The 9/11 Forum

This is to stake out a place for crane failures, really only one in particular I have in mind right now. The case is extreme and the analogy to anything being discussed regarding WTC1/2/7 is debatable, but I feel it is one small bit in gaining an understanding of propagation of failure in structures. This is a steel structure, however the mode of failure is quite different from that of a building, obviously. The relevance is at least in the unequivocal demonstration that there is a very small range of displacement over which any resistance to load descent is applied; once the integrity is lost, it's not a fraction of the as-built capacity, it's effectively zero.


http://www.youtube.com/watch?v=LWlj5nEqRCE

Hmm. From all the pre-fall-over cracking and popping, I'd say the failure(s) were pretty slow.

I'd compare that kind of crane to the little cylindrical bit of metal that keeps the rotor on a helicopter. If that shears (and it can happen to any bit of metal if it gets just the right amount of force in the wrong direction) all of the safety systems surrounding it are not going to stop the helicopter coming down sharpish.

Looked like all forces were focussed on a small section at the base. Lots of tension going through it. One member failed. Other members followed. The thing was only held taught by balancing on the ground links, which also created all the wire tension to keep it stable. Once the *master bolt* failed, the thing goes kersproing, falls over, hits stuff on the way and breaks up.

Don't quite get the context of the final sentence...

Suppose it may be interesting to see how close to it's rated maximum load it was carrying, and whether there was any human error in configuring the thing...?

femr2 wrote:Hmm. From all the pre-fall-over cracking and popping, I'd say the failure(s) were pretty slow.

I never said that initiation was fast in either the relative or absolute sense. Since you bring it up though, there's about 10 seconds between the first pop and confirmed descent. By the time the camera pans up it's well beyond creep. Ten seconds is only about 2.5x as long as WTC1. But, a more important measure of rapidity, in my mind, is the fact that it goes from as-built capacity to a wet noodle in essentially that short period of time. By comparison, that's not slow compared to an hour plus, is it?

The real point is there was probably displacement of less than 5% of the height before the effective restraining force on the load was negligible.

I'd compare that kind of crane to the little cylindrical bit of metal that keeps the rotor on a helicopter. If that shears (and it can happen to any bit of metal if it gets just the right amount of force in the wrong direction) all of the safety systems surrounding it are not going to stop the helicopter coming down sharpish.

Yes, rather like a single column failing, and not a web of interconnected components. The purpose in this example was to be extreme in that regard, so as to make the phenomena isolated and easy to see.

Don't quite get the context of the final sentence...

Basically: that capacity within design parameters means nothing after a certain point, a point which can come rapidly.

Suppose it may be interesting to see how close to it's rated maximum load it was carrying, and whether there was any human error in configuring the thing...?

Either overloading or misuse, but it was supporting the load successfully less than 15 seconds before it was freefalling to the ground.

So, yeah, the analogy is limited. Take from it what you will.

Unfortunately that video has been removed for copyright reasons and I couldn't watch it.

I don't know if it was the same as this one:
http://www.youtube.com/watch?v=yG1JrEdt3Mg

In any case, this one has two interesting features:

- Negligible resistance of the arm after it starts to buckle, as described in this thread.
- An over-g mechanism similar to the most likely one in WTC7.

It's not too clear from the images, but it seems that the cause of the release is that the connection of a cable (independent of the crane, hooked to the top of the tower) or the cable itself breaks while the crane's cable is not tight. The cable to the crane tightens during the fall and the dynamic load makes the arm fail.

The risks associated with imposition of dynamic overload on a loose crane cable reminds me of a real world experience of mine. Circa 1978.

The situation which arose was that a rapidly rising flood on the main water storage dam serving Sydney left us with a "stop log" jammed in front of the main crest gate which releases the floodwaters.

The "stop log" was the bottom one of a set. 90 feet long, five feet high and dead weight about 24 tons. We should have removed the stop logs. Had removed the top two but the flood waters arrived far too soon for us. And the flood flow passing over the 400ft high dam had twisted and jammed the bottom stop log. We had no idea what additional force over the 24 tons would be required to loosen the friction and compression induced jamming. And the meteorological conditions threatened to keep the flood rising. If the rising flood beat us and broke the stop log it could damage the crest gate and instantly reduce Sydney's water storage by about 50%. With no available recovery options. And me considering future employment options.

Above the dam was a 2000 ft span cableway capable of lifting 24 tons - Actually 18 tons with full dynamic impacts and 24 tons if taken very cautiously. A maintenance left over from the construction activities when the dam was built - 1946-61.

But we needed a lot more than 24 tons and we needed it in a hurry. So two mobile cranes were available at short notice with IIRC 25 tons and 15 tons capacity. So one onto each end and we tried the lift. No go. The overhead cableway with another 24 tons sitting temptingly overhead and I was getting lots of advice to put it into the mix with the two cranes.

No way. To get 24 tons lift from the cableway the main cable would pull down about 50feet of extra catenary sag. So far so good. BUT when the stop log came clear it would allow the cableway to rebound upwards taking the load off the mobile cranes with their fixed length near rigid jibs and cables. Still no problem. BUT once it reached highest upwards movement the cableway would then oscillate downwards until it dropped the full load back onto the mobile cranes. With full dynamic effect. And the 15 ton crane would have folded just like the one in the video -- which reference puts me back "on topic".

We actually went hunting for a bigger mobile crane - found one of 125 tons capacity which was the biggest in our side of Australia at that time. Many bigger ones now.

The next problem is where do you put the crane on a dam crest where you can place the outriggers. Cut to the chase they had a crane scale on that big crane and it touched 60 tons when the stop log came free.

This is as good as I can find - the twin 18 ton cableways and about half of the length of the stop log visible in the foreground. You will have to imagine the rest.

pgimeno wrote:I don't know if it was the same as this one:

It wasn't, but that one's nice too.

In any case, this one has two interesting features:

...
- An over-g mechanism similar to the most likely one in WTC7.

I thought I'd let this lie fallow for a while, in the hopes you wouldn't notice my response. I really have no interest in reviving the pissing contest we had some time back in another thread. This case is the same as the last... whatever mechanism for over-g was present in WTC7, if any was present, this was probably not it.

It is established by direct observation that the period of high acceleration in the early descent of WTC7 was accompanied by insignificant global tilt. The mechanism of over-g in this video is the same as any possible over-g condition in the tipping brick wall and the tipping entertainment center we were discussing in the other thread. The over-g condition results from kinematic constraint and is apparent in only a portion of the total mass - that which is effectively leveraged, away from the pivot through the center of mass.

Since it's quite obvious that global tilt was not present in WTC7, the only means by which this type of mechanism could be responsible is if there is local tilt; that is, distortion in building geometry which results in planar sections undergoing the same sort of rotation about a pivot. All things considered, there is only one way for that to happen and that is rotation of the outer about the inner (i.e. toroidal) where the inner provides the pivot.

There are several problems with this, the most severe being the interior portion is not fixed relative to ground! It is instead presumed to be in collapse progression as well, with some lead time. A moment's reflection on this situation - that the alleged pivot about which the perimeter could swing down at over g - is not stopped but instead already descending with greater velocity, suggests an entirely different and mutually exclusive effect: the already descending interior is dragging the exterior portions down with it, even mildy 'slingshotting' it downward (with some elastic response coming from god knows where).

The latter mechanism is by far the most typically proposed mechanism for over g from those who either believe it did or can occur in the case of building seven. The mechanism you cite, however, is used to demonstrate over-g is possible in the general case. I don't think anyone else is proposing this mechanism is present in WTC7's collapse, that was the point I was trying to make to you when you posted a tipping wall video at JREF and I called you on it here. Subsequent to calling you on it, you made some pretense of how you weren't trying to say it applied to WTC7, just that you had some (invalid) point to make about one of Major_Tom's comments.

Well, here you are explicitly claiming it is the likely mechanism for over-g in B7, so I guess we can dispense with the pretense of that other thread. My point remains the same: no, this is not the likely mechanism of over-g, if there was over-g. I can get a very good third degree fit to about 90% of g over that period. With all due respect to femr2 (and much more is due than you've probably ever given), over-g is not proven. I think femr2 makes an excellent case, and I wouldn't fault anyone for going with what appears to be the most solid empirical evidence he's provided.

However, those who DO go with this evidence, and that's everyone from femr2 to his biggest detractor (tfk) cite instead drag-down from the interior as the most likely cause. I have to agree, it is the most likely cause if indeed there's any effect which needs explanation. But there are problems when examining it from the other perspective, namely first principles. Kinematically, the perimeter MUST draw in for over-g to occur from simple inelastic drag-down when the interior remains coupled. Additionally, the core-floor and perimeter-floor connections must not only survive rotation and ductile elongation well in excess of anything which seems reasonable, the connections and the floor assemblies must remain sufficiently intact to mediate tension to accomplish this pull down.

I'm not saying the above can't happen, WTF do I really know? I'm saying it's highly speculative and seems unlikely to me, given the existing literature on small angle failures of all types of connections, welds and fasteners. This is why I've suggested the notion of a debris train providing sustained impulse against the floor assemblies (near the walls), which permits things to break apart as they are wont to do in a collapsing piece of shit like that building apparently was, yet accounts for the appearance of a continuous force applied over a substantial period of time. A series of discrete impulses which do result in disintegration but result in acceleration in total. The differential speed between the moving debris and the impulsively entrained perimeter would be the cause, and would still pretty much require negligible resistance below. No hard coupling like the drag down - more like the action near stall speed in a torque converter within an automatic transmission, or a clutch while partially engaged.

Or, one could take the less speculative route which is that sub-g acceleration for all descent is within error bands of measurement, so maybe there's no over-g to explain.

I suppose I could be misinterpreting; is it the crane boom getting the jerk from the falling piece?

If so, no to that, too. This is a short duration impulse, the cable can be seen to go slack again immediately after the discrete jerk. After that, any over-g of the boom is simply the pivoted tipping described above.

People keep overlooking the simple fact that a short discrete impulse does not a sustained period of over-g make!

OneWhiteEye wrote:The mechanism of over-g in this video is the same as any possible over-g condition in the tipping brick wall and the tipping entertainment center we were discussing in the other thread.

The mechanism of over-g in this case is a pull of the structure downwards. Should the cable have been more elastic, as were presumably the girders and beams in WTC7, elasticity which was also favored by the horizontal disposition, instead of a very quick progress from zero to more than zero (the "discrete" jerk you mention), the progression would have been slower. Ignoring the magnitude of the elasticity of the cable, which basically only influences the magnitude and duration of the acceleration peak, the mechanism is most likely the same.

I never claimed that tipping over was a mechanism in WTC7. That was just a misinterpretation of a side point I made.

Furthermore, in this case it is very unclear if over-g was reached at all due to tipping, because even if it's apparent that the arm accelerates as it falls, it also seems that there is some resistance offered by the hinge. Not enough to prevent it to accelerate due to its own weight or pulled by the cable (which still exerts force even if not tight), but enough as to make the tip hit the ground not at the same speed as if it had free-fallen. That being purely estimated by eyeballing and therefore subject to error.

The hinge still opposed negligible resistance, though, in that it can't hold its own weight, so it would definitely be overwhelmed if it was carrying a heavy weight, whose fall would probably have been indistinguishable from free fall if that bending was the only force opposing it.

Back to the elasticity issue, if the cable was more elastic, it would probably have dampened the dynamic effect, possibly to the point of having prevented failure. But for WTC7 we're in the case where the façade was most likely unable to hold its own weight plus that of the failing core pushing downwards through the girders and beams. So the similarity, albeit limited, is valid, disregarding the fact that in the case of the crane the failure is due to a dynamic effect while in WTC7 it's less clear how much of the effect was dynamic and how much static.

Based on these last explanations, I concede.

It is on the basis of this statement that I concede:

pgimeno wrote:But for WTC7 we're in the case where the façade was most likely unable to hold its own weight plus that of the failing core pushing downwards through the girders and beams.

This is more in line with what I lean towards, where elastic response is secondary to a series of impulses from a debris train. Not that I don't have problems with this, too, but at least it seems possible.

Please see this post for more thoughts on the subject.

The point on elasticity was directed at addressing this:

OneWhiteEye wrote:I suppose I could be misinterpreting; is it the crane boom getting the jerk from the falling piece?

If so, no to that, too. This is a short duration impulse, the cable can be seen to go slack again immediately after the discrete jerk. After that, any over-g of the boom is simply the pivoted tipping described above.

People keep overlooking the simple fact that a short discrete impulse does not a sustained period of over-g make!

In the crane case, a much heavier falling object with the same elasticity of the cable would probably have caused a very similar acceleration profile, perhaps with a slightly sharper peak, not sure. A more elastic cable with a much heavier object would have caused a longer over-g period given enough distance to the floor.

Regardless of the point on elasticity, this video remains a good example to illustrate a documented case of a structural collapse where over-G is reached clearly due to pulling down, which is the probable mechanism in WTC7, and I've already used it several times for such purpose.

Interesting set of demolitions in China. You have to endure an advertisement first:

http://www.itn.co.uk/And%20Finally/54597/twin-skyscrapers-blown-up-in-china



Interesting rotational movement on one of them. That is a great example of an actual hinge rotation.