Hacker News new | past | comments | ask | show | jobs | submit login
Are we underestimating the vulnerability of bridges? (stanford.edu)
92 points by smb06 on May 1, 2017 | hide | past | favorite | 58 comments



That's the beauty of facts and engineering.

We may overestimate or underestimate. Nobody has built the vast network of bridges that we have before to the same standards, so operating experience should result in changes in our understanding of what and how we assess them!

If we look at the bridges that have survived 100, 150 or more years, they are mostly massively overbuilt due to limitations of the era they were built in. Perhaps some of the more modern spans are less overbuilt and are thus more vulnerable to environmental or other changes?


The biggest threat is politicians (and voters) who don't understand that "structurally deficient" means we need to invest in inspections, repairs, and replacement.

The I-35W collapse was near my house at the time and the fact that Governor Pawlenty had long pushed for reduced government spending (inspections in particular) was not lost on many people. His replacement, Governor Dayton, was partly elected due to promises to rebuild/repair even if taxes had to be raised to fund it (what a radical theory!).


Even with unlimited budgets for inspections, repairs, and replacement, if we do not understand and anticipate the failure mechanisms, we will have unanticipated failures.

Specifically for I-35W, "photos from a June 2003 inspection of the bridge showed gusset-plate bowing" but apparently the inspectors didn't realize they needed to be watching for that.

"...inspectors did not routinely check that safety features were functional."

https://en.wikipedia.org/wiki/I-35W_Mississippi_River_bridge...


Absolutely. I didn't mean to imply that Governor Pawlenty was responsible for the collapse - it was a combination of engineering, not funding inspections/repairs/replacement adequately, and bad luck.


Mainly it was that they had parked 270 tons of construction equipment and concrete on top of the bridge, and that they had added 2 inches of pavement over the years. In addition the original design or construction was deficient and some metal plates that should have been 1 inch were 1/2 inch.

I drove over that bridge not long before it collapsed, and marveled at how they had concrete and trucks lined up from one end to the other in a closed lane.


> added 2 inches of pavement over the years

yeah that's what scares me most, infrastructure is built to the cheapest bidder that can satisfy safety requirement that often don't account for aging

I've recently passed under two (!) overpass that failed at different times due to lack of maintenance, and every time I'm in queue on top of a bridge it scares the heck out of me watching everyone tailgating each other instead of leaving some margin to the car ahead.


>but apparently the inspectors didn't realize they needed to be watching for that.

Could we use neural nets to help the inspectors?


Sorry but I have to: https://xkcd.com/1831/


Hey, with some domain specific knowledge maybe we could teach it what the various failure modes of the materials used look like.

Have drones with webcams circle the bridge. It would surely speed up the process and may alert to something the humans overlooked.


If I build a big tower, and it falls down and injures or damages, I'm liable, right?

If government-owned infrastructure falls down (because of lack of funding) and injures or damages, is anyone liable?

Not a rhetorical question


IIRC for 35W there were lawsuits against the engineering company and department of transportation.

http://www.mprnews.org/story/2012/11/14/law/i35w-bridge-sett...

"After the structure fell into the Mississippi River during rush hour Aug. 1, 2007, killing 13 people and injuring 145 more, Minnesota lawmakers set up a compensation fund for victims. As part of that legislation, they nullified a statute of limitations that would have prevented the state from seeking damages. In May, the nation's highest court declined to hear Jacobs' challenge to that law, and the company had little recourse.

State Sen, Ron Latz sponsored the legislation that set up the compensation fund. While the state recovered from Jacobs and another firm less than half of the nearly $37 million paid out in compensation to victims, Latz said it was the right thing to do, not just for the victims and their families but also for Minnesota taxpayers."


Sure, why not? The engineering and construction companies don't have special immunity. Not sure if you can sue the government or individuals in in it. For the i35 bridge case, lets see...

From Wikipedia "URS Corporation [...] settled for $52.4M [...] URS had performed fatigue analysis consulting on the bridge for the Mn/DOT."

and "Jacobs Engineering Group, the successor of Sverdrup & Parcel, the firm that designed the bridge [...] paid $8.9 million in Nov. 2012 to settle the suit"


Not good enough. If some bridge is decades old, I don't think it's fair to expect some firm that is the descendant of whatever firm designed it so long ago to be liable, especially when all the people ever involved in design and building it are probably dead or close to it.

The real fault is with the government, and especially the elected politicians who don't run the government properly and get these things inspected and fixed (and worse, run on platforms of shrinking government). Why aren't they ever personally liable? They should be.


I'm sure limitations on liability is a heavily legislated and litigated area of law. I have no idea about the details, and am thus am not supporting any particular position, merely relaying some info.

I don't think engineers are being prosecuted for criminal negligence in cases like this - rather, the firms have civil liability. As far as holding politicians personally liable, they can be prosecuted if they break the law. Other than that, I envision it would be chaos if individuals could sue politicians for how they do their job.


So you're saying it's unreasonable to expect maintenance continuity over multiple decades from a private company and its employees, but it is resonable to expect the very same thing from a party that's less knowledgeable and less financially invested in the project?

To me that feels lacking, unless the government bodies responsible for infrastructure maintenance get proper resources and time and expertise to do this on a large scale. Can people today really expect to understand how things were built a 100 years ago? (honest question, imagine the same for IT)


It's unreasonable to hold the original engineer responsible for changes after construction that should require review/stamp/signature from another engineer, such as increasing the depth of the pavement on the roadway. If the bridge didn't meet the code at the time it was built, that's one thing. But that's not the case here.

Also, the private engineering firm likely had nothing to do with the maintenance of the bridge after construction -- That's the job of the state DOT and whoever they contract out to.

There's a lot of knowledge and experience in State DOTs, they'll do some bridges in house, some they contract out for. (WA state, states may vary, depends on budgeting conditions and workload.) The older engineers there are generally really smart, and have been doing it for a long time. In stark contrast to the computer industry, experience really does mean something.


Less knowledgeable and less financially interested? The politicians are the ones who make the ultimate decisions about these projects: whether maintenance will be done (by controlling the budget), whether a replacement will be done, etc. They're the MOST invested in the project, and should be held the most accountable. Why should decision-makers get to run away scot-free after their bad decisions wreak havoc?


They may take some of the decisions, but that doesn't make them knowledgeable or financially invested. How would a politician make money from a construction project? The company that engineered it does, sure, but the guy who commissioned it? He benefits from being in office.


> Can people today really expect to understand how things were built a 100 years ago?

Yes. This can be done.

The same as today, 100 years ago designs were based in calculations of the bridge response to the expected loads. Today, we know better what are those loads and how to calculate the response of the structure, but it is the same process.

Construction itself has also changed with the years. We have ways to asset materials and machine them that are way ahead of what we had 100 years ago, we can assembly the bridge with much more precision, and the efficiency of our tools has vastly improved, but the principles are the same.


Hmmmm. You might be right, but I'm just thinking about what I read about difficulties replicating the Saturn V rockets, which obviously were built much later than 100 years ago, under extreme scrutiny of top-of-the-line engineering of their time, and yet, we can't build them again from original blueprints.

The issue seems to be that despite having extremely detailed blueprints a lot of the actual assembly was left to skilled welders and technicians on site, and they used techniques which were not documented. So You can look at the blueprint, you can look at the actual remaining Saturn-V rocket, and still have no idea how two elements were welded together to withstand launching into space, because it's just not documented anywhere. The person who did the weld is long dead, so in a way, an extremely important element of assembly is lost and has to be replicated by experimentation.

I'd imagine it's the same with bridges - blueprints might say that a column of specific durability was used, but it won't say how it was made, and therefore you won't know what can break inside it.

Edit: corrected Atlas to Saturn V


The other factor I read regarding the difficulties of resurrecting the Saturn V engines, was that they used a lot of mechanical feedback control systems to regulate fuel flow etc.

Modern engineers can understand them in theory, but they'd just use a micro-controller for anything that complicated, so they don't have the same fluency in those systems as the original engineers.


Bridges are a lot simpler than a Saturn 5 :-)


Yes when I did my mech eng course in the 80's we learnt how to design a belt feed system - as used in 18th/19th century water powered factories


The government can be sued. If they loose the suit they (or rather) taxpayers have to pay. So in theory if you vote for politicians and policies that cause these kinds of failures you the voter are held liable.


Maybe we should have a system where we no longer have secret ballots, and instead your votes are recorded, but in an encrypted way so that only a proper impartial investigation board (perhaps at the federal level) can have access to it, in cases where it's decided that the voters for bad politicians need to be punished personally for their contribution to the problem. So when the government is sued, everyone who voted for those politicians gets to pay. I really don't see why those who voted against the negligent or criminal politician should be on the hook for it.


If you vote against them you're held liable, too. It would be wonderful if I could pick what portion of expenditures the government is allowed to use my individual taxes for though.


Then use simpler language like: "it's falling apart", and, "Please sign here confirming that you understand it's going to fall over any minute now and it's going to be your fault when people die."


Better yet. Amend the state constitution(s) so that the state is obliged to close to public traffic any bridge, span, or roadway that falls below a certain level of structural sufficiency.


It looks like our bridges are getting batter as we replace aging bridges. See http://usafacts.org/metrics/20697

24% were classified struct def in 1990 vs 10% in 2015.


When the budget is small, you don't build for situations that you cannot imagine. You get what you pay for. Now climate is changing, these situations are suddenly reality.


As of writing, this links to university news release.

The actual paper is doi:10.1061/(ASCE)IS.1943-555X.0000354 [1].

Abstract

Predictions of the risk to built infrastructure posed by climate and land-use change have suggested that bridge collapses may increase due to more frequent or intense flooding. Assessments of the United States often assume that bridges may collapse when the 100-year flood (i.e., a flood with 1% annual frequency of exceedance) occurs, but this assumption has not been fully tested because of a lack of comprehensive collapse records. Thirty-five bridges for which a stream gauge on or near the bridge recorded the flow during total or partial collapse were identified and used to test this assumption. Flood frequency analyses, other statistical analyses, and structural reliability methods were used to quantify the return periods of collapse-inducing flows, identify trends linked to event and site characteristics, and evaluate the potential importance of collapse return period variability in assessing the impact of climate and land-use change on hydraulic collapse risk. The results indicate that the collapse-inducing flow return periods varied considerably (range: 1 to >1,000  years) and were frequently lower than values considered in many climate impact assessments: 23 of the 35 bridges were estimated to have collapsed during flows with return periods of lower than 100 years. Annual failure probabilities computed using the full distribution of return periods of the collapse-inducing flows, as opposed to central values (e.g., means), were more sensitive to an assumed increase or decrease in the underlying frequency of flooding. These results suggest that linking bridge collapse to only the 100-year flow does not capture significant variability associated with collapse return periods, potentially reducing sensitivity to flood frequency changes and reducing the robustness of assessments of the impact of climate, land-use, and streamflow-regulation change on hydraulic bridge collapse risk.

[1] http://ascelibrary.org/doi/abs/10.1061/%28ASCE%29IS.1943-555...


What exactly are "modern design standards" required by bridges? When did the modern era start? Is there really a major moment in bridge design where the standards changed? Is there even broad agreement on what these standards are?

(Side note: the bridge pictured in the article is the Bridge of the Gods, just upstream of Bonneville Dam on the Columbia River. There's going to be more than bridge problems if the flow of the river takes out that bridge.)


Yes. There are standards, they're the Building Codes. There are several versions of them that have come out over time, and they've been evolving with the advances in the science and engineering. When I was in Engineering school, we were generally taught off of a newer code than was in common use in the wild, but that transition was occurring in the couple of years that I was actually doing bridges.

In the case of bridges, you're going to get the basic Steel or Concrete code, maybe some federal guidelines if it's an interstate, definitely state DOT standards.

As for modern vs not -- There are definitely classes of bridges that I'd not consider modern. Anything with rivets, or sections that are built up by lots of little pieces (like the old style riveted trusses). Those aren't done anymore for several reasons: onsite labor is more expensive than shop labor, rivets are inconsistent and weak compared to bolts or welds, and the ongoing painting to control corrosion is a nightmare. In places like the West Coast, the older bridges are not built to current seismic standards, which have been evolving as quickly as the rest of the codes (e.g. tiedowns, base isolation, column jacketing). A lot of the older bridges also have details that aren't terribly good for fatigue performance.

There was definitely an inflection point, probably in the 80's where precast prestressed concrete started being the go-to design for more than just a plain old girder bridge over the interstate.


AASHTO publishes the standard specification for bridges and other road structures. It can be purchased at https://bookstore.transportation.org/item_details.aspx?ID=22...

"The AASHTO LRFD Bridge Design Specifications are intended for use in the design, evaluation, and rehabilitation of bridges, and are mandated by the Federal Highway Administration for use on all bridges using federal funding."

It is over 2000 pages long without errata. It is updated yearly with major updates less frequently.


As a layman, the difference I notice between "modern" and older bridges is, less exposed steel and more concrete.


The Forth crossing (north of Edinburgh) seems to show this progression quite nicely.

https://commons.wikimedia.org/wiki/File:Forth_Replacement_Cr...

Three bridges built in 1890, 1964 and 2017 respectively (from right to left in the picture above).


> more concrete

That will be heavily spalled 50 years from now. Almost as dumb as building thin flexible skyscrapers with reinforced concrete.


I fail to see what's dumb about using concrete for skyscrapers. Tall buildings have been built this way for over 50 years without issue.


It's fine for skyscrapers. The Empire State Bldg isn't going anywhere. It's dumb for thin flexible skyscrapers that will accumulate microfractures in a relatively short time span and whose steel structure is insufficient to support the building in the face of significant concrete degradation.


Microfractures =/= global strength loss. Additionally, the concrete itself tends to strengthen over time. I guess I just don't understand what failure mode you are worried about.

There are thousands of slender concrete structures that have been in service (thus far) without issue.


Lured in for a spanning tree protocol CVE, left with an equally important civil engineering discussion.

It would be interesting to see an evolution of the concept of 100-year event over the few decades that its existed. Given the crazy rate of change compared to more stable engineering challenges like the national bureau of standards definition of the second or length standard, the endlessly changing definition of 100 year events look barely more scientific than rolling DnD dice.



Most Georgia bridges have fifty years before rebuild or replacement. We had our 1000 year flood not to long ago and some were replaced quicker. It is not uncommon to be traveling country roads and come across a recently rebuilt or new bridge.

Now this I85 debacle is purely in the camp of mismanagement and vandalism. The vandalism of course being that a person set a fire sufficient to ignite supposed non flammable building materials that should never have been stored under a bridge. Doubtful anyone in the State will suffer consequences due to sovereign immunity.

What many of us who watched all this unfold is to find out exactly what was stored there and in what quantity. While not flammable doesn't mean it won't burn it should not have been this simple to ignite.

Credit the contractors rebuilding the bridges, damn they are fast when they need to be. Of course another contractor managed to buckle the I20 interstate by filling pipes improperly


As some one who lives very close to the bridge in Atlanta, it is hard to believe how disruptive it has been.

While I thought traffic in the past was bad, it recently took me an hour to go 1.5 miles. A trip that usually takes 10 minutes at most. People don't go to restaurants or out at all, anytime close to rush hour. I use to pop up to corner to grab a pizza, now it takes me 20 minutes just to get to the end of my street! Glad I can bike places.


The UK had a similar fire, though the viaduct didn't collapse.

I think the result was a reduction in what was permitted to be stored under bridges and viaducts, in common with a similar restriction that was already implemented on major railways. I can't find a source for that.

http://www.bbc.com/news/uk-16425491


>damn they are fast when they need to be

They are fast when the situation is dire enough that the politicians don't care what it costs and the NIMBYs and nit-pickers shut their mouths.


Not really the kind of vulnerability the article discusses. Most every aspect of our infrastructure is vulnerable to maliciousness.


I'm not sure if that situation is strictly relevant. We don't normally design bridges to withstand heat directly on the supporting structure of potentially a thousand degrees or more.


I don't think it's that bridges need to be able to withstand that much heat, but that no one should ever be able to apply that much heat to it. If a pipeline runs under it and is able to be accessed by anyone, that inherently means the infrastructure is strongly vulnerable to attack and it just takes one person mere seconds to destroy everything.

I'm not blaming the designers of the freeway though. My house has a giant plate glass window right next to my deadlocked front door. Our entire society is built around trusting that most people don't want to destroy the nice things we've collectively built. But if someone smashed my window and stole my TV, I'd probably make sure that window was a little more secure. Incremental improvements.


At least you aren't being stupid and weighing the window/door risks backwards leading to installing a double-keyed knob or deadbolt. That's the insanity that happens from poor risk analysis.


Totally irrelevant unless you're going to argue the Oklahoma City building was also under engineered.


How many bridge collapses are caused by floods versus other kinds of failure?

In the Northwest, we're pretty worried about the next big earthquake, whenever it happens. I also assume that bridges sometimes just fall down on their own due to corrosion, metal fatigue, low quality materials, improper construction, or uneven thermal expansion of various materials.


The thing is, we've been underestimating bridge failure potential for decades. At least with reinforced concrete bridges, corrosion is deadly and silent, and some substances took decades to slowly corrode the steel. It took unexpected catastrophic failure for us to understand how to build them better.


The real killer here is salt on the concrete in winter. It slowly leaches through the top level of the concrete and starts eating away at the steel. Very effective, totally invisible. Also lots of balconies at risk because of this.


Do people salt their balconies? I've never seen nor heard of this.


Balconies and galleries in apartment buildings to avoid falls in winter.

There's a huge flap around that theme here in NL right now with 1000's of apartment buildings having to be inspected before mid-june because a building shed it's balconies.

https://www.nrc.nl/nieuws/2015/12/22/ministerie-laat-vloeren... (Dutch, sorry, no English source)


I thought everybody uses epoxy-coated rebar on bridges now.


"Now". Not all bridges in use now were built now. In fact, most are quite old, which is the problem.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: