The 9/11 Forum

Additional Comment, October 4th 2008:

Here I wish to return to the problem of collapse initiation of WTC 7 and the motion of the roofline of the building in the time period bounded by the collapse of the East Penthouse and the downward motion of the entire building. This time interval corresponds to NIST’s computer simulation times, at least in its LS-DYNA model, from ~ 15 to 25 seconds - beginning with the failure of core column No. 79, 14.7 seconds into the simulation, and ending with a global downward motion of the roofline of the building, (with velocities in excess of 10 m/s), 24.5 seconds into the simulation. – see NCSTAR 1-9A, Executive Summary, pages xlii to xliv.

NIST’s “best” description of the destruction of WTC 7 on the afternoon of 9/11/01 found that global collapse initiation of the building occurred approximately 23.5 seconds into its computer simulation. However, the same simulation shows that already, at 22.2 seconds, the eastern end of the north face roofline of WTC 7 was moving downwards at ~ 1 m/s – see Figure 4-52 of NCSTAR 1-9A. Nevertheless, Figure 4-53 of NCSTAR 1-9A, which NIST claims represents the condition of the north face of WTC 7 approximately 23.5 seconds into the simulation, shows that there was no significant downward movement of the north west corner of the building at this time. Thus NIST’s computer model of the collapse of WTC 7 requires that the downward motion of the north east corner of the building started 1.3 seconds before the north west corner.

This result of NIST’s FEA analysis should be compared with videos of the collapse of WTC 7 which show that, in spite of a well-defined kink near the upper center of the north face, the top half of the building descended with a precisely maintained horizontal roofline for at least 3 seconds of downward motion. Indeed, measurements show that as building 7 collapsed, deviations from a perfectly horizontal roofline were well within ½ a floor height, (or 2 meters), for at least the first few seconds of downward motion.

If we also consider that the roofline of building 7 descended with an acceleration of about 80 % of g, we must conclude that the eastern and western ends of the roofline started to move downwards within 0.7 seconds of each other. Unfortunately for NIST, this is contrary to its FEA model which predicts that the north east corner of the north facade began descending 1.3 seconds before the north west corner. While such differences in timing may appear to be small, they are quite enough to seriously discredit NIST’s FEA model. Thus in 1.3 seconds, at an acceleration of 80 % of g, the north east corner of WTC 7 would have dropped 6.6 meters before the north west corner started moving – enough to cause the roofline to fall in a visibly non-horizontal fashion. Indeed, if the north east corner of the north face of WTC 7 started dropping 1.3 seconds ahead of the north west corner, the roofline would be almost 15 degrees off horizontal within 3 seconds of the collapse – a degree of asymmetry in the collapse of WTC 7 that was never observed.

One final point of note: NIST states that, (see page xliii of the Executive Summary of NCSTAR 1-9A), global vertical motion of WTC 7 started 23.5 seconds into its FEA simulation and, at 24.5 seconds, the roofline of WTC 7 was falling with a velocity of approximately 10 m/s to 15 m/s. Now even if we take the lower velocity of 10 m/s, this means, (at least according to NIST), that the acceleration of the roofline of building 7 during the first second of its collapse averaged 10 m/s^2 which looks a lot like free fall!

Thanks for the PM's, ...... (yes, I did get them), .... but have been too busy to reply - for which I apologize!

No problem.

Photographic record: 1 point,

NIST: 0 points.

Yep. Pretty solid case you have. There seems to be no valid rebuttals, or any rebuttals at all anymore.

This is the stuff I'm interested in, for sure, but I hope your other points in the comment are not getting short shrift for the sake of the fall times and perimeter deformation. I'm not in a position to evaluate these points, but they would seem to be at least as, if not more, important.

How does this work? Will NIST ever speak to you on these issues, or do you just wait and see if they pay any attention by examining their next draft?

OneWhiteEye:

The first step would be for NIST to post (on its very own website?) all the "Comments on the WTC 7 Draft Report" it received before its self-imposed deadline (Sept 15th, 2008, if I remember correctly ...... ).

And by the way, NIST did indeed post the "Comments" it received on its WTC 1 & 2 Report - I know, I am on the list, and can still pull up my contribution.......

However, if NIST cannot publish the "Comments" it received on its WTC 7 Report, why did it ask for public comments?

The hollow shell is also one of my humble objections. I also don't see the link with the simulation.

In the simulation it looks like a complete part of the building fails. In the video we see that the penthouses collapse first, we also see that a couple of stories lower some glass breaks, but the internal structure at that point is not completely falling within an empty shell. That's not possible, first of all the shell is connected with the building and secondly video observation (eyeballing) also doesn’t show this. The almost horizontal roofline and lines of the stories indicate that it fell with keeping its shape. There is something to add and that is that the building caves in a little bit at each position just before the global collapse, I mean the horizontal lines become a very little bit U-shaped.

My conclusion is (from video observation) that before the global collapse (starting at t0 [roofline starts moving down]) all columns where simultaneously broken on at least one level.

There are videos around in which columns are blow up in a sequence, for example from left to right. There you see that at the parts where the columns are cut it starts immediately falling as a whole, no horizontal roofline. I have to say that the NIST simulation looks more like a natural collapse than wtc7’s collapse. I’ve seen other simulations, also one of the Oklahoma building, in that one they say also one column failed and a progressive collapse occurred, the simulation also looked natural, the building was half gone but the remaining columns kept it in the air. It was of course no 47 story building but will that differ ? The taller it is the more difficult it is to get it into its footprint isn’t it ?
A simulation of wtc7’s collapse is believable if it can reproduce the crush-up hypothesis.

Einsteen,

Nice post!

However, when you say:

My conclusion is (from video observation) that before the global collapse (starting at t0 [roofline starts moving down]) all columns where simultaneously broken on at least one level.

I agree, except I would make it:

"All columns were broken within ~ 0.5 seconds of each other on at least one lower level."

By the way, the WTC 7 rubble pile was remarkably compressed. Take a look at this image:

http://911research.wtc7.net/wtc/evidenc ... pile8.html

That's all that was left of 47 floors!

Where does that number come from, I didn't calculate it but should that not be much lower if the roofline stays horizontal ? Did you assume a resolution of 0.3 meter or something like that or used the difference in the graphs ? I forgot a lot what I've read...

If you look at the roofline its shape changes from

y=0 on [-w/2,w/2]

to

y=kx^2 on [-w/2,w/2], with a small k

in a fraction of a second. It could be an other function but this is the lowest order approach.

Since I see it with the bare eyes that implies that k>= resolution/w^2

The penthouses of course imply that some parts fail first but at a certain moment it goes very fast, of course I'm not sure about it like nothing is sure...

Einsteen:

I am ignoring (in my argument for now) the bowing of the roofline and simply considering the motion of the east and west ends of the roof. I think we can agree that for the roofline to descend in a perfectly horizontal manner, the east and west ends of the roof must begin their downward motions at exactly the same time, i.e. they must have the same t(0).

But we need to consider how "horizontal" the roofline really was. Now I would say, (and I think NIST does too),that the east corner of WTC 7 "let go" a tad before the west corner. This is confirmed by looking at some of the video frames reproduced in the Draft Report such as Figures 5-207 and 5-209. In these images, taken about 1 second into the collapse, we can see that the east side of WTC 7 is marginally lower than the west side by about 1/2 a story height or let's say 2 meters. But this was ~ 1 second into the collapse - the drop difference will be less at the start of the collapse. So let's assume the east side dropped 1 meter before the west side moved. Then we have:

s(west) = 0

s(east) = 1 = 1/2.a.t^2

Hence, taking a to be ~ 8 m/s^2, we have t^2 = 2/8, or t = 0.5 seconds.

This is why I suggested that the observed "horizontal" roofline requires the east and west ends of WTC 7 to start moving within about 0.5 seconds.

I don't understand that, if the difference is 1 meter, but taken 1 second after collapse initiation, should
you then not write

s(west)=(1/2)a(T-t)^2
s(east)=(1/2)aT^2

=> s(east)-s(west)=(1/2)a[T^2-(T-t)^2]

and then use T=1, s(east)-s(west)=1, a=8 ==> t= 0.118 sec

But I still don't understand what an assymetry far after collapse initation has to do with the speed columns fail at collapse initiation. Even if they fail exactly at the same time then due to chaos it is going to tilt. Look at the Landmark tower, that one is also going to tilt later although it is timed to fall symmetrically.

ps. the bowing of the roofline implies that all internal columns failed at about the same time except the ones at the perimeter.

Einsteen,

If you read my last post you will see that I am suggesting that the drop difference was about 2 meters 1 second into the collapse, (just as a rough estimate!). I then extrapolate to t(0) and guess that the drop difference was ~ 1 meter at collapse initiation. This gives a t = 0.5

Or you can calculate, as you do (which I agree is a better way!), from a 2 meter drop difference at T = 1. In this case Ithink you get t = 0.29 seconds.

Ok, that's what you mean. But at collapse initiation the drop difference is zero of course, by definition. It's cool that nist also used the method of substracting the images from different frames, there seems also a kind of motion vector filter for video that should be able to visualize the difference between frames, looks nice, maybe for later http://compression.ru/video/motion_esti ... ex_en.html

From the new video of WTC7, the first significant visible deformation of the west face occurs late and is essentially the top breaking:

OneWhiteEye:

Yes, that's a very interesting video. I wonder why it took so long to appear, ....... and why now? The image you have captured looks to be about 6 seconds into the collapse when the drop was about 80 meters and the tilting to the south, (of the upper 6 floors), was at least 6 degrees ....

However, I'm having problems with the rate of fall of the building in this video. As best I can tell, it appears to take about 7 seconds for the roofline to drop about 100 meters, which implies an average acceleration of only about 4 m/s^2.

EDITED ON NOV 4:

Looks like I was using a version of this new video that was in slow motion! (Sorry, I should have spotted that sooner!). There are two clips of the WTC 7 collapse on YouTube and the 2nd, which is the one I used to make some timing estimates, is in slo-mo. Using the real-time version I now would say that OWE's frame capture is about 4 seconds into the collapse, in which case the average acceleration is at least 8 m/s^2. However, I believe the descent quickly moves to a relatively constant (terminal?!?) velocity.

Dr. G:

I, too, wonder about the release of this (and the WTC1) video. I suppose there is a lot more out there that will never become public.

As for the fall rate... timing, timing, timing. This video is excellent for providing a close view from a new angle but not very good for measurements because of the panning and zoom.* Makes it all the more difficult to establish a t(0). Because the top is not visible much in the crucial time period, I fixed on a spot on the NW corner edge adjacent to the building roofline on the left. Now, if you ignore the dropping going on elsewhere - which is proper to do for the NW corner - then the t(0) for the corner is around 2.2 seconds into the footage. About four seconds have elapsed when 17 regular stories hit that same location. Understand this is very rough, but that's an average of 8.4 m/s^2. It does seem to be reasonable, but it's all about catching the corner t(0), much harder in this video than the other two we've been dealing with lately.

It may be slowing down more noticeably at this point. Not that I see that in this video yet but there is that manual data I took last year with an obvious tail-off. While this video may not yield suitable measurements in the first half of the visible displacement, they could be really good in the last half. Eventually, with sufficient accuracy of measurement and sync points, this video could extend the data we have.

The only obvious deformations visible are the roofline/face kink (or fold) and the upper part kinking over as seen above. While NIST's simulations did demonstrate solutions leading to catastrophic failure in their model, the dynamics depicted by those simulations most assuredly do not resemble the actual dynamics. This new angle leaves little doubt that your comment in this matter is solid.

* I do not favor stabilization as an intermediate step towards acquiring hi-res data. For that matter, I favor no transformations or filtering in this context. For quick and dirty measurements, though, anything goes.

http://www.nist.gov/public_affairs/techbeat/tbx2008_1120_wtc7.htm

.

http://wtc.nist.gov/comments08/

The comments.