Merged Discussion of femr's video data analysis

[femr2]

A couple of questions I hinted at earlier, but then forgot to actually ask...

1. What derivative do your SS_err values relate to: velocity or acceleration ?
   
2. If acceleration, did you simply apply another central difference approximation to the derived velocity data ?
   
3. Did you use adjacent samples in my data when performing central difference approximation, or a wider band equivalent to the NIST ~0.25s sample interval ?

W.D.Clinger, I assume you'll get around to responding, but in the meantime (and assuming your answers) a few observations...

Assumptions:

• Your SS_err values relate to velocity.
• You used samples n-1 and n+1 in my raw displacement data to perform your central difference approximation.
• You integrated your reverse engineered degree-5 polynomial for acceleration to produce your "femr2 Poly(10)" velocity curve.

Observations:

• You'll clearly have lost detail during the integration.
• You made assumptions about the methods by which the acceleration curve femr2 Poly(10) was generated.
• You performed central difference approximation using n-1 and n+1 on 59.94Hz sample data ! resulting in an effective inter-sample interval of ~0.05s. NIST used a slightly variable inter-sample interval, with most being ~0.23s.
• You'll have applied your previously highlighted T₀ choice to alignment between NIST and femr2 data.

The following graph contains my velocity curve Poly(10), the NIST velocity curve (Shifted) and symmetric difference data using an effective 0.23s wide window. My T₀ is still 11.88s of course..

...and yours...

• In order to be equivalent to the NIST central difference data, I have used a 13 sample symmetric difference (n-6, n+6).
• My T₀ is still set to 11.88s of course, even though the NIST curve has been shifted ~1s earlier to force generalised alignment. Use of my curve before that T₀ is inappropriate.
• The symmetric difference point data between 12-13s and 15.5-17s is matched by my curve much better than the NIST curve in contradiction to your previous assertions (which suffered from an unreasonable T₀ for my data).
• Your point data is extremely noisy for obvious reasons (3 sample wide window on 59.94Hz sample data).

I'll sort out SS_err data later (though a picture tells a thousand words), but if my assumptions above are correct, it is clear that your previous conclusions...

NIST's models are considerably more accurate than femr's near the beginning of the collapse (from 11 to 13 seconds on femr2's time scale), but are considerably less accurate near the end of femr2's data (at 17.2 seconds on femr2's time scale).

...are unfortunately based upon inappropriate method and assumption.

Even before performing SS_err calculations I'm confident in suggesting that my Poly(10) velocity curve is more accurate than NISTs throughout any >=0.5s period following my T₀...just by looking at the graph.

 

[femr2]

Femr, is this a place to discuss the interpretation of the drop and acceleration data?

The meaning of the "greater than g" hump in the data?

I don't see why not, though I guess it depends upon what interpretation is suggested, what meaning is suggested and whether it's finalised, sound and productive dialogue or ongoing thought process.

I suggest the latter is likely to cause lots of noise, which I hope has just settled down a bit.

 

[W.D.Clinger]

If you've been following this thread, then you know that some people have misinterpreted part of NIST's NCSTAR 1-9 Volume 2 section 12.5.3 as a claim that WTC 7's north wall descended at exactly 1g for 2.25 seconds.

You also know that the nonlinear model described in NIST's section 12.5.3 describes

• average acceleration of almost 1g for those 2.25 seconds
• instantaneous acceleration of 1g or greater for almost 1 second out of those 2.25 seconds

In the acceleration graph below, which uses femr2's time scale, the 2.25 seconds in question run from about 12.65 to 14.9 seconds, and the interval of acceleration at 1g or greater runs from about 13.56 to 14.53 seconds.

NIST's section 12.5.3 never states the average acceleration given by its model for the 2.25s interval in question. Although it's easy to calculate, and several recent posts have suggested that its value might contradict parts of NIST's section 12.5.3, I don't think anyone has yet done the calculation.

After I correct an error in one of my responses to femr2, I will perform the calculation, quote the relevant passage of from NIST section 12.5.3, and draw a few conclusions.

My error referred to NIST's Figure 12-77, which femr2 has extracted for us:

The red line is NIST's linear approximation to part of NIST's nonlinear approximation.

No, the red line is NIST's linear approximation to NIST's data points for Stage 2. The purpose of that red line was to provide an empirical test (sanity check) of NIST's model during Stage 2, whose formula for velocity is shown in the boxed legend at the upper left. The formula for the red line is shown in the boxed legend at the bottom right.

(The rest of the paragraph in which my error appeared remains correct.)

Here's what NIST's section 12.5.3 says about Figures 12-76 and 12-77:

Figure 12–76 presents a plot of the downward displacement data shown as solid circles. A curve fit is also plotted with these data as a solid line. A function of the form z(t) = A{1 – exp[–(t/λ)^k]} was selected because it is flexible and well-behaved, and because it satisfies the initial conditions of zero displacement, zero velocity, and zero acceleration. The constants A, λ, and k were determined using least squares fitting. The fitted displacement function was differentiated to estimate the downward velocity as a function of time, shown as a solid curve in Figure 12–77. Velocity data points (solid circles) were also determined from the displacement data using a central difference approximation. The slope of the velocity curve is approximately constant between about 1.75 s and 4.0 s. To estimate the downward acceleration during this stage, a straight line was fit to the open-circled velocity data points using linear regression (shown as a straight line in Figure 12–77). The slope of the straight line, which represents a constant acceleration, was found to be 32.2 ft/s2 (with a coefficient of regression R² = 0.991), equivalent to the acceleration of gravity g. Note that this line closely matches the velocity curve between about 1.75 s and 4.0 s.

To review: the velocity curve (black curve in Figure 12-77) comes from NIST's nonlinear model, which was obtained by least squares fitting over the entire interval from 0 to 5.4 seconds on NIST's time scale (which corresponds to about 10.9 to 16.3 seconds on femr2's time scale). The straight red line in Figure 12-77 comes from least squares fitting of a linear model to NIST's data points for the Stage 2 interval from 1.75 to 4.0 seconds on NIST's time scale (which corresponds to about 12.65 to 14.9 seconds on femr2's time scale).

How closely does the red line match the velocity curve between 1.75 s and 4.0 s? The red line corresponds to a constant acceleration of about 32.2 ft/s². The velocity curve corresponds to an average acceleration of about 30.24 ft/s² during Stage 2. So NIST was saying that 32.2 is close to 30.24.

The value of 30.24 ft/s² comes from NIST's nonlinear model, as displayed by the black lines in Figures 12-76 and 12-77. By the fundamental theorem of calculus, integrating NIST's nonlinear model for acceleration over the interval from 1.75 to 4.0 seconds is equivalent to subtracting the value given by NIST's nonlinear model for velocity at 1.75s from the value given by that model at 4.0s. That tells us that, according to NIST's nonlinear model, the change in velocity was about 68.0 ft/s during that interval. Dividing by the 2.25s duration of the interval gives us the average acceleration: about 30.24 ft/s².

With better data, as has been provided by femr2, we can do a better job of estimating the average acceleration during NIST's Stage 2. Fitting a second degree polynomial to femr2's vertical displacement data for the NW corner of WTC 7 from 12.65s to 14.9s, and differentiating twice, I got 33.34 ft/s², which is slightly higher than NIST's estimate of 32.2 ft/s² for the entire north wall.

Some people might wonder how the average acceleration of the NW corner could have been greater than 1g during NIST's stage 2. From physics, we know that when a moderately rigid object is undergoing some rotation as it falls, the largest differences between the acceleration of its center of mass and the acceleration of a particular point on the object are likely to occur at the object's most extreme points. The NW corner is an extreme point of the north wall.

Before we leave the subject of femr2's data, let's calculate the average acceleration during Stage 2 for a version of NIST's nonlinear model whose parameters have been recalculated using femr2's data for the NW corner. We would expect that recalculated model to provide a better match for the acceleration calculated from femr2's data, and it does: applying the fundamental theorem of calculus as before, the change in velocity during Stage 2 is about 76.65 ft/s, and the average acceleration is about 34.1 ft/s².

In conclusion, NIST's linear regression (the red line in Figure 12-77) served as a sanity check on its nonlinear model (the black curve in Figure 12-77). NIST found the agreement between its linear regression and its nonlinear model to be close but not exact. Using femr2's data for the NW corner to estimate the average acceleration during Stage 2 and to recalculate the parameters of NIST's nonlinear model improves the agreement between NIST's nonlinear model and observation.

 

[femr2]

I'll sort out SS_err data later (though a picture tells a thousand words)

Pretty conclusive I'd say.

 

[Major_Tom]

Some people might wonder how the average acceleration of the NW corner could have been greater than 1g during NIST's stage 2.

If people had a clue of how to study real movement, this would have been the fundamental question. A transient surge over g could easily be a unique signature of a special kind of motion.

From physics, we know that when a moderately rigid object is undergoing some rotation as it falls, the largest differences between the acceleration of its center of mass and the acceleration of a particular point on the object are likely to occur at the object's most extreme points. The NW corner is an extreme point of the north wall.

The interpretation of the greater than g surge is important for those who want to understand the early motion. I think that when people complain about femr too much and totally overlook the fact that there is a greater than g surge in the acceleration, they show that they have no clue how to study motion.

Not you so much since your comments are much better than others.

It is like preparing the bath water and then putting the baby in the garbage. The correct data is only the bath water. Where is the analysis? Forgetting something?

We will verify greater than g surge and then then hand-wave it off as normal? Like NIST, aren't you overlooking a big freaking clue?

We know the greater than g surge is not from ordinary falling as a "block" or a point of mass would do since we know that there is a "g" acceleration limit to natural falling.

None of us disagrees with Newton'ts concept of gravity so the hump must mean something else.

Both femr and WD CLinger mention tilt. Good idea, but it is well worth checking further.

 

[NoahFence]

How do all these graphs and charts change things?

 

[Major_Tom]

They are called "measurements". They have proven useful when analyzing motion for hundreds of years.

For some reason, debunkers tend to be uninterested in them.


Analysis of them is called "physics". Another weak spot for debunkers and many, many truthers.

If you don't use them, you are analyzing your own dreams only.

 

[NoahFence]

They are called "measurements". They have proven useful when analyzing motion for hundreds of years.

For some reason, debunkers tend to be uninterested in them.


Analysis of them is called "physics". Another weak spot for debunkers and many, many truthers.

If you don't use them, you are analyzing your own dreams only.

And yet my question remains unanswered.

 

[femr2]

If people had a clue

That kind of dialogue isn't going to keep the noise down here, is it

Both femr and WD CLinger mention tilt. Good idea, but it is well worth checking further.

I suggest attachment to the core could be the primary factor.

Attachment to the core could explain numerous behaviours.

 

[Myriad]

We know the greater than g surge is not from ordinary falling as a "block" or a point of mass would do since we know that there is a "g" acceleration limit to natural falling.

Such a limit applies only to the center of gravity of the entire mechanically connected falling mass. It does not hold for arbitrary points on that mass. Knock a butter knife off a table, and parts of it at times will accelerate downward at greater than g. (And that's about as ordinary as falling can get, isn't it?)

Respectfully,
Myriad

 

[femr2]

You also know that the nonlinear model described in NIST's section 12.5.3 describes...average acceleration of almost 1g for those 2.25 seconds

As you note later, we haven't actually performed that calculation yet...

NIST's section 12.5.3 never states the average acceleration given by its model for the 2.25s interval in question. Although it's easy to calculate, and several recent posts have suggested that its value might contradict parts of NIST's section 12.5.3, I don't think anyone has yet done the calculation.

...and, as yet have still not done so.

After I correct an error in one of my responses to femr2, I will perform the calculation

Thanks for the correction, however, I have written numerous posts recently highlighting numerous issues.

I hope you are not one of those folk who chooses not to respond to criticism, even if it requires admission of error or confirmation of alternate point of view.

In particular I think I have provided more than enough additional detail regarding your recent assertions that the NIST acceleration curve (which they don't use within the report) was more accurate than my Poly(10) to warrant a response from you.

A significant focus of this thread has been *is femr2's data accurate*, which includes *is femr2's acceleration graph more accurate than NISTs* and also *how much more accurate is femr2's displacement/velocity/acceleration data than NISTs*.

It would be appreciated if you address my responses to your earlier posts.

The purpose of that red line was to provide an empirical test (sanity check) of NIST's model during Stage 2

As has been said on other sides of different coins, I don't think there is enough information to second-guess what NIST were doing if they didn't state it.

In my opinion NIST didn't touch their equation of motion beyond deriving for velocity. I see no indication that they touched deriving it to acceleration in the section at all.

Your opinion may be correct. It may not. May I suggest making opinion distinct from fact.

The red line corresponds to a constant acceleration of about 32.2 ft/s².
The velocity curve corresponds to an average acceleration of about 30.24 ft/s² during Stage 2.
So NIST was saying that 32.2 is close to 30.24.

I see nothing within the section to support your opinion that is what NIST was saying I'm afraid.

I think rather that asserting such attempts to lend weight to your opinion that NIST utilised their equation of motion derived to acceleration. I don't think they did.

We can of course disagree.

Fitting a second degree polynomial to femr2's vertical displacement data for the NW corner of WTC 7 from 12.65s to 14.9s, and differentiating twice, I got 33.34 ft/s², which is slightly higher than NIST's estimate of 32.2 ft/s² for the entire north wall.

A more accurate peak acceleration is much nearer to 40 ft/s², as you can see from the numerous acceleration profiles provided. I'd eye-ball estimate around 37ft/s² for the NW corner.

Before we leave the subject of femr2's data

See requests above.

NIST found the agreement between its linear regression and its nonlinear model to be close but not exact.

You found that, NIST may not have performed the checks you have at all.

 

[Major_Tom]

That kind of dialogue isn't going to keep the noise down here, is it

I am sorry. I am not myself in this forum. Too many cow pies to dance around.

I suggest attachment to the core could be the primary factor.

Attachment to the core could explain numerous behaviours.

Yup. It is a big clue telling us something.


These are my first reactions to your data:

1) Is it real? We reached this same point a few years ago on the other forum before you even joined. We saw the same thing but didn't believe it was real (within errors). We wrote the phenomenon off to a problem locating t=0.

2 years later, this is the first time the hump needs to be treated as real


2) Once we take it to be real, what the &$#^ is that greater than g hump?

I'm kind of shocked that nobody else seems interested. I watched research develop for 2 years and you do this wonderful job now.

And then nobody cares about the unique signature of the hump? Whoa, Nelly!

I agree it probably demonstrates a core-perimeter interaction. I think Achimspok has come the closest to capturing the early global interaction between penthouse drop and perimeter "take-down" through the greater than g hump.

Of course it requires careful analysis of the kind in which you have proved quite talented.


3) Does a tight core-perimeter coupled model with a transient greater than g whipping action match the NIST description or model of collapse?

Not even close. There is clearly much more internal structure between core and perimeter than any NIST model comes close to acknowledging.

>>>>>>>>>>>>>>>>>>>>>>>>

And so my first reality check for the NIST is this:

Since they themselves measured the greater than g hump, why do they seem to have no clue what it implies for collapse initiation mechanics?

It is as if they made an important discovery and then completely ignored its implications in their own analysis.


Next reality check: Their own data shows them they cannot ignore horizontal components of early motion. The greater than g hump is a big warning sign they are missing an important element in the mechanics.

Their first reaction should have been: "What the %@#& is that greater than g hump?" "Maybe it is a signature of a whip-down action within the perimeter shell?"

Instead they issue it, and many people read it, as if greater than g moments are pretty normal during the initial movements of a building.

To me that is like there is no reality in the discussion at all.

 

[alienentity]

That kind of dialogue isn't going to keep the noise down here, is it


I suggest attachment to the core could be the primary factor.

Attachment to the core could explain numerous behaviours.

Indeed. That is implicit in NCSTAR 1-9, and with regard to the NW corner it is interesting to note the wording:

'As the interior columns buckled at the lower floors and the corresponding upper column sections began to move downward, the exterior columns buckled inward at the lower floors as a result of floor pull-in forces caused by the downward movement of the building core. The floor connections to the columns had not yet failed in this region, as there were no fires
observed on the west side of Floors 10 through 14 at any time during the day, so the floors were intact and able to pull the exterior columns inward. '
1-9 Vol 2, pp 606-607

 

[Myriad]

I suggest attachment to the core could be the primary factor.

Attachment to the core could explain numerous behaviours.

And I agree completely.

From nearly two years ago: http://www.internationalskeptics.com/forums/showthread.php?postid=4912626#post4912626

So, having established that something that's attached to something that's already falling may accelerate faster than g, I re-examined the available information on building 7 and found considerable evidence that the wall in question was indeed attached to other parts of the building, and that other parts of the building did indeed begin falling first.

Therefore, I conclude that a brief period of free fall or even faster than free fall acceleration of a portion of a building experiencing progressive collapse, while it might be an unusual phenomenon, is not unexpected or unexplainable given the known circumstances.

It seems to have been a very long run for a very short slide,* but as long as we got there eventually I guess I shouldn't complain.

Respectfully,
Myriad

*Colloq. phrase derived from baseball, meaning approximately "a lot of effort for relatively little reward," such as a long explanation to establish a simple or already apparent point. Interestingly, the most frequent current usage appear to be in medicine, to describe invasive or expensive diagnostic tests that make minimal difference in the final diagnosis or treatment. Could also apply to the use of a "pithy" phrase in an Internet post which then requires a paragraph of explanation.

 

[alienentity]

I
2) Once we take it to be real, what the &$#^ is that greater than g hump?

I'm kind of shocked that nobody else seems interested. ..

And then nobody cares about the unique signature of the hump? Whoa, Nelly!

Whatever point it is you're trying to make, you're overselling it. What do you think is so important about the hump?

..

3) Does a tight core-perimeter coupled model with a transient greater than g whipping action match the NIST description or model of collapse?

Not even close. There is clearly much more internal structure between core and perimeter than any NIST model comes close to acknowledging.

hmmm. A completely unsupported statement which comes across as a bias against NIST. Just sayin'.

 

[beachnut]

Femr, is this a place to discuss the interpretation of the drop and acceleration data?

The meaning of the "greater than g" hump in the data?

Wow, impressive engineering terms. How does this fit with your view of 911? How does this help your claim of this paranoid nonsense?

These are just some of the factors which, when studied in depth, show that the supposed "gravity-driven collapse" is a mere illusion to mask an intentional act so barbaric, so inhumane and morally impoverished that the fabled characteristics of Satan come to mind.

With gravity being the primary energy source in CD, how did you come up with this failed claim, and how does femr's video data analysis, analysis goals, and analysis conclusions dovetail with this claim. Any explanation to tie this all together into one single integrated coherent story? Any goals where this is going, any thesis? What does your hump mean?

 

[Major_Tom]

Indeed. That is implicit in NCSTAR 1-9, and with regard to the NW corner it is interesting to note the wording:

'As the interior columns buckled at the lower floors and the corresponding upper column sections began to move downward, the exterior columns buckled inward at the lower floors as a result of floor pull-in forces caused by the downward movement of the building core. The floor connections to the columns had not yet failed in this region, as there were no fires
observed on the west side of Floors 10 through 14 at any time during the day, so the floors were intact and able to pull the exterior columns inward. '
1-9 Vol 2, pp 606-607

Yes, and then they used camera 3 data while totally ignoring horizontal movement.

Not smart.

They also use an FEA that totally ignores this integral connection and flexing between core and perimeter. Not smart.

Femr has shown that the NIST incorrectly stretched "stage 1" by not considering the early horozontal motion.

Stage 2 is telling us that a tight whip relation between core and perimeter is a good possiblity, and the discovery of flexure in the perimeter, shown in a gif earlier, is further proof.

Reinterpretation and correction of the data shows us a probable collapse initiation mechanism. VIsual evidence confirms the mechanism.


So how can the NIST sell you such a crappy model that totally ignores it or the real geometry of the building?

Where is a whip mechanism? Where the characteristic greater than g hump? We need tightness in the west architecture for a greater than g hump.

So whip it! Whip it good! Go for it! Move ahead!

And whip it!

 

[NoahFence]

I'm still trying to figure out what the purpose of this over-analysis is.....

 

[Major_Tom]

Like with WTC 1 information, it shows that many of you are living in a dream world, but it does so mathematically.



The real motion can only be expressed through visual evidence and correct measurement. WIthout them, I have no idea what people imagine they are studying.

Echoes of ones own mind?

 

[NoahFence]

Like with WTC 1 information, it shows that many of you are living in a dream world, but it does so mathematically.



The real motion can only be expressed through visual evidence and correct measurement. WIthout them, I have no idea what people imagine they are studying.

Echoes of ones own mind?

For crying out loud....

Does it in any way shape or form change what happened that day? Does it prove or disprove anything aside from NIST possibly being incorrect in some measurements?