Kostack Studios WTC2 (South Tower) Collapse Simulation

[Thomas B]

I would be very interested to read a detailed account of what is happening from 2:15 to 2:20 in the video. This seems to be the collapse initiation, where the building goes from a stable state to one where it sure seems that

global collapse was inevitable".

When I say "a detailed account" I mean one that explains what is happening to each of the elements and their connections (in the model) and what forces they are responding to (not what was happening to the actual WTC).

 

[Oystein]

I would be very interested to read a detailed account of what is happening from 2:15 to 2:20 in the video. This seems to be the collapse initiation, where the building goes from a stable state to one where it sure seems that

When I say "a detailed account" I mean one that explains what is happening to each of the elements and their connections (in the model) and what forces they are responding to (not what was happening to the actual WTC).

As I explained earlier in this thread, Kostack makes no apparent effort to figure out the actual initiation sequence, as he does not model any fires nor any heating of anything - not dynamically, not statically. Instead of thermal inputs he implements "a spatially decaying gradient that reduces element strength and stiffness with distance from the centre of impact", gradually, until something fails. We get no information what temperature(s) these reduced strengths and stiffnesses might correspond to, nor by what mode of failure initiation came about.

We can be quite certain however that this simulation initial failure is NOT the same location nor mode nor temperature (or rather steel strength and stiffness) as was the case in reality (arguments have been presented that "centre of impact" is unlikely to be the same as "center of fuel loads"), nor is there any good reason to presume that the vicinity of the initial failure, the elements "next in line" to fail, are at the same strength and stiffness sim vs reality, and so might react differently to the changed load that the initiating failure imposes on them.

As @econ41 has never tired to remind us, the broad mechanisms of collapse progression can be understood without making any assumptions about the ultimate cause and mechanism of initiation. So, if "Truther" has trouble understanding both initiation and progression (and, if that concept is helpful, transition between the two), you can explain (unassisted) progression while allowing, for the sake of discussion, that a few or a lot of explosive charges, or nano-thermite bundles, or platoons of midgets with hacsaws, or alien space rays were the cause.
The same way, we can learn from Kostack's visualization of collapse progress without looking at initiation at all [1].

So, knowing that Kostack's initiation sequence is almost certainly different from that of the actual WTC, why is an in-depth narrative of it of interest?
He might be missing / not be able to model some relevant mechanism that played a role in reality. (For example, he seems to imply that he did NOT consider thermal expansion / contraction. [EDIT to add] Also, I am sure he does not calculate creep, which may occur to the cold structure where plane impact+load redistribution brings elemens too close to yield stress, and which increases with temperature; because Kostack has no temperatures in his model, I don't see how he could account realistically for temperature-induced creep. Creep is not a simple function of strength and/or stiffness, it's also time-dependent, adding another dimension that Kostack probably does not model with values equal to that in the actual WTC. [/EDIT])
His simulation might result in mechanisms relevant to the simulated initiation that however did not happen in reality or are even impossible [2]. (For example, do strength and stiffness degrade proportionally to one another or not in his model, do they in reality? If his arbitrary gradients result in proportions that differ from steel's behavior in reality, then looking at the initiation event is unlikely to teach us much of worth.)

Instead of implementing "a spatially decaying gradient that reduces element strength and stiffness with distance from the centre of impact" to artifically creep to an initiation event, he could have arbitrarily have removed or cut one element after the other around that "centre of impact", as if caused by explosives / nanothermate sulphur attack / midgets / space rays, until collapse ensues, and very probably, a total collapse with many or all of the features this one has would have resulted. Would it then be of interest to be told in detail just HOW a nano-thermite pack, or a midget with a hacksaw, made an element fail?

---------------
Footnotes:
[1] That is a lot like Hulsey's model of WTC7. Except Hulsey does it even more simplistically: He just conjures away all columns, all over 8 floors - just makes them magically disappear, whithout proposing any mechanism. They are just plain gone, removed from the universe, in an instant. That does not seem to have bothered any "Truther" at all: They know they are supposed to insert "explosives!" where Hulsey deliberately remains a no-claimer.
[2] Like some of the deformations in Hulsey's simulations are entirely impossible in the real physics of our current universe: He has beams passing through each other without interaction, or the entire building falling through the solid ground without deformation.

 

[Oystein]

I am a bit disppointed that no one has commented yet on the FACT that Kostack's simulation exhibits some serious degree of "Crush-Down before Crush-Up" behavior"

Look: This is moments after the initiation event - a couple of non-consecutive floors detaching from the core and falling on one another, creating an initial stack of 6 pancaked floors:

About three seconds later, we see that 6 floors from the lower block have been acreted to the first 6, as the first, lowest floor of the upper block just impacts a core column and gets thus dislocated.:

It must be noted here that the crush-down is not primarily floor-stack-on floor, but driven by the upper perimeter passing inside the lower perimeter and thus massively loading the stack of pancaked floors.

Another about three seconds later, only 5 upper-block floors have been crushed, primarily through the action on the core. We cannot see how many more floors are accreted to the floor stack / pile-driver, but it's a lot more than 5.

Another three seconds later: While the entire collapse is accelerating and crushing down more and more floors per second, the crush-up has eaten only another ca. 5 more floors - and the top, with 14 (?) floor still in place and in good shape, is leaning 45° and predominantly outside the original footprint. The core no longer pokes into the upper block:

Result: An almost comlete upper 10% of the tower falls to the ground intact, as ca. 80% of all floors have been crushed down into one huge stack:

Now, I realize that this is quite different from Bazant's models, where the top stays on top and rides the crush zone down almost undamaged. I realize that Bazant was thinking about crushing columns, not pancaking and certainly not about columns loading floors by poking them from above or below.

However, I think this simulation can teach us how some degree of "crush-down before-crush up" is almost inevitable:
Because the growing crush zone below the top is getting accelerated in the same direction as the top, while the bottom stands still. This way, the speed-difference between acretion layer and bottom is soon much larger than that between top and acretion layer. Consequently, far more floors are affected per second on the crush-down than on the crush-up, and they are impacted with far greater force.

 

[LennyHipp]

wow. this is amazing to watch. but i saw the black spots on each tower, presumably damage from landing gear and engines. (obviously) But what i'd like to know, are there any photos of either tower with the punch out damage (exit holes) caused by landing gear on both buildings. Also, are there any photos of damage on the north tower, after the collapse of WTC2? thanks

 

[Mendel]

wow. this is amazing to watch. but i saw the black spots on each tower, presumably damage from landing gear and engines. (obviously) But what i'd like to know, are there any photos of either tower with the punch out damage (exit holes) caused by landing gear on both buildings. Also, are there any photos of damage on the north tower, after the collapse of WTC2? thanks

The landing gear was retracted.
Also, "punch-out" doesn't really describe it, because a) the girders were bent, but the simulation doesn't care because the bent ends weren't connected to anything and thus had no structural function, and b) the facade consisted of staggered panels, some of which dropped out as a whole when their connections failed.

 

[Thomas B]

The landing gear was retracted.
Also, "punch-out" doesn't really describe it,

I think @LennyHipp meant the smaller exit holes on the other sides of the buildings. If I remember correctly, these were indeed caused by the relatively heavy, solid landing gear and engine assemblies in the planes, which were otherwise lightweight and mostly air.

 

[Mendel]

I think @LennyHipp meant the smaller exit holes on the other sides of the buildings. If I remember correctly, these were indeed caused by the relatively heavy, solid landing gear and engine assemblies in the planes, which were otherwise lightweight and mostly air.

The aircraft (767-200ER) weighs 80t empty and can load 70t of fuel.
"Mostly fuel" describes it better than "mostly air".

My source writes:

So these exit holes exist, and there's probably some study reconstructing where they would be, but I don't know that offhand.

 

[econ41]

As I explained earlier in this thread, Kostack makes no apparent effort to figure out the actual initiation sequence,

The first concern that I identified.

As @econ41 has never tired to remind us***, the broad mechanisms of collapse progression can be understood without making any assumptions about the ultimate cause and mechanism of initiation. So, if "Truther" has trouble understanding both initiation and progression (and, if that concept is helpful, transition between the two),

*** Hence my limited input in this discussion. It is drifting back into issues we have discussed at length on previous occasions. Arguably without clear resolution of "What is the purpose of this model".

Failing to distinguish the two main stages has been a persistent failing on arguments from both "sides" of the debate. And IMNSHO (bare assertion at this stage) not distinguishing the stages, or worse conflating elements from one stage onto the other, nullifies any attempt at argument. Also, in my preferred explanation, posted on previous occasions, a comprehensive explanation requires four stages. The Kostack model conflates/overlaps three of them - he may be correct. I have never claimed clear separation between the stages.

Back to the topic - the Kostack moddel .

To my eye, preliminary inspection only, it looks like the model is a graphics animation representation of what Kostack imagines were the details.
Hence my persistence with another of my long held themes. "Is the model a graphic animation OR a valid simulation of engineering applied physics?" And I suspect Kostack's goal plus the interpretation of one or two members is that some aspects of applied physics have been included. (References by Mendel and Thomas B if my memory is correct.) My position Unless the simulation is rigorously based in engineering physics it remains a visual graphics animation. If the foundation is assumptions as to what may have happened - represented by graphics animation - THEN adding bits of physics onto the graphic foundation does not impart rigor of structural analysis.

So, knowing that Kostack's initiation sequence is almost certainly different from that of the actual WTC, why is an in-depth narrative of it of interest?

Exactly

If his arbitrary gradients result in proportions that differ from steel's behavior in reality, then looking at the initiation event is unlikely to teach us much of worth.)

 

[econ41]

I am a bit disppointed that no one has commented yet on the FACT that Kostack's simulation exhibits some serious degree of "Crush-Down before Crush-Up" behavior"

I've been biting my tongue. I've never won that argument.

It must be noted here that the crush-down is not primarily floor-stack-on floor, but driven by the upper perimeter passing inside the lower perimeter and thus massively loading the stack of pancaked floors.

The main supporting assertion of my preferred explanation - supported by the graphic I borrowed from Achimspok. No need to repeat it here.

Another three seconds later: While the entire collapse is accelerating and crushing down more and more floors per second, the crush-up has eaten only another ca. 5 more floors - and the top, with 14 (?) floor still in place and in good shape, is leaning 45° and predominantly outside the original footprint.

We need Major_Tom's visual researching skills.

Result: An almost comlete upper 10% of the tower falls to the ground intact, as ca. 80% of all floors have been crushed down into one huge stack:

I place on notice - I will need persuading. I have serious doubts - another one for Major_Tom?

Now, I realize that this is quite different from Bazant's models, where the top stays on top and rides the crush zone down almost undamaged. I realize that Bazant was thinking about crushing columns, not pancaking and certainly not about columns loading floors by poking them from above or below.

However, I think this simulation can teach us how some degree of "crush-down before-crush up" is almost inevitable:

Plausible - a point for possible further consideration. However, remember that Bazant & Verdure's CD/CU was based on quantifying the energies required for column crushing. (From vague memories, the energies favoured crush down because acting in the direction is assisted by gravity. Hence, crush down occurred before crush up. It was pure mathematical theory. No attempt to explain a mechanism or consider visual evidence.)

 

[econ41]

I would be very interested to read a detailed account of what is happening from 2:15 to 2:20 in the video. This seems to be the collapse initiation, where the building goes from a stable state to one where it sure seems that

When I say "a detailed account" I mean one that explains what is happening to each of the elements and their connections (in the model) and what forces they are responding to (not what was happening to the actual WTC).

I think we (You, Oystein, and I) are querying the same issue. The Kostack model does not seem to explain "initiation". Oystein and I both mean the actual collapse initiation and whether or not the model represents it. You seem to be questioning "...what is happening from 2:15 to 2:20 in the video." Different topic.

 

[Scaramanga]

This is a very detailed physical simulation that shows (without the smoke and dust) the various aspects of the collapse that have puzzled people like AE911Truth. It's shown from various angles, with some cutaways.

I'm curious in that the video section at 2:18 shows the left hand side of the top floor section rotate through 90 degrees and then decelerate as it crashes into another building, with the top floor eventually ending up upside down near that. So at least for this section it is not in free fall all the way but is slowed down. Whether it is enough deceleration that anyone might have initially survived is another matter.


Source: https://youtu.be/WpHXj62Ylw0?t=138

 

[Mendel]

To my eye, preliminary inspection only, it looks like the model is a graphics animation representation of what Kostack imagines were the details.
Hence my persistence with another of my long held themes. "Is the model a graphic animation OR a valid simulation of engineering applied physics?" And I suspect Kostack's goal plus the interpretation of one or two members is that some aspects of applied physics have been included. (References by Mendel and Thomas B if my memory is correct.) My position Unless the simulation is rigorously based in engineering physics it remains a visual graphics animation. If the foundation is assumptions as to what may have happened - represented by graphics animation - THEN adding bits of physics onto the graphic foundation does not impart rigor of structural analysis.

There is no imagination involved.
Bullet Constraints Builder implements DEM (discrete elements method), as does e.g. ANSYS. The physics model starts with the WTC with impact damaged elements removed, then initiates collapse in a bit of a slapdash fashion by progressively weakening structures around the impact zone until the collapse commences. As Oystein has remarked earlier, initiation could just as well be by cutting charges or whatever: the model is really focused on collapse progression.

This tallies with the chosen method and Kostack's research interest: he had proposed that this method may be able to predict cavities in the rubble where survivors might be located. For this end, he needs a method that can model the collapse sequence accurately, i.e. track the interactions of loose parts. This is in contrast to Finite Element Methods that examine the forces in connected structures.

The upshot is that, for this animation, Kostack creates the collapse initiation conditions ex machina, but the collapse itself is determined wholly by the DEM physics model of Bullet Constraints Builder. Blender is merely used to visualize the behaviour of that model (camera placement, transparency of walls, etc).

 

[Scaramanga]

I think we (You, Oystein, and I) are querying the same issue. The Kostack model does not seem to explain "initiation". Oystein and I both mean the actual collapse initiation and whether or not the model represents it.

I doubt if there is a recoverable sequential initiation traceable to one beam or another. Whilst one can argue from numerous videos that external beams buckled or bent at specific times...there's no video of inside for which one can tie down what happened internally right down to the order in which things buckled or broke. I take the Kostack model as being specific where it can be, and extrapolating as closely as possible where it can't.

 

[econ41]

I doubt if there is a recoverable sequential initiation traceable to one beam or another.

No problem. Neither Oystein nor I are after low level detail. I, possibly more than Oystein, am looking at the macro level understanding.

We have had similar discussions on several previous occasions here on Metabunk and before that on two other high level 9/11 forums which have been the main focus of WTC collapse discussion. This Kostack model is new. The debate goes back to about 2010 on the issues of the stages of collapse.

What is probably closest to the prevailing wisdom explaining Twin Towers collapse initiation is that the collapses - both of them - were dominated by cascading sequenced failure of columns in axial overloading. The trigger for the cascade in the case of WTC2 was near certainly floor joist sagging, causing (starting) inward bowing of perimeter columns. There is strong evidence for both those assertions in the visual record.

The Kostack model shows the sagging and disconnecting floors - i.e. the inward pulling catenary force which started the column bowing but does not seem to include the column failure aspect.

The concern is that Kostack only shows one aspect of the initiation, as several members have commented.

Whilst one can argue from numerous videos that external beams buckled or bent at specific times...there's no video of inside for which one can tie down what happened internally right down to the order in which things buckled or broke. I take the Kostack model as being specific where it can be, and extrapolating as closely as possible where it can't.

Understood.

 

[Thomas B]

I think we (You, Oystein, and I) are querying the same issue. The Kostack model does not seem to explain "initiation". Oystein and I both mean the actual collapse initiation and whether or not the model represents it. You seem to be questioning "...what is happening from 2:15 to 2:20 in the video." Different topic.

Yes, for me, the key is to explain how any structure with roughly WTC-like proportions (a 1:6 aspect ratio and a 4:1 ratio of dead load to weight of the structural elements) can crush itself using the top fifth (roughly 20%) of the structure. Once that's done, it's just a matter of fine tuning the details to approximate the actual WTC. At that point, I think the truthers' main challenge has been met.

 

[econ41]

Yes, for me, the key is to explain how any structure with roughly WTC-like proportions (a 1:6 aspect ratio and a 4:1 ratio of dead load to weight of the structural elements) can crush itself using the top fifth (roughly 20%) of the structure.

Sorry but you are changing topic and reverting to debate of a topic already answered many times.

That explanation had been presented many times. It is the same macro explanation Oystein and I are accepting as the basis of queries about "initiation". i.e. the presumption that there are at least 2 and probably 4 stages.

Stated briefly these are the four stages of collapse:

Spoiler: 4 Stage Hypothesis

(1) "Initiation" - driven by cascading failure of columns until top block starts to "drop" AKA move bodily downwards. That process necessarily ensures that the top parts of broken columns miss their bottom parts.

(2) "Transition" - a complex process but the only feature of significance to the macro level explanation is that tilting of Top Block vastly increases the "missing distance" between top part and bottom part of each broken column. i.e. falling top part must miss bottom part.

(3a) Early transition - the period where ROOST shearing of floor plates starts concurrently with mutual break up of Top block and upper levels of lower tower. (The Kostack model disagrees - showing the Top Block not involved in mutual break up.)
(3b) Established progression - ROOSD the dominating process.

Those four define what happened in macro. And shows why the falling 2-% destroyed the remainder. Independent of both the confusing term "crush" and recent ambiguous references to "crush down/crush up" which is used differently to the Bazant & Verdure false hypothesis.

Once that's done, it's just a matter of fine tuning the details to approximate the actual WTC.

It has been done many times. What more details are needed?

And the topic of this thread is the Kostack model including how it matches the actual event. NOT re-debating the known features of the actual event. Unless, somehow, Kostack presents a valid fresh insight. And I'm not seeing valid or relevant fresh insights for reasons already posted

At that point, I think the truthers' main challenge has been met.

You are too optimistic. Most currently still active truthers are dedicated conspiracy theorists. By definition, they are determined to never learn.
They will avoid any attempt at explanation no matter how carefully framed to be understandable.

The challenge is in behavioural psychology - not technical explanation.

 

[Mendel]

The Kostack model shows the sagging and disconnecting floors - i.e. the inward pulling catenary force which started the column bowing but does not seem to include the column failure aspect.

Yes.
I think this model includes the column damage from the aircraft impact (which is obviously already uncertain), both exterior and core column failures.

But it appears to me that, in this model, floor failures cause loss of rigidity which then causes buckling on the facade.

 

[econ41]

Yes.
I think this model includes the column damage from the aircraft impact (which is obviously already uncertain), both exterior and core column failures.

Yes.

But it appears to me that, in this model, floor failures cause loss of rigidity which then causes buckling on the facade.

Which is exactly the feature I explicitly defined in my preferred explanation. And one that was over several years subject of both sides of debate going down a false trail.

Explained briefly. (I hope. ) Sagging floor trusses started inward bowing. The false trailers argued that floor sagging could not account for the full observed extent of inward bowing. From memory 51 inches but I won't guarantee that. The missed point was that bending the column very rapidly reduces its axial load carrying strength. And beyond a small deflection critical point the bowing will continue under the applied vertical load. Sagging floor catenary pull-in no longer needed. And the whole of the online debate missed that point. On one occasion on JREF/ISF I intervened in a one on one argument between two engineers. One Truther "enik" and one debunker. Both were wrong. The debunker never agreed nor forgave me. Enik, one of my respected truther colleagues, took the debate to another forum where we came to agreement.

Excuse the meandering drift. But it is the déjà vu of all these debates going round in the same circle we have navigated multiple times.

Bottom line I have no doubt that floor sag was a trigger mechanism that would cause column inwards bowing and could in turn trigger the initiation stage for WTC2. Note my position (a) ZERO doubt it could be the trigger (b) highly persuaded that it WAS the trigger.

 

[Oystein]

Yes, for me, the key is to explain how any structure with roughly WTC-like proportions (a 1:6 aspect ratio and a 4:1 ratio of dead load to weight of the structural elements) can crush itself using the top fifth (roughly 20%) of the structure. ...

A) On "1:6 aspect ratio":
That is of no importance whatsoever. This structure would have collapsed completely regardless of whether there had been 10 floors below the level where collapse initiated (so total 40 floors - 1:2.2 aspect ratio), or 80 floors (as was the case) or 190 (so total 220 floors - 1:12 aspect ration)

B) On "4:1 ratio of dead load to weight of the structural elements"
I don't understand what that is supposed to mean. Can you explain? Maybe give an example of a specific structural element, its dead load and its weight?

C) On "crush itself using the top fifth (roughly 20%) of the structure"
First, where do you get that number from? 1/5th (20%) of the structure is 22 floors. Do you claim SOuth Tower collapse started around the 88th floor? Isn't it closer to a quarter of the structure?
Anyway: Again, there would have been total collapse no matter how many floors there are below the collapse initiation zone. That "20%" figure is irrelevant.
Suppose they had scaled down the project at the last minute for economic reasons to merely 20 floors per tower by essentially just cutting off the 90 floors above the mezzanine level and putting the top - what is it, 18? - floors as they were evetually built on top of the mezzanine level.
Do you suppose 20% of THAT (4 floors) could ave crushed all the rest (16 floors)?
But how about 28 floors, as was the case at the South Tower? Could 28 floors crush the remaining 12?
My point is: You are falling for one of the oldest "Truther" errors of thinking. The top 20% did not crush 80% in one fell swoop. Rather, as in any collapse, this happens one element by one element.
Have you already forgotten what I taught you (again! And now I do it again!!!!) about the vertical load bearing strength OF THE FLOOR SYSTEMS??!? And #18 in the NIST FAQ? You know, these FACTS that Truthers have been schooled about for oh so many years again and again and again and again? That each ONE (1) floor system was bound to fail because the dynamic weight of more than 6 floors impacted them?
The twin towers collapsed because each single floor in succession was grossly overloaded. A load that continued to increase with every new floor down the tube. While load bearing capacity of the floors systems stayed the same all the way to the ground.
Do you remember that?
You ought to have learned from this that the important determinant ratio is not the aspect ratio nor the dead load : element weight ratio (whatever that is) but the ratio between the dynamic load of the initial falling "block" exerted on the first floor system vs. its load bearing capacity.
As "Truthers" have been taught for many many years time and time again.
As I have personally taught YOU not even two days ago.
That important ratio is 1:6 between weight of one floor system and number of floor systems that it can bear when loaded dynamically.

More than 6 floors dropped onto the next below in reality as in Kostack's sim.
Ergo, that first impacted floor was doomed.
And because that floor would be added to the weight
and because that increased weight will impact the next floor with increased velocity
the 2nd floor below is also doomed.
and the 3rd, too
and the 4th
and the 5th
and the 80th
and the 800th
and the millionth.

The "20%" ratio of yours is entirely irrelevant.
What IS relevant is that 30 top floors started to move down and began loading structural elements not designed to resist anything even remotely in the vicinity of such a load.

It's not the ratio of too floors to bottom floors.
It's the number of top floors.
(And the specifics of the particular design. Which is why this number is of course not going to be the same for all tall buildings.)

 

[econ41]

That each ONE (1) floor system was bound to fail because the dynamic weight of more than 6 floors impacted them?
The twin towers collapsed because each single floor in succession was grossly overloaded. A load that continued to increase with every new floor down the tube. While load bearing capacity of the floors systems stayed the same all the way to the ground.

I have a vague recollection of a crude graphic illustrating that point.
.
.
Here, try this:

 

[Thomas B]

Bottom line I have no doubt that floor sag was a trigger mechanism that would cause column inwards bowing and could in turn trigger the initiation stage for WTC2. Note my position (a) ZERO doubt it could be the trigger (b) highly persuaded that it WAS the trigger.

This is something I might want to take into account when modelling. The floor-to-column connections need to be strong enough to pull the perimeter inwards to the point of collapse initiation. That is, there's an interesting ratio between the (lateral) strength of the perimeter structure and the strength of the floor assembly and its connection to the perimeter.

That's partly where they intuition that the buildings were "held together" by the total structure (including the floors) comes from. This strength was of course completely destroyed by the falling upper section impacting the floor assemblies (as @Oystein explains). It's just a matter of building something (at a manageable scale) that actually behaves this way.

 

[Thomas B]

It's not the ratio of too floors to bottom floors.
It's the number of top floors.
(And the specifics of the particular design. Which is why this number is of course not going to be the same for all tall buildings.)

I do actually understand this. I just want to build something that demonstrates it.

 

[Oystein]

The floor-to-column connections need to be strong enough to pull the perimeter inwards to the point of collapse initiation.

That depends on where in time, or with what event, you define the term "collapse initiation".
I am not sure you actually understand what @econ41 explained: That the sagging floor trusses did NOT need to pull so strongly on the perimeter that the perimeter would already fail just because of the pull. That the sagging trusses did NOT need to pull in the perimeter the full 50 inches, or what was it, that were observed. That instead, after some initial tug-in, the already bowed columns, an account of having lost some of their capacity due to this bowing, would continue to bow because of the weight they already beared to begin with.
They were able to bear that weight as long as they were straight.
And still able to bear that weight as long as they were pulled in only a leeeeeeettle.
But at some point - when pulled in by 1/4 inch? 1 inch? 4 inches? - their capacity has dropped to the point where they no longer can carry the weight they had carried for 40 years.
And then, that weight drops, and pushes down, and THAT will continue to bow the columns even further. No more tugging from the trusses needed.

Do you understand that ANY pulling inward with ANY force at all, no matter how small and trivial, results in a corresponding inward bowing of the perimeter columns?
One Newton of force pulling inward (1 N is e.g. the weight (gravity force) of 0.1 liters of water - a force that most small children or frail elders can easily exert) deforms the perimeter columns by X meters of inward bowing. This is because every solid material has a stress-strain ratio where, starting at the origin 0|0, each positive stress causes a positive strain. The stress-strain curve is not linear, but steadily increasing at least as long as the strain is elastic. X is of course very small for small forces such as 1 N.
Now, the vertical load bearing capacity of a column is premised on it being a) vertical and b) straight. These two properties known jointly as "being in plumb". A simple, symmetric column, has its maximum load bearing capacity when it is in plumb. You pull it out of plumb by any amount, and load bearing capacity decreases accordingly. Do you underdstand that?

So, pull on a column laterally with even a small force, and you will bend it, and as a result decrease its load bearing capacity.
Of course, a small force resulting in a tiny bow should decrease capacity by just a tiny bit and be of no concern in a practical building.

Now, have multiple trusses sag and each pull inward with a force far greater than children or seniors could.
Do this even on two or more consecutive floors.
What are the possibilities? Let me enumerate them completely:

1. Columns still do NOT visibly bow inwards, and DO retain sufficient capacity to hold the weight above
2. Columns DO visibly bow inwards, but still DO retain sufficient capacity to hold the weight above
3. Columns do NOT visibly bow inwards, but do NOT retain sufficient capacity (because inward boing, which is there but small enough to not be visible, is already so significant that column capacity drops below load)
4. Columns DO visibly bow inwards, and do NOT retain sufficient capacity

(By "visibly" I mean measurable in the photos and videos that we have, with their limited resolution, contrast etc.)

We do not see the floors sag.

We only see columns. We assume that floor trusses at some point start to sag and tug. Here is a sequence of observations on video when that happens:

1. At first (before truss sagging), the columns do not bow inwards
2. then they start to bow inward but we can't see it yet (truss sag has begun)
3. then, as they continue to bow inward, we start to see it
4. then they continue to bow inward even more
5. and finally we see them yield

You say "collapse initiation". In which of the 5 phases described above do YOU figure "collapse initiation" occurs, and what IS this "collapse initiation" event?

econ41 explained to you that at first, columns bow inward because of truss sag.
We can't see the inward bowing at first.
But the moment ANY, even invisible, inward boing occurs, the load bearing capacity of the column decreases
This causes a change in the stress:strain curve of the column - it gets steeper (the same stress causes a greater strain)
So, with the same vertical load on the column from above, because the strain in the column increases, that load drops in height. This drop in height is due to one or both of two factors: Column gets compressed vertically OR column bows more.
Initially, as the trusses start to tug inward ever so gently, I'd assume the column strain is wholly in axial compression.
But as the trusses tug more and more, and the columns bows more and more due to being tugged, and its strain due to vertical stress goes up and up, at some point the vertical stress will begin to cause some increased bowing in addition to the bowing caused by the sagging trusses.
This may happen before, just as, or after the inward bowing first becomes apparent on video.
And so it now continues that BOTH truss sag and vertical load cause the column to bow further.
And bow further...
And bow further...
Until "finally we see [it] yield" - my phase 5 above.

I would argue that the latest possible moment that could pass as THE collapse initiation event is when the vertical load on the already bowing columns starts straining the columns by bowing them inward further.
Again, this may happen before we even notice any inward bowing, as we first notice inward bowing, or at any moment during the time interval that we see inward bowing up until the moment before the columns yield.

From mere observation of videos, we can't pinpoint that exact moment of collapse initiation. We can't see how much of the bowing is caused by lateral tugging, and how much by vertical stress.

So where, exactly, do YOU place that moment of collapse initiation?

The point I am trying to make is that, if you put that moment as late as the moment where columns yield, or at which inward bowing is so pronounced we can measure its progress, then NO, the "floor-to-column connections" do NOT "need to be strong enough to pull the perimeter inwards to the point of collapse initiation". They only need to be strong enough to pull the perimeter inwards just enough for the perimeter's capacity to drop just enough so that the already existing vertical load would continue to strain them out of plumb.

As econ41 said: The trusses do not need to pull the perimeter 50 inches inward. Perhaps a pull-in of as little as half an inch will do.

That is, there's an interesting ratio between the (lateral) strength of the perimeter structure and the strength of the floor assembly and its connection to the perimeter.

That's partly where they intuition that the buildings were "held together" by the total structure (including the floors) comes from. This strength was of course completely destroyed by the falling upper section impacting the floor assemblies (as @Oystein explains). It's just a matter of building something (at a manageable scale) that actually behaves this way.

I do not disagree with this, but want to caution that the real structure of course reacts with more complexity to the changing stresses.

You say "the buildings were "held together" by the total structure (including the floors)". Totally correct of course.
This "total structure (including the floors)" is indeed what pushed the "strength of the perimeter structure" to what it was: The floor structure supported the perimeter structure laterally every 12 feet of height.
The perimeter structure had its own in-plane lateral support structure - the spandrels on each floor between any two columns. But it had no lateral support in the perpendicular direction - inward and outward. That was the structural purpose of the floor trusses. The cold steel trusses would prevent the perimeter both from bowing inward and from bowing outward.
If one floor truss was taken out, or even started to sag and pull inward (that would happen already if you merely cut the bottom chord of a floor truss - if you want to develop a realistic hypothesis of intentional demolition, that is the exact point where I would start!), then the trusses to its left and right, and above and below, would still be there to resist that local inward bowing, and I am certain that this is then enough redundancy to not cause any trouble.
Quite possibly, the structure would also survive if all trusses on one floor sagged.
But the reality of 9/11 was that there were fires raging on half a dozen and more floors at once, weakening several consecutive floor systems, making perhaps more than one sag and tug, and even those that did not sag would provide less support in the inwards-direction on account of being heated. And so that simultaneous failure and/or weaking of several adjacent floors left the perimeter laterally unsupported over not one, not two but multiple floors.

And I am sure you know about Euler buckling: Increasing the unsupported length of a column by a factor N decreases its capacity by a factor of N².
Have one floor sag instead of acting as lateral support increases unsupported column length by a factor of 2, and decreases column capacity by a factor of 4.
Have one more adjacent floor fail its lateral support role, even while not yet itself actively sagging, and unsupported length increases by another factor of 1.5, and capacity drops by another factor of 2.25.
This effect comes on top of the increasing strain:stress ratio due to pulling things out of plumb - and further complicates the task of pinpointing the "collapse initiation" event.

Perhaps a single floor sagging would have to pull in the perimeter by 2 inches to cause collapse, but 2 floors sagging would need to pull in by only half an inch, and three floors by mere millimeters. After that, increased strain caused by equal weight testing doubly decreased capacity becomes dominant.

-------------------

The simple, real world experiment you can do is still the empty beverage can that you can stand on if you are careful and not too heavy, and that also your child, weighing not even half of what you weigh, coulds stand on, but if you tap the side of the can ever so lightly, it will collapse and crumple up completely instantly.
(Caveat: I have crushed plenty of soda cans that way in my teenage years in the 1980s. It was a fun thing to do, required a bit of balancing skill and was thus something to show off on the schoolyard. However, I am pretty sure that today's soda cans, in a quest to save costs and material and amount of garbage, have much thinner, lighter walls than they used to in the 1980s. I have not tried this experiment in decades. As a matter of fact, I have not even drunk from a beverage can in years, because for environmental reasons, such cans are hardly in use anymore here in the EU. I personally haven't bought any since the 1990s I am sure. The only ones I ever get to see nowadays seem to be for energy drinks - yikes! )

 

[Oystein]

I do actually understand this. I just want to build something that demonstrates it.

Okay - have at it.

B) On "4:1 ratio of dead load to weight of the structural elements"
I don't understand what that is supposed to mean. Can you explain? Maybe give an example of a specific structural element, its dead load and its weight?

 

[Thomas B]

Okay - have at it.

B) On "4:1 ratio of dead load to weight of the structural elements"
I don't understand what that is supposed to mean. Can you explain? Maybe give an example of a specific structural element, its dead load and its weight?

Each WTC tower weighed about 500,000 tons, 100,000 of which was the structural steel. The rest was the load carried by that steel.

 

[Thomas B]

The simple, real world experiment you can do is still the empty beverage can that you can stand on if you are careful and not too heavy, and that also your child, weighing not even half of what you weigh, coulds stand on, but if you tap the side of the can ever so lightly, it will collapse and crumple up completely instantly.

Simplifying somewhat, what I want to build is a structure that can replicate this effect with only the weight of 4 cans on top (compressed into a single disk). That means designing the "can" to be a good deal weaker than an actual pop can. (Imagine cutting slits and holes in it.) It must first carry that load easily and then, when a "sagging" *within the structure* itself pulls the walls of the can inwards, it buckles under its own weight, i.e., the weight of rougly four times that of the structure itself. That initiating event requires connections and structures inside the can to provide the "tap", pulling in from the inside, rather than pushing from the outside. And that gives us some sense of how strong those connections have to be. First, to hold the "floor" in place (which wouldn't require a very strong connection) but then, when it sags, not simply to disconnect from the inside of the outer wall.

 

[Thomas B]

Do you understand that ANY pulling inward with ANY force at all, no matter how small and trivial, results in a corresponding inward bowing of the perimeter columns?

This is something that we've gotten into the weeds about before. But it really is my understanding that the "tube action" of the perimeter walls diverted most the windloads to the corners of the buildings and then into the perpendicular face, which now acted as a shear wall. That is, the walls were very resistant to lateral forces; they did not depend on transferring those forces, via the floor pans, to the cores (though I imagine some such transfer must have happened in order for the whole building, including the core, to sway in the wind.)

I'm guessing that for the floors to pull those walls inward, the connections would have had to be strong enough to handle a pulling force greater than a hurricane force wind pushing on the area of any given floor (about 12') of the facade. [Edit: yes, that really is a guess at this point. Feel free to improve my estimate with some math and knowledge.]

 

[Mendel]

But it really is my understanding that the "tube action" of the perimeter walls diverted most the windloads to the corners of the buildings

Yeah, you may want to revisit that. Elsewhere.

 

[Thomas B]

Yeah, you may want to revisit that. Elsewhere.

This still from Guy Nordenson's lecture "Patterns and Structure" (cued here at 14:56, the image appears at 15:59) gives a good sense of what puzzles the sincere truther. That lightweight floor structure, seriously weakened by fire, is supposed to have delivered a strong enough lateral force to those massive perimeter columns to pull them inwards enough to buckle, remembering also that, "in a tube structure, the idea is that the entire perimeter of the structure is mobilized, so the front and back faces of the buildings become like flanges of a box beam" that have been specially engineered to "get the stresses all the way around the corners and into the center [of what would be like the web]" (17:55-18:15).

I want to find out if there is something in the Kostack simulation that can be used to explain this and, like I say, guide me towards an instructive physical model.

 

[benthamitemetric]

This still from Guy Nordenson's lecture "Patterns and Structure" (cued here at 14:56, the image appears at 15:59) gives a good sense of what puzzles the sincere truther. That lightweight floor structure, seriously weakened by fire, is supposed to have delivered a strong enough lateral force to those massive perimeter columns to pull them inwards enough to buckle, remembering also that, "in a tube structure, the idea is that the entire perimeter of the structure is mobilized, so the front and back faces of the buildings become like flanges of a box beam" that have been specially engineered to "get the stresses all the way around the corners and into the center [of what would be like the web]" (17:55-18:15).

I want to find out if there is something in the Kostack simulation that can be used to explain this and, like I say, guide me towards an instructive physical model.

This was one of the many subjects discussed at length in NIST's NCSTAR 1 report (and the applicable appendices thereto), which was published 20 years ago. People who are sincerely interested in the topic have had two decades to read this report. There is nothing in the Kostack simulation that can or will explain it any better as the Kostack simulation did not simulate the events leading up to the collapse initiation; it merely assumed a collapse initiation and then generated a high-level approximation of a collapse sequence.

[...]

 

[econ41]

I do actually understand this.

Great!!

I just want to build something that demonstrates it.

Why?

To whom?

Because that is the fundamental question. Definition of objective.

And we have been round that circle many times. Changing the objective.

 

[Mendel]

This still from Guy Nordenson's lecture "Patterns and Structure" (cued here at 14:56, the image appears at 15:59) gives a good sense of what puzzles the sincere truther. That lightweight floor structure, seriously weakened by fire, is supposed to have delivered a strong enough lateral force to those massive perimeter columns to pull them inwards enough to buckle, remembering also that, "in a tube structure, the idea is that the entire perimeter of the structure is mobilized, so the front and back faces of the buildings become like flanges of a box beam" that have been specially engineered to "get the stresses all the way around the corners and into the center [of what would be like the web]" (17:55-18:15).

I want to find out if there is something in the Kostack simulation that can be used to explain this and, like I say, guide me towards an instructive physical model.

well...

Nordenson talks about vertical stresses that are in-plane with the facade.
The wind pressure per se is not a lot, the windows can resist it, so can thd steel, the structural problem is that the colkective force acting on the structure has a lot of leverage because the building is so tall, and the tilting moment causes the forces in the facade. But these are up and down and sideways in the facade, not inward or outward.

Because the facade is made up of a network of thin sticks, this only works as long as the sticks are straight: once they start to buckle, they can no longer transfer load.

The sincere truther has had this explained to him several times before, but failed to take this to heart. This greatly diminishes the value of his protestations of sincerity: what is the value of him wanting to learn if he does not learn?

You also completely ignore that being strong in one direction does not imply being strong in a different direction (what is the easiest way to topple a domino and why? and how can it support so much weight?).

A structure that is more complex than something built from kid's toys can be engineered to be only as strong as it needs to be, in any direction. If it's strong against lateral loads, that's independent of its strength against vertical loads.

So, pull on a column laterally with even a small force, and you will bend it, and as a result decrease its load bearing capacity.

The simple, real world experiment you can do is still the empty beverage can that you can stand on if you are careful and not too heavy, and that also your child, weighing not even half of what you weigh, coulds stand on, but if you tap the side of the can ever so lightly, it will collapse and crumple up completely instantly.
(Caveat: I have crushed plenty of soda cans that way in my teenage years in the 1980s. It was a fun thing to do, required a bit of balancing skill and was thus something to show off on the schoolyard. However, I am pretty sure that today's soda cans, in a quest to save costs and material and amount of garbage, have much thinner, lighter walls than they used to in the 1980s. I have not tried this experiment in decades. As a matter of fact, I have not even drunk from a beverage can in years, because for environmental reasons, such cans are hardly in use anymore here in the EU. I personally haven't bought any since the 1990s I am sure. The only ones I ever get to see nowadays seem to be for energy drinks - yikes! )

Have you re-read these quotes? Have you understood that the load factor of a stack of beverage cans far exceeds 4:1? and that the can is therefore a practical demonstration of a structure strong against vertical loads, where a small lateral force can compromise that strength?

The Kostack simulation, specifically the 5 seconds you highlighted, can be understood when these principles are applied.

Play a bridge building game if you want to learn about load paths.

 

[Thomas B]

People who are sincerely interested in the topic have had two decades to read this report.

As I've said elsewhere on this forum, science communicators have also had two decades to popularize this mechanism (cf., e.g., The Titanic, Tacoma Narrows, Hyatt Regency, Challenger, Concorde) so that curious people could learn how the buildings collapsed without having to interact directly with specialists, conspiracy theorists, and debunkers.

As Tucker Carlson's recent doc reminds us, the same old questions remain open in the popular imagination.

This Kostack video could be very helpful with a little recutting and narration -- first explaining how the enormous strength of the building was modeled (including the strength of the floor connections) and then how that strength was overcome by the unique circumstances of the 9/11 attacks. I hope Kostack either does this himself or lets a competent science vlogger do it.

I think it would be worth the effort even from a strictly monetary point of view.

Watching the building collapse from various angles for ten minutes is one thing. Zooming in and explaining the various points of failure (i.e., the actual structural elements and how they failed) would be quite a lot more instructive.

 

[Mendel]

first explaining how the enormous strength of the building was modeled

strength. is. directional. (like a beverage can)

 

[Thomas B]

Because the facade is made up of a network of thin sticks, this only works as long as the sticks are straight: once they start to buckle, they can no longer transfer load.

This is a great image and exactly the one that I want to physically model. If the facades behave like "a network of thin sticks" then we should be able to build one, attach some floors to it, and then cause them to sag, excerting sufficient force on the facade to buckle it. So long as the facade and floors seem realistically stable before we introduce the sagging, we should have something convincing for a sincere inquirer to learn from.

And yes, I don't think merely "having it explained" works. I say that simply because I empathize with people who don't understand the explanations in the absence of a demonstration of (indeed, refusal to demonstrate) the mechanism in a physical model.

The sincere truther has had this explained to him several times before, but failed to take this to heart. This greatly diminishes the value of his protestations of sincerity: what is the value of him wanting to learn if he does not learn?

I think it is pointless to argue with someone whose sincerity you don't presume. (It's best just to stop arguing with them once you drop the presumption of good faith.) So I continue to imagine a sincere truther with valid concerns (perhaps in their late teens or early twenties with a lot of physics and engineering still to learn, lots of ignorance to overcome) as my audience. Essentially someone who is beginning the journey I started twenty years ago (and am still on).

The Kostack video may end up being a turning point for this debate if the model actually simulates the failure mechanisms, not merely illustrates the collapses (as @econ41 also emphasizes).

 

[Mendel]

As I've said elsewhere on this forum, science communicators have also had two decades to popularize this mechanism

you have had 5 years to experiment with a structural engineering simulation (aka a bridge building game)

yet you still demand people do what you want

 

[Mendel]

I think it is pointless to argue with someone whose sincerity you don't presume.

you misunderstood my point

 

[Thomas B]

In attempt to refocus discussion on the Kostack video, here are two questions that I think we're all interested in.

1. What does this new simulation/animation add to our understanding of the collapses?
2. How might it help us to better lead people away from or out of the rabbit hole of controlled demolition theories?

That is, what new arguments does it give us and how might these be effectively presented?

And here's a bonus question that apparently only I am in interested in (feel free to ignore it):

3. What new small-scale physical models does it suggest that would be more effective than the familiar pop cans and book shelves that have so far failed to correct the naive physical intuitions that truthers appeal to?

 

[econ41]

I agree that those are three valid questions. The first two are the two we should be asking about the simulation that is the topic of this thread:

In attempt to refocus discussion on the Kostack video, here are two questions that I think we're all interested in.

1. What does this new simulation/animation add to our understanding of the collapses?
2. How might it help us to better lead people away from or out of the rabbit hole of controlled demolition theories?

In brief (1) Nothing; and (2) nothing presented so far that is persuasive.

Spoiler: More Detailed Explantions

1. It adds nothing to "our" understanding. Where "our" is the set of members of Metabunk discussing the simulation. Its status is analogous to the physical model of the ROOSD progression stage that Mick West worked with. i.e. is shows "US" nothing we didn't know because we know enough to critique whether the simulation or model legitimately represents the real event. It does nothing new for "us". There is a contention from you, Thomas B, that it MAY be of some utility for informing truthers and other interested parties. As I have said many times, I remain unpersuaded on that purpose.

2. It adds one more "looks like" representation which may have supporting benefit in a discussion with truther or interested person. I remain convinced that there are simpler, and more effective ways of supporting explanations with visual aids. It is not a suitable tool for instructing in technical detail. And it does not clearly indicate the four stages of collapse, which are necessary for comprehensive overview.

That is, what new arguments does it give us and how might these be effectively presented?

None therefore moot.

And here's a bonus question that apparently only I am in interested in (feel free to ignore it):

Not just you. I'm sure I must have made the same observation in previous discussions.

3. What new small-scale physical models does it suggest that would be more effective than the familiar pop cans and book shelves that have so far failed to correct the naive physical intuitions that truthers appeal to?

You correctly identify that small-scale models could be a useful tool for explaining physics. Especially when the model is either identical or closely resembles the true detail of physics being explained. "Pop can" and "book shelf" analogies are more distant and present an additional "translation barrier" for many people.

 

[Oystein]

Each WTC tower weighed about 500,000 tons, 100,000 of which was the structural steel. The rest was the load carried by that steel.

a) I am pretty sure these numbers are quite off. As for the weight of the towers, always go with a figure of 300,000 tons. 90,000 of which were steel (including rebar; all numbers above ground level). That's a steel:nonsteel ratio of 90:210 = 1:2.33

b) I still don't get why that ratio is relevant.