The entire 53m long main span structure was built off-site and then moved into its final position on mobile transporters where it was set on temporary shims.


Changing support system


Ahead of the move, the two end diagonals of the truss were tensioned to ensure they remained in compression while in transit (see diagram).

In the casting yard there was full contact support across the deck and its two ends, yet while resting on the pier the shims were placed either side of the central truss – meaning there was no direct support under the truss location.

This induced unexpected stresses and strains in the structure which were exacerbated when the end diagonal truss member was de-stressed after it was positioned. This triggered the significant cracking which should have given warning that collapse was imminent. Such a critical construction stage should have been modelled by the checking engineer, Malik concludes.
A second key lesson, says Malik, is the old-adage of the devil being in the detail. Investigators have overwhelmingly concluded that the trigger for the collapse was punching shear failure at the node where the end vertical and diagonal truss members met. The NTSB has been unequivocal in stating that Figg’s load and capacity calculation errors at this key node probably led to the failure. Post-collapse calculations showed that the load demand on the node was twice that designed.

 

Error went unnoticed

 

Louis Berger, as design checker, should have noticed this error.

In a post-collapse interview, the Louis Berger engineer who conducted the peer review told the NTSB that the agreed budget did not stretch to this level of analysis.
“In the beginning, I suggested to do this kind of analysis, to analyze the connections,” he said. “I’m talking about the nodes, or the joints to analyze the connections. However, the budget and time to do this actually was not agreed upon with the designer.”

“Clearly in this case no-one looked carefully enough at what was a really critical location,” says Malik.
Malik points to a host of obvious deficiencies, including decisions to locate a drainage pipe and run several cable ducts through the critical 11/12 nodal region and construction decisions to cast the structure with cold joints at the base of the diagonals. He also questions the quality of preparation of the construction joints.

“You have to look at the construction assurance – was it wise to have the cold joint in such a critical zone, that’s one question, and was it prepared correctly? And we’ve got [on-site supervisor] Bolton Perez Associates – did it do its job? These are all questions that are a little bit up in the air,” he says.

Retensioning to close cracks

 

One of the most alarming aspect of this collapse was that the only action taken to stabilise the structure – even when cracks up to 100mm deep starting appearing – was to re-tension cables in the end diagonal in the hope that this would close the cracks.

Again, here there are clear lessons to learn, says Malik. “In this case it appears that Figg didn’t recognize that the structure was on its way to collapsing.”
“It carried out calculations and those calculations seemed to show that it was working fine and that the cracks should not be occurring.

“So it freely admitted that it didn’t understand why the cracks were occurring, but on that basis still decided to restress the bars to try and arrest what was happening – even though it didn’t understand it. This was poor decision making,” he says.

Malik suggests that engineers should be educated in the pre-collapse behaviour of structures. “Do we need to include in our design processes an additional question: ‘What are the ways in which this structure could fail?’”
Malik’s final observation concerns the safety management of the remedial action being taken – restressing of the diagonal member above live traffic.
“So… [site workers] were restressing the bars in a critical member on a non-redundant structure and were doing so over live traffic. Obviously it was not a good decision in hindsight but even in foresight it was not a good decision,” he asserts.

And here he believes all parties should have stepped in. But why didn’t they? “Are we becoming too specialized?” he asks. Builders build, designers design, supervisors supervise. Is this preventing them from “exercising their own independent professional judgement”?

Figg and MCM, Louis Berger and Bolton Perez have been contacted for comment.

Investigators View: Bruce Landsberg - NTSB vice chairman

 

A bridge-building disaster should be incomprehensible in today’s technical world. We have been building bridges in the United State for over 200 years, and long before that in other parts of the world. The science should be well sorted out by now – and for the most part, it is.

The investigation clearly highlighted basic design flaws and a complete lack of oversight by every single party that had responsibility to either identify the design errors or stop work and call for a safety stand-down, once it was clear that there was a massive internal failure.

The “what” is very clear but the “why” is more elusive. Despite the public’s anger, distress, and disappointment, none of the responsible organizations had any intent for this tragic event to occur or to cause any injury or loss of life. Sadly, good intentions do not suffice for competence and diligence.

Engineering schools will use this as a landmark case study for years – and they should.

The Engineer of Record employed by Figg was experienced, but his calculations were erroneous. Reflection on this event should go far beyond merely a technical review. The checks and balances that were required by the Florida Department of Transportation (FDOT) and American Association of State Highway and Transportation Officials guidance and incumbent upon Louis Berger, the peer-reviewing organization, were completely lacking.

Louis Berger lowered its bid to review the project by 43% in order to get the business, but also reduced the scope of the review. The reason given was there wasn’t enough money in the project to cover its efforts. That’s both disingenuous and unconscionable. It also was in violation of FDOT’s requirement that there be an independent second set of eyes to review everything – not just what was economically convenient.

It is likewise incomprehensible and unethical that Louis Berger would even bid on a job for which it lacked the requisite qualifications (see main feature). That FDOT, which was supposed to review the plans, did not know, or think to ask, about its qualifications is more than just an oversight. It’s just plain sloppy. Ditto for Figg.

FDOT claimed a technical error on the FDOT website and then, after the collapse, fabricated a disclaimer that it is not responsible for the data that it posts. That’s unacceptable in my view – either ensure the information is accurate or don’t post it.

The bridge was not properly designed, and there was no qualified oversight on that design. When the inevitable began to happen – a creeping, catastrophic material failure, nobody did anything, despite what NTSB chairman [Robert] Sumwalt accurately described as the “bridge screaming at everyone that it was failing.” Why?

Once the cracking became evident, not one of the organizations involved was willing to take the essential and unpopular step to call a halt and close the road.

This is similar to the circumstances of the space shuttle Challenger disaster where the decision was made to launch in extremely cold weather. The engineers recommended against it because the O-Rings that were critical to fuel system integrity would be operating outside their design parameters. Rationalization, optimism and schedule pressure contributed to what has been described in management training circles as “Group Think.” Strong and confident personalities persuade everyone that everything will be OK. Despite misgivings and technical expertise that advise against such action, the team moves forward as a group.

It appears that the same mindset was in play here, in every organization: Figg, Louis Berger, MCM (the construction company), Bolton Perez (the engineering firm overseeing the bridge construction), FDOT, and finally, Florida International University. It also appears that every organization absolved themselves of responsibility by rationalizing that if the Engineer of Record says it’s OK, it must be OK, and if anything bad happens – it’s on him. That is not the intent of peer review or safety oversight, and certainly fails the system of checks and balances in place to prevent catastrophes like these.

The NTSB’s stated role is not to lay blame, but some would say that’s exactly what we do when we apportion causation. The human failing that affects all of us is complacency. It is not a term the NTSB uses often, but in my opinion, it is present in nearly every accident and crash. We are creatures of habit, and when we become comfortable through long repetitive experience, the guard often comes down – periodically with disastrous results. This is precisely what safety management systems are designed to prevent – to trap errors in process before they become catastrophes. While disasters may be perfectly clear in hindsight, the best organizations take proactive measures – constantly.

Schedule pressure, economics, overconfidence, and complacency all work to counter good intentions and too often create tragedy.
It is my fervent hope that the organizations involved will take the NTSB recommendations (see box) seriously and quickly implement them. The lives lost and the families disrupted deserve at least that much.

 

 

Key recommendations from the NTSB report

 

The National Transportation Safety Board made the following key safety recommendations, primarily to the Florida Department of Transportation:

  • Require that the independent peer review for category 2 bridge structures includes checking and verifying the design calculations used for all nodal forces

  •  Require the independent peer-reviewed to submit a prequalification letter showing that it is qualified in accordance with Florida Administrative Code

  •  Specify that when structural cracks are initially detected during bridge construction, the engineer of record, construction engineering inspector, design
    and build contractor, or local agency that owns or is responsible for the bridge construction immediately close the bridge to construction personnel and close the road underneath; fully support the entire bridge weight using construction techniques that do not require placing workers on or directly under the bridge during installation; and restrict all pedestrian, vehicular, and construction traffic on the bridge until the complete support is in place and inspected

  •  Require personnel to monitor and inspect all local agency bridge projects determined by the department to have uncommon designs

  •  Add a discussion about redundancy to the Structures Manual, Structures Design Guidelines, emphasizing uncommon bridge designs


Additionally it recommended that the American Association of State Highway and Transportation Officials and Federal Highway Administration:

  •  Develop a requirement that concrete bridge structures be designed with reasonable estimates for interface shear demand

And to Figg:

  •  Train staff on the proper use of the permanent net compressive force normal to the shear plane when calculating nominal interface shear resistance.


This website was originally developed by Charles Camp for CIVL 3121.
This site is maintained by the Department of Civil Engineering at the University of Memphis.
Your comments and questions are welcomed.

striped line
Department of Civil Engineering | 104 Engineering Science Bldg | Memphis, TN 38152 | Phone: 901/678-2746 | Last updated: 10/23/2024