Catastrophic bridge collapse

[AustinBoston]

If they are going to grade these bridges (which they do) they should be required to post that grade ON the bridge so that people can make a decision whether to use it or not. In California, we grade all of our restaurants and the grade is posted. I will NOT eat in any restaurant that is not graded "A" and you better believe that I would not drive on a bridge that is graded "Structurally deficient".

I posted earlier that the DOT (this is national) rated hundreds of bridges as "structurally deficient."  That was based on somethign someone told me.  When I found documentation, it's more like 70,000 bridges rated "structurally deficient".  There is a good chance you drive on one or more every day.  

On our commute home yesterday, I pointed out visible rust on every bridge we drove under.  You know what? a little paint goes a long way...

Both sides of that bridge started falling at exactly the same time.

At less than two frames per second, you can not know that from the video.  (Watch the woman walking around after the collapse - she takes a step between frames.) A co-worker and I have repeatedly gone over that video frame by frame, and there is simply no way to tell.  The best copy may have more resolution, but the versions on the web are not clear enough.

At 1/2 second, the bridge could easily fall 8 feet between the frame before and the frame after it started to fall, which is more than enough to pull the other end off it's mounting.

If the collapse started at the other end, it could have been falling for at least two frames before it would be clearly visible.  At the distance of the other end of the bridge (1,000 feet), one image pixel is about 8 feet, or 1/2 second of falling.

It also assumes that the bridge wasn't shifted lengthwise before it started to fall, which is a distinct possibility.  Construction workers on the bridge reported significant vibration before the collapse.  That went on long enough for a foreman to order everyone off the bridge.  Based on the reports, I don't think it was long enough for workers to begin complying.

And then there is the design of the bridge.  It was held up at the very ends only.  The bridge relied on both ends having full support at all times.  If one end started going, the other was also going down - NOW.  I have been told (but I have not seen images or read about this) that the main span's steel structure was on rollers in order to deal with expansion & contraction.  I have seen other bridges built this way, so I know it is done.  If there was lengthwise movement because of the initial failure, the other end was going down NOW.

In the end, I believe the facts will show that this bridge fell because of a combination of bad/underdesign, overweight traffic, improper maintenance, inadequate inspection, improper materials, excess vibration, government delays, and other forms of incompetence.

Austin

[AustinBoston]

This video clearly shows the rust that Wayne is talking about...and clearly shows it is a point of failure (not the same as cause, but it sure looks bad).  It shows before and after images of a specific joint that has clearly failed.

http://www.wcco.com/video/?id=29695@wcco.dayport.com

The other thing these images show is that significant parts of the structure are box beams.  Box beams are fairly strong, but very difficult to properly inspect, and nearly impossible to properly maintain.  How do you paint the inside of what is essentailly a 30 foot long tube with both ends partly blocked?  How do you remove loose rust, and prevent water and salt entry?

This collapse started before it was even built.

Austin

[wavery]

But in the end, I believe the facts will show (whether you want to accept them or not) that this bridge fell because of a combination of underdesign, overweight traffic, improper maintenance, inadequate inspection, improper materials, excess vibration, government delays, and other forms of incompetence.

I think that's probably right. I heard this morning that the bridge may have been rated to carry a much heavier load than it should have been, when it was built.

I find it interesting that we have several city streets that have signs that say "No Trucks".....over a certain weight. There are bridges with similar signs. I think that they should not allow trucks over a certain weight to cross bridges that are rated "structurally deficient". It just seems like a no brainer but it's all about politics and $$.

In California, we have arranged a "Shake test" :yikes:  every year or so for our bridges :p .

[AustinBoston]

I think that's probably right. I heard this morning that the bridge may have been rated to carry a much heavier load than it should have been, when it was built.

I find it interesting that we have several city streets that have signs that say "No Trucks".....over a certain weight. There are bridges with similar signs. I think that they should not allow trucks over a certain weight to cross bridges that are rated "structurally deficient". It just seems like a no brainer but it's all about politics and $$.

In California, we have arranged a "Shake test" :yikes:  every year or so for our bridges :p .

Did you arrange the shake test, or did someone of a somewhat higher authority?  :p

We do "stress tests" of our bridges.  It consists of watching and/or measuring stress at key points while the stress (weight) is applied and removed.  The stress...three fully loaded tractor trailers are driven over an empty bridge at the same time.  Hello?  This is an eight lane bridge...how about driving eight tractor trailers (with tandems) at the same time?

I write software that undergoes stress testing.  Stress testing for us (in part) consists of appplying between two and ten times the expected load...and lives don't depend on my software.

Austin

[dademt]

There are many factors involved including how much total weight was on the bridge and how much of that was stationary or near stationary.  On the Del Mem Bridge that I drive ever day, they tend to monitor for excessive truck traffic on days where there are backups.  If needed they stop traffic coming onto the bridge to reduce the load.
 
Understand also, as someone that sells steel, even prime steel can have "bad spots" which sometimes are called inclusions which if in the right spot can undermine the best engineering.  Furthermore, some forms are steel, namely A871 weathering steel are designed to rust initially.  Rust alone is not a sign of bad material.
 
Utimately, it was a great loss of life and hopefully they will learn from and correct future mistakes out of this.

[AustinBoston]

Understand also, as someone that sells steel, even prime steel can have "bad spots" which sometimes are called inclusions which if in the right spot can undermine the best engineering.

I believe the kinds of defects you mention are unavoidable.  That is the reason redundancy is so important in large structures.  The concept is that there might be (or might develop) a defect in a critical part.  If the odds of a defect are 1 in 10,000, that's the best you can do.  If the structure uses 1,000 of those parts, then there is a 1 in 10 chance of the structure containing a defective one.  But in a redundant structure, that changes dramatically.  Now you need two defective parts in the same place.  The odds of two together being defective is 1 in (10,000 x 10,000), or 1 in 100,000,000.  Because there are twice as many parts, the odds of having a defective one go up, so that it would now be 2 in 10 (or 1 in 5), but the odds of two defective parts together would be very small, about 1 in 100,000.  Now you have a bridge that is (in theory) ten times safer for twice the price.  It also simplifies maintenance, because any one part can be removed and replaced without compromising the whole structure.

Of course, there is a limit to this.  Improper maintenance can adversely affect all of the parts, putting the whole redundancy scheme in jeopardy.  Two parts can rust through as easily as one.

This bridge had a number of potential single points of failure.

aw738 mentioned the Silver Bridge over the Ohio river.  This bridge collapsed in 1967 in a large part because of a lack of redundancy.  Even though the "eyebars" were significantly over-engineered, if one failed, the nearby "eyebars" would not be able to take the added weight and would fail.  A defect of one tenth of an inch in one eyebar brought the whole thing down at rush hour, killing 46 people.

Another example where redundancy worked (for a time) was in the World Trade Center.  If redundancy had not been designed and built in to the buildings, they both would have come down immediately, instead of standing for long enough for most occupants to be able to evacuate.

Furthermore, some forms are steel, namely A871 weathering steel are designed to rust initially.  Rust alone is not a sign of bad material.

Absolutely not!  But, in most cases, excess rust is a sign of inadequate maintenance, and rust can become a cause of failure if allowed to continue too long.

I can remember seeing some new oil storage tanks in Boston.  The tank pieces were put in place and carefully welded.  Then the completed tanks sat there and rusted for some time - I think about 6 months, but it could have been much longer or shorter - in the Massachusetts salt air.  Eventually a co-worker asked why, if the existing tanks were all painted in the company colors, weren't the new tanks being painted?  

Apparently, the question had been asked many times, because the receptionist had a sheet for him that explained it.  The way the tank sections were manufactured, they ended up with thin layers of steel on the surface that could easily peel off.  If they were painted when new, the paint would just peel off in places.  By allowing the tanks to rust, this microscopically thin layer would be removed (where it existed).  In addition, the microscopic pits formed by the rusting formed a much better bonding surface for the new paint.  The rust was a deliberate part of the design.

Utimately, it was a great loss of life and hopefully they will learn from and correct future mistakes out of this.

While I would not have considered this bridge to be a diamond (more like a piece of quartz), I think what one former member of the NTSB said about airline crashes is appropriate.  He said "Diamonds reveal their hidden flaws when broken."  What he meant was that we learn things from failures (even things that did not cause the failure) that can make other airliners safer.  That really ought to apply to bridges as well.

One other thing - if they were starting from scratch today, I suspect that bridge (specifically the long span that crossed the river) would be built in 3-6 spans, not one.  There was no operational need for a single span, and several other bridges in the area are set on multiple pylons (as many as 15).  With shorter spans, even if one fails - hopefully not, but always a possibility - it can be replaced with a temporary structure in days or weeks.  This thing is going to take years to replace.

Austin

[wavery]

I believe the kinds of defects you mention are unavoidable.  That is the reason redundancy is so important in large structures.  The concept is that there might be (or might develop) a defect in a critical part.  If the odds of a defect are 1 in 10,000, that's the best you can do.  If the structure uses 1,000 of those parts, then there is a 1 in 10 chance of the structure containing a defective one.  But in a redundant structure, that changes dramatically.  Now you need two defective parts in the same place.  The odds of two together being defective is 1 in (10,000 x 10,000), or 1 in 100,000,000.  Because there are twice as many parts, the odds of having a defective one go up, so that it would now be 2 in 10 (or 1 in 5), but the odds of two defective parts together would be very small, about 1 in 100,000.  Now you have a bridge that is (in theory) ten times safer for twice the price.  It also simplifies maintenance, because any one part can be removed and replaced without compromising the whole structure.

Redundancy does not have to be the answer. Where $ is concerned, like oil pipe lines or aerospace, every single piece of metal that goes into these structures are x-rayed for cracks and material defects. It is an expensive procedure but it can't be any more expensive than building a bridge with 100% redundancy (which is what it would take).

My son-in-law owns a company that builds huge x-ray machines http://www.willick.com/index.html . He built the x-ray machine for inspecting whole component parts for the latest space shuttle assembly. They can put an entire wing in this machine at once. There is no reason why they couldn't do the same with bridge construction materials (today, not in 1967).

The part that really ticks me off is that there is a company that has been trying to sell devices to the DOT and the States for about 2 years now. This device is placed in critical areas of any structure (like a bridge) and it measures small movements and can tell of a potential failure minutes before it takes place in 90% of cases. This devise is relatively inexpensive and I will bet that every bridge in the US will have them within a year or 2 (or immediately after the next tragedy).

[dademt]

Absolutely not! But, in most cases, excess rust is a sign of inadequate maintenance, and rust can become a cause of failure if allowed to continue too long.
 
Hey Austin, with all due respect, I am right about this one.  Here is the information from the ASTM spec ( abbreviated version )
 

ACTIVE STANDARD: ASTM A871/A871M-03 Standard Specification for High-Strength Low-Alloy Structural Steel Plate With Atmospheric Corrosion Resistance

Developed by Subcommittee: A01.02
See Related Work by this Subcommittee

Book of Standards Volume: 01.04






1. Scope


1.1 This specification covers high-strength low-alloy steel plate intended for use in tubular structures and poles or in other suitable applications. Two grades, 60 and 65, may be provided as-rolled, normalized or quenched and tempered as required to meet the specified mechanical requirements.

1.2 The atmospheric corrosion resistance of this steel in most environments is substantially better than that of carbon structural steels with or without copper addition (see Note 1). When properly exposed to the atmosphere, this steel can be used bare (unpainted) for many applications.



Note 1

[AustinBoston]

Redundancy does not have to be the answer. Where $ is concerned, like oil pipe lines or aerospace, every single piece of metal that goes into these structures are x-rayed for cracks and material defects. It is an expensive procedure but it can't be any more expensive than building a bridge with 100% redundancy (which is what it would take).

While it is less expensive, it is also less effective.  One of the things that was drilled into us when I worked in Quality Assurance was that "100% inspection is not 100% effective" (100% inspection is actually only 85% effective, and we proved it repeatedly).  200% inspection is actually worse that 100% inspection.  Testing is more effective than inspection because it takes the human element out of detection, but still is not 100% effective.  I have to assume the x-ray system you suggest is somewhere in between testing and inspection.

I don't know if the pieces of the alaskan oil pipeline were all X-rayed, but I know some of it was, and it has had failures.

Obviously, not everything can be made 100% redundant.  Making a plane with an entire second set of wings (either of which could fly the plane) would make the plane effectively worthless.  But when you have thousands of identical devices, with each one being mission critical, redundancy is the only way to go.

My son-in-law owns a company that builds huge x-ray machines http://www.willick.com/index.html . He built the x-ray machine for inspecting whole component parts for the latest space shuttle assembly. They can put an entire wing in this machine at once. There is no reason why they couldn't do the same with bridge construction materials (today, not in 1967).

Remember, 100% inspection is only 85% effective...

The part that really ticks me off is that there is a company that has been trying to sell devices to the DOT and the States for about 2 years now. This device is placed in critical areas of any structure (like a bridge) and it measures small movements and can tell of a potential failure minutes before it takes place in 90% of cases. This devise is relatively inexpensive and I will bet that every bridge in the US will have them within a year or 2 (or immediately after the next tragedy).

While I've pointed out flaws with this (or any) inspection system, please don't think I don't think it's a bad idea.  There is NO 100% effective solution (including redundancy).  It becomes a matter of risk versus cost (or risk vs. risk).  For example, we could drastically reduce the risk of rust-related failure by making bridges from aluminum.  The cost of the bridge would skyrocket, and the risk of metal fateague would go up.

The real solution lies in a number of areas, including:

• Designs that are not prone to single points of failure (this could be accomplished through redundancy, but there are other ways.
• Avoiding designs where failure is so catastrophic. For example, this would have been less devastating if this collapse had been contained to just the main span.  But after the main span fell, additional spans fell on either side, and on one end, additional spans after that, until all spans of the bridge had failed.
• Using designs that lend themselves to simple inspection and maintenance.  This is another area assisted by redundancy.  In a fully redundant system, you can replace any part while in service without a service disruption.  But even more important is avoiding construction that ends up doing things like create hidden recesses that require exotic mechanisms to inspect or maintain.
• Using appropriate means to reduce or eliminate defects in materials and workmanship.  Examples would include buying high quality materials (I believe in forcing the vendor to prove to me that it's good - via x-ray, via statistical testing, via vodoo - as long as he proves it is good) and using things like X-rays to inspect welds and rivets at the site.
  
• Building in a sliding weight scale, so that as a bridge ages, it's weight rating automatically goes down.  I believe that no matter how careful we are in design, and no matter how careful we are in inspection, there are always going to be types of fateague that we don't anticipate or don't understand and therefore potential types of failures that are totally unexpected.  If the bridge is designed for a useful life of 100 years, the weight rating should meet the anticipated traffic weight in 100 years.  At that point, the bridge would need to be replaced or would have to have enforced weight restrictions.  If traffic weight risies faster than anticipated (which was the case with the Silver Bridge), that point comes sooner (as it should).  I think the same will be found with this bridge; I believe it was never expected to be an 8-lane bridge, but a 6-lane bridge with right-side shoulders.  That is a 33% increase in traffic weight with no structural improvements to the bridge!
• Diligence, diligence, diligence when it comes to inspection and maintenance.  This is where my mistrust of the government comes in.  The bridge was being resurfaced (probably because of complaints).  But comparatively, that resurfacing was simply cosmetic.  People see potholes and complain, so that's where the money goes.  But the real problem was only 25 feet lower, and may have cost no more to correct.  Meanwhile, MN/DOT had a report (that they commisioned) that said certain joints (I'll bet we know which ones) should be reinforced with steel plate.  Instead, they decided to "fix the cracks."

All of this, of course, cost $$$.  In the end, I am one who feels rail systems are far more cost effective than road systems.  Here in the Twin Cities, we have far fewer passenger rail systems than comparable metro areas in the U.S.  (And the U.S. is way behind most other countries.)  Of course, don't ask about rail bridge maintenance...  :yikes:

Austin

[AustinBoston]

Hey Austin, with all due respect, I am right about this one.

With all due respect, where I went to school, if you said "Rust alone is not a sign of bad material." and I said "Absolutely not!" it would be an agreement, not a disagreement.  If there was any ambiguity, the rest of the paragraph (and the next two) should have made my intent obvious.  Did you read them?

Austin

[dademt]

With all due respect, where I went to school, if you said "Rust alone is not a sign of bad material." and I said "Absolutely not!" it would be an agreement, not a disagreement. If there was any ambiguity, the rest of the paragraph (and the next two) should have made my intent obvious. Did you read them?
 
Austin

My appologies.  I do agree with you also, people should stop doing the work with the lest effective materials to pass the current test with the attitude that it can be someone elses problem tomorrow.  You don't know how many times someone has called asking for a grade of material only to hear the price and comment that they need to get the engineer to figure out if something less expensive will do.  If frustrates me royally.  I wish you had supervised some of the engineers in western pa when I use to deal with them.

[AustinBoston]

I do agree with you also, people should stop doing the work with the lest effective materials to pass the current test with the attitude that it can be someone elses problem tomorrow.  You don't know how many times someone has called asking for a grade of material only to hear the price and comment that they need to get the engineer to figure out if something less expensive will do.  If frustrates me royally.  I wish you had supervised some of the engineers in western pa when I use to deal with them.

Thinking a third time, I could have been clearer in my first response.

This should have been one of my bullet points, and is at the core of the problem here:

The government does business in only one of three ways: 1) Lowest bidder; 2) Highest briber; 3) Overwhelming public pressure.  Don't assume because #3 exists that #1 and #2 aren't still going on.

I don't object to using a lower grade of material if you use enough more of it to make up the difference (in both rating and risk).  But as I am sure you know, that seldom actually saves money.  If I use two or three of grade "B" when I could have used one of grade "A", now everything supporting that has to be upgraded for the extra weight..."Why is it so expensive?  Can't we find a cheaper way?"

Austin

[haroldPE]

what an incredible journey of fabrication to get to your point of currupt government.

Thinking a third time, I could have been clearer in my first response.
 
 This should have been one of my bullet points, and is at the core of the problem here:
 
 The government does business in only one of three ways: 1) Lowest bidder; 2) Highest briber; 3) Overwhelming public pressure.  Don't assume because #3 exists that #1 and #2 aren't still going on.
 
 I don't object to using a lower grade of material if you use enough more of it to make up the difference (in both rating and risk).  But as I am sure you know, that seldom actually saves money.  If I use two or three of grade "B" when I could have used one of grade "A", now everything supporting that has to be upgraded for the extra weight..."Why is it so expensive?  Can't we find a cheaper way?"
 
 Austin