Debunked: NIST computer simulation of Building 7 collapse is inaccurate

[Marc Powell]

Conspiracy theorists claim that the NIST computer simulation of the collapse of Building 7 is inaccurate because it shows walls folding in as the interior of the building crumbles away. And since the NIST simulation looks nothing like videos of the actual event where the building walls remained relatively vertical as they fell, it is asserted that the NIST hypothesis for the collapse mechanism must be wrong. In the 2014 David Hooper film, The Anatomy of a Great Deception, (viewable in its entirety at youtube.com/watch?v=l0Q5eZhCPuc ), the NIST computer simulation is presented several times. The first is at the 20:37 mark, where it is speeded up for comic effect, and the second is at the 32:34 mark where it is repeated several times interspersed with a repeated video clip of the actual global collapse. Below is a screen capture of the NIST simulation from the video clip presented in Hooper's film:

The narrator (David Hooper) presents a list of presumably provable issues denied by NIST and then says, "Even the NIST computer animation of Building 7's collapse was inaccurate showing the outer structural walls crumpling in instead of coming straight down." However, that is not true at all. NIST conducted two collapse simulations for Building 7, one that included damage due to debris impact from the collapse of the WTC North Tower and one that did not include such damage. The simulation with debris impact damage closely resembles the actual observed collapse event. The simulation shown in Hooper's film is the other simulation that differs significantly from actual observations. Below is the NIST simulation with impact damage that Hooper should have presented:

For comparison, below is the NIST collapse simulation without impact damage (the one Hooper chose to present):

By presenting the wrong computer simulation Hooper's film leaves its audience with the false impression that either, (1) NIST scientists are incompetent and their report unscientific or, (2) that the report is a fantasy produced strictly for political purposes. One would expect that, with all their exhaustive research, sincere seekers of truth and justice like David Hooper and his Technical Director, Richard Gage, would know about the two NIST computer simulations and make an effort to present the correct one.

 

[Thomas B]

NIST conducted two collapse simulations for Building 7, one that included damage due to debris impact from the collapse of the WTC North Tower and one that did not include such damage.

Thanks for this. It's a very good point. Do you have a reference to the part of the NIST report (or a FAQ) where this is explained?

 

[Marc Powell]

Thanks for this. It's a very good point. Do you have a reference to the part of the NIST report (or a FAQ) where this is explained?

Chapter 12 of the NIST Building 7 investigation talks about the collapse simulations. The simulation with debris impact damage is covered in Section 12.4.4 and the one without impact damage is covered in Section 12.4.6. Section 12.5 compares the two simulations with actual observables and Section 12.6 summarizes findings.

 

[Landru]

Chapter 12 of the NIST Building 7 investigation talks about the collapse simulations. The simulation with debris impact damage is covered in Section 12.4.4 and the one without impact damage is covered in Section 12.4.6. Section 12.5 compares the two simulations with actual observables and Section 12.6 summarizes findings.

Perhaps you can provide links.

 

[Marc Powell]

Perhaps you can provide links.

I was referencing NIST NCSTAR 1-9. It can be downloaded at https://tsapps.nist.gov/publication/get_pdf.cfm?pub_id=861611. Section 12.4.4 starts on page 571 and Section 12.4.6 starts on page 591. In Section 12.5, Table 12-2 on page 599 compares the two simulations with actual observed events. Section 12.6, starting on page 603, summarizes findings for the collapse simulations NIST performed.

This information is also presented in a more abbreviated form in Section 3.4.6 of NIST NCSTAR 1A, available at https://nvlpubs.nist.gov/nistpubs/Legacy/NCSTAR/ncstar1a.pdf .

 

[Mendel]

Thanks for this. It's a very good point. Do you have a reference to the part of the NIST report (or a FAQ) where this is explained?

Googling for "nist wtc7 collapse simulation version" gives you the FAQ, and answer 29 adresses this issue.

 

[Oystein]

Googling for "nist wtc7 collapse simulation version" gives you the FAQ, and answer 29 adresses this issue.

As per no-click policy:

29. The simulation of the collapse modeling of WTC 7 does not match the video footage of the collapse. In particular, the large inward deformations of the upper exterior walls after the beginning of global collapse are not visible in the video footage. Can NIST explain the difference between the results of its computer model of the collapse and the available video evidence?

NIST conducted two global collapse analyses, one that included damage due to debris-impact from the collapse of WTC 1, and one that did not include any debris-impact damage. These two analyses were conducted to determine the influence of the debris-impact damage on the response of WTC 7 when subjected to the effects of the fires that burned on floors 7 to 9 and 11 to 13. In its comparison of the two analyses (see NIST NCSTAR 1A Section 3.5), NIST showed that the analysis with the debris-impact damage better simulated the sequence of observed events, and it is this simulation that is considered here.

NIST believes that the simulation of the collapse, based on the analysis with debris-impact damage, does capture the critical observations derived from the digital video recording. The critical observations and corresponding failures identified in the structural analysis include: 1) east-west motion of the building beginning at approximately the same time as failure of floors 6 through 14 around Column 79, 2) the formation of the "kink" in the roofline of the east penthouse approximately one second after Column 79 was found to buckle, 3) window breakage on the east side of the north face as the buckling of Column 79 precipitated the failure of upper floors, and 4) the beginning of global collapse (vertical drop of the building exterior) within approximately one-half second of the time predicted by analysis. Both measured time and analytically predicted time, from the start of failures of floors surrounding Column 79 to the initial downward motion of the north face roofline, was 12.9 seconds (see NIST NCSTAR Report 1A, Table 3-1). The collapse observations, from video analysis of the CBS News Archive video, are covered in detail in NIST NCSTAR Report 1A Section 3.5 and NIST NCSTAR Report 1-9, Section 8.3. Only in the later stages of the animation, after the initiation of global collapse, do the upper exterior wall deformations from the NIST analysis differ from the video images.

Uncertainties associated with the approach taken by NIST are addressed in NIST NCSTAR Report 1A, Section 3.5, where it is noted, "Once simulation of the global collapse of WTC 7 was underway, there was a great increase in the uncertainty in the progression of the collapse sequence, due to the random nature of the interaction, break up, disintegration, and falling debris." The contribution to stiffness and strength of nonstructural materials and components, such as exterior cladding, interior walls and partitions, was not considered in the analyses conducted by NIST. It is well known that such non-structural components can increase the stiffness and strength of a structural system, but their contribution is difficult to quantify. Given these factors, disparities between the video and the animation in the later stages of collapse would be expected.

Source: https://www.nist.gov/topics/disaster-failure-studies/faqs-nist-wtc-7-investigation

 

[econ41]

Isn't the problem even more simple and fundamental. Two problems actually:

First the type of simulation NIST presents exaggerates the magnitude of strains to show the direction of movements. It is not intended to "look like" the real event rather to allow further engineering analysis. It is a class of simulation which intentionally does NOT "look like" the real event. That is not its purpose.

Second truthers, esp those influenced by AE911, have been taught that simulations MUST visually "look like" the actual. That is not true of all engineering modelling - some types of simulations "look like". Many dont. Both are legitimate.

THEREFORE the reason for the trickery with the Hulsey "simulations". They were deliberately "fudged" to meet the pre-set expectation of the AE911 audience. Viz Simulations which "look like" are correct - simulations which do not "look like" are wrong.

So in the broader picture "Simulations must look like the real event" is merely another of those handful of false premises which underpin much of truther lore. Other ones being "Free fall proves CD" >> it doesn't. "Falling in own footprint proves CD" >> it doesn't. "Symmetry proves CD" >> It doesn't - whatever is meant by "symmetry". And setting aside that for all three the core claim (free fall, footprint et simile) is usually false.

 

[Mendel]

As per no-click policy:

I was hoping to "teach a man to fish", since Thomas seemed to be aware that the NIST FAQs exist, but also appeared unable to find them. That's why I didn't provide a direct link.
I wasn't discussing the information, I was discussing how to find it.

 

[Mendel]

First the type of simulation NIST presents exaggerates the magnitude of strains to show the direction of movements. It is not intended to "look like" the real event rather to allow further engineering analysis. It is a class of simulation which intentionally does NOT "look like" the real event. That is not its purpose.

Well, there's the simulation and the visualization of the simulation.
The simulation itself is calculating the physics and the movement that results from it, and that should reflect the event.
The visualization aims to make the results of the calculations apparent.

The problem is that when the simulation results that interest you are too small to notice, you need to exaggerate them. Kinda like all maps that show "profile" (or even all profile graphs, like for a mountain section of the Tour de France) scale height more because otherwise you wouldn't really notice the hills that much.
The profile is still based on real data, it's just visualized in a distorted way because that's more useful that way.

Is that what you're referring to?

 

[econ41]

The problem is that when the simulation results that interest you are too small to notice, you need to exaggerate them. Kinda like all maps that show "profile" (or even all profile graphs, like for a mountain section of the Tour de France) scale height more because otherwise you wouldn't really notice the hills that much.
The profile is still based on real data, it's just visualized in a distorted way because that's more useful that way.

Is that what you're referring to?

Exactly. There are broadly two main classes of simulations or modelling viz "look like" or "exaggerated to facilitate an engineering analysis purpose".

The gross distortions of the NIST model grossly exceed the ductility limits of the materials. Engineers and others familiar with the methodology are not confused. But whether intended or not the failure to "look like" the real event has been used by truther advocates to mislead followers. So that is a fundamental problem - the other details are probably of less importance.

 

[Oystein]

...the type of simulation NIST presents exaggerates the magnitude of strains to show the direction of movements. ...

I am fairly certain that this is NOT the case for the two global collapse animations in question here.
Could you please provide a citation for your claim?

 

[econ41]

I am fairly certain that this is NOT the case for the two global collapse animations in question here.

Interesting. Given the level of ductility needed to allow the large distortions.

Could you please provide a citation for your claim?

None. Other than general engineering understanding. Since that doesn't meet the criteria for this forum I won't pursue the issue.

 

[Oystein]

Interesting. Given the level of ductility needed to allow the large distortions.

Right.
Which would (if I am correct and you are wrong) mean that your "general engineering understanding" fails you in this regard, or that this NIST simulation is "wrong" (a point I have seen some Truthers make: "Look at how unrealistically distorted this is, how can you believe this? And in extrapolation: How can you believe anything NIST published?" (<- obviously a paraphrase, not actual quote)) - or both.

None. Other than general engineering understanding. Since that doesn't meet the criteria for this forum I won't pursue the issue.

Should be in the sub-sections of Chapter 12 of NIST NCSTAR 1-9 that @Marc Powell referenced above.
I won't pursue it either - your claim, your burden of proof.

 

[econ41]

Right.
Which would (if I am correct and you are wrong) mean that your "general engineering understanding" fails you in this regard,...

A rare but real possibility.

.... or that this NIST simulation is "wrong" (a point I have seen some Truthers make: "Look at how unrealistically distorted this is, how can you believe this? And in extrapolation: How can you believe anything NIST published?" (<- obviously a paraphrase, not actual quote)) - or both.

NIST is only "wrong" in the minds of truthers who are looking for a "look like" simulation. Yes I agree (have done many times over quite a few years) that the NIST modelling does not "look like" the real event.

f course if I am wrong then the whole field of argument is open for debate....

I won't pursue it either - your claim, your burden of proof.

I have no interest in proving it. If I'm right it only adds one more example to the history of identifying fundamental issues which are rarely picked up in debates which are focused on details. If I'm wrong - I'm wrong. Add one more to the list.

By the way can I take it that you don't disagree with my starting premise - that there are two broad types of simulations/models viz "look like" and "not intended to look like"?

 

[Oystein]

...
By the way can I take it that you don't disagree with my starting premise - that there are two broad types of simulations/models viz "look like" and "not intended to look like"?

Yes.

 

[Jeffrey Orling]

I think in this day of computer animation many assume that an animation can be created which is accurate and faithful in the details to the actual collapse. This IS possible with adequate and accurate input variables and enough computing power and time. Lots of things these days use computer simulations. Even film simulates and mixed the simulations with real "footage".
So the input variables require accuracy. I don't think this is / was available. Whomever builds these simulations has to limit the number of variables and make a lot of assumptions. Lot of hit or miss and with such complexity the desired result may elude.

So what is the purpose of such a simulation?

 

[JarJar]

Does NIST specify if the simulation was created with a particular piece of software or engine? Or maybe by a contractor? I don't see that info anywhere.

Companies like Dassault Systèmes have used design/testing simulations without the need for a physical prototype for more than 30 years. Any variables can be programmed.

Edit: at least part of it was done with ANSYS who has a similar history of simulations. They even list the model details for the different materials in the appendix of their report.

 

[Marc Powell]

So what is the purpose of such a simulation?

My understanding is that the NIST computer simulations were intended merely to demonstrate the plausibility of the premise that a single connection failure (resulting from the effects of fire) could initiate a progressive collapse that could bring down the entire structure. The detailed discussion about beam A slipping out of connection B and slamming into floor C, causing buckling of column D, etc. is only in reference to what the model did, not to what actually happened. There were far too many assumptions made and much too much chaos in the collapse for the model representing it to provide anything more than a general idea of what may have been possible. And the farther into the collapse the simulations were carried, the more unreliable the results became. That is why NIST cut them off when they did… a few seconds into the beginning of the global collapse. And that is also why any simulation that looks exactly like the actual events (Hulsey) should be suspect. Please forgive me if I am stating the obvious. Computer modeling of building failures is not my forte.

 

[Oystein]

... And the farther into the collapse the simulations were carried, the more unreliable the results became. ...

If there is anything of real value to be learned from NIST's two LS-DYNA runs, it is this:

The outer appearance of the collapse progresses is significantly dependent on (changes significantly with varying) starting conditions.

It is very much apparent to both the amateur and the expert viewer that the collapse simulations a) with, b) without structural damage to the South face from the North Tower collapse are quite different.

Next step to understand: NIST (nor anyone else) did not know the exact extent and severity of that initial structural damage.
Also, there is of course a real degree of uncertainty about the exact extent and severity of damage accrued by the many hours of fire progression.

It follows that we must expect the collapse simulation to differ from reality, given that it is more likely than not that real damage differed from model input damage.
I would expect (and a close reading of Chapter 12 no doubt will reveal this) that NIST's assumptions for the with-initial-damage run were rather on the conservative side - that they input only the damage which has sufficient evidence from observation going for it; there may have been more damage, or more severe damage, in reality.

The with-initial-damage run still did a decent job of recreating a number of the observed significant collapse features - more than the silly Hulsey model. Such as:

• early collapse of EPH, with correct kink in EPH itself (Hulsey has the bottom of the EPH's sides swinging outward, instead of the corners sinking inward)
• East-to-West progression of WPH collapse
• the kink in the Eastern portion of the North wall
• the building twisting counter-clockwise

 

[Jeffrey Orling]

Seems to me that there were way too many variables especially as failure progresses through the structure to produce a simulation which closely resembles real world other than in gross ways.
NIST felt they had to demonstrate that a local failure at a column in the NE quadrant could produce a total collapse. They did.

 

[Marc Powell]

It follows that we must expect the collapse simulation to differ from reality, given that it is more likely than not that real damage differed from model input damage.
I would expect (and a close reading of Chapter 12 no doubt will reveal this) that NIST's assumptions for the with-initial-damage run were rather on the conservative side - that they input only the damage which has sufficient evidence from observation going for it; there may have been more damage, or more severe damage, in reality.

According to firefighter reports, Building 7 had been popping, creaking and showing external deformations as the un-fought fires spread that afternoon. These were indications of internal failures that NIST had no way of quantifying and so, were not considered in its modeling of the initial damage conditions.

The with-initial-damage run still did a decent job of recreating a number of the observed significant collapse features - more than the silly Hulsey model. Such as:

• early collapse of EPH, with correct kink in EPH itself (Hulsey has the bottom of the EPH's sides swinging outward, instead of the corners sinking inward)
• East-to-West progression of WPH collapse
• the kink in the Eastern portion of the North wall
• the building twisting counter-clockwise

The Hulsey model also did not include the east-west swaying of the building that NIST determined to have started six seconds before the falling of the EPH and continued until the beginning of global collapse. The swaying, which NIST referred to as "vibration," can actually be seen in the CBS video of the collapse taken from West Street.

 

[Jeffrey Orling]

According to firefighter reports, Building 7 had been popping, creaking and showing external deformations as the un-fought fires spread that afternoon. These were indications of internal failures that NIST had no way of quantifying and so, were not considered in its modeling of the initial damage conditions.

The Hulsey model also did not include the east-west swaying of the building that NIST determined to have started six seconds before the falling of the EPH and continued until the beginning of global collapse. The swaying, which NIST referred to as "vibration," can actually be seen in the CBS video of the collapse taken from West Street.

The observations you note WERE the signs of the fire warping the frame which eventually failed. NIST sims are the collapse phase one the frame went past the point of no return structurally. Their models "begin" with the beam / girder drop on floor 13.

 

[Abdullah]

The main deviations from reality I can see are

1. Slow collapse of east penthouse (2.7 Vs 2.0)

2. Inward deformation of the northeast corner

3. Premature collapse of west penthouse (before the screenwall)

4. Inward deformation of the southwest corner

Some of these may be causally connected

1. Slow descent in the east core may have allowed floor connections to persist for longer and therefore impart larger impulse on the wall

2. Early descent in the West core may have generated pull in forces absent in reality

The slow descent in the east may be because

1. NIST underestimated fire damage

2. The initialization sequence that applied all connection failures simultaneously caused premature column 79 buckling before floor damage had completed.

NIST speculated that the early collapse in the West was due to overestimated impact damage.

 

[Abdullah]

Two strange things.
For simevreasin Kai kostack is feeding this myth. Watch him compare Hulseys results with the wrong NIST analysis


Source: https://youtu.be/jVE3YwRgU9k


2. Meanwhile, Dr. Hulsey has set the record straight

 

[Abdullah]

Has southern sway in excess of 1 meter been observed in the West half of WTC7 before global collapse began?

We can overlook gross deformations in unimportant locations, but we really need to verify this because it sets the mechanism of global collapse

 

[Jeffrey Orling]

Sway would be a back and forth movement... not merely a displacement (of the top section).

 

[Abdullah]

Are you sure the idea that the NIST simulation is inaccurate is debunked? I think it is better to say that the claim that they do not remotely resemble the real thing has been debunked.

 

[Jeffrey Orling]

Are you sure the idea that the NIST simulation is inaccurate is debunked? I think it is better to say that the claim that they do not remotely resemble the real thing has been debunked.

The NIST sim doesn't look much like real world. You don't know what's going on inside in the video except clues from "deformation" of the curtain wall telegraphing something dropping inside. Do you think we can see the insides "coming apart" pre collapse?

 

[Abdullah]

The NIST sim doesn't look much like real world. You don't know what's going on inside

I kind of agree with the first sentence, but not for the reason in the second. The inward deformation at the northeast and southwest corners is unrealistic. It suggests that NIST got something wrong.

It is already startling that the real world building showed no inward bowing even as all the floors pulled in at it. But NIST's failure to explain why is frustrating. This is especially true for the northesst corner.

Perhaps the floors were not pulling in, because the specified floor connections did not exist? Ie the building was built wrong? Perhaps this was why the fires were able to collapse floors in the first place?

We also do not know, as you have stated, what happened beyond the curtain walls. The the fires did not cause loss of bracing around column 82? Maybe column 81 therefore did not buckle initially? The more rigid boundary conditions on the falling floors could explain the lack of inward deflection.

 

[Jeffrey Orling]

I kind of agree with the first sentence, but not for the reason in the second. The inward deformation at the northeast and southwest corners is unrealistic. It suggests that NIST got something wrong.

It is already startling that the real world building showed no inward bowing even as all the floors pulled in at it. But NIST's failure to explain why is frustrating. This is especially true for the northesst corner.

Perhaps the floors were not pulling in, because the specified floor connections did not exist? Ie the building was built wrong? Perhaps this was why the fires were able to collapse floors in the first place?

We also do not know, as you have stated, what happened beyond the curtain walls. The the fires did not cause loss of bracing around column 82? Maybe column 81 therefore did not buckle initially? The more rigid boundary conditions on the falling floors could explain the lack of inward deflection.

I can only guess.... I suspect there was a NE failure of beams and girder (cols 79,80, 81...) that caused partial floor slabs to break and collapse. This provided enough "impulse" to disturb the transfers on flrs 5-7. And that failure propagated west through the structure (via girders)... involving much of the internal steel. This led to further partial floor collapses around the core... ultimately leaving the tower largely hollowed out. The axial support at the base was "disturbed" by the impulse from the floor collapse leading to the moment frame and attached curtain wall to collapse more or less intact. It kinked at the girders between the NE corner of the core and the exterior moment frame.

 

[Henkka]

I agree with the OP that the film should have presented the simulation with debris impact and fire damage considered. That's the one NIST presented at their press conference, and is supposedly most representative of what happened in reality. However, I don't agree with the OP that that simulation "closely resembles" the real collapse. This is kind of a dead horse, I've talked about this in many threads before this one, but the biggest problem is of course that the simulation arrests before the building even properly falls. It also doesn't look like the simulation could possibly achieve 9.8 m/s^2 downward acceleration of the roofline.

Another thing, though. The simulation that is presented in the film does show what 9/11 debunkers often contend happened... That the entire interior progressively collapsed east to west. But without internal support, the perimeter walls become floppy and fold inward, unlike what happened in real life. This leaves you with two choices:

1) The simulation is accurate, and therefore the interior did not collapse, as the walls did not fold in

or

2) The simulation is inaccurate, the entire interior did collapse in real life, but there's something wrong with the simulation that makes the walls behave inaccurately

If 1), that essentially proves controlled demolition. You're just left with an entire building suddenly going into symmetrical free fall, without the fig leaf of a preceding interior collapse. If 2), that makes it hard to say what exactly the simulation tells us about what really happened, if anything.

 

[Oystein]

...This leaves you with two choices:

1) The simulation is accurate, and therefore the interior did not collapse, as the walls did not fold in

or

2) The simulation is inaccurate, the entire interior did collapse in real life, but there's something wrong with the simulation that makes the walls behave inaccurately

If 1), that essentially proves controlled demolition. You're just left with an entire building suddenly going into symmetrical free fall, without the fig leaf of a preceding interior collapse. If 2), that makes it hard to say what exactly the simulation tells us about what really happened, if anything.

So much wrong...

First, without thinking too hard about this, you present what is almost certainly a false dilemma: It's not just these two distinct options, there could be more (e.g. that the assumptions you IMAGINED before coming up with the choice are just false), and that both could play a part - there is some gray area here.

Second, the simulation of course IS "inaccurate" - it could not possibly be accurate - and that is NOT a problem! For you never define the level of accuracy you require. What makes you think it is relevant for the model to show the walls behaving to some unspecified degree of accuracy like in the real world? Wall collapse was the last stage of many in the sequence, and obviously, the building was doomed to total collapse when wall collapse commenced.

Third, no, it does not follow from 1) that it "proves CD" - that's several jumps of logic and waaaaaaaay too much trust in your IMAGINED assumptions

Fourth, no "entire building" in reality went into freefall on 9/11 - how many hundreds of times must this be pointed out before you stop repeating this FALSE allegeation??

Fifth, nothing about the collapse can objectively be termed "symmetrical" while making any engineering sense at all. - how many hundreds of times must this be pointed out before you stop repeating this NONSENSE allegeation??

Sixth, there was no free fall! Observing that some point on a falling, rotating, twisting object averages freefall acceleration does NOT mean it IS in "free fall" - how many hundreds of times must this be pointed out before you stop repeating this DECEPTIVE allegeation??

Seventh: The particular area of north wall roofline that DID exhibit a couple of seconds of, on average, did not go into that sort of acceleration "suddenly" - The failure of almost all of the building preceded that time interval, which occurred very late in the collapse sequence - how many hundreds of times must this be pointed out before you stop repeating this OBVIOUS falsehood??

Eighth, the "preceding internal collapse" is not a "fig leaf" - what gross dishonesty!! It is a matter of direct observation: The penhtouses collapsed and disappeared, East to West, before the walls went down. Have you ever opened your eyes and payed attention while looking at any of the collapse videos? Ever???

Ninth, you discuss your (false) choice as if the two options have about equal a priori likelihood of being true - but there is a very powerful arbiter of the likelihood of things happening in reality: Reality! It is a matter of observed reality that the interior started to collapse before the walls did, and it is a matter of reality that the walls behaved (flexed or didn't, deformed or didn't) the way they did - because those are events observed and recorded in the past. The likelihood of an event to happen that already happened in the past is very precisely 100%. In contrast, the likelihood of a simulation of this complexity to be accurate to arbitrary degrees of accuracy picked by people with a years long interest in dissing the team that did the simulation, on the other hand, is very very close to 0%. This is not a problem of the simulation, it's a problem of the dishonest critics misrepresenting what a simulation can and should accomplish.

---

Here is what the simulation was all about.
The question was: Could the predicted local failures that arose from the preceeding steps of simulation (fire spread over time -> structural heating over time -> structural response/failures over time) initiate the collapse sequence OBSERVED in REALITY.
That collapse sequence was:
Kinking and fall of the east penthouse -> several second later rapid East-to-West drop of the west penthouse -> fractions of a second later, all walls start descending

NIST found that YES, the failure of that c79-c44 girder on floor 13, in conjunction with all the other structural damage, would indeed result in a collapse sequence that I just described.

That not EVERY detail of the simulation animation is a perfect match for the observed collapse is as incidental as it is expected and unavoidable.

The MAIN features arose in the simulation as a result of a physics simulation that started with a plausible set of conditions (building as built with "normal" fuel load, plus some structural breaches from WTC1 collapse, plus fires started in areas where fires were observed early on), and then was carried on in an uninterrupted chain of physics calculations to result in the dynamic total collapse. So, NIST actually EXPLAINED the collapse sequence as the physical result of realistic initial conditions.

Contrast that with the Truthers' champion, Mr. Hulsey, who skipped all the preliminary initial conditions and modelling steps and went straight to conjuring away, as if by magic, lots and lots of major structure at times where he just needed them to disappear.
And then STILL, almost nothing of his simulation resembled reality - some of it was even physically impossible.
Yes, I know, different topic for a different thread. I just wonder, Henkka: Did you tear apart Hulsey's simulation and conclusions with the same Call to Perfection you go after NIST?

 

[Abdullah]

The lack of inward bowing may be a clue that the floor connections were not up to spec. Scandalous.
Will AE911Truth investigate this? Probably not.

 

[Henkka]

Second, the simulation of course IS "inaccurate" - it could not possibly be accurate - and that is NOT a problem! For you never define the level of accuracy you require. What makes you think it is relevant for the model to show the walls behaving to some unspecified degree of accuracy like in the real world? Wall collapse was the last stage of many in the sequence, and obviously, the building was doomed to total collapse when wall collapse commenced.

Your post was very long, I'm just gonna respond to this so we're not talking about a dozen different things at once… I'd just point out that your objection here also applies to NIST, that the degree of accuracy required is arbitrary and undefined. Someone at NIST, presumably Sunder, looked at the simulation as it was and determined that although it's not an exact match, it's "close enough". But how close is "close enough"? It's arbitrary. Another director might've determined that no, it's not yet close enough, and they need to keep working on it until they figure out why the walls didn't behave like that in reality.

But besides the walls folding, of course the real problem is that the simulation cuts halfway. Anyone presuming that the simulated building continued to fall smoothly is doing so exclusively on faith. I'm a skeptic, so I don't take such things on faith. A skeptic worth their salt would demand NIST render out the entire collapse and release it.

 

[benthamitemetric]

But how close is "close enough"?

Oystein ended his post with 5 paragraphs addressing this exact question. It is odd for you to choose this one point out of all of the points Oystein raised and then either not recognize or ignore that Oystein addressed your question about it.

 

[Mendel]

. A skeptic worth their salt would demand NIST render out the entire collapse and release it.

the reason I "ignore" you on this forum is that you ignore most things people tell you

so here's a repeat, with some new info re: chaos added, because it's on topic:

the finite element analysis software that NIST used 20 years ago can not model a collapse. It's data reduces the building to sticks and joints, and it can fairly accurately model how the forces are distributed in such a stick structure when a load is applied, and where it is getting overloaded and may break.

However, this software is not able to model solid objects falling and colliding with other objects, because that cannot be easily expressed as sticks and connections.

This is not a shortcoming of the software. This kind of software is used to analyse structures which are expected not to fail, and to verify that they won't fail under the loads they are rated for.

Any actual collapse simulation, if it existed, would have to contend with the fact that manufacturing tolerances exist, i.e. the exact strength of each connection as built is not known to an arbitrary precision (and neither is the load in each office on each floor):

The collapse is likely a chaotic process, in the "chaos theory" sense, in that certain systems depend highly on initial conditions. Imagine a strong steel bar, horizontal, suspended at each end by a weak connection. If you overload it, one of the connections will break, but which one it is depends on manufacturing tolerances; but this small difference determines which end of the steel bar goes down first, which may change how the collapse sequence unfolds from then on. You may believe that a mechanical system that's essentially deterministic always has a predictable outcome, but chaos theory proves mathematically that this isn't true.

Source: https://m.youtube.com/watch?v=6nhzrq4ALMc

[1:00] each time you swing the pendulum, the motion you get is basically unique.

That is why no engineer "worth their salt" demands that "NIST render out the entire collapse": they know it'd differ from the actual collapse, and that this difference wouldn't mean anything: it would merely demonstrate the shortcomings of the software when used for a purpose it wasn't designed for, and the fact that the collapse depends on initial conditions that are not accurately known.



The actual purpose of the simulation was to make the case that fire could cause total collapse in a steel-framed high-rise, an occurence you don't tire to point out was unprecedented. The argument is similar to the one made after the Tacoma Narrows bridge collapse: it said, hey, when you run the numbers and consider the fire (the wind-induced oscillations in case of the bridge), it turns out this thing will come down; this isn't as safe as we believed it was. For that purpose, the simulation only needs to run to the point where total collapse appears inevitable to any engineer who looks at it.

Your demand, that NIST produce a video simulation that mimics the actual collapse, is not something NIST set out to do, nor is it necessary for engineers in order to to understand how the collapse occurred and how to design safer buildings.

 

[Oystein]

Your post was very long, I'm just gonna respond to this so we're not talking about a dozen different things at once… I'd just point out that your objection here also applies to NIST, that the degree of accuracy required is arbitrary and undefined. Someone at NIST, presumably Sunder, looked at the simulation as it was and determined that although it's not an exact match, it's "close enough". But how close is "close enough"? It's arbitrary. Another director might've determined that no, it's not yet close enough, and they need to keep working on it until they figure out why the walls didn't behave like that in reality.

While is is obviously true that somewhere, someone has to judge whether the simulation is "close enough", I think your IMAGINATION of how it is done - the director eyeballing the outer appearance of the simulation video and either giving thumbs up or down - is informed by nothing (...)

But besides the walls folding, of course the real problem is that the simulation cuts halfway. Anyone presuming that the simulated building continued to fall smoothly is doing so exclusively on faith. I'm a skeptic, so I don't take such things on faith. A skeptic worth their salt would demand NIST render out the entire collapse and release it.

No, that is NOT "the real problem" at all. Not even close! I would propose a contrary judgement: That a (minor) problem is that NIST ran the simulation so far.

When you have hundreds of thousands of nodes in a simulation interacting in millions of ways with neighboring nodes, then, once a little bit of inaccuracy arises, that can very quickly be amplified to total chaos.
Consider a game of billiard: When you hit the first ball with the queue, it can be computed with great accuracy how that ball will move - speed, direction, spin perhaps, and how it interacts with the surface of the table. The first collision, with ball 2, likewise, can be simulated with great accuracy - the momentum transfer, the angles at which the two balls roll relative to initial angle, etc. The second collision, with ball 3, or third collision, with ball 4, start to introduce a tiny bit of inaccuracy, the numerical representation of values in a computer is not perfectly smooth but discreet. It can be shown that what happens to ball 8 ends up being entirely unpredictable, as even the tiniest of influences like the gravity of players (how heavy are they? Do they move after playing the first ball?) introduce enough inaccuracy as to make the simulation impossible.

This does NOT mean that simulating dynamic interactions of many parts cannot and should not be done, period. It means you need to be mindful of the perils, and understand what your particular simulation can and cannot recreate. This requires experience and intelligence.
I trust, with excellent reason, that the NIST team (not just Sunder) possessed FAR more experience and intelligence than you. Or any Truther who ever looked at the issue. Case in point again, just look at the total lack of criticism of the OBVIOUSLY flawed Hulsey "simulations" (where structural failures arose not as a result of fires, but as a result of Hulsey manually removing them willy-nilly) from the entire "Truth" movement. You would think that there are hundreds of engineers on the AE911T petition who understand that NIST's work, even if flawed, was a million times better than Hulsey's.

Which is why I asked you - and notice you chose not to answer:

Did you, Henkka, tear apart Hulsey's simulation and conclusions with the same Call to Perfection you go after NIST with?

 

[Henkka]

Did you, Henkka, tear apart Hulsey's simulation and conclusions with the same Call to Perfection you go after NIST with?

I didn't bring up Hulsey, and almost never do, since that report has very little bearing on me believing what I do. So I don't really know why you keep bringing it up.

You're implying I want the simulation to be an absolutely perfect match, and nothing less will do. That's not what I think, so let me clarify what I would want. I'd like to see a simulation that:

1) Runs from the start of the collapse to the end, not cutting in the middle. The fall should be relatively smooth, like it was in real life.

2) Can be shown to achieve some amount of g acceleration of the roofline. It doesn't have to be the exact 2.25 seconds NIST identified as their "Stage 2", I'm fine with less than that.

3) Has all the input data files released so universities and other knowledgeable people can run the simulation themselves and see how it was built. I don't buy the argument that this would jeopardize public safety at all.

I don't think these are too unreasonable.

 

[Mendel]

I'd like to see a simulation that:
[...]
I don't think these are too unreasonable.

Who would pay for it? Who else would want to see it?

Why doesn't Kostack's simulation do what you want?

Why are you derailing the topic?