Tacoma Narrows bridge (was: MySQL v. PostgreSQL ...)

Ben Scott dragonhawk at gmail.com
Thu Aug 2 17:06:23 EDT 2007


On 8/2/07, Bill Sconce <sconce at in-spec-inc.com> wrote:
> Amateurishness includes succumbing to the temptation to not think
> through to consequences when reality offers new data.  In the Tacoma
> Narrows case: "However, the mass of the bridge was considered sufficient
> to keep it structurally sound."[3]  And then the torsional mode appeared,
> and the longitudinal resonance morphed into aerodynamic flutter, which
> was certainly understood at the time even if no bridge had yet been
> built to demonstrate flutter occurring in steel and concrete at 0.2 Hz.

  I've heard it claimed that, before the first Tacoma Narrows bridge
failed, aerodynamics had never been considered significant in bridge
design.  If you look at the bridges of the time, I can believe that.
Most of them were simple truss/arch designs: Solid, squat, heavy
structures.  Suspension bridges were a new technology at the time.  It
may well be that they should have seen it coming, but who can say?
I'm not in a position, with my 20/20 hindsight, to judge.

  What may be more apropos to your point is the fact that, once the
oscillation was observed, the bridge was apparently kept in active,
public service.  The collapse did not occur right away, but there were
still cars on it when the bridge fell.  If you build a bridge, and you
do not expect it to move, and then see that it *does* move, that
should be a red flag, I would think.  I haven't seen much about what
decisions were made between the first sign of trouble and the
collapse.  Did engineers dismiss the problems out of hand?  Did
"management" override engineering concerns?  I would be interested in
seeing reliable sources on this aspect of the TN incident.

> "No one could have connected the dots" is not a professional excuse.

  If the quote is accurate, then it is.  If *no one* could have seen
in coming, then by definition, it was unavoidable.  Sometimes that
does happen.  But if the speaker really means "We just didn't think
about it", well, that's another story entirely.  And a depressingly
common one.

>     A more recent example, and the event which made me write this,
>     the I35 bridge in Minnesota, which collapsed yesterday.

  I think it is rather too early to start making proclamations about
the I-35 incident, given that it happened less than 24 hours ago.
I've seen reports that stated the bridge was considered to be in
acceptable condition, and that a near-majority of bridges in the US
reportedly are in similar condition.

  I can see quite a few possibilities:

  (1) The reported conditions were accurate.  Said conditions led
directly to the collapse, and would have done so without any other
factors.  That would indicate the present standards for bridge safety
are woefully inadequate.  It might also mean that there is significant
risk of similar incidents (it would depend on the specific failure
mode).  Worrying about this is, I think, justified at this point.
Drawing conclusions, however, seems premature.

  In particular, calling for immediate re-inspections of other bridges
(something many will no doubt do) isn't going to help if this is the
situation.  It's not the bridges which are at fault, but the
inspection process itself.  Root cause analysis matters.

  For the record, I actually feel this is a very likely possibility,
but my "gut feeling" certainly isn't a valid engineering method.

  (2) The reported conditions were accurate, but were not sufficient
to cause the collapse.  They combined with some other event to trigger
a collapse.  This is really just a variation of #1, as either way, the
standards are likely inadequate.  But it may be that the confluence of
events was unusual.  This matters mainly in terms of the scope of
response (the difference between a nation-wide overhaul and a
corner-case which must be added to the checklist).  The dead are still
dead, and the fact that it's a one-in-a-million thing doesn't make it
any less tragic.  (Arguably the opposite.)

  (3) The reported conditions were inaccurate.  Conditions were in
fact much worse, and that led directly to the collapse.  If so, it
would hardly be the first time that safety inspection results were
falsified.  This would demand not stricter standards, but stricter
oversight (not an engineering problem, but a problem of law).

  (4) The reported conditions were irrelevant.  The bridge collapsed
for some other reason.

  (4)(a) In particular, the bridge was apparently undergoing
renovation at the time.  If construction standards were not followed,
or something was otherwise done inappropriately,   the condition of
the bridge may matter not.  It might still be an engineering problem,
but one which matters only to bridges undergoing construction (making
it an engineering problem, and a rather critical one).  Or maybe
construction standards were shorted, which would again make it a
matter of law.

  Point being: Premature judgment is even worse than premature optimization.

-- Ben


More information about the gnhlug-discuss mailing list