What went wrong technically to cause the bridge collapse is only the second half of the story.
The first half of the story is that the stakeholders wanted an over-the-top bridge design, complete with fake cable stays, and out of hubris decided they would use an all-concrete single-truss Accelerated Bridge Construction (ABC) design because all-concrete would look better and the university's (FIU's) engineering department sort of specialized in promoting ABC.
What could have been a much cheaper nice looking normally-designed bridge (which would be open today) ballooned (or you could say, was hijacked) into a $10mm+ bridge disaster that would be comedy if people hadn't been hurt and killed. A majority of it was even paid for by federal grants.[1]
This is why we can't have nice things. They were so focused on building a nice looking bridge — and meeting deadlines — that they didn't/couldn't allocate sufficient resources to verify the design, verify it was built and moved into place correctly.
> stakeholders wanted an over-the-top bridge design [...]
Nothing wrong with that!
> They were so focused on building a nice looking bridge — and meeting deadlines
These are also good things.
To me sounds like the civil engineers started behaving more like software engineers. The independent peer reviewer they hired was not qualified to do the review.
> Louis Berger was not qualified by the Florida Department of Transportation to conduct an independent peer review and failed to perform an adequate review of the FIGG Bridge Engineers design plans and to recognize the significant under-design of the steel reinforcement within the 11/12 node, which was unable to resist the horizontal shear between diagonal 11 and the bridge deck.
This whole business of the engineers ignoring the cracks that the contractors kept needing reassurance about is also insane.
> To me sounds like the civil engineers started behaving more like software engineers.
Things like these were common in construction before software engineering was a thing. It always makes me chuckle when someone defers to "if we built buildings like we build software" trope. We do, and worse.
Lots of constructions standing around the world are not up to the codes in zillion different ways. They are still standing because redundant nature of physical construction is more forgiving than execution on a Turing machine.
> Things like these were common in construction before software engineering was a thing
I read in an architect's memoirs that when he was a student back in the 1940s whenever his Architecture faculty teacher and his students visited a construction site to maybe change some small details (or at least they were perceived as "small" by said architects) oftentimes they were met with very angry construction site workers who would make the above architects think again about trying to force the implementation of those small changes. A similar story was told to me by my father (also a civil engineer) as happening in the 1980s, though the level of implied violence coming from the construction workers was a little more subdued. Wish that we in the software industry would be like those construction workers from time to time.
You have the metaphor wrong. You're the software engineer. Your toolset is the angry (or not) construction worker. Build a toolset that yells at you when you want to do something stupid/dangerous.
> Wish that we in the software industry would be like those construction workers from time to time.
A cluster of burly construction workers holding spanners is probably more physically intimidating than whatever we software developers could typically muster.
A friend who worked in construction in NY city told me a story of how, on a project, they had one section of rebar sitting on the ground. When the construction inspector came to inspect the site, they would put the piece of rebar near where they were pouring concrete to give the impression that they were forming the concrete with rebar in it. They were not (I presume to save time and/or money).
We have that 12-story municipality building here, standing since 1970s. An inspection last year has found it doesn't have even half of rebar necessary in the pillars, so they had to evacuate it.
>> stakeholders wanted an over-the-top bridge design [...]
> Nothing wrong with that!
Well, in this case, apart from the structural problems, the bridge just looked cheap and kitschy because of the fake cable stays. So yes, there was something very wrong with that, but this problem is so common in contemporary architecture that we don't seem to recognize it anymore. There is just no way for an over-the-top designed building to have any dignity.
Except that it's not private money. It's public money.
This is the bridge equivalent of a town using a state grant to buy a Cadillac for the code inspector to drive around in. There is no place for that kind of behavior with other people's money.
If the university wants to build a bridge with all the bells and whistles with their own money that's fine. When they do that with public money there is an opportunity cost of the other improvements to society that money could have bankrolled. To claim that it is ok to spend public money on frills like this is to claim that the net difference between a boring pedestrian bridge and a fancy one (i.e, the "frills") is more important to society than the other things that money could have done.
"What could have been a much cheaper nice looking normally-designed bridge (which would be open today) ballooned (or you could say, was hijacked) into a $10mm+ bridge disaster..."
This exactly parallels the disaster that is the new span of the (SF-Oakland) Bay Bridge. What would have been a $400M fully-functional replacement instead became an $8 billion art piece that is less safe and needs far more maintenance.
> The first half of the story is that the stakeholders wanted an over-the-top bridge design, complete with fake cable stays
To be fair, you can make that criticism of almost every monument and famous building in the world. If the Sagrada Familia didn't have all those over-the-top arched ceilings and all those do-nothing decorative carvings, it'd have been finished decades ago!
"The structure of the chapel, built in a great hurry in the 1470s, had given trouble from the start, with frequent cracks appearing. At Christmas in 1525 a Swiss Guard was killed when entering the chapel with the pope when the stone lintel to the doorway split and fell on him. The site is on sandy soil, draining a large area, and the preceding "Great Chapel" had had similar problems."
This bridge collapse was featured in an episode of Engineering Catastrophes a few weeks ago.
According to the narrator, although FIU has a department for ABC they weren't consulted nor involved with the bridge project (and were not happy about it?). No idea how accurate that statement by the narrator is.
I attended a US engineering university that built a new engineerng building with extensive solar panels. The building was fine, in general, but the solar panels never worked as intended. The school's engineering professors always made it a point to emphasize that they were not consulted in the planning of the project. :-)
This is similar to the San Francisco Bay Bridge replacement plan. Instead of the basic designs that were floated a radical design was chose for aesthetic reasons and ended up being damaged during the tensioning process. Now an over the top design that cost at least twice what it should have is unlikely to serve its purpose of providing for earthquake safe transportation over the long term.
Municipalities really need to carefully consider the balance of aesthetics, costs, and risks.
One of the problems was the original design Caltrans developed without any input from Oakland and San Fransisco was aesthetically an ugly piece of shit. Unsurprisingly the mayors of SF and Oakland wouldn't sign off on it.
Part of the reason for that is Gray Davis wanted the bridge construction to start as soon as possible because he wanted to show the voters he was competent and effective instead of the hapless fool he is.
The originally proposed design was very basic but also pragmatic and likely to endure. When that design was rejected teams were assembled to come up with a range of designs and pick one that would be suitable. The group that chose the winning design later claimed that they had no comprehension of the design risks and had only renderings and optimistic initial cost projections to work from.
There was a process that involved many people and it failed.
The bridge was supposed to self-support without them during assembly, without a hurricane or marching band. This does not mean the cable stays were entirely decorative. They had a structural purpose, primarily to change the resonance frequency to be outside the range of ordinary human motion. They were intended to bear some of the load.
To me there were obvious problems, particularly the sharp non-articulated corners in the the concrete at the truss edges (high stress points), but of course my concerns would have been dismissed prior to the collapse. Lots of people, with even minimal understanding of mechanical design, should have been alarmed. Maybe that was the point of the design: make something that looks alarming and thus dramatic.
You could go back even further. A bridge collapse was made possible by the building of a bridge. So you could go back to the moment when the do-nothing alternative (cost: $0) was rejected in favor of building a bridge. Well why was that decision made? Let's keep going back. This is where you start running into certain pachyderms in the room. The problem being solved with a bridge, is the existence of 8th Street, a soul- and flesh-destroying 8-lane traffic monstrosity that apparently is sufficiently inconvenient, intimidating and dangerous for human beings to cross unassisted, that a bridge starts to look like a good idea. Seems like the people who allowed 8th St. to happen deserve some of the blame, namely backward-thinking urban "planners" who think it's still 1950, haven't heard of (lalala I can't hear you) induced demand, peak oil or climate change, and who have collectively been sold a bill of goods, and who don't have the imagination to think up any ways of creating "economic activity" in the town that don't destroy the town. This species unfortunately infests the entire continent and kills 35,000 people a year, so I don't expect them to lose sleep over these 6.
Expertise only matters if you're hired, at least hired to review the calculations of the engineers: "the probable cause of the … bridge collapse was … calculation errors made by FIGG Bridge Engineers, Inc." and "Factors in the collapse included … inadequate peer review".
The reason we can't have nice things is that when idiots come along and take a budget that can cover a boring bridge and boring testing then piss it all away on a fancy bridge nobody needs they don't get punished. If anything they get rewarded.
The only reason this got noticed is because people died and bodies create opportunities for all the parties that should have stopped this stuff before it started to score virtue points and look like they're doing something cleaning up the mess.
If the (metaphorical or literal) tar and feathers came out every time taxpayer money was spent on a boondoggle like this you wouldn't have nearly as many boondoggles like this.
> The construction and inspection firms working on the bridge were aware of the cracks and reported the cracks to the design firm, asking for guidance. The engineer of record at the design firm repeatedly indicated that the cracks were of no safety concern.
Welp, that's a lawsuit laid out cold right there, setting aside even the fact that
> the probable cause of the ... bridge collapse was the load and capacity calculation errors made by FIGG Bridge Engineers, Inc.
> Errors in design may occur, but systems should be in place to catch those errors when they do occur. In this case, a firm was hired to independently review the bridge design for errors. However, [...] Although the design reviewer recognized that he should have examined the nodes and stages, he indicated that there was not enough budget or time to evaluate those factors. Contributing to this review failure was the reviewing firm’s lack of qualification to do this work.
Like even with software there's a point where I tell folks "Look if we're not going to do this X far, it's not of any value to do it half way and claim we did anything here."
You'd think there would be some engineering standard where "No this isn't a legitimate review because we didn't do enough." and that might set off some alarms...
There is. After the NTSB final report comes out next month I expect both investigations for taking away PE licenses and probably also criminal charges for gross negligence.
The engineer of record at the design firm repeatedly indicated that the cracks were of no safety concern.
I've heard many other examples of viciously denying that anything is wrong, all the way until something becomes very very wrong, to wonder whether this is an inherent part of human nature or a futile effort to avoid liability for knowingly screwing it up.
It's called 'the Normalisation of Deviancy' e.g. [0]; there's quite a lot of recent literature about it. Basically a risk is taken and nothing goes wrong; as a result the risk gets repeated more and more often with less and less concern until things do actually go wrong.
It very much seems it is, though what specifically drives this may be complex.
Effectively, playing with apparent risk/reward, in the sense of overstating or understating the potential benefits or costs/hazards of an activity or phenomenon, is a form of rent-seeking. It allows an asset to appreciate (or a risk to be discounted), without doing any real work.
Whether that's done through a deliberate calculation, through inherent psychological bias, through group or peer pressure, or other mechansism, seems to vary. But you'll find this across activities, technical domains, space, and time.
It's a major element of Bernhard J. Stern's "Resistances to the Adoption of Technological Innovations", from 1937, which looks at the history of numerous innovations across multiple areas. You'll find this repeatedly in all kinds of technical and technological disasters: Chernobyl, Fukushima, Three Mile Island, numerous aircraft disasters, military campaigns, entire wars, business decisions, and any number of mundane accidents. In the use of lead, asbestos, tobacco, CFCs, and CO2.
One common element is that risks, whilst occasionally somewhat apparent, are not unambiguously manifest. Failures are almost always highly nonlinear. At one moment, things are mostly OK, though there are possible signs of trouble. Moments later, everything is gone to hell.
Up until that point, it's possible to argue otherwise. Occasionally even past it.
Human nature to avoid liability for screwing up - knowingly or not. You can see it even in little children who, having broken something, will deny they did so and invent excuses up until they realize the parent isn't buying it.
It also makes sense both intuitively and mathematically. If you screwed up and don't admit it, there's a chance (say X) nobody will find out (that it was you, or that anything was screwed up in the first place). Conversely, there's a (1-X) chance it'll be found out and you'll pay the costs - which my be huge and very personal, including tarnished reputation, monetary damage or even jail time. Now, if you admit to the error yourself, you get 100% certainty of slightly lower punishment. As long as the "slightly lower" isn't lower than (1-X), it makes sense to keep quiet mathematically. And of course, once you add fear into this, people tend to overestimate X.
Consequences to others turn this from a small to huge moral issue, but the mechanism is still arguably dominated by fear. That's why punishments are a tricky thing. Threat of losing license is a good deterrent from screwing up, but also a good motivator to hide your failure if you screwed up.
I suspect that the engineer did think along these lines, although there was probably also some irrational denial that there was anything wrong. What's missing from this "calculus of professional risk", however, is any consideration of the danger to other people, and that is what makes it so unethical.
There is no practical alternative to taking away a licence in cases like this, at least temporarily in less egregious cases. What meaning would the licence have if the person responsible here retained it? He has demonstrated an appalling lack of judgement which goes beyond this one disaster. Removing the threat of revoking the licence would not fix the incentives issue, as even with that off the table, the engineer is to blame for the cost and delays of his failure, even if he does not care about the risk to other people.
Here's an example of an engineer doing the right thing - though the prospect of failure was so appalling that it probably focused his mind on what mattered most.
None of these names nuclear power plants that are "operating beyond their design life and keep (sic) increasing their allowable crack tolerance". There's no mention of crack tolerance at all, just references to containment vessels and caps requiring maintenance for radiation damage.
I just posted these from a quick 30s search. I'm not sure of the link for the post I've read about the concrete cracks in specific, but I believe it was about a year ago that I read it. Regardless, you can see the state of affairs and how the need for power generation and the desire to keep costs low leads to compromises in safety precautions.
After watching Seconds from Disaster and Air Crash Investigation for many years, USCSB videos are now my favorite accident investigation documentaries.
I don't get this. There were large, steadily growing cracks in supporting elements, observed and documented over multiple days. Why on earth wasn't the street closed down?
The bridge slowly and visibly announced its impending collapse, which - as I understand it - is exactly what engineers want. The thing nobody wants is a bridge like the Morandi bridge in Genua, where the signs of impending collapse were hidden under thick layers of concrete, until the structure suddenly and unexpectedly gave way.
In all seriousness though, I remember watching that video soon after the event. I have to give him credit, he came pretty close to the right conclusions just from a few photographs
A fake cable-stayed 175-foot span pedestrian bridge for $9 million.
For less money, they could have had a real cable-stayed bridge, like the one over I-280 near Apple HQ.[1] That cost $14 million, but it's a 375-foot clear span over the freeway.
iirc all the bids for a concrete version (ala Florida) were twice as much for the steel design that was used, otoh California version also has to be earthquake resistant. I gotta give credit to all the engineering firms and Caltrans oversight. Very little downtime on 280, that would have been a traffic nightmare for entire peninsula.
And then there's this accidental destruction of the phone of the engineer who stamped the plans for the bridge.
"Denney Pate, who stamped the plans for the bridge, told a judge June 12 that his wife accidentally put his pants with the phone in the pocket into a washing machine, inflicting damage that destroyed any call records or images."
10. Train your staff on the proper use of Pc (the permanent net compressive force normal to the shear plane) when calculating nominal interface shear resistance.
Pretty humiliating. I wonder how bad the damage is to FIGG.
I lived on the corner of 7th Terrace and 109th AVE during my senior year. I could see the bridge being built from my bedroom window; I would cross 8th Street to go to school everyday.
The bridge collapsing was one of the most surreal experiences of my life.
It sometimes fascinates me (and scares the shit out of me) to wonder at all of the little errors/assumptions like this that are baked into the millions of spreadsheets that form the foundation of so much of what we build and do in the world.
Dead comment reply by 'FakeComments' that I am not sure deserves to be dead (what happened there?): "I’d say the better part of engineering is knowing how to design in the face of small errors creeping into the project."
The world is held together with nothing but human thought. Man must be free so that his mind is free to think. A doctor working under compulsion should terrify those under his scalpel. Likewise for the engineer, and his customer, who designs a bridge, an elevator, or even a simple toaster.
FIGG is one of the premier bridge designers in the country. You can probably get on one of their bridges within a couple hours no matter where you live.
If I'm reading the report correctly, the bridge was barely sufficient to support its own weight, which is extremely surprising since I remember reading a very long time ago that bridges and other structures are designed to be 10x stronger than they normally need to be; a vague reference in the article https://en.wikipedia.org/wiki/Factor_of_safety supports that. It seems like this bridge was designed for minimal cost...
No, bridges (usually, but it depends on countries/national regulations and on the nature of the structure) have an overall design factor of safety between 1.2 and 1.8, in some cases 2, but the actual presumed loads are already increased (i.e. the calculation loads are more than what will actually happen) and there are some "implicit" safety factors in the use of materials, so that the overall "real" factor of safety usually is between 2 and 3.
The highest required factor of safety in engineering is AFAIK 5 and is used for things like lifts, cablecars and similar.
Factor of safety in the way that most people understand it has gotten increasing difficult to tease out since the switch to load and resistance factor design (LRFD) over allowable stress design (ASD) for most structures. I think timber is the only one which still defaults to ASD for permanent construction.
But the short version of the story is that for LRFD there are two main factors which affect “factor of safety” in the traditional sense (eg - structural capacity / structural demand) which are load factors and resistance factors (as you’d guess). Load factors increase applied forces while resistance factors decrease assumed strengths. Load factors vary depending on the type of load but typically they’re between 1.2 to 1.6. Resistance factors vary based on failure mechanism under consideration and are between 0.6 to 0.9 for structural work.
So the range is 1.2/0.9 to 1.6/0.6 or ~1.35 to ~2.75.
There are secondary sources of additional factors of safety. For example, material properties assumed in design are the minimums that are allowed and so chances are you’ll get something on site that exceeds by 25-50% or so because nobody wants to deliver material that doesn’t meet the minimums and open themselves to claims liability. There’s also a redundancy factor which is touched on in the report where loads are increased if no redundant load path exists, for concrete there’s some additional caps/limitations on a couple values due to time effects like creep, etc.
I would say most modern construction has an overall factor of safety of 3-5 with bleeding edge construction sitting at around 2 and public works construction (utilities or highways - things where architects aren’t involved) at over 5.
This is totally hearsay, but I've heard from multiple older engineers that as computer models have been improved over time the safety factor has been reduced as we have more accurate simulations. If the safety factors from the '50s had been used it's possible this bridge would never have been designed the way it was.
The (approximate) rule of the thumb is that anything in reinforced concrete calculated by "hand" has an "implied" additional safety factor of 5-10%, mainly due to roundings by excess in the calculations, at least this has been my experience in the mid 1990's, when for a couple of bridges existing hand made calculations had been re-made with computers.
It is a small (but good) thing, though it would have IMHO not been enough to save this particular bridge, here the issue is evidently a wrong static model, or - less likely - a wrong assessment of loads.
Besides the initial (huge) calculation error, what is really preoccupying is that they didn't consider as serious the cracks, which as seen in the video/photos on youtube are the typical (severe) shear cracks, that should have been ringing all alarms.
The poor structure, as often happens, did all it could to "signal" the issue, what I find bordering being criminal is ignoring or severely underestimating those clear signals.
I've noticed from videos of recent bridge failures (FIU and the Taiwain Arch Bridge collapse a few weeks ago) how fast the failures occur. Contrast with, say, the Tacoma Narrows bridge failure in the 1940s.
I suspect current designs are far nearer critical tolerances. Even under the ridiculous dynamics of the Tacoma Narrows incident, ultimate failure took hours.
I study at FIU... I just don't see the point this bridge. Why would you spend 10+ million dollars on a bridge to cross a street, when the same problem can be solved by a simple crosswalk? We already have one of these bridges... and the majority of people don't bother to use it and simple take the cross walk. It just such a waste of money.
To put it into perspective, 8th street is an 8 lane street at that intersection, with cars driving past 50 and 55 mph all the time. Not to mention that there are several on/off ramps for the Florida Turnpike right next to it. I would be scared of crossing the street there to be honest. If I remember correctly there had been several hit and runs of students too.
The bridge was connecting FIU with a large student residence. Negligence during construction notwithstanding, to say that a simple crosswalk was enough is disingenuous.
This is the meta story. The campus is my city is the same way, bordered by a six-lane arterial with poor pedestrian connectivity so cars can speed as fast as possible through the high-pedestrian area. As long as North American cities put the car first, people will die whether from bridge collapses or car crashes.
As I’ve posted before, there’s nothing special about a guy with a PE other than him hanging out there with liability for the lawsuit.
Someone with a PE license has met the minimum requirements from the state. Nothing else can be garnered from someone possessing a license. The minimum requirements have, historically, resulted in engineers who put out public facing designs that, more often than not, do not bring harm to the public. I can tell you first hand that sometimes the guy with a license who is actually at “the minimum” does not always put out something that is safe.
Indeed. You can browse through the disciplinary actions taken against engineers by professional associations and you will find all sorts of incompetence and malfeasance. I find the discipline tends to be on the lenient side—lots of fines, suspensions and reexaminations with only the occasional permanent loss of license.
Still, I'm glad professional licensure is required. That's what will prevent these folks from quietly rebranding and designing another disasterous bridge tomorrow.
After employing various PEs on various aspects of new buildings, I can't say I'm surprised. Never a straight answer when a convoluted one is available. Never a helpful creative suggestion. Rarely a thought for the needs of the end user. Rarely a plan drawing that clearly explains what contractors are supposed to do. Even worse, never a clear answer to the questions inevitably inspired by those inadequate drawings.
Yes, and (disclaimer: this is part of my job) often structural engineers have a very "narrow" view which summed up with the (as well often "narrow" views) of the architects, of the plant engineers and of other technicians/designers that are involved in a project create the need of a coordinator/facilitator that helps in the communication among these people, checks the (sometimes not-so-obvious) possible conflicting parts of the various drawings and "translates" the design for the contractor.
Particularly lately I have seen an increase of (otherwise very good, smart and qualified) engineers and architects with no or very little on-site experience, that produce beautiful and valid drawings/designs that cannot be "as they are" realized in practice.
Agree. Seems a PE should be required to have a few years experience in the construction field before they are licensed. So much friction in the process would be avoided if the professionals writing and drawing the instructions for a structure had actually performed the work, or physically seen day to day challenges prior to having responsible charge. Structural engineers would be less arrogant as well. I see this happening on a daily basis.
They should care because https is not just about protecting secrets. Thinking that is the case is a common misconception about https. Among other things having https also makes it harder for a bad actor to deliver false content while intercepting requests to the site.
The first half of the story is that the stakeholders wanted an over-the-top bridge design, complete with fake cable stays, and out of hubris decided they would use an all-concrete single-truss Accelerated Bridge Construction (ABC) design because all-concrete would look better and the university's (FIU's) engineering department sort of specialized in promoting ABC.
What could have been a much cheaper nice looking normally-designed bridge (which would be open today) ballooned (or you could say, was hijacked) into a $10mm+ bridge disaster that would be comedy if people hadn't been hurt and killed. A majority of it was even paid for by federal grants.[1]
This is why we can't have nice things. They were so focused on building a nice looking bridge — and meeting deadlines — that they didn't/couldn't allocate sufficient resources to verify the design, verify it was built and moved into place correctly.
[1] see page 9 of https://www.transportation.gov/sites/dot.gov/files/docs/brie...