I suspect a lot of Dopers have seen footage of this, but for anyone who hasn’t:
I grew up only a few miles from there, and a few decades later, so I’ve heard and read various things about it over the years. The bridge earned the nickname Gallopin’ Gertie during construction because of the way it would rise and fall in strong winds. The bridge was completed and open for four months, and efforts were underway to find a find a fix when it collapsed.
So, could it have been fixed if they’d had more time, and knowing what we know now? The replacement bridge was made with an open truss for the road deck (instead of the solid girders of the original) and grating in surface to allow wind to blow through.
Was there any fix that could have been applied to the original, adding structure to stiffen it, or cutting holes to let air flow through? Maybe remove the road deck, keep the towers and main cables, and build a new deck? What would it have taken to fix the original bridge, or was there no choice but to start all over?
Since the structure experienced considerable vertical oscillations while it was still under construction, several strategies were used to reduce the motion of the bridge. They included[11]
attachment of tie-down cables to the plate girders, which were anchored to 50-ton concrete blocks on the shore. This measure proved ineffective, as the cables snapped shortly after installation.
addition of a pair of inclined cable stays that connected the main cables to the bridge deck at mid-span. These remained in place until the collapse, but were also ineffective at reducing the oscillations.
finally, the structure was equipped with hydraulic buffers installed between the towers and the floor system of the deck to damp longitudinal motion of the main span. The effectiveness of the hydraulic dampers was nullified, however, because the seals of the units were damaged when the bridge was sand-blasted before being painted.
The Washington State Toll Bridge Authority hired Professor Frederick Burt Farquharson, an engineering professor at the University of Washington, to make wind tunnel tests and recommend solutions in order to reduce the oscillations of the bridge. Professor Farquharson and his students built a 1:200-scale model of the bridge and a 1:20-scale model of a section of the deck. The first studies concluded on November 2, 1940—five days before the bridge collapse on November 7. He proposed two solutions:
To drill holes in the lateral girders and along the deck so that the air flow could circulate through them (in this way reducing lift forces).
To give a more aerodynamic shape to the transverse section of the deck by adding fairings or deflector vanes along the deck, attached to the girder fascia.
The first option was not favored, because of its irreversible nature. The second option was the chosen one, but it was not carried out, because the bridge collapsed five days after the studies were concluded.[8]
After the Tacoma Narrows disaster, other bridges were retrofitted with devices to prevent the same. One in New England had plates fitted as outriggers to prevent the same type of wind oscillation from building up.
As the OP seems to have been fairly well answered, I’ll just add that according to a documentary I recently watched on this, later suspension bridges were designed with a cross-section similar to an upside-down aerofoil, so as to convert lateral wind force into downward pressure on the bridge - eliminating the problem. Or at least, alleviating it sufficiently - the Severn Bridge just upriver from where I am has remained apparently unaffected in its 54-year lifespan so far, and it can get pretty breezy around here.
I guess the engineering community didn’t deem it the best use of time to further investigate how the original design could have been ‘fixed’ once most of it was already underwater. Instead, they determined how the problem had occurred, and then used this knowledge for new designs. I’m not an expert in the field by any means, I just don’t see how you’re going to get much more info than what has been posted already. No harm in asking, of course - as always, I’d be very happy to see other perspectives from those with more knowledge than I have.
Based on the documentary I mentioned in my previous post, refitting the deck to make it an inverted aerofoil could have worked, but presumably would have cost nearly as much as starting over (certainly much more than drilling holes or adding fairings).
Nope, that’s a common myth or misunderstanding. The cause was aeroelastic flutter, which is a different phenomenon.
Any amount of twist in the bridge created vortices, or areas of low pressure, in locations that actually amplify the twisting motion. As the bridge returned to its natural state, its momentum twisted it in the other direction where the wind could catch it and continue the twisting. This phenomenon is called aeroelastic flutter. It’s the same reason that a strap or sheet of paper vibrates in the wind. It’s a completely separate mechanism than resonance from vortex shedding, because the periodic forces are self induced from the naturally unstable aerodynamic shape of the bridge. This torsional flutter eventually created too much stress in the suspension cables, and the bridge failed.
Yeah, I guess the question became a lot less urgent all of a sudden. But it’s also a famous failure, and I gather it was taught for years in Engineering classes. Maybe it still is. I thought maybe some clever professor somewhere might have put some effort into teaching a fix, as well.
It wasn’t the first, or the longest suspension bridge ever designed/built, but it wasn’t exactly routine, either. I saw one documentary on it which said that suspension bridge designs were progressing toward slimmer road decks. The less steel you have to use, the less it costs, and slimmer decks were also considered an aesthetic improvement. Instead of an open, lattice-like truss, the Narrows Bridge used solid girders.
Not the first bridge, but the first one to make that particular mistake.
I’ve heard this correction (i.e., “it’s not resonance, it’s aeroelastic flutter”) a number of times now and never been very satisfied with it. Aeroelastic flutter is obviously (to me) a type of resonance, just a self-induced one. The flutter is at the resonant frequency of the bridge (in the torsional mode), and the same design features that would improve other resonant effects would improve this one (increased stiffness, dissipating elements, etc.).
The WIkipedia article cites some authors that seemingly agree, at least in principle:
To some degree the debate is due to the lack of a commonly accepted precise definition of resonance. Billah and Scanlan[3] provide the following definition of resonance “In general, whenever a system capable of oscillation is acted on by a periodic series of impulses having a frequency equal to or nearly equal to one of the natural frequencies of the oscillation of the system, the system is set into oscillation with a relatively large amplitude.” They then state later in their paper “Could this be called a resonant phenomenon? It would appear not to contradict the qualitative definition of resonance quoted earlier, if we now identify the source of the periodic impulses as self-induced, the wind supplying the power, and the motion supplying the power-tapping mechanism. If one wishes to argue, however, that it was a case of externally forced linear resonance, the mathematical distinction … is quite clear, self-exciting systems differing strongly enough from ordinary linear resonant ones.”
It is definitely not a vortex shedding effect, nor was it induced by any other external oscillation (like an army marching in step). Nevertheless, it seems rather silly to call this not resonance even though it is using external energy to pump an oscillation at the resonant frequency.
The 20,000 foot answer is that designing bridges is hard, and to err is human.
Bridges do collapse. A bridge collapsed last month in Norway:
Just going by what Wikipedia says, price was a factor in Washington, and they used a cheaper alternative. When the key design decisions were made, we were in the midst of terrible economic conditions.
I worked for a WSDOT contractor that handled tolls. Internet access on our computers was restricted, but their own website was accessible. This includes an extensive website chronicling everything about the Narrows Bridge.
The bridge failed because of human hubris on the part of designer Leon Moisseiff. He completely disregarded the effects of wind and the design was driven first by aesthetics. It was doomed. Even if it hadn’t fallen, they would have spent at least the same ten years implementing the numerous fixes needed.
“Resonance”, to me, says that there’s some key frequency, and something else just happened to be at that frequency, and that’s what caused the failure. If something with no inherent frequency at all, like a steady wind, can cause the oscillation, it’s not a resonance, because a resonance is always between something and something else.
Is a pendulum a resonant mechanism? It sure seems to be so, even though the energy input is constant. Of course, there’s a mechanism–the escapement–that provides the little pushes at the resonant frequency. Or rather, those little pushes happen at key points along the swing, so that they naturally happen at whatever frequency the pendulum is at, which is by definition the resonant frequency.
Same principle here. The air may be pushing constantly, but the aerodynamic configuration of the bridge is such that the air has a more significant effect when the bridge has twisted into a particular shape, and so it ends up pushing more at just the right points in the cycle.
I’m over my skis a bit, but I think resonance could describe what happened to the bridge, but harmonic resonance suggests two separate things with the same frequency (thus the harmony).
I agree with that–it’s not harmonic resonance, in the way that a guitar string might cause a nearby one to vibrate if they share an overtone. It’s more like a pendulum or a child’s swing where some mechanism (the escapement, the person intentionally timing their pushes) causes the energy inputs happen at the resonant frequency. And I agree that this is really just a matter of definition. I’d define resonance to be something like “amplification of the resonant frequency caused by energy inputs happening at the same phase in the cycle”. That would include harmonic, self-excited, and other types.
eg Wheeling, West Virginia… At time of opening , it was the largest span cable suspension bridge was finished in 1851, but badly damaged in 1854.
A list of early suspension bridges contains a lot of early suspension bridge failures…
Back to Tacoma Narrows, Moisseiff and Frederick Lienhard, the latter an engineer with what was then known in New York as the had published a paper asserting that using wire cable instead of wrought iron chain systems, meant that the bridge could be pre-stressed into being stiff enough. He was under pressure to build Tacoma Narrows bridge cheaper than Golden Gate Bridge, and so pushed the boundary for how much they relied on the suspension cables for stiffness.
Moisseiff had street cred from Golden Gate, but he didn’t have Strauss or Ellis to consult with . He made an assumption about why the Golden Gate’s deck had to be made so stiff… for the horizontal wind load… He assumed it was the horizontal wind load that was the problem, see ? And having decided that it was well beyond requirements, made Tacoma Narrows far less stiff than Golden Gate.
Previous suspension bridges had failed due to snake wobble (bouncing side to side … ) and bouncing vertically, but this just hides that they would fail to the third mode of wobble … torsion…
