To add comments or start new threads please go to the full version of: Physics Of 9/11 Events - Part 3
PhysForum Science, Physics and Technology Discussion Forums > General Sci-Tech Discussions > Other Sci-Tech Topics
Pages: 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, 100, 101, 102, 103, 104, 105, 106, 107, 108

OneWhiteEye
12th December 2007 - 05:22 PM
QUOTE (David B. Benson+Dec 11 2007, 09:45 PM)
However, this run using C449 suggests that the assumption of a constant is ever so slightly better:

Would you mind summarizing your current results with all the adjustments, tweaks, and corrections?
David B. Benson
12th December 2007 - 06:44 PM
Same five hypotheses using C447 data:

CODE
BV-ZS-F-exp-pow-stretch   dB= 0.0 sd= 0.118

Sef-K+Z+ZSS-F-exp-pow-stretch dB= 1.3 sd= 0.179
BV-K+Z+SS-F-exp-pow-stretch dB= 2.7 sd= 0.223
BV-K+Z+ZSS-F-exp-pow-stretch dB= 3.3 sd= 0.236
Sef-K+Z+SS-F-exp-pow-stretch dB= 6.1 sd= 0.310


Here only the last is substantially disconfirmed.

Summary coming up shortly! smile.gif
David B. Benson
12th December 2007 - 07:12 PM
Summary (2007 Dec 12)

The original, simple hypothesis of a constant force with
constant stretch is decisively disconfirmed in favor of
X-ZS-F-exp-pow stretch in which the force function is
of the form

kZS

with k constant (or very slightly increasing with Z), and
X is either BV or Sef. This holds for
other resistive force function forms with a constant stretch.

Alternatives which do almost as well and so are not (yet)
disconfirmed by the naive Bayes factor method include
BV-K+Z-F-exp-pow-stretch
BV-K+Z+SS-F-exp-pow-stretch
BV-K+Z+ZSS-F-exp-pow-stretch
Sef-K+Z+ZSS-F-exp-pow-stretch.

However, when either the BV form or the Sef form is
disconfirmed, this throws a shadow on the other member of the
pair.

Tentative Conclusion
The resistive force function of the form

kZS

summarizes a wide variety of dissipative paths by which energy
was consumed. Rather than attempting to discover the form and
amount of these paths, along the lines of

k0 + k1*Z + k2*S*S + ...

we instead have a simple summary form, with only one parameter
to estimate. However, there is no particular reason to suspect
that this summary form describes what happened beyond the first
3.82 seconds of the progressive collapse.
einsteen
12th December 2007 - 08:53 PM
It would still be interesting to see what the range of possible results would be. If the function provided by OneWhiteEye is called f(t) then an error margin leads to two functions f_min(t) and f_max(t), for which

f_min(t) <= f(t) <= f_max(t) for t in [0s,3.82s]

the correction for tilt leads to g_min(t) and g_max(t)
the correction for perspective/etc leads to h_min(t) and h_max(t)
and so on. Shouldn't an error margin be part of the result.

If L_min(t) and L_max(t) are the Landmark functions then it is possible that the intersection of those ranges lead to a couple of possible functions, let's say
p1(t),p2(t), etc, then if there is no physical reason to pick p_DBB(t) then why would that be the right one. And I don't get the single column data. Maybe I will understand it if your final version is available...
David B. Benson
12th December 2007 - 09:22 PM
QUOTE (einsteen+Dec 12 2007, 01:53 PM)
Shouldn't an error margin be part of the result.

And I don't get the single column data.

The error is expressed as the weighted standard deviation between the measured data and the computed data.

'Single column data'? I don't follow what you are referring to...
David B. Benson
12th December 2007 - 10:44 PM
Using C447 data, the 'vertical avalanche' equation for the force is tested:

CODE
BV-ZS-F-exp-pow-stretch   dB= 0.0 sd= 0.133

BV-ZSS-F-exp-pow-stretch  dB=29.4 sd= 0.688
Sef-ZSS-F-exp-pow-stretch dB=44.0 sd= 0.856


But then I noticed a problem (parametr estimation tecchnique may find a poor local minimum), so I changed the starting values, and

CODE
BV-ZSS-F-exp-pow-stretch  dB= 0.0 sd= 0.109

BV-ZS-F-exp-pow-stretch   dB= 0.2 sd= 0.120
Sef-ZSS-F-exp-pow-stretch dB=65.1 sd= 0.856


so the resisting force function form

kZSS

is the winner, with the B&V equation. (Possibly the problem with the Seffen equation is not starting well, dunno yet.)
David B. Benson
12th December 2007 - 11:59 PM
Vertical avalanches with C449 data:

CODE
BV-ZSS-F-exp-pow-stretch  dB= 0.0 sd= 0.069

Sef-ZSS-F-exp-pow-stretch dB= 0.4 sd= 0.084
BV-ZS-F-exp-pow-stretch   dB= 0.7 sd= 0.098


Wowser! wink.gif smile.gif
OneWhiteEye
13th December 2007 - 05:01 AM
QUOTE (David B. Benson+Dec 12 2007, 07:12 PM)
Summary (2007 Dec 12)
...

Thanks. It's good to get an overview of what's going on.
OneWhiteEye
13th December 2007 - 05:11 AM
QUOTE (David B. Benson+Dec 12 2007, 09:22 PM)
(to einsteen)'Single column data'?  I don't follow what you are referring to...

I think einsteen is wondering why there are so many runs now using only a single column of data. Originally, when I asked you:

QUOTE
Could you run each of C447-449 independently and produce three tables like that?


your response was:

QUOTE (->
QUOTE
Could you run each of C447-449 independently and produce three tables like that?


your response was:

I could, but am unlikely to do so.


What made you change your mind?
OneWhiteEye
13th December 2007 - 06:39 AM
QUOTE (David B. Benson+Dec 12 2007, 09:22 PM)
The error is expressed as the weighted standard deviation between the measured data and the computed data.

einsteen refers to propagation of error throughout the analysis. Perhaps you are addressing this and I'm missing it; perhaps, through some operation I don't understand, it doesn't apply in this case.

In that vein...

Given a dataset which has an associated error of +/- x units, how can any method distinguish between the infinite number of curves that lie within that band over that interval? Simply being close doesn't always count.

Hypothetical: suppose a data recorder has been rated as accurate to +/- 1 unit. It is fed a calibration signal and the captured data is compared to the known input. If the calibration signal is a straight line defined by, say, f(x) = x + 1, then that input, along with a tolerance band of +/- 1, looks like this:

User posted image
http://i6.tinypic.com/8676ogp.png

Say the recorder produces the following output:

User posted image
http://i15.tinypic.com/6ldofvo.png

which, when compared with the input and acceptable error range, is seen to be within the rated error.

User posted image
http://i3.tinypic.com/6y4ej4h.png

In this scenario, the input is known and the measurement is judged against an error band on the known value. This situation is somewhat reversed from what we have, but is quite instructive nevertheless.
OneWhiteEye
13th December 2007 - 06:49 AM
...because I'm going to take away the calibration signal. Now, given an arbitrary signal, we no longer have the benefit of a calibration curve with which to transform away the systematic error of the recorder, we just have to trust it to be accurate within +/- 1 pixel (oops, did I say pixel? I meant unit)

Incidentally, the curve I used to generate the 'recorder output' was

0.75*exp(0.05*x)*x - 0.025*x^2 + 1.5

Hardly a straight line, but well within a straight line +/- 1 unit.

Please stay with me.
OneWhiteEye
13th December 2007 - 07:10 AM
If our hypothetical recorder just happens to be used to measure a signal that is truly a straight line, we'll get the output shown above. Given the recorder is known to be accurate to +/- 1, but only knowing that, we construct a graph showing the output with the error band:

User posted image
http://i16.tinypic.com/834zokj.png

Say you want to use the Bayesian method to distinguish among competing hypotheses H1 and H2 offered to explain the nature of the signal captured:

H1: f(x) = Ax + B
H2: f(x) = A*exp(Bx)x + Bx^2 + C

Uhh, which one wins?
OneWhiteEye
13th December 2007 - 07:16 AM
The latter, I'd hope, since it would be absurd to conclude the values represent a straight line, given the input data.

Now think of that error band as noise, but you have many, many points. Apply the Bayesian methods again; what do you get? I'd wager you get the same result, with discrimination according to how bad the noise is versus the number of points. But, no matter how many points you add, if they're randomly distributed around THAT curve, you will not select the correct hypothesis, f = Ax + B.

On the other hand, if you had even a few points from a thousand recorders, each with the same error band, but random variations in their skew, I'd bet Bayes selects the straight line.

Let's have a look at all the curves together, one more time:

User posted image
actual and recorded with bands

Do you see what I'm getting at, Dr. Benson?
OneWhiteEye
13th December 2007 - 08:06 AM
In case the answer is no, please understand that what I've given you is like that hypothetical data recording. I know -

- there is error up to 2 pixels
- it is not necessarily small compared to the signal
- it is not random

There is zero noise in that data, for all practical purposes. I don't mean a little, I mean none. Drift, yes. Some of it from my eye/hand coordination, but not much. This data is already smoothed by virtue of the process and my eyes are NOT 2 pixels off.

A plot of the C447-449 data:

User posted image
http://i11.tinypic.com/8evh1cl.png

A close-up of the latter portion:
User posted image
http://i13.tinypic.com/7y3wp69.png

The first clue is that the curves differ by as much as almost two pixels... does that even make any sense for adjacent columns? Well, yes it does.

I don't know how much this affects your results, but it should be of concern until it's resolved. IMO.
OneWhiteEye
13th December 2007 - 06:08 PM
Correction:

H2: f(x) = Aexp(Bx)x + Cx^2 + D
David B. Benson
13th December 2007 - 07:50 PM
I'll start replying to questions in a bit, but just now I post the vertical avalanche hypotheses using the C448 data:

CODE
BV-ZSS-F-exp-pow-stretch  dB= 0.0 sd= 0.123

Sef-ZSS-F-exp-pow-stretch dB= 0.1 sd= 0.133
BV-ZS-F-exp-pow-stretch   dB= 0.3 sd= 0.146


and later I'll rerun using the C447 data now that I have good starting values for these hypotheses. I am rather confident that the vertical avalanche hypotheses represent a highly likely (and simple) explanation for the progressive collapse. More on this later.
David B. Benson
13th December 2007 - 09:34 PM
Here is C447 with decent starting estimates for the parameters:

CODE
Sef-ZSS-F-exp-pow-stretch dB= 0.0 sd= 0.098

BV-ZSS-F-exp-pow-stretch  dB= 0.3 sd= 0.110
BV-ZS-F-exp-pow-stretch   dB= 0.5 sd= 0.120


So, vertical avalanches are the best for all three sets of data. wink.gif smile.gif
David B. Benson
13th December 2007 - 09:52 PM
QUOTE (OneWhiteEye+Dec 12 2007, 10:11 PM)
What made you change your mind?

At first? Just to satisfy your curiosity.

I learned something about proper data exploration from that.

You were right and I was wrong.
David B. Benson
13th December 2007 - 10:26 PM
QUOTE (OneWhiteEye+Dec 13 2007, 12:10 AM)
Say you want to use the Bayesian method to distinguish among competing hypotheses H1 and H2 offered to explain the nature of the signal captured:

H1: f(x) = Ax + B
H2: f(x) = A*exp(Bx)x + Bx^2 + C

Uhh, which one wins?

To apply the naive Bayes factor method, there has to be some estimate of the standard deviation of the random errors. I'll use your +/1 one unit to mean a standard deviation of 1.0.

With that, in the situation you describe, neither is substantially better than the other at describing the signal. It is buried too far down in the noise for either hypothesis to be better than the other.

=========================
In similar cases, I invoke the principal of parsimony, Ockham's Razor. By that criterion, H1 is to be preferred.
David B. Benson
13th December 2007 - 10:41 PM
QUOTE (OneWhiteEye+Dec 13 2007, 12:16 AM)
Do you see what I'm getting at, Dr. Benson?

If the issue is systematic error in the recording device, then the naive Bayes factor method will not be adequate to compare two hypotheses regarding the signal (assuming poor signal-to-noise ratio).

If the nature of the systematic error is known, it can usually be compensated for.
David B. Benson
13th December 2007 - 11:17 PM
QUOTE (OneWhiteEye+Dec 13 2007, 01:06 AM)
- there is error up to 2 pixels
- it is not necessarily small compared to the signal
- it is not random

Drift, yes.

I don't know how much this affects your results, but it should be of concern until it's resolved. IMO.

2 pixels out of 170 or so? Seems small to me. How do you know it is not random?

Explain your notion of drift further, please.

My opinion as well.

As for einsteen's original question, another way to consider the degree of consistency of the results is to compare parameter values for the three different pixel columns of data. I'll just give the parameter for the resisting force.

Sef-ZSS-F-exp-pow-stretch
C447: 3.3424
C448: 3.2569
C449: 3.1709

BV-ZSS-F-exp-pow-stretch
C447: 3.3795
C448: 3.3262
C449: 3.1088

To me, with up to 9% differences, it seems that the differences in the three data sets ought to be understood and removed. If the differences were all below say 5% I wouldn't bother.
David B. Benson
14th December 2007 - 01:31 AM
Then again, maybe it does not matter. I just finished running Sef-ZSS-F-exp-pow-stretch with the artificial constraint that the acceleration must always be positive (downwards). This resulted in the resistance force parameter being only 0.9942, with changes in the stretch shape parameters compensating enough so that the standard deviation remained a respectable 0.109.
OneWhiteEye
14th December 2007 - 05:52 AM
QUOTE (David B. Benson+Dec 13 2007, 09:52 PM)
You were right and I was wrong.

Thanks for saying that. Such a statement is an uncommon occurrence in message boards. It speaks volumes for your integrity.
OneWhiteEye
14th December 2007 - 06:01 AM
QUOTE (David B. Benson+Dec 13 2007, 10:26 PM)
To apply the naive Bayes factor method, there has to be some estimate of the standard deviation of the random errors.  I'll use your +/1 one unit to mean a standard deviation of 1.0.

There's something important here, but I'm going to bag it for later.

QUOTE
With that, in the situation you describe, neither is substantially better than the other at describing the signal.  It is buried too far down in the noise for either hypothesis to be better than the other.

Fabulously good answer.

QUOTE (->
QUOTE
With that, in the situation you describe, neither is substantially better than the other at describing the signal.  It is buried too far down in the noise for either hypothesis to be better than the other.

Fabulously good answer.

In similar cases, I invoke the principal of parsimony, Ockham's Razor.  By that criterion, H1 is to be preferred.

As well, another good answer.
OneWhiteEye
14th December 2007 - 08:47 AM
QUOTE (David B. Benson+Dec 13 2007, 11:17 PM)
2 pixels out of 170 or so?  Seems small to me.

It depends on when this 'drift' occurs. Here's a chart of the ratio of 2 pixels' error to the signal for the last 3.8 seconds of columns 447-449:

User posted image
http://i15.tinypic.com/82m2fis.png

It can't be like that, but that's the worst-case boundary over time. Without more precise quantification, it's impossible to rule out the potential for (say) 5 - 10% error over a 1-2 second span.

QUOTE
How do you know it is not random?

Because I know where it comes from, how it got there. Executive summary:

When you're plotting a smear that represents ONE pixel's width - the ONLY way you'll get a correct result is if the feature you're looking at STAYS at that same horizontal pixel location. If it moves laterally, then you are looking at something else over time. The dark band MOVES to the left 6 pixels by the end of the dataset, probably a lot of it early on. The band is shaped like an inverted 'U' so that curve is imposed on the real curve. And it is a curve, not a random deviation from true. Like a drift.

In recent posts I explained this in excruciating detail, probably too much. But I could not characterize it any better, so if you need more detail, I refer you to this post for an overview of the problem and this post for the examination which arrives at an estimate of 2 pixels' error.

QUOTE (->
QUOTE
How do you know it is not random?

Because I know where it comes from, how it got there. Executive summary:

When you're plotting a smear that represents ONE pixel's width - the ONLY way you'll get a correct result is if the feature you're looking at STAYS at that same horizontal pixel location. If it moves laterally, then you are looking at something else over time. The dark band MOVES to the left 6 pixels by the end of the dataset, probably a lot of it early on. The band is shaped like an inverted 'U' so that curve is imposed on the real curve. And it is a curve, not a random deviation from true. Like a drift.

In recent posts I explained this in excruciating detail, probably too much. But I could not characterize it any better, so if you need more detail, I refer you to this post for an overview of the problem and this post for the examination which arrives at an estimate of 2 pixels' error.

Explain your notion of drift further, please.

Slowly varying deviations. Practically speaking, very slow noise - one or two 'zero crossings' if any in a dataset. The way a car travels in a lane: the deviations from centerline plotted over a few miles might look like noise, but over a hundred meters - it's drift. Just what you might think it is.

OneWhiteEye
14th December 2007 - 09:47 AM
You know, I was terribly green at all this when I posted that data. I didn't know all I know now and I hadn't examined this video frame-by-frame many hundreds of times or blown up pixels to an inch across. Smear-o-grams are soooo easy. They're fine when:

P1) feature motion is collinear with the pixel slice (the 1D axis)
-OR-
P2) the feature is a thin STRAIGHT boundary, orthogonal to the slice and longer than any DRIFT off-axis
-OR-
P3) extreme accuracy is not required (quick-look curves, rough comparisons)

Either of the first two cases allow for accurate representation of motion, the last simply doesn't require that much accuracy. The C447-449 data, sadly, fails all three.

R1) the antenna moves left as it drops
R2) the band is not even close to a horizontal line
R3) even two pixels of deviation mixed up between three datasets causes significant rearrangement of rank in the Bayesian analysis

As to R3, the error almost assuredly causes the jambalaya effect in the results. It's about the only thing that could! And it should! Your analysis is very sensitive - it needs to be. The transcription is probably accurate to far better than a pixel. But the transcriptions are of curves that are already distorted. Like perspective in nature, but from artifact.

QUOTE
If the issue is systematic error in the recording device, then the naive Bayes factor method will not be adequate to compare two hypotheses regarding the signal (assuming poor signal-to-noise ratio).

Makes perfect sense, I think you can see that's been my take since we started discussing this. All I need to do is convince you the error is systematic, find out how large it is, then...

QUOTE (->
QUOTE
If the issue is systematic error in the recording device, then the naive Bayes factor method will not be adequate to compare two hypotheses regarding the signal (assuming poor signal-to-noise ratio).

Makes perfect sense, I think you can see that's been my take since we started discussing this. All I need to do is convince you the error is systematic, find out how large it is, then...

If the nature of the systematic error is known, it can usually be compensated for.

Indeed. I thought it wouldn't be too hard to characterize the deviation over time for a correction, but I no longer feel that way for a variety of reasons. Let's hash out the matters above, then discuss this. Honestly, had I spent as much time working towards automation as I had explaining why the C447-449 set is questionable, you'd be able to run the best data already.

QUOTE
I'll use your +/1 one unit to mean a standard deviation of 1.0.

I'm still not ready to deal with that seemingly simple statement. It's really the crux of the matter but I'm talked out right now. ('thank god!' says spook 1, 'fat chance' says spook 2)
OneWhiteEye
14th December 2007 - 10:25 AM
When you try to draw a line with a straightedge and a ball-point pen, the pen must be held at a constant angle or the tip DRIFTS and you don't get a straight line. Little kids in pre-school experience this! This is like the error in C447-449! Your analysis does not tolerate systematic distortion, even small amounts for some hypotheses, nor should it. OBJECTS IN MIRROR ARE CLOSER THAN THEY APPEAR!

Soon enough, you'll have 114 points for each feature over the same 3.82 seconds, 10x the position resolution, and probably only marginally greater standard deviation than placing points by hand. Full 2D tracking, ONLY perspective correction needed. It will look a bit noisy but, as we know, that's OK. Stochastic methods prefer stochastic input and lots of it.

Laser precision, in the final analysis, is my prediction. Your rankings may look a bit different than they do when tested against the clean, but slightly warped, datasets of columns 447 - 449.

('told ya' says spook 2)
adoucette
14th December 2007 - 01:10 PM
The National Construction Safety Team (NCST) Advisory Committee (Committee), National Institute of Standards and Technology (NIST), will meet via teleconference Tuesday, December 18, 2007, from 1 p.m. to 3 p.m. The meeting will be audio webcast so that the public may listen to the meeting as it takes place.

http://wtc.nist.gov/media/NCSTACmeetingDec18_2007.htm

Webcast URL:
http://origin.eastbaymedia.com/~nist/asx/nist-wtc-121807.asx


The primary purpose of this meeting is for the NCST Advisory Committee to
discuss its annual report to the Congress and for NIST to update the Committee
on the status of the investigation of World Trade Center 7. The meeting will
be conducted via teleconference with a live audio webcast. The final agenda
will be posted on the NIST Web site at http://www.nist.gov/ncst.
Individuals and representatives of organizations who would like to offer
comments and suggestions related to items on the Committee’s agenda for
this meeting, are invited to request a place on the agenda. Approximately
one-half hour will be reserved for public comments, and speaking times will be
assigned on a first-come, first-served basis. The amount of time per speaker
will be determined by the number of requests received, but is likely to be 5
minutes each. Questions from the public will not be considered during this
period. Speakers who wish to expand upon their oral statements, those who
had wished to speak but could not be accommodated on the agenda, and those
who were unable to attend in person are invited to submit written statements to
the National Construction Safety Team Advisory Committee, National Institute
of Standards and Technology, 100 Bureau Drive, MS 8611, Gaithersburg,
Maryland 20899–8611, via fax at (301) 975–6122, or electronically by e-mail to
ncstac@nist.gov.
David B. Benson
14th December 2007 - 07:23 PM
QUOTE (OneWhiteEye+Dec 14 2007, 01:47 AM)
Slowly varying deviations. Practically speaking, very slow noise - one or two 'zero crossings' if any in a dataset. The way a car travels in a lane: the deviations from centerline plotted over a few miles might look like noise, but over a hundred meters - it's drift. Just what you might think it is.

Drift is then an example of red noise. Red noise is some effect treated as random but which is correlated with itself over 'short' periods of time.

One possible cause of red noise in the C447--9 data is the lensing effect of hot air pulses rising. Another possible effect, towards the end, is that the tower top is moving at about 27 m/s. That ought to cause turbulence in the air from the roof line to the antenna tower. Since the vortexes have pressure gradients, again (I think) there will be a lensing effect.

But more: the flowing avalanche resistive force implies that the equation I am solving, but also the tower itself may have slow, fairly small, instabilities in the solution and in nature. An extreme example, much greater than the effect here, is in a pulse jet engine, which I hope suffices to understand the nature of this less than perfectly smooth flow. This effect might explain some of the red noise deviations between the data and the calculated points.

Strictly speaking, it is something of a cheat to assume Gaussian noise when we know that it is red noise. However, if the red noise correlation with itself is of short duration (the case here), then naively assuming Gaussian noise only contaminates the Bayesian factor method slightly.

Taking everything said so far into account, later today I'll do a computer run of the favored hypotheses using the combined data, just to see what may be seen.
David B. Benson
14th December 2007 - 07:53 PM
QUOTE (OneWhiteEye+Dec 14 2007, 02:47 AM)
R3) even two pixels of deviation mixed up between three datasets causes significant rearrangement of rank in the Bayesian analysis

Let's hash out the matters above, then discuss this.

Rank rearrangement is actually a non-issue. What matters is the number of decibans down hypotheses stand in comparison to the best fit. That appears to change little enough that hypotheses disconfirmed by one data are (mostly) disconfirmed by all three. The one hypothesis for which this was not so was then discarded as too sensitive to exact data values.

I'm a bit lost. Which matters are those? Or did my previous post address (at least some of) them?
OneWhiteEye
15th December 2007 - 06:50 AM
QUOTE
Drift is then an example of red noise. Red noise is some effect treated as random but which is correlated with itself over 'short' periods of time.


I sincerely regret throwing out that example. I've had a grueling week and that was not my best writing. It's foolish to offer such a poor example when I've already rigorously and thoroughly characterized the effect in previous posts. It's insane to offer yet more explanations.

I'm exhausted, and next week promises to be worse. There's part of me that wants to carry this through, but the greater part of me says...

OK
David B. Benson
15th December 2007 - 06:13 PM
Another possible cause is the hat truss acting as a spring with the antenna tower riding 'up-n-down' on top of it, a motion superimposed on the general downtrend.

In any case, I am now going to s6tart working on determining the tilt. I have a plan. smile.gif

OneWhiteEye --- Don't over do it. There is no rush.
OneWhiteEye
15th December 2007 - 06:35 PM
QUOTE (David B. Benson+Dec 15 2007, 06:13 PM)
Another possible cause is the hat truss acting as a spring with the antenna tower riding 'up-n-down' on top of it, a motion superimposed on the general downtrend.

Distinct possibility. As well:

QUOTE
One possible cause of red noise in the C447--9 data is the lensing effect of hot air pulses rising. Another possible effect, towards the end, is that the tower top is moving at about 27 m/s. That ought to cause turbulence in the air from the roof line to the antenna tower. Since the vortexes have pressure gradients, again (I think) there will be a lensing effect.


No doubt. (edit: but I know they're all a lot less than the distortion I've been describing)

QUOTE (->
QUOTE
One possible cause of red noise in the C447--9 data is the lensing effect of hot air pulses rising. Another possible effect, towards the end, is that the tower top is moving at about 27 m/s. That ought to cause turbulence in the air from the roof line to the antenna tower. Since the vortexes have pressure gradients, again (I think) there will be a lensing effect.


No doubt. (edit: but I know they're all a lot less than the distortion I've been describing)

In any case, I am now going to s6tart working on determining the tilt.

Perhaps a little later I can post something on this, might be helpful. The first and simplest approximation is to assume the target motion deviates only infinitesimally from the optical axis, then use a calculation similar to your perspective correction.

The basic distinction is in using the differential form of the tangent relation (that is, use theta + delta_theta). First, take the difference band - dish for each frame. Then, for the resulting set, obtain the set of differences (p(t+i) - p(i)); this is simply the first difference wrt time, which is the result obtained when considering delta_y = Z(tan(theta + delta_theta) - tan(theta)).

Obviously, this is a crude approximation, but one does not need to know the actual distance between the antenna features, as the work is with differences. Knowing the rotation about a pivot below the roofline results in a translation (down and away) and equivalent rotation about the lower feature, it's a good guess that the translation will not matter too much as compared with the angle change. One can solve for the delta_theta for each frame and add it to the initial angle.
Trippy
15th December 2007 - 06:37 PM
Spending 12 hours a day crushing and grinding (Gold) ore samples for assaying has taught me a few things.

Like it doesn't take a whole lot of pressure to make schist fail explosively (What does seem to matter is power and impulse).

I do have a question though, based on something I observed last night. What was the average size of the debris in the dust cloud? (The dust particles I mean, I'll go into more detail later.

(edited to translate post into english).
David B. Benson
15th December 2007 - 08:57 PM
QUOTE (OneWhiteEye+Dec 15 2007, 11:35 AM)
The basic distinction is in using the differential form of the tangent relation (that is, use theta + delta_theta). First, take the difference band - dish for each frame. Then, for the resulting set, obtain the set of differences (p(t+i) - p(i)); this is simply the first difference wrt time, which is the result obtained when considering delta_y = Z(tan(theta + delta_theta) - tan(theta)).

Obviously, this is a crude approximation, but one does not need to know the actual distance between the antenna features, as the work is with differences. Knowing the rotation about a pivot below the roofline results in a translation (down and away) and equivalent rotation about the lower feature, it's a good guess that the translation will not matter too much as compared with the angle change. One can solve for the delta_theta for each frame and add it to the initial angle.

Interesting idea. smile.gif

But it would be helpful to have your estimate of the distance from the roofline to the dish(es).
lozenge124
16th December 2007 - 03:54 AM
NIST releases additional faq supplement:

http://wtc.nist.gov/pubs/factsheets/faqs_12_2007.htm

Question 1 is particularly interesting:
QUOTE
1. Was there enough gravitational energy present in the World Trade Center Towers to cause the collapse of the intact floors below the impact floors?  Why was the collapse of WTC 1 and 2 not arrested by the intact structure below the floors where columns first began to buckle?

Yes, there was more than enough gravitational load to cause the collapse of the floors below the level of collapse initiation in both WTC Towers.  The vertical capacity of the connections supporting an intact floor below the level of collapse was adequate to carry the load of 11 additional floors if the load was applied gradually and 6 additional floors if the load was applied suddenly (as was the case).  Since the number of floors above the approximate floor of collapse initiation exceeded six in each WTC Tower (12 and 29 floors, respectively), the floors below the level of collapse initiation were unable to resist the suddenly applied gravitational load from the upper floors of the buildings.  Details of this finding are provided below:

Consider a typical floor immediately below the level of collapse initiation and conservatively assume that the floor is still supported on all columns (i.e., the columns below the intact floor did not buckle or peel-off due to the failure of the columns above).  Consider further the truss seat connections between the primary floor trusses and the exterior wall columns or core columns.  The individual connection capacities ranged from 94,000 lb to 395,000 lb, with a total vertical load capacity for the connections on a typical floor of 29,000,000 lb (See Section 5.2.4 of NIST NCSTAR 1-6C).  The total floor area outside the core was approximately 31,000 ft2, and the average load on a floor under service conditions on September 11, 2001 was 80 lb/ft2.  Thus, the total vertical load on a floor outside the core can be estimated by multiplying the floor area (31,000 ft2) by the gravitational load (80 lb/ft2), which yields 2,500,000 lb (this is a conservative load estimate since it ignores the weight contribution of the heavier mechanical floors at the top of each WTC Tower).  By dividing the total vertical connection capacity (29,000,000 lb) of a floor by the total vertical load applied to the connections (2,500,000 lb), the number of floors that can be supported by an intact floor is calculated to be a total of 12 floors or 11 additional floors.

This simplified and conservative analysis indicates that the floor connections could have carried only a maximum of about 11 additional floors if the load from these floors were applied statically.  Even this number is (conservatively) high, since the load from above the collapsing floor is being applied suddenly.  Since the dynamic amplification factor for a suddenly applied load is 2, an intact floor below the level of collapse initiation could not have supported more than six floors.  Since the number of floors above the level where the collapse initiated, exceeded 6 for both towers (12 for WTC 1 and 29 for WTC 2), neither tower could have arrested the progression of collapse once collapse initiated.  In reality, the highest intact floor was about three (WTC 2) to six (WTC 1) floors below the level of collapse initiation.  Thus, more than the 12 to 29 floors reported above actually loaded the intact floor suddenly. 


Note, no reference to Bazant, no discussion of core & perimeter column strength, no discussion of upper block tilting, the column to column bolts, no discussion of "wedging", etc... According to NIST, you just need to consider 2 parameters to determine the inevitability of collapse past initiation: truss seat capacity & floor load! To anyone paying attention this is MIND-BOGGLING! Apparently, you can make a tube in tube structure with core & perimeter support columns of infinite capacity, and as long as you overload the truss seat connections of 1 floor the whole thing will come crashing down (progressively of course).

And just in case, you are thinking there has to be more to this, NIST cannot simply be waving this problem away like this! Check question 10.
QUOTE (->
QUOTE
1. Was there enough gravitational energy present in the World Trade Center Towers to cause the collapse of the intact floors below the impact floors?  Why was the collapse of WTC 1 and 2 not arrested by the intact structure below the floors where columns first began to buckle?

Yes, there was more than enough gravitational load to cause the collapse of the floors below the level of collapse initiation in both WTC Towers.  The vertical capacity of the connections supporting an intact floor below the level of collapse was adequate to carry the load of 11 additional floors if the load was applied gradually and 6 additional floors if the load was applied suddenly (as was the case).  Since the number of floors above the approximate floor of collapse initiation exceeded six in each WTC Tower (12 and 29 floors, respectively), the floors below the level of collapse initiation were unable to resist the suddenly applied gravitational load from the upper floors of the buildings.  Details of this finding are provided below:

Consider a typical floor immediately below the level of collapse initiation and conservatively assume that the floor is still supported on all columns (i.e., the columns below the intact floor did not buckle or peel-off due to the failure of the columns above).  Consider further the truss seat connections between the primary floor trusses and the exterior wall columns or core columns.  The individual connection capacities ranged from 94,000 lb to 395,000 lb, with a total vertical load capacity for the connections on a typical floor of 29,000,000 lb (See Section 5.2.4 of NIST NCSTAR 1-6C).  The total floor area outside the core was approximately 31,000 ft2, and the average load on a floor under service conditions on September 11, 2001 was 80 lb/ft2.  Thus, the total vertical load on a floor outside the core can be estimated by multiplying the floor area (31,000 ft2) by the gravitational load (80 lb/ft2), which yields 2,500,000 lb (this is a conservative load estimate since it ignores the weight contribution of the heavier mechanical floors at the top of each WTC Tower).  By dividing the total vertical connection capacity (29,000,000 lb) of a floor by the total vertical load applied to the connections (2,500,000 lb), the number of floors that can be supported by an intact floor is calculated to be a total of 12 floors or 11 additional floors.

This simplified and conservative analysis indicates that the floor connections could have carried only a maximum of about 11 additional floors if the load from these floors were applied statically.  Even this number is (conservatively) high, since the load from above the collapsing floor is being applied suddenly.  Since the dynamic amplification factor for a suddenly applied load is 2, an intact floor below the level of collapse initiation could not have supported more than six floors.  Since the number of floors above the level where the collapse initiated, exceeded 6 for both towers (12 for WTC 1 and 29 for WTC 2), neither tower could have arrested the progression of collapse once collapse initiated.  In reality, the highest intact floor was about three (WTC 2) to six (WTC 1) floors below the level of collapse initiation.  Thus, more than the 12 to 29 floors reported above actually loaded the intact floor suddenly. 


Note, no reference to Bazant, no discussion of core & perimeter column strength, no discussion of upper block tilting, the column to column bolts, no discussion of "wedging", etc... According to NIST, you just need to consider 2 parameters to determine the inevitability of collapse past initiation: truss seat capacity & floor load! To anyone paying attention this is MIND-BOGGLING! Apparently, you can make a tube in tube structure with core & perimeter support columns of infinite capacity, and as long as you overload the truss seat connections of 1 floor the whole thing will come crashing down (progressively of course).

And just in case, you are thinking there has to be more to this, NIST cannot simply be waving this problem away like this! Check question 10.
10. Why didn’t NIST fully model the collapse initiation and propagation of WTC Towers?

The first objective of the NIST Investigation included determining why and how WTC 1 and WTC 2 collapsed following the initial impacts of the aircraft (NIST NCSTAR 1).  Determining the sequence of events leading up to collapse initiation was critical to fulfilling this objective.  Once the collapse had begun, the propagation of the collapse was readily explained without the same complexity of modeling, as shown in the response to question #1 above.
(italics added)

unbelievable.
adoucette
16th December 2007 - 04:27 AM
QUOTE (lozenge124+Dec 15 2007, 10:54 PM)
According to NIST, you just need to consider 2 parameters to determine the inevitability of collapse past initiation: truss seat capacity & floor load! To anyone paying attention this is MIND-BOGGLING! Apparently, you can make a tube in tube structure with core & perimeter support columns of infinite capacity, and as long as you overload the truss seat connections of 1 floor the whole thing will come crashing down (progressively of course).


What's strange about that?

Its BY FAR a simpler description than Bazant (which does deal with the columns) and simply shows how ONE MANNER of progressive collapse of a structure built like the WTC towers can occur.

If you are going to argue that progressive collapse is impossible (as many CT'ers do) then explain what is wrong with this version of how a progressive collapse can occur.

Arthur
David B. Benson
16th December 2007 - 06:35 PM
QUOTE (lozenge124+Dec 15 2007, 08:54 PM)
According to NIST, you just need to consider 2 parameters to determine the inevitability of collapse past initiation: truss seat capacity & floor load!

Yes. I did a version of this as did shagster. NIST's seems simpler and quite clearly explained.
David B. Benson
16th December 2007 - 06:45 PM
QUOTE (lozenge124+Dec 15 2007, 08:54 PM)
Apparently, you can make a tube in tube structure with core & perimeter support columns of infinite capacity, and as long as you overload the truss seat connections of 1 floor the whole thing will come crashing down (progressively of course).

No. Only the office floors, i.e., outside the core. Bageling would occur and leave the exterior walls and the core intact.

Of course, that is not what occurred.
Nor what NIST is describing in the answer to question #1.
Trippy
16th December 2007 - 06:59 PM
QUOTE (lozenge124+Dec 16 2007, 04:54 PM)
Note, no reference to Bazant, no discussion of core & perimeter column strength, no discussion of upper block tilting, the column to column bolts, no discussion of "wedging", etc... According to NIST, you just need to consider 2 parameters to determine the inevitability of collapse past initiation: truss seat capacity & floor load! To anyone paying attention this is MIND-BOGGLING! Apparently, you can make a tube in tube structure with core & perimeter support columns of infinite capacity, and as long as you overload the truss seat connections of 1 floor the whole thing will come crashing down (progressively of course).

Think about what you're saying for a moment.

If I have a floor that weighs 10 ton, has a 100 ton capacity, and a 10% saftey margin. If I then load that floor up to 111 ton, and let it collapse on to the floor below it.

The floor below it now has 121 ton of weight on it, and only a 100 ton capacity, and is 11 ton over it's 10% saftey margin..

What, precisely do you think is going to happen next? (Especially if the second floor is carrying any load of it's own).
Grumpy
16th December 2007 - 07:06 PM
David B. Benson

QUOTE
Of course, that is not what occurred.
Nor what NIST is describing in the answer to question #1.


Sorry Doc., but that seems to be EXACTLY what happened, as I have describe numerous times before. I'm not arguing with your work in the energy equations but it does match the appearances of most of the floor supports(striped vertically), the expulsions of dust below the level of the crush front, the funneling of the floor debris inside the still intact outer frames into the basements and the banana peeling of the outer frames. Bagelling did not initiate collapse, but WAS a result of that initiation, leaving both the core and the outer frame without lateral support.

Grumpy cool.gif
David B. Benson
16th December 2007 - 07:14 PM
QUOTE (Grumpy+Dec 16 2007, 12:06 PM)
... but that seems to be EXACTLY what happened, ...

No. Neither the exterior walls nor the core columns were infinitely strong, as was lozenge124's supposition. The walls pealed apart and most of the core was crushed, in both cases.

Which is why I said that destruction of only the office floors was not observed.

We are in fundamental agreement, I am sure.
Grumpy
16th December 2007 - 07:46 PM
David B. Benson

Sorry, my bad.

Grumpy huh.gif
David B. Benson
16th December 2007 - 09:06 PM
QUOTE (Grumpy+Dec 16 2007, 12:46 PM)
Grumpy huh.gif

De nada.
David B. Benson
17th December 2007 - 01:03 AM
More realistic resistance force function does slightly better on C447 data:

CODE
Sef-SS+ZSS-F-exp-pow-stretch dB= 0.0 sd= 0.079

BV-SS+ZSS-F-exp-pow-stretch dB= 0.4 sd= 0.094
Sef-ZSS-F-exp-pow-stretch dB= 0.5 sd= 0.098


SS+ZSS-F denotes a force function form of

k0SS + k1(Z-Z0)SS

where S is the speed. The first term, with k0 quite small, is the air movement force. The second term, wih k1 rather large, is similar to the flowing snow avalanche resistance, but the mass (Z-Z0) is only the mass of the crushed materials. This term is due, in part, to the continued re-crushing of concrete and other materials.
OneWhiteEye
17th December 2007 - 02:42 AM
David B. Benson:

I've overlooked something simple in perspective correction; can't say if you have, also. I suspect so since I don't think you have the optical axis angle. Please inspect this diagram and see if you get what I mean without explanation, as my fatigue prevents giving one right now.

User posted image
http://i13.tinypic.com/6odidth.png
OneWhiteEye
17th December 2007 - 03:04 AM
User posted image
http://i3.tinypic.com/8bqei2o.png
David B. Benson
17th December 2007 - 06:34 PM
QUOTE (OneWhiteEye+Dec 16 2007, 08:04 PM)
User posted image

Yes, thank you. I should correct for that.
adoucette
17th December 2007 - 06:39 PM
QUOTE (OneWhiteEye+Dec 16 2007, 10:04 PM)
User posted image
http://i3.tinypic.com/8bqei2o.png

What was your estimate of the elevation of the camera?

Arthur
David B. Benson
17th December 2007 - 07:53 PM
Arthur --- The best estimate (so far) is that the camera was at (World Trade Center) ground level. This means the up-angle is about 13--15 degrees of arc.
Chainsaw,
17th December 2007 - 09:45 PM
Dear Dr. Benson do you know anyone who knows if the live decking of the floor pans was insulated?
Or where information can be found on the wiring system?
David B. Benson
17th December 2007 - 10:17 PM
QUOTE (Chainsaw,+Dec 17 2007, 02:45 PM)
... if the live decking of the floor pans was insulated?

Or where information can be found on the wiring system?

There is nothing in the NIST report (that I have seen) that suggests that it was. Considering the design, corrugated floor pans filled with concrete, it seems most unlikely.

No one single place, I fear. But generally, it was a mess. The main reason was the personal computers and servers, not planned for originally. So more wiring had to be added. Underneath the concrete (I don't know whether above or below the floor pans) were sheet metal conduits. Holes were drilled in the concrete to provide a wiring path between the workstation computers and the conduits.
RealityCheck
17th December 2007 - 10:34 PM
QUOTE (David B. Benson+Dec 17 2007, 10:17 PM)
There is nothing in the NIST report (that I have seen) that suggests that it was.  Considering the design, corrugated floor pans filled with concrete, it seems most unlikely.

No one single place, I fear.  But generally, it was a mess.  The main reason was the personal computers and servers, not planned for originally.  So more wiring had to be added.  Underneath the concrete (I don't know whether above or below the floor pans) were sheet metal conduits.  Holes were drilled in the concrete to provide a wiring path between the workstation computers and the conduits.



Hi DBB, Chainsaw, all!

How MANY such 'holes' were drilled through those concrete floor 'membranes'?

It's starting to sound like those floors were like SIEVES even before any cracks/voids due to the plane impact damage.

And we all know by now what firemen think about flames issuing from cable risers! hehehe.

RC.
.
David B. Benson
17th December 2007 - 10:48 PM
QUOTE (RealityCheck+Dec 17 2007, 03:34 PM)
How MANY such 'holes' were drilled through those concrete floor 'membranes'?

And we all know by now what firemen think about flames issuing from cable risers!

I don't know. At most one per workstation, but more likely is one for each four workstations.

The cable risers in the 1975(?) fire in WTC 1 were completely open holes in the corners. After that fire, surely insulating covers were placed so that there was no free air path from one floor to the next.
RealityCheck
17th December 2007 - 11:19 PM
QUOTE (David B. Benson+Dec 17 2007, 10:48 PM)
I don't know.  At most one per workstation, but more likely is one for each four workstations.

The cable risers in the 1975(?) fire in WTC 1 were completely open holes in the corners.  After that fire, surely insulating covers were placed so that there was no free air path from one floor to the next.



Yeah, you'd think so. I wonder (assuming they did put them in) if they would have stayed in place over the years of office re-wiring etc. And of course, whether they would have been easily dislodged on many floors at/near the impact levels.

Thanks for your response, DBB!

Cheers all!

RC.
.
Chainsaw,
17th December 2007 - 11:22 PM
QUOTE (David B. Benson+Dec 17 2007, 10:48 PM)
I don't know. At most one per workstation, but more likely is one for each four workstations.

The cable risers in the 1975(?) fire in WTC 1 were completely open holes in the corners. After that fire, surely insulating covers were placed so that there was no free air path from one floor to the next.


Thank You Dr. Benson,
This is what I was looking into.

"The floor structure was then installed between the outer perimeter wall and the inner core. The floors also came in pre-assembled sections, consisting of 32-inch-deep (81-cm) trusses topped with a corrugated metal surface. To finish each floor, the crew would pour concrete over the metal surface and top it off with tile. The floor sections included pre-assembled ducts for phone lines and electrical cable, to make things easier for the electricians who would come in later. After the steel structure was in place, the crew attached the outer "skin" to the perimeter -- anodized aluminum, pre-cut into large panels. "
http://people.howstuffworks.com/wtc4.htm

When I modeled them in a fire I got a result like, this.

http://www.chiefmontagna.com/Articles/manhole%20fires.htm

The arcing, burning wire generates various toxic and combustible gases including high concentrations of carbon monoxide and neoprene gas.

The ignition can be explosive, sending the 300-pound manhole cover flying into the air. Manhole covers have been blown onto the roofs of six-story buildings and have gone up in the air only to come crashing down through the roofs of passing vehicles.

http://www.hhrobertson.com/4-code.cfm

For example, the Port Authority of New York and New Jersey prohibits the use of PVC in the World Trade Center. If installed, removal by the tenant is required.

I have done a few experiment on what would happen if these structures were non insulated and exposed to fire.

Arthur as usual thinks I am nuts.

David B. Benson
18th December 2007 - 12:28 AM
Ok, tomorrow I can put in the improved pixels-to-meters conversion, which will make a few percent difference over the current hack.

With the best resistive force function forms, the progressive collapse had a terminal speed of close to

67 m/s

assuming uniform density and no mass loss. I'm a bit surprised at this low rate. huh.gif
metamars
18th December 2007 - 01:45 AM

NYC Window Washer Survives a 47-Story Fall

Unfortunately, his brother died in the same fall.

http://www.foxnews.com/wires/2007Dec07/0,4...ccident,00.html
QUOTE
NYC Window Washer Dies in 47-Story Fall

NEW YORK —  A window washer fell 47 stories to his death and his brother was critically injured Friday when the scaffolding on a high-rise apartment building gave way, authorities said.

The brothers were getting onto the scaffolding from the roof of the 47-story building when the platform gave way, Fire Department spokesman Seth Andrews said.

"They apparently fell all the way from the top," Fire Department spokesman John Mulligan said.

Edgar Moreno, 30, of Linden N.J., was pronounced dead at the scene. His 37-year-old brother, whose name was not released, was in critical condition at a New York Weill Cornell Medical Center, officials said.




Anyone care to calculate his terminal velocity, and compare to the estimated terminal velocity in the WTC 1 & 2 collapses?

I wonder if the brother who died has his bones reduced to chips?

Well, not really.
gnik_isrever
18th December 2007 - 01:59 AM
It was a controlled demolition of all three buildings, and a missile which hit the pentagon.

It was under orders by George Bush and *** Cheney to prevent the National Economic Security And Reformation Act from being announced and implemented, and then to make us think the "ARABS" did it, so that we would support them in their choice to murder millions of innocent people.

If you think otherwise, I suggest you watch "loose change" videos, parts one, two, and the final cut. That video tells it all.

I do apologize if my reply is not really about the "physics" calculations you guys are doing, but I just thought it would be good for you guys to know the truth of the matter you are discussing.
OneWhiteEye
18th December 2007 - 07:51 AM
QUOTE (adoucette+Dec 17 2007, 06:39 PM)
What was your estimate of the elevation of the camera?

Arthur

Very interesting question. I don't know. There is more building footprint than ground in the region of possible locations for the camera. The buildings around there don't look too tall (the arrow is to the right of the building that also appears in the video):

User posted image
http://i5.tinypic.com/8avf9zl.png

User posted image
http://i4.tinypic.com/6q1m7f5.png

The camera could be in a window or on a roof. Maybe no higher than 80 feet?

I've done nothing with elevation as a parameter, yet. Got a pretty decent rendering with zero elevation a while back, but the aspect ratio was probably screwed up, squashing the vertical, so who knows? Eventually I want to revisit that using the floor heights you provided for vertical calibration. Once everything looks as close as possible from the ground, I'll play around with height, too. Brute force, but it will probably provide a fairly tight 3D bounding box for the camera location.

For now, camera x,y,z location, view angle, optical axis vector and even camera vertical orientation are all yet to be established with reasonable precision. Even if the location were documented, there would still be some uncertainty.

About the only thing I can say with some certainty is the optical axis intersects the NE corner at a point 123 feet below the roofline, give or take a few feet.
einsteen
18th December 2007 - 12:06 PM
QUOTE (adoucette+Dec 16 2007, 04:27 AM)
What's strange about that?

Its BY FAR a simpler description than Bazant (which does deal with the columns) and simply shows how ONE MANNER of progressive collapse of a structure built like the WTC towers can occur.

Arthur

I disagree, if you ignore it then it is a matter of pure momentum transfer. On the one hand they reject pancaking for initiation, but what does this story mean ? They talk about the load of 6 floors and add a factor 2 for dynamic load. They want to explain that the static load of the top section is already sufficient for a non-arrestable collapse, but that static load was already there for 30 years (and I understand this sounds like BS because there was no damage but that's not the reason I mention it). If one floor fails then it is clear that it will be stopped even with their factor 2, that's why they rejected pancaking for initiation. They say that the whole top section falls down, which we also observe from video, but let's look at it from a static mass's point of view. How do you want to place that top section on the next floor in order to let that one fail ? The perimeter columns and core columns will touch each other. Real funneling is only possible if the initiation is also pancaking. Their answers are blatant nonsense in my humble opinion.
einsteen
18th December 2007 - 12:26 PM
QUOTE (David B. Benson+Dec 17 2007, 06:34 PM)
Yes, thank you. I should correct for that.

I thought you already had that script for perspective correction, or am I confusing parallax ?
adoucette
18th December 2007 - 04:48 PM
QUOTE (einsteen+Dec 18 2007, 07:06 AM)
How do you want to place that top section on the next floor in order to let that one fail ? The perimeter columns and core columns will touch each other. Real funneling is only possible if the initiation is also pancaking. Their answers are blatant nonsense in my humble opinion.

No, its quite clear that one whole side of the building bowed in. The building TILTS and the upper side where the bowing occurs hits the floors whose supports fail.

Arthur



David B. Benson
18th December 2007 - 07:00 PM
QUOTE (gnik_isrever+Dec 17 2007, 06:59 PM)
... I just thought it would be good for you guys to know the truth of the matter you are discussing.

I'll bet you think the earth is flat, too. dry.gif
David B. Benson
18th December 2007 - 07:10 PM
QUOTE (OneWhiteEye+Dec 18 2007, 12:51 AM)
There is more building footprint than ground in the region of possible locations for the camera.

I'm sure you have already thought of this, but moving about a quarter block to the west, towards Varick St., puts the location between two parking lots.
newton
18th December 2007 - 07:19 PM
QUOTE (adoucette+Dec 18 2007, 04:48 PM)
No, its quite clear that one whole side of the building bowed in. The building TILTS and the upper side where the bowing occurs hits the floors whose supports fail.

Arthur

that is clear in one case.
not the other.

and, anyway, if the core was the first thing to drop(which it was, because that's how it was controlled), then it would bow in the weakened perimeter through the floor trusses(which it did).
collapse would have arrested there, if it weren't for the rapid succession of timed explosions that followed.

too many people know.

the truth will out.

David B. Benson
18th December 2007 - 07:22 PM
QUOTE (einsteen+Dec 18 2007, 05:26 AM)
I thought you already had that script for perspective correction, ...

Yes, but part of the proper correction is missing. Yesterday I took the time to learn about tangents, chords and secants in geometry and apply these to finishing up the complete correction.
David B. Benson
18th December 2007 - 07:34 PM
QUOTE (newton+Dec 18 2007, 12:19 PM)
that is clear in one case.
not the other.

Perfectly clear in both cases. For WTC 1, the NIST report states that the south, east and west walls bowed in. OneWhiteEye noticed the north wall buckling. Thus the entire load was transferred by the hat truss to the core, quite suddenly. This overloaded the core far beyond its (reduced) capacity and global collapse ensued.
einsteen
18th December 2007 - 08:07 PM
QUOTE (David B. Benson+Dec 18 2007, 07:34 PM)
Perfectly clear in both cases. For WTC 1, the NIST report states that the south, east and west walls bowed in. OneWhiteEye noticed the north wall buckling. Thus the entire load was transferred by the hat truss to the core, quite suddenly. This overloaded the core far beyond its (reduced) capacity and global collapse ensued.

But then they should talk about that and not only the connections of the floors. They should take the energy to destroy the core into account. But IMHO that model also leads to huge problems.

David B. Benson
18th December 2007 - 08:11 PM
QUOTE (einsteen+Dec 18 2007, 01:07 PM)
But then they should talk about that ...

But IMHO that model also leads to huge problems.

The NIST report does.

huh.gif Columns loaded far beyond rated capacity buckle.
einsteen
18th December 2007 - 09:04 PM
But do they also give quantities for the core and perimeter columns, i.e. with an explicit energy value ? And I mean not the connections.
Gehn
18th December 2007 - 09:09 PM
QUOTE (gnik_isrever+Dec 18 2007, 01:59 AM)
It was a controlled demolition of all three buildings, and a missile which hit the pentagon.

It was under orders by George Bush and *** Cheney to prevent the National Economic Security And Reformation Act from being announced and implemented, and then to make us think the "ARABS" did it, so that we would support them in their choice to murder millions of innocent people.


You obviously don't know f*** about 9/11, and are some stupid conspiracy theorist. I can't stand people like you. 9/11 was a terrible incident, and it disgusts me that you think it was an inside job. You have no actual proof, and you are obviously delusional. The current American government may be corrupt, be it would never make a decision which would cause the deaths of so many civilian American citizens. It has never done so in recent times, and never will.

How could a missile have imitated an aircraft? And what would be the point of flying two aircraft into the Twin Towers if it could be blamed on a bomb? There are accounts from everywhere of the incident, and they all contradict your story. You were probably watching "Why the Earth is Flat" when it happened. The NESARA was suppressed, but destroying the WTC would have been a very stupid decision to help suppress it. In conclusion, you are being paranoid. You seem to believe everything you see on the Internet. If you continue to act like this, no one will ever respect you or your ideas.


- Gehn biggrin.gif
David B. Benson
18th December 2007 - 09:31 PM
QUOTE (einsteen+Dec 18 2007, 02:04 PM)
But do they also give quantities for the core and perimeter columns, i.e. with an explicit energy value ? And I mean not the connections.

The NIST report gives the summarized loads on each of the four walls and on the core for specific floors. In addition, the report states that the aggregate DCR of the core was about 0.5 and in a briefing mentioned that the aggregate DCR for the exterior walls was about 0.2.

That is enough data to determine that the DCR for the surviving core columns when forces to also bear the loads of the walls above the failure level was about 2--3, vastly more than the critical load for failure by buckling.
RealityCheck
18th December 2007 - 10:08 PM
QUOTE (einsteen+Dec 18 2007, 09:04 PM)
But do they also give quantities for the core and perimeter columns, i.e. with an explicit energy value ? And I mean not the connections.



Hi einsteen!

Ignore the connections at your own risk, hehehe.

The core was nothing BUT 'connections'.

Without all the steel 'wickerwork' which was fixed on SHEER VERTICAL COLUMNS, there would have been nothing to stop everything sliding right to the ground even AS BUILT.

And hwen the planes hit the place like a big bomb, and fatally (eventually) compromised the 'integrated (read 'inter-CONNECTED) design/structure, then the inevitable happened as the UNcontrolled fires and heating/cooling weakenings/distortions and resultant failings/load-shifting sent the whole INTEGRATED-CONNECTION-DEPENDENT structure into a mechanical/thermodynamical 'death spiral' of events that led to collapse.

In any serious analysis, ignoring the WEAKEST LINK factors of INTEGRATED-SUPPORT and INTEGRAL CONNECTIONS that FAILED under the circumstances would be just a wee bit too 'politically/vestedly/ingenuously' EXPEDIENT, I think.

Cheers all!

RC.
.
OneWhiteEye
19th December 2007 - 09:21 AM
QUOTE (David B. Benson+Dec 18 2007, 07:10 PM)
I'm sure you have already thought of this, but moving about a quarter block to the west, towards Varick St., puts the location between two parking lots.

Yes, tried that general location. I started from Varick St. and moved incrementally east to roughly the location I've marked.

There isn't much play in the lateral placement (azimuthal angle wrt tower) because 101 Ave of the Americas (the bldg next to green arrow) has to occupy a certain portion of the left side of the frame. Could move back some distance and reduce the view angle accordingly, there is some clear area further back from that vertex, as well.

I need to fix the aspect ratio problem first, which is not a big deal, just haven't gotten around to it. I think I made the view angle for vertical the same as horizontal, but the image is 3:2, which would explain why the floor heights appeared to be way too short, despite reasonably careful dimensioning. And a good match on the horizontal.

What will help is to get a scaled facade on my mock-up of 101 Ave of A; this will further clarify the minimum view angle by checking window sizes. Getting the location from rendered scene geometry is backwards, as I've said, but the rules of physical optics are obeyed so it's not as ridiculous as it might at first seem.

I posted this before a while back. It's a blend of a rendered frame and one from the real video:

User posted image
http://i5.tinypic.com/8ayf2p2.png

You can see it's not perfectly aligned on either building, but much closer than my first pass, trust me. Even the moire patterns almost match. The next pass will probably be much better. Once the best match in all dimensional respects is obtained, then I'll vary the camera position and view angle about that ideal until a sensible uncertainty range is defined.
lozenge124
19th December 2007 - 03:36 PM
Interesting 300 foot tower demolition:

The tower drops a few floors, but the "crush-up" is arrested and it topples over.

http://www.youtube.com/watch?v=xgE9S3bV6QM&NR=1
einsteen
19th December 2007 - 05:27 PM
QUOTE (lozenge124+Dec 19 2007, 03:36 PM)
Interesting 300 foot tower demolition:

The tower drops a few floors, but the "crush-up" is arrested and it topples over.

http://www.youtube.com/watch?v=xgE9S3bV6QM&NR=1

But Lozenge...think about the scale. It will always be arrested for a scale 1 inch below the wtc scale, but for the wtc scale, it will ba a 100% total complete collapse, probably with a symmetrical debris zone ejected in all directions, no toppling at all, even no symmetrical initiation zone is needed. The wtc was a one off construction, Think about it ;-)!
einsteen
19th December 2007 - 09:24 PM
Anyone in the mood for a wtc7 smearogram 29.97 frames/second

User posted image
http://i6.tinypic.com/7w9ius5.jpg

If the collapse of the left part of the penthouses means that the structure is already broken, then the value that Shagster and I arrived at independently is an indication of the E1/M ratio of a broken building. I'm wondering how this fits with the initial movement of the top of the building. The difference between g and a_wtc7 could say something about that value also.
OneWhiteEye
20th December 2007 - 06:26 AM
QUOTE (einsteen+Dec 19 2007, 05:27 PM)
But Lozenge...think about the scale. It will always be arrested for a scale 1 inch below the wtc scale, but for the wtc scale, it will ba a 100% total complete collapse, probably with a symmetrical debris zone ejected in all directions, no toppling at all, even no symmetrical initiation zone is needed. The wtc was a one off construction, Think about it ;-)!

Good one, einsteen. Cracked me up, you did.

lozenge124:

Thanks for the video, I always appreciate interesting collapse videos. Nowadays. For some reason.
OneWhiteEye
20th December 2007 - 06:35 AM
QUOTE (einsteen+Dec 19 2007, 09:24 PM)
Anyone in the mood for a wtc7 smearogram 29.97 frames/second

OK.

User posted image
http://i2.tinypic.com/8100lkg.png


Not my best work. Very quick, probably not too accurate. NOT TO BE USED TO AID IN DESIGN OF COLLAPSE-RESISTANT BUILDINGS!!!

9 floors = 70 pixels (this is the weakest link)

47 stories, or so they say.

I don't know the floor height of WTC7.

Emporis:
570 feet = 174m
uniform => 12 ft/story max

NIST:
610ft = 186m
uniform => 13 ft/story max

Looks like Emporis might have used (floor height x #floors) to arrive at the total. Assume a range on the uniform floor height of 12 - 13 ft, or 3.66 - 3.96 m, with a possible skew to the low side. I think we can take 3.96m to be the ceiling on the floor height, if you'll pardon the pun.

Columns are
- Tsec: Time in seconds
- Yfl: Position in floor units
- Ymin: Minimum position in meters
- Ymax: Maximum position in meters

CODE

0.0,0.0,0.0,0.0
0.107734300967634,0.0,0.0,0.0
0.219183550216884,0.0,0.0,0.0
0.334347747747748,0.0286232142857143,0.104760964285714,0.113347928571428
0.449512012012012,0.100191342857143,0.366700314857143,0.396757717714286
0.562818985652319,0.221864914285714,0.812025586285714,0.878585060571428
0.676125291958625,0.3721716,1.362148056,1.473799536
0.778286786786787,0.558262285714286,2.04323996571429,2.21071865142857
0.880448948948949,0.758694728571428,2.77682270657143,3.00443112514286
0.971465298631965,0.966247585714286,3.53646616371429,3.82634043942857
1.06433983983984,1.18097087142857,4.32235338942857,4.67664465085714
1.15535685685686,1.40282601428571,5.13434321228571,5.55519101657143
1.24265849182516,1.65335387142857,6.05127516942857,6.54728133085714
1.33367550884218,1.8752283,6.863335578,7.425904068
1.4154049049049,2.11858572857143,7.75402376657143,8.38959948514286
1.49899215882549,2.36908401428571,8.67084749228571,9.38157269657143
1.58257874541208,2.6410743,9.666331938,10.458654228
1.66430814147481,2.9130543,10.661778738,11.535695028
1.74232248915582,3.19934558571429,11.7096048437143,12.6694085194286
1.81847931264598,3.48563815714286,12.7574356551429,13.8031271022857
1.89277861194528,3.77909087142857,13.8314725894286,14.9651998508571
1.96522038705372,4.06538215714286,14.8792986951429,16.0989133422857
2.03208992325659,4.36598601428571,15.9795088122857,17.2893046165714
2.10453169836503,4.66659887142857,17.0797518694286,18.4797315308571
2.17511561561562,4.96721301428571,18.1799996322857,19.6701635365714
2.24384234234234,5.28213715714286,19.3326219951429,20.9172631422857
2.31071154487821,5.59706258571428,20.4852490637143,22.1643678394286
2.37758108108108,5.94062087142857,21.7426723894286,23.5248586508571
2.44073556890224,6.26985758571429,22.9476787637143,24.8286360394286
2.50574691358025,6.59194315714286,24.1265119551429,26.1040949022857
2.56332916249583,6.9283413,25.357729158,27.436231548
2.62462645979313,7.26473815714286,26.5889416551429,28.7683631022857
2.68406556556557,7.61545658571428,27.8725711037143,30.1572080794286
2.73607457457457,7.9804773,29.208546918,31.602690108
2.78994177510844,8.32403558571428,30.4659702437143,32.9631809194286
2.84195145145145,8.67474501428571,31.7495667522857,34.3519902565714
2.89767550884218,9.03261458571428,33.0593693837143,35.7691537594286
2.94968485151819,9.39048415714286,34.3691720151429,37.1863172622857
3.00169419419419,9.75551515714286,35.7051854751429,38.6318400222857
3.04998848848849,10.1205461571429,37.0411989351429,40.0773627822857
3.10014064064064,10.4784247285714,38.3510345065714,41.4945619251429
3.15029279279279,10.8362955857143,39.6608418437143,42.9117305194286
3.19858708708709,11.2013253,40.996850598,44.357248188
3.24502419085752,11.5591948714286,42.3066532294286,45.7744116908571
3.2896034367701,11.9385371571429,43.6950459951429,47.2766071422857
3.33604020687354,12.3035655857143,45.0310500437143,48.7221197194286
3.38062012012012,12.6757413,46.393213158,50.195935548
3.42705655655656,13.0479298714286,47.7554233294286,51.6698022908571
3.47535151818485,13.4129570142857,49.0914226722857,53.1153097765714
3.52178795462129,13.7779841571429,50.4274220151429,54.5608172622857
3.57008291624958,14.1358498714286,51.7372105294286,55.9779654908571
3.61837787787788,14.5008770142857,53.0732098722857,57.4234729765714
3.66667283950617,14.8587427285714,54.3829983865714,58.8406212051428
3.71125075075075,15.2237698714286,55.7189977294286,60.2861286908571
3.7614042375709,15.5887970142857,57.0549970722857,61.7316361765714
3.81155772439106,15.9466627285714,58.3647855865714,63.1487844051428
3.85799416082749,16.3116898714286,59.7007849294286,64.5942918908571
3.90257540874208,16.6695555857143,61.0105734437143,66.0114401194286
3.94901184517851,17.0345827285714,62.3465727865714,67.4569476051428
3.99544828161495,17.3924613,63.656408358,68.874146748
4.04002619285953,17.7574884428571,64.9924077008571,70.3196542337143
4.08089039039039,18.1225155857143,66.3284070437143,71.7651617194286
4.12918201534868,18.4875427285714,67.6644063865714,73.2106692051429
4.16818768768769,18.8668798714286,69.0527803294286,74.7128442908571
4.20719336002669,19.2462170142857,70.4411542722857,76.2150193765714
4.25548832165499,19.6112313,71.777106558,77.660475948
4.29820770770771,19.9690970142857,73.0868950722857,79.0776241765714
4.33536152819486,20.3484470142857,74.4753160722857,80.5798501765714
4.37993943943944,20.7206355857143,75.8375262437143,82.0537169194286
4.42823440106773,21.0928113,77.199689358,83.527532748
4.47095712379046,21.4578384428571,78.5356887008571,84.9730402337143
4.51367984651318,21.8371755857143,79.9240626437143,86.4752153194286
4.56011628294962,22.1878927285714,81.2076873865714,87.8640552051428
4.60469419419419,22.5457584428571,82.5174759008571,89.2812034337143
4.64555839172506,22.9107855857143,83.8534752437143,90.7267109194286
4.69013630296964,23.2829741571429,85.2156854151429,92.2005776622857
4.73285902569236,23.6480013,86.551684758,93.646085148
4.7792954621288,24.0130284428571,87.8876841008571,95.0915926337143
4.82387337337337,24.3780427285714,89.2236363865714,96.5370492051429
4.87402352352352,24.7287470142857,90.5072140722857,97.9258381765714
4.91674290957624,25.1009227285714,91.8693771865714,99.3996540051429
4.95740690690691,25.4627355857143,93.1936122437143,100.832432919429
5.00631247914581,25.8055198714286,94.4482027294286,102.189858690857
5.04595211878545,26.1689398714286,95.7783199294286,103.629001890857
5.08722672672673,26.5418741571429,97.1432594151429,105.105821662286
5.12583533533534,26.9012698714286,98.4586477294286,106.529028690857
5.17345628962296,27.2460084428571,99.7203909008571,107.894193433714
5.22107724391058,27.5907598714286,100.982181129429,109.259409090857
5.27302585919253,27.9243898714286,102.203266929429,110.580583890857
5.33074691358025,28.2413313,103.363272558,111.835671948


Please note I did not propagate the errors associated with the floor height measurement and the data point measurements, which each have to be at least +/- 1 pixel for point placement. So, minimum +/- 2 pixels or more for the scale and +/- 1 for the points, probably much more.

Edit: And, sorry for all the digits, I couldn't be bothered to format it
OneWhiteEye
20th December 2007 - 07:31 AM
After removal of the first two records, polynomial fits for the first 4.5 seconds of minimum and maximum displacement, WTC7:

User posted image
YMin 2nd order term only

User posted image
YMin 3rd order

User posted image
YMax 2nd order term only

User posted image
YMax 3rd order

The second order alone gives a good fit. Here, the coefficients are:

YMin : 3.98
YMax : 4.31

So, 7.96 - 8.62 m/s^2, or 81 - 88% of g.

Of course, it could all be wrong.
OneWhiteEye
20th December 2007 - 08:50 AM
At least part of it was wrong. I removed the first two records but did not offset the time to zero.

The 2nd order fit for the first 4.2 seconds gives coefficients of:

YMin : 4.52
YMax : 4.89 (~g!)

The fits are not good, however. The 3rd order fits are good, but who cares?
einsteen
20th December 2007 - 09:47 AM
thanks (I hope it didn't take too much of your time) I will save this page when I'm home. The distance should be known to correct the parallax (that will be difficult), the original video was taken from "the third tower".
metamars
20th December 2007 - 02:31 PM
I participated in an interesting thread at JREF,

http://forums.randi.org/showthread.php?t=100488

starting at p. 16

The most important post I made was on p. 17, where I quote Graff:


QUOTE

In Wave Motion in Elastic Solids by Karl F. Graff, p. 103, the ratio of initial impact force is given (eq. 2.4.25) for elastic-on-elastic vs. rigid-on-elastic collisions. The elastic body being impacted is a thin rod. This is:

1 / (1 + Z1/Z2).


Where Z1 is the impedance of the elastic impacting object, and Z2 is the impedance of the impacted elastic object.

Z is given by rho*A*c0, where
rho is density
A is area
c0 is the speed of sound

In the special case of the collision of rods of exactly the same dimensions and density, the impedances are equal, and thus the peak force will be 1/2 times that of the analogous rigid-on-elastic impact.


You can see the effect of assuming a rigid impacting upper block on the lower block on Bazant and Zhou's Eqn #1. Taking the limit as h -> 0, you get twice the static load.

Bazant and Zhou also make no allowance for elastic energy leaving the impacted column section under consideration, during any reasonable time period it would take for a buckling to occur. If the supporting object under the column section were an ideal, rigid body, incapable of being accelerated, then it would not allow transfer of energy from the column to itself.

The reality is very different.
scimethod
20th December 2007 - 06:07 PM
QUOTE (wcelliott+Aug 31 2007, 03:26 PM)

OK, just some quick answers.

The molten orange flow was probably lead from Uninterruptable Power Supplies on floor 81, which use banks of lead-acid batteries.  It *couldn't* have been thermite/thermate, because it has no effect on the metal sheathing of the building, also clearly visible on the same videos showing the flowing lead.  If it had been thermite/thermate, it would've melted straight through the material, and it didn't.

No possibility of it being thermate.  None.  Zero.  Give it up.

Regarding signatures of accelerants, there was only one accelerant, it was called "jet fuel", and there was lots of that present, both planes were loaded with it.

Welcome back to reality.


> The molten orange flow was probably lead from Uninterruptable Power
> Supplies on floor 81, which use banks of lead-acid batteries.

That is a hypothesis that would have some weight with me if it had experiments behind it. You do not even know that these were actually batteries, as they were never used.


> It *couldn't* have been thermite/thermate, because it has no
> effect on the metal sheathing of the building,

An iron-rich alloy is the result of a thermate reaction. Is it possible that thermate could have melted the steel, producing an iron-rich allow that flowed out of the building? Surely you understand that water can flow over ice without completely melting it. You say you saw "no effect" because you analyzed the metal sheathing? Perhaps some melting did occur. Your opinion does not have evidence to back it up.


> also clearly visible
> on the same videos showing the flowing lead. If it had been thermite/thermate,
> it would've melted straight through the material, and it didn't.

As stated above, this was likely an iron-rich alloy in liquid form.

You appear to be quite unaware of the actual physical evidence that has been recovered. Iron-rich spheres were found by the USGS ( pubs.usgs.gov/of/2005/1165/graphics/IRON-04-IMAGE.jpg ) RJ Lee Group ( www.nyenvirolaw.org/WTC/130%20Liberty%20Street/Mike%20Davis%20LMDC%20130%20Liberty%20Documents/Signature%20of%20WTC%20dust/WTC%20Dust%20Signature.Composition%20and%20Morphology.Final.pdf
) and others.


Of particular interest is the FEMA BPAT report is www.fema.gov/pdf/library/fema403_apc.pdf which shows a picture of melted steel and discusses evidence of controlled demolition like sulfidation. (Sulfidation is a byproduct of a thermate reaction.)


> No possibility of it being thermate. None. Zero. Give it up.
>
> Regarding signatures of accelerants, there was only one accelerant,
> it was called "jet fuel", and there was lots of that present, both
> planes were loaded with it.
>
> Welcome back to reality.

Your inflammatory debate style might make you feel good, but it does nothing to account for the actual evidence.

Perhaps you should also check the melting temperature of steel (2750-2800F) and the open-air burning temperature of jet fuel (500-599F) before you start declaring you know reality.


> An adult is responsible for the foreseeable consequences of his
> deliberate actions. Deliberate inactions are deliberate actions.


A scientist is responsible for using the scientific method. The scientific method avoids the confusing lunge and parry style of the debate. Credible hypotheses are those with supporting evidence and which suggest experiments that provide ways to verify the hypotheses.
David B. Benson
20th December 2007 - 07:56 PM
QUOTE (metamars+Dec 20 2007, 07:31 AM)
The reality is very different.

Yes.

There were no impacts at collapse initiation. Instead, there was a sudden load of long duration imposed on the core via the hat truss holding up the rest of the structure above the buckling points of the wall. In the case of WTC 1, this was a sudden load of 2--3 times the ability of the core to hold it up.
David B. Benson
20th December 2007 - 08:07 PM
QUOTE (scimethod+Dec 20 2007, 11:07 AM)
The scientific method avoids the confusing lunge and parry style of the debate.

Credible hypotheses are those with supporting evidence and which suggest experiments that provide ways to verify the hypotheses.

As a practicing scientist, I doubt your claim. wink.gif

What experiments would you suggest? The credible hypothesis, with supporting evidence, is that in the NIST report.
David B. Benson
20th December 2007 - 08:50 PM
QUOTE (OneWhiteEye+Dec 20 2007, 01:50 AM)
(~g!)

Yes, I expected something in the range 0.9--1.0 g.
shagster
21st December 2007 - 01:07 AM
Here's a smeargram I made using the CBS footage of WTC7. The slice doesn't include the screenwall or penthouses. The top of the smear is the roofline of the west side of the building. The smear doesn't include the screenwall or penthouses.

User posted image

http://i134.photobucket.com/albums/q91/sha...tc7/wtc7cbs.jpg

It looks like the smear Einsteen posted includes the screenwall.

shagster
21st December 2007 - 01:11 AM

This pic shows the location of the slice for the smeargram.

user posted image

http://i134.photobucket.com/albums/q91/sha...wtc7/wtc7sm.jpg

shagster
21st December 2007 - 01:23 AM
OneWhiteEye,

Notice in your plots of the WTC7 smear that the acceleration is close to freefall early in the collapse and then a tail forms near the end of collapse where the velocity apparently decreases. That's expected in a crush-up collapse because there is less falling mass near the end of collapse. That means less GPE is released over a drop through one story height toward the end to overcome the E1 of the remaining upper stories. That slows the collapse near the end. There may be camera angle effects that also contribute to the tail in the curve.

The observed shape of measured collapse curve with the slowing tail near the end suggests that the final global collapse of WTC7 was indeed a crush-up with a local collapse front near the ground level. That's different than a situation where all the stories of the building collapse starting at the same time with all stories falling near the free-fall rate throughout the entire collapse with no spatially local collapse front.

The tail where the velocity apparently decreases appears to be more prominent on the smear that I did. I selected the slice such that the roofline could be seen close to the ground between two buildings in the foreground. Einsteen's smear may not go down that far and may not show as much of the tail where the velocity apparently decreases.
shagster
21st December 2007 - 01:29 AM
I haven't yet downloaded the video of the Landmark that Einsteen posted on Megaupload. If the view in that video is relatively straight-on, it would be a good video for a smear. I would expect a relatively slow crush-up collapse like the Landmark to show a signficant tail where the collapse slows near the end.


OneWhiteEye
21st December 2007 - 05:17 AM
QUOTE (einsteen+Dec 20 2007, 09:47 AM)
thanks (I hope it didn't take too much of your time)

No, not at all. Took about 15 minutes to manually place the points and no time to get the numbers. That's why I started wondering what was up when I hit the 20 minute mark with shagster's smear (see next post).

QUOTE
The distance should be known to correct the parallax (that will be difficult), the original video was taken from "the third tower".

Yes, have to look in to that.
OneWhiteEye
21st December 2007 - 05:48 AM
QUOTE (shagster+Dec 21 2007, 01:23 AM)
Notice in your plots of the WTC7 smear that the acceleration is close to freefall early in the collapse and then a tail forms near the end of collapse where the velocity apparently decreases.  That's expected in a crush-up collapse because there is less falling mass near the end of collapse.  That means less GPE is released over a drop through one story height toward the end to overcome the E1 of the remaining upper stories.  That slows the collapse near the end.  There may be camera angle effects that also contribute to the tail in the curve.

Yes, I saw the tail. I think it's real velocity reduction and not artifact. It seems quite abrupt to be perspective, but I've been surprised before.

QUOTE
The tail where the velocity apparently decreases appears to be more prominent on the smear that I did.  I selected the slice such that the roofline could be seen close to the ground between two buildings in the foreground.  Einsteen's smear may not go down that far and may not show as much of the tail where the velocity apparently decreases.

Is this a full speed video? I shouldn't have to ask, I should be able to tell you that at this point, but I'd rather be careful than wrong. The collapse portion is about twice as wide (in smearspace, for those who don't know, one dimension of the image is time instead of space - in this case, horizontal) as einsteen's smear. Either the video is reduced speed or this captures about 10 seconds of collapse.

I placed the points but I'll wait for some word on the time scale before running too far with it.

User posted image
http://i4.tinypic.com/81ro3mg.png

Little green halos in SVG make for quick work:

User posted image
http://i13.tinypic.com/8g5rnyd.png

I bet this is half-speed video.
OneWhiteEye
21st December 2007 - 05:59 AM
I thought it was cool that the curve in shagster's was exactly square, 311x311 pixels:

User posted image

einsteen's curve, though, is 156x221:

user posted image
http://i7.tinypic.com/86480lf.png

Those rascals at CBS! Kenneth, what is the frequency?

About half the real frequency.
OneWhiteEye
21st December 2007 - 06:48 AM
Ran it anyway at 60fps. 7 floors = 66 pixels. Fat trimmed.

CODE
Tsec,Yfl,Ymin,Ymax
0.0000,0.0000,0.0000,0.0000
0.0525,0.0295,0.1079,0.1167
0.1019,0.0786,0.2876,0.3112
0.1544,0.1080,0.3954,0.4278
0.2053,0.1768,0.6471,0.7001
0.2562,0.2456,0.8988,0.9724
0.3056,0.3143,1.1504,1.2447
0.3550,0.3733,1.3661,1.4781
0.4044,0.4617,1.6897,1.8282
0.4523,0.5501,2.0132,2.1783
0.5001,0.6287,2.3009,2.4895
0.5495,0.7662,2.8042,3.0340
0.5974,0.8939,3.2716,3.5397
0.6483,0.9823,3.5951,3.8898
0.6915,1.1493,4.2063,4.5511
0.7394,1.2671,4.6377,5.0179
0.7872,1.4243,5.2130,5.6403
0.8305,1.5717,5.7522,6.2237
0.8814,1.7190,6.2915,6.8072
0.9292,1.8663,6.8308,7.3907
0.9740,2.0333,7.4420,8.0520
1.0188,2.1905,8.0172,8.6744
1.0651,2.3575,8.6284,9.3356
1.1068,2.5441,9.3115,10.0747
1.1500,2.7209,9.9586,10.7749
1.1932,2.9076,10.6417,11.5140
1.2333,3.1040,11.3608,12.2920
1.2750,3.2907,12.0438,13.0311
1.3151,3.4871,12.7629,13.8090
1.3584,3.6639,13.4100,14.5092
1.3969,3.8604,14.1290,15.2872
1.4340,4.0667,14.8840,16.1040
1.4757,4.2631,15.6031,16.8820
1.5158,4.4792,16.3940,17.7378
1.5528,4.7052,17.2209,18.6324
1.5914,4.9114,17.9759,19.4493
1.6300,5.1276,18.7668,20.3051
1.6686,5.3338,19.5218,21.1220
1.7041,5.5598,20.3487,22.0167
1.7396,5.7955,21.2116,22.9502
1.7767,6.0116,22.0025,23.8060
1.8122,6.2277,22.7934,24.6617
1.8492,6.4536,23.6203,25.5564
1.8832,6.6894,24.4831,26.4899
1.9156,6.9349,25.3819,27.4624
1.9495,7.1707,26.2447,28.3959
1.9835,7.4064,27.1076,29.3295
2.0175,7.6324,27.9344,30.2242
2.0514,7.8779,28.8333,31.1966
2.0854,8.1235,29.7321,32.1691
2.1224,8.3691,30.6308,33.1416
2.1579,8.6048,31.4937,34.0751
2.1903,8.8406,32.3565,35.0087
2.2243,9.0861,33.2553,35.9811
2.2567,9.3317,34.1541,36.9536
2.2876,9.5969,35.1247,38.0038
2.3200,9.8425,36.0235,38.9763
2.3509,10.1077,36.9942,40.0265
2.3817,10.3631,37.9290,41.0379
2.4142,10.6087,38.8277,42.0103
2.4497,10.8444,39.6906,42.9439
2.4821,11.0900,40.5894,43.9164
2.5145,11.3257,41.4522,44.8499
2.5454,11.5713,42.3510,45.8224
2.5793,11.8169,43.2497,46.7948
2.6102,12.0723,44.1845,47.8062
2.6411,12.3178,45.0832,48.7786
2.6735,12.5732,46.0180,49.7900
2.7043,12.8286,46.9527,50.8013
2.7337,13.0742,47.8514,51.7737
2.7645,13.3197,48.7502,52.7461
2.7862,13.6144,49.8287,53.9130
2.8186,13.8698,50.7635,54.9244
2.8541,14.1154,51.6623,55.8969
2.8819,14.3806,52.6330,56.9471
2.9127,14.6360,53.5677,57.9585
2.9421,14.8914,54.5024,58.9698
2.9714,15.1468,55.4372,59.9812
3.0023,15.4120,56.4079,61.0315
3.0300,15.6772,57.3786,62.0817
3.0594,15.9424,58.3493,63.1320
3.0918,16.1880,59.2480,64.1044
3.1227,16.4434,60.1828,65.1158
3.1489,16.7086,61.1534,66.1660
3.1767,16.9738,62.1241,67.2163
3.2060,17.2390,63.0948,68.2665
3.2338,17.5141,64.1014,69.3557
3.2600,17.7891,65.1081,70.4448
3.2894,18.0445,66.0428,71.4561
3.3218,18.2900,66.9415,72.4285
3.3326,18.5945,68.0560,73.6344
3.3542,18.8991,69.1705,74.8403
3.3820,19.1643,70.1412,75.8905
3.4082,19.4393,71.1479,76.9797
3.4375,19.7045,72.1185,78.0299
3.4669,19.9795,73.1251,79.1190
3.4977,20.2251,74.0239,80.0915
3.5240,20.5002,75.0306,81.1806
3.5533,20.7654,76.0013,82.2309
3.5780,21.0404,77.0079,83.3200
3.6073,21.2958,77.9426,84.3313
3.6367,21.5610,78.9133,85.3816
3.6644,21.8360,79.9199,86.4707
3.6938,22.0914,80.8546,87.4821
3.7262,22.3370,81.7534,88.4545
3.7586,22.5826,82.6522,89.4269
3.7848,22.8576,83.6588,90.5161
3.8157,23.1326,84.6654,91.6052
3.8435,23.4077,85.6721,92.6944
3.8744,23.6631,86.6068,93.7057
3.9052,23.9086,87.5055,94.6781
3.9361,24.1738,88.4762,95.7283
3.9670,24.4292,89.4109,96.7397
3.9963,24.6944,90.3816,97.7899
4.0272,24.9400,91.2803,98.7623
4.0565,25.2150,92.2870,99.8515
4.0843,25.4901,93.2936,100.9406
4.1152,25.7356,94.1923,101.9130
4.1460,25.9910,95.1271,102.9244
4.1738,26.2660,96.1337,104.0135
4.2031,26.5214,97.0684,105.0249
4.2386,26.7670,97.9672,105.9973
4.2711,27.0126,98.8659,106.9697
4.3019,27.2679,99.8007,107.9811
4.3313,27.5233,100.7354,108.9924
4.3621,27.7885,101.7061,110.0426
4.3976,28.0341,102.6048,111.0151
4.4285,28.2895,103.5396,112.0264
4.4594,28.5449,104.4743,113.0378
4.4918,28.8003,105.4090,114.0491
4.5196,29.0655,106.3798,115.0994
4.5520,29.3111,107.2785,116.0719
4.5844,29.5567,108.1773,117.0443
4.6153,29.8022,109.0761,118.0168
4.6477,30.0478,109.9749,118.9892
4.6801,30.2933,110.8737,119.9617
4.7063,30.5782,111.9162,121.0897
4.7372,30.8238,112.8150,122.0621
4.7696,31.0792,113.7497,123.0734
4.8067,31.3051,114.5766,123.9681
4.8422,31.5507,115.4754,124.9406
4.8761,31.8060,116.4101,125.9520
4.9039,32.0713,117.3808,127.0022
4.9286,32.3463,118.3875,128.0913
4.9626,32.6017,119.3222,129.1027


Not very good agreement between the two sets, but I'm only guessing at the time scale for this one. The floor scaling was not so good on either.
OneWhiteEye
21st December 2007 - 07:15 AM
Oh, I see. It's been resized. Never mind.
OneWhiteEye
21st December 2007 - 07:55 AM
When Larry Silverstein said 'pull it' I had no idea he literally meant pull it to the ground faster than gravity could.

User posted image
http://i9.tinypic.com/85f9y8m.jpg
OneWhiteEye
21st December 2007 - 08:40 AM
user posted image

Listen to my heart beat.
einsteen
21st December 2007 - 01:06 PM
I've not much time, this will be a house painting weekend, yes I took that penthouse also, the right part
PhysOrg scientific forums are totally dedicated to science, physics, and technology. Besides topical forums such as nanotechnology, quantum physics, silicon and III-V technology, applied physics, materials, space and others, you can also join our news and publications discussions. We also provide an off-topic forum category. If you need specific help on a scientific problem or have a question related to physics or technology, visit the PhysOrg Forums. Here you’ll find experts from various fields online every day.
To quit out of "lo-fi" mode and return to the regular forums, please click here.