WTC Collapse Simulation using Unity/Besiege

[Jeffrey Orling]

I am thinking... and I would like Econ to weigh in on this...
That the simulation of column failures was inside the core... the facade was the last to fail...
I suspect the load redistribution was to ALL undamaged core columns not one by one... and that the hat truss was the structure that did the load redistribution because it "linked all main load bearing core columns together structurally". Many core columns were really mostly forming elevator shafts and supporting some corridors in the core. The 24 perimeter core columns were carrying floor loads outside the core. It was these 24 were the ones which failed essentially at the same instant which left the facade carrying all the floor loads above the crash zone... which effectively doubled the load... as well as likely introduced some lateral translation. Down the tops came.

 

[econ41]

I am thinking... and I would like Econ to weigh in on this...

As you know I can explain the Twin Towers failure and collapse mechanisms in as much detail as is needed to support rigorous argument at the level we encounter on these online forums (or FBook Groups.)

But I'm conscious of this Forum's rules and the need to stay on topic. And the focus of this thread is software simulation. So IMNSHO the explanation of the real mechanisms only needs to go as far as is necessary to help @dylbie get his simulation "looking right". If that is even possible given the limitations of low-cost gaming style software.

So I will limit my comments until @dylbie has an opportunity to respond. Please note the advice I offered @dylbie:

Take care with any assumptions that core columns or perimeter columns somehow dominated the collapse and sort of acted as individual groups. Also take care with assumptions of "simultaneous collapse".

The driving process was a "cascading failure" which is essentially sequential and not simultaneous. Even tho the separate events may be very close and many of them overlapping in time. The whole process was very complex - essentially 5D. Geometrically 3D plus dynamic plus a different dynamic of heat movement. Hence effectively 5D. Hence also my recommendation is to start by understanding a simple model - this one:

IF you and @dylbie or any other interested members can understand how column "C" of THAT simple model is overloaded it should help to comprehend the several degrees more complicated real event mechanism.

 

[Jeffrey Orling]

Sequential which is extremely rapid is effectively simultaneous for practical purposes... as in the bridge failure video posted.

 

[econ41]

Sequential which is extremely rapid is effectively simultaneous for practical purposes... as in the bridge failure video posted.

Jeffrey I am very patient but I do know what I am talking about. And on this aspect of engineering I am correct. And your use of weasel wording "is effectively simultaneous for practical purposes" is misleading BS. If you want to disagree with me - don't rely on "bare assertions". Prove me wrong.

Here is the proof even tho I'm aware it is not your preferred way of reasoning:
1) We can know conclusively that the dominating process of the initiation stage was a sequenced failure of columns. The movement of both "Top Blocks" took place over many minutes. With the speed of motion accelerating as collapse became imminent.
2) That failure must have been a cascading sequence - one where load-redistribution causes columns to overload in sequence. As any column (or small group of columns) fails to carry its load the load must transfer to other columns.
3) The load cannot transfer until after it is dumped by the original column. AND until it transfers the next column doesn't have the load needed to cause failure. THEREFORE the process is sequential.
4) IF the mailing of adjacent columns was to be "effectively simultaneous for practical purposes" there would be no increased load to cause the failure of any column which is allegedly "failing simultaneous". The extra load goes onto the "second" column - the next column in the failure sequence - because it has already been "dumped" by the "first" - the preceding one in the sequence. No matter how small the interval there must be an interval.

DISCLAIMERS:
I have deliberately explained it in simple terms of 'first>>second' or 'preceding<>succeeding' for ease of understanding. In practice it was more likely small groups of columns dumping their load onto other groups of columns. NOT one column at a time. Also at the time scale of milli-seconds>fractions of seconds, the explanation is more complicated involving factors such as the timing of elastic compression or rebounding. BUT none of that detail changes the reality that the process is sequenced and NOT simultaneous.

.

 

[dylbie]

I'll respond to your post more directly later as I'm at work...

But a side note: have you heard the testimony of Tim Brown who was underneath the South Tower as it started to fall? He somehow miraculously (luckily) managed to survive by running into the Marriot Hotel. Anyway, he states he heard an almighty 'crack' first which echoed around the streets of Manhattan, then followed by the 'boom boom boom etc' of the top section hitting and crushing lower floors.

I know eye witness accounts are a dreadful source of evidence, but he's an experienced firefighter who has been in previous building collapses, and his testimony is probably some of the most detailed I've ever heard.

This loud 'crack' which he heard before the main collapse does suggest a sequential event. I'm guessing it was the failure of the floor to exterior column failure which he heard which then caused the runaway collapse.

Anyway, back to work...

 

[Jeffrey Orling]

Jeffrey I am very patient but I do know what I am talking about. And on this aspect of engineering I am correct. And your use of weasel wording "is effectively simultaneous for practical purposes" is misleading BS. If you want to disagree with me - don't rely on "bare assertions". Prove me wrong.

Here is the proof even tho I'm aware it is not your preferred way of reasoning:
1) We can know conclusively that the dominating process of the initiation stage was a sequenced failure of columns. The movement of both "Top Blocks" took place over many minutes. With the speed of motion accelerating as collapse became imminent.
2) That failure must have been a cascading sequence - one where load-redistribution causes columns to overload in sequence. As any column (or small group of columns) fails to carry its load the load must transfer to other columns.
3) The load cannot transfer until after it is dumped by the original column. AND until it transfers the next column doesn't have the load needed to cause failure. THEREFORE the process is sequential.
4) IF the mailing of adjacent columns was to be "effectively simultaneous for practical purposes" there would be no increased load to cause the failure of any column which is allegedly "failing simultaneous". The extra load goes onto the "second" column - the next column in the failure sequence - because it has already been "dumped" by the "first" - the preceding one in the sequence. No matter how small the interval there must be an interval.

DISCLAIMERS:
I have deliberately explained it in simple terms of 'first>>second' or 'preceding<>succeeding' for ease of understanding. In practice it was more likely small groups of columns dumping their load onto other groups of columns. NOT one column at a time. Also at the time scale of milli-seconds>fractions of seconds, the explanation is more complicated involving factors such as the timing of elastic compression or rebounding. BUT none of that detail changes the reality that the process is sequenced and NOT simultaneous.

.

Econ... in the case of a single column failure... the loads would seem to transfer to the adjacent columns. I get that. And then as those columns are "weakened" by heat they would do similar... transfer load to the columns adjacent to them. I get that.

My point is that... and I am assuming for the sake of argument... that the heat was not localized at a single column... but widespread and weakening multiple columns... I don't know if one of the columns being weakened failed first and its load added to the adjacent ones.... It seems possible that multiple columns of different capacity were might have failed "at the same time" and this led to a very rapid runaway failure progression which occurred in a blink of an eye and they all failed.

I am thinking this was the core not the facade... the so called core led collapse. A core without columns that had failed on a few levels would leave all the floor loads above those failures "hanging" from the hat truss... for an "instant" and all those loads would be transferred lickity split to the facade columns which likely failed together... the top block losing support but the cage still intact... and the outside the core floor slabs breaking free and falling inside the footprint... and that became the "ROOSD" mass.

What would a structural failure collapse which was not rapid look like?

 

[econ41]

I'll respond to your post more directly later as I'm at work..

Take your time.

But a side note: have you heard the testimony of Tim Brown who was underneath the South Tower as it started to fall? He somehow miraculously (luckily) managed to survive by running into the Marriot Hotel. Anyway, he states he heard an almighty 'crack' first which echoed around the streets of Manhattan, then followed by the 'boom boom boom etc' of the top section hitting and crushing lower floors.

Yes. I have heard (read) it. Two points:
First - like many claims it identifies a single issue or factor which may need explanation.
Second - the "boom boom boom etc" is "progression stage" - not within the scope of the initiation stage which is where the sequenced cascading failure of columns occurred.

I know eye witness accounts are a dreadful source of evidence, but he's an experienced firefighter who has been in previous building collapses, and his testimony is probably some of the most detailed I've ever heard.

Except he doesn't have a full context, coherently argued explanation for the 'almighty "crack"'. The usual truther presumption is false: "If you debunkers cannot explain it it proves it was CD."

This loud 'crack' which he heard before the main collapse does suggest a sequential event. I'm guessing it was the failure of the floor to exterior column failure which he heard which then caused the runaway collapse.

The potential source of the loudest "bang" at that stage was the impact as the Top Block overwhelmed the last remaining surviving columns. So all those survivors essentially did fail near enough simultaneously. But I would expect the noise to be "softer" - more of a "whoomp" than a "crack".

Anyway, back to work...

0930 Lima TZ here. I've finished my morning work shift .. having breakfast as I type.

 

[econ41]

What would a structural failure collapse which was not rapid look like?

Just like what actually happened with both WTC Twin Tower collapses on 9/11. The dominant aspect which took most of the time of the initiation stage was not rapid. It took about one hour. Less for WTC2 more for WTC1. The process we are discussing - sequential cascading failure of columns - started with the aircraft impact and fires. It ended with the "release" - the point when there was no longer sufficient remaining capacity to support the Top Block. And it "dropped" or "fell" although both of those terms can be ambiguous as we are both aware after many years of discussion. What actually happened at that time was that the surviving columns were caused to fail rapidly. As you actually describe it in this paragraph of your post:

It seems possible that multiple columns of different capacity were might have failed "at the same time" and this led to a very rapid runaway failure progression which occurred in a blink of an eye and they all failed.

Yes, The cascading failure of columns was a slower process that took place over about one hour. Very slowly at first. Getting faster as it continued. UNTIL the threshold point when there was no longer sufficient capacity to hold up the Top Block. THEN near simultaneous failure of all the still surviving columns. Very fast - fractions of a second.

Then:

My point is that... and I am assuming for the sake of argument... that the heat was not localized at a single column... but widespread and weakening multiple columns... I don't know if one of the columns being weakened failed first and its load added to the adjacent ones....

Yes. It was complicated. Hence my repeated advice to get to understand the base principle BEFORE trying to add on all the other complicating aspects.

 

[Jeffrey Orling]

I am wondering if it is possible that many (most) core columns (the ones not severed) were losing capacity in the same time frame.... but had capacity to carry the loads... That is to say perhaps there was no load redistribution... but and overall weaking/loss of capacity until capacity was below service loads and the tops dropped.

So the columns were not the same... the loads on them was not the same.. the weakening on them was not the same (heat from fires). I don't see why the failure could not have been compressed into a blink of an eye... not an hr long process where one column fails at a time with its load being transferred to another column?

Or maybe the process was a combination of the two mechanisms... "creeping failures" and "simultaneous) failures?

My sense is that creeping failures would should warping or partial collapse (maybe). At least in 1WTC the facade (most of it) seems to let go at once (aside from the bowed portion on the south east).

I suspect the "creeping failure" scenario would attend with a "moving fire" and the "instantaneous failure" would attend from a fire which was "attacking" all the core columns at the same time. Of course the fire was organic and so it was likely not "uniform" over the core footprint. But it's reasonable to say that the fire/heat was not acting on one column at a time either.

 

[dylbie]

We're any simulations done by NIST of the fires? And in particular, the affect of wind on the fires at 1000ft?

You can clearly see there must have been a reasonably strong North Westerly wind by the smoke pattern. This would have created a furnace type effect and would have caused the fires to be more severe in the South and East sides. In fact I think this is backed up by the images of 1WTC South facade just before collapse which appear deformed, and also by the people who were seeking refuge in the North Western corner which must have been the coolest place.

 

[Jeffrey Orling]

We're any simulations done by NIST of the fires? And in particular, the affect of wind on the fires at 1000ft?

You can clearly see there must have been a reasonably strong North Westerly wind by the smoke pattern. This would have created a furnace type effect and would have caused the fires to be more severe in the South and East sides. In fact I think this is backed up by the images of 1WTC South facade just before collapse which appear deformed, and also by the people who were seeking refuge in the North Western corner which must have been the coolest place.

Comparing the path of the planes through the towers it seems reasonable to assume that the fuel was distributed and spread in a pattern related to the path. So for 1WTC we might assume that the fuel was concentrated in the center of the building... through the core... and less on the East and West side. Of course being liquid it would flow over the level floor.
In 2WTC the fuel would be concentrated on the path going through the East side.
But.... is the fuel was quickly spread over the entire footprint... do we assume the entire footprint was burning? Would this not be visible through the facade or not? Where would the smoke be exiting? What would that tell about the location of the fires?

Then there is the question of what the fires weakened/impacted the most? Would be be the bracing beams in the core? The belt girder that supported the core side of the trusses? The bar trusses outside the core? The core columns?
How would the cross sectional area of the steel matter when under siege by fire? Would the large sections have the least impact from the fires... and the bar trusses the most?
If the fuel was spread over the entire footprint.... would not all the floor trussed be affected in a similar manner? And if so... would not all 4 facades show almost a simultaneous inward bending?

As 2WTC tipped to the SE is it a reasonable that that side lost most support and offered the least resistance to the falling top block? Does this mean that the NE had the stronger columns and the last to fail?

What would "column failure" from heat look line? Would the columns reduce in height and bulge / buckle? Height reduction would mean warping and distortion of the geometry of the top block. How much would be noticeable?

Any asymmetry of the top's drop would suggest asymmetry of the fires and the loss of capacity.

Can we understand the fires' behavior from the motion of the dropping tops?

 

[dylbie]

All good questions. I'd like to see a simulation that is capable of simulating your above points. Sadly I think we're still some way off that at the moment. I'm hoping that Kostack does some different scenarios with different column failures to see if he gets different results.

In the mean time I've been running some Version 2 simulations on Besiege from various different camera angles. I've tried to roughly replicate the impact damage, and also removed some of the interior columns where I think fire damage weakening would be most severe.

I have noticed a new issue which is that the floor to column connections are too strong. The floors appear to be able to hold the above sections on their own. So I'm gonna need to come up with a new design and somehow try and make them weaker. I'm not entirely sure how I'm going to do this yet. Another issue is it seems to be collapsing slightly too fast, but I have 10 floors less at the top, and 7 floors less at the bottom so the 17 missing floors could explain the speed.

There's two things I've noticed in the latest simulations which are interesting...

Firstly the tilt of 2WTC must have shifted the whole core inside the building as it was collapsing. You can see this in my video on the close up angle and it's easier to see the core as the sections are yellow. I'd imagine this would have been responsible for a tremendous amount of damage to the floors either side of the core as it shifted, but also might explain how some of the core was still standing, if what was once above it shifted laterally.

Secondly, I think this may be a flaw with the software, but the sudden impact of the top section on the lower section overloads the columns near ground level. I think this is not realistic, but I guess in real life the whole lower section was obscured by dust so it was difficult to see what exactly was going on down there.

Anyway here's the video...

 

[Jeffrey Orling]

All good questions. I'd like to see a simulation that is capable of simulating your above points. Sadly I think we're still some way off that at the moment. I'm hoping that Kostack does some different scenarios with different column failures to see if he gets different results.

In the mean time I've been running some Version 2 simulations on Besiege from various different camera angles. I've tried to roughly replicate the impact damage, and also removed some of the interior columns where I think fire damage weakening would be most severe.

I have noticed a new issue which is that the floor to column connections are too strong. The floors appear to be able to hold the above sections on their own. So I'm gonna need to come up with a new design and somehow try and make them weaker. I'm not entirely sure how I'm going to do this yet. Another issue is it seems to be collapsing slightly too fast, but I have 10 floors less at the top, and 7 floors less at the bottom so the 17 missing floors could explain the speed.

There's two things I've noticed in the latest simulations which are interesting...

Firstly the tilt of 2WTC must have shifted the whole core inside the building as it was collapsing. You can see this in my video on the close up angle and it's easier to see the core as the sections are yellow. I'd imagine this would have been responsible for a tremendous amount of damage to the floors either side of the core as it shifted, but also might explain how some of the core was still standing, if what was once above it shifted laterally.

Secondly, I think this may be a flaw with the software, but the sudden impact of the top section on the lower section overloads the columns near ground level. I think this is not realistic, but I guess in real life the whole lower section was obscured by dust so it was difficult to see what exactly was going on down there.

Anyway here's the video...
high staggered panels and peeled off as the inside of the building collapsed

this fails to properly depict the 3 story tall 3 column wide staggered facade panels... they peeled off and away as the floor collapse progressed downward

 

[econ41]

@Jeffrey. I'm sorry to see that, despite my several explanations of the points you raise, you once again engage in diverging speculations without responding to what I have already posted to address those points. So let me briefly explain once again and taking it paragraph by paragraph:

I am wondering if it is possible that many (most) core columns (the ones not severed) were losing capacity in the same time frame.... but had capacity to carry the loads... That is to say perhaps there was no load redistribution... but and overall weaking/loss of capacity until capacity was below service loads and the tops dropped.

Those comments simply restate the known facts of the starting point of discussion. EXCEPT the bit I struck out "That is to say perhaps there was no load redistribution." which is utter nonsense for reasons I clearly explained in previous posts. The main mechanism was "load redistribution" and there could not have been any collapse without load redistribution.

So the columns were not the same#... the loads on them was not the same##.. the weakening on them was not the same### (heat from fires*). I don't see why the failure could not have been compressed into a blink of an eye**... not an hr long process where one column fails at a time with its load being transferred to another column?***

# Correct - a known starting premise.
## Also correct - a known starting premise.
### Also correct - same reason.
* Not necessarily that one cause.
** Then why not join in discussion to improve your understanding rather than circling and repeating the same points?
*** The "hour long process" is well proven by the video and written record. Why do you need to deny it? And the process of sequenced failure is the only plausible mechanism for reasons given on numerous previous occasions.

Or maybe the process was a combination of the two mechanisms... "creeping failures" and "simultaneous) failures?

There is no "maybe" about it. The collapse involved both slow and fast processes as (a) has been explained many times and (b) you have agreed on previous occasions. Nothing has changed. Both occurred.

My sense is that creeping failures would should warping or partial collapse (maybe).$ At least in 1WTC the facade (most of it) seems to let go at once (aside from the bowed portion on the south east).$$

$ Unclear.
$$ It was NOT simultaneous for the same reasons posted multiple times including previous posts in this thread.

I suspect the "creeping failure" scenario would attend with a "moving fire" and the "instantaneous failure" would attend from a fire which was "attacking" all the core columns at the same time. Of course the fire was organic and so it was likely not "uniform" over the core footprint. But it's reasonable to say that the fire/heat was not acting on one column at a time either.

Why don't you EITHER explain what really happened OR read the many explanations including mine which explain it step by step?

 

[Jeffrey Orling]

The hour long process... could be:
a) individual columns failing... from heat driven loss of capacity...
or
b) all core columns failing... from heat driven loss of capacity...

In (a) there would obviously be a load redistribution (driving the column closer to failure)
aggregate capacity is lower while load remains unchanged... this process could continue until there was insufficient capacity

In (b) fire is "weakening" the steel in the core and the aggregate capacity of ALL the steel columns and beams are being driven down by heat until the collection of columns fail all at once...

Why would the fire not be attacking the entire core? Why would it weaken on column at a time or two or three?... Would the fire then move to the remaining columns which were not changed by fire? Would the fire act like a "front" moving through the core?

How do we know or how can we tell? There was no "creep" or visible warping.... there was an instant release.... driven below capacity. (all remaining columns).

 

[econ41]

The hour long process... could be:
a) individual columns failing... from heat driven loss of capacity...
or
b) all core columns failing... from heat driven loss of capacity...

Why speculate "could be"? We know that each column failed. We don't know and never will know the exact sequence. Some of it was heat-driven. Again we will never know how much. We do know that every column that failed due to load redistribution had sufficient load imposed to cause failure AND sufficient heat (rise in temperature) to reduce strength as much as that column needed its strength reduced. We will never know "how much" We do know "enough".

We do know all core columns failed.

AND a) b) are NOT "ORed" - it is AND

 

[Jeffrey Orling]

Why speculate "could be"? We know that each column failed. We don't know and never will know the exact sequence. Some of it was heat-driven. Again we will never know how much. We do know that every column that failed due to load redistribution had sufficient load imposed to cause failure AND sufficient heat (rise in temperature) to reduce strength as much as that column needed its strength reduced. We will never know "how much" We do know "enough".

We do know all core columns failed.

AND a) b) are NOT "ORed" - it is AND

We don't know that every column failed from load redistribution. Some surely failed from loss of capacity... some maybe have been pushed or pulled out of position by expanded beams.
In the end the axial failure was due to loads exceeding capacity... loads were no increasing... so capacity (aggregate) was driven down.

 

[econ41]

@dylbie. I"m not sure how I can help you further at this stage. The latest videos are quite good as a general introduction to Twin Towers collapses. Even tho far from perfect to satisfy technical pedants like me.

One word of general advice. Try to keep clear in your thinking the distinction between what your "simulation" shows and what actually happened in the real event. The history of 9/11 debate is littered with false arguments and much confusion which resulted when persons mixed up "model" with "real event".

So - with that point in mind - a few comments:

I'm hoping that Kostack does some different scenarios with different column failures to see if he gets different results.#

I've tried to roughly replicate the impact damage,## and also removed some of the interior columns where I think fire damage weakening would be most severe###.

# and ### Be aware that you are committing to Hulsey style "fudged graphics" to produce a "looks like" graphic and not a true simulation. Not a problem provided you keep the distinction clear.

## Good move.

I have noticed a new issue which is that the floor to column connections are too strong. The floors appear to be able to hold the above sections on their own. So I'm gonna need to come up with a new design and somehow try and make them weaker. I'm not entirely sure how I'm going to do this yet. Another issue is it seems to be collapsing slightly too fast, but I have 10 floors less at the top, and 7 floors less at the bottom so the 17 missing floors could explain the speed.

The speed problem is a feature of the software modelling. It would not have occurred in the real event if the Towers had been shorter in the way you describe. So - "speed" is not an issue - feel free to adjust the graphics given that the model is a "looks like" and not a rigorous simulation.

Firstly the tilt of 2WTC must have shifted the whole core inside the building as it was collapsing. .... I'd imagine this would have been responsible for a tremendous amount of damage to the floors either side of the core as it shifted, but also might explain how some of the core was still standing, if what was once above it shifted laterally.

It is an artefact of your software modelling. In the real event the "tilting" was a consequence of the failure mechanism. Not a cause. More explanation available if you need it.

Secondly, I think this may be a flaw with the software, but the sudden impact of the top section on the lower section overloads the columns near ground level. I think this is not realistic, but I guess in real life the whole lower section was obscured by dust so it was difficult to see what exactly was going on down there.

Correct - it is another artefact of the software. It is not realistic. In the real event the columns below the impact and fire damage zones were never overloaded. And we can understand the actual physics so the "obscured by dust" is not relevant.

 

[Jeffrey Orling]

The only simulation which would interest me is one of the interior which modeled the fire destroying the frame... But this pretty much unknown in the details... One would have to make assumptions for heat (fire spread) and see what happens to the frame.

As I am a proponent of the core led collapse initiation... once capacity was lost and the top blocks start coming down... the rest in not of much interest.

I would be interested to see how the ROOSD "mass" was created as it eventually... rapidly actually... involved the entire footprint... but I don't think it was entire slabs dropping like square donut shapes. I suspect somehow there was local break up of the slabs on multiple levels over the entire footprint.

Another driver might have been the heavy mech equipment on the uppermost floors (below WOW).

If the entire core "failed"... the floors above the failed core zone would "hang" from the hat truss. If this core area of more than 15 floors broke free the floors outside the core would be pulled down as well... and break up and perhaps be the ROOSD mass. But the hanging core would mean extra loads on the facade.

 

[econ41]

As I am a proponent of the core led collapse initiation...

For which neither you nor any other person has ever put forward a persuasive valid hypothesis.

I would be interested to see how the ROOSD "mass" was created as it eventually... rapidly actually... involved the entire footprint... but I don't think it was entire slabs dropping like square donut shapes. I suspect somehow there was local break up of the slabs on multiple levels over the entire footprint.

Please stop implying that you don't know how ROOSD was started. It has been explained many times. The falling perimeter of the Top Block applied a linear concentrated overload to shear off those first floors thereby creating the necessary weight of debris. (And break up of the Top Block started concurrently.) See where the perimeter columns - yellow arrows - moved through the area of the OOS floor joists - blue lines. This is the most extreme pair of examples - two out of eight "sides" - the other six are analogous for reasons that have also been explained.

Another driver might have been the heavy mech equipment on the uppermost floors (below WOW).

If the entire core "failed"... the floors above the failed core zone would "hang" from the hat truss. If this core area of more than 15 floors broke free the floors outside the core would be pulled down as well... and break up and perhaps be the ROOSD mass. But the hanging core would mean extra loads on the facade.

How does this collection of wild speculations help EITHER @dylbie's development of his simulation OR understanding of the actual collapse? Since it disagrees with both.

 

[Jeffrey Orling]

For which neither you nor any other person has ever put forward a persuasive valid hypothesis.

Please stop implying that you don't know how ROOSD was started. It has been explained many times. The falling perimeter of the Top Block applied a linear concentrated overload to shear off those first floors thereby creating the necessary weight of debris. (And break up of the Top Block started concurrently.) See where the perimeter columns - yellow arrows - moved through the area of the OOS floor joists - blue lines. This is the most extreme pair of examples - two out of eight "sides" - the other six are analogous for reasons that have also been explained.
View attachment 47940


How does this collection of wild speculations help EITHER @dylbie's development of his simulation OR understanding of the actual collapse? Since it disagrees with both.

I did put forward a core driven scenario I called sink hole top drop. I'll post it again tomorrow perhaps.

 

[econ41]

I did put forward a core driven scenario I called sink hole top drop. I'll post it again tomorrow perhaps.

Why post it again? Does it help @dylbie's simulation project? Is it even compatible with what we know actually happened on 9/11?

 

[Jeffrey Orling]

Why post it again? Does it help @dylbie's simulation project? Is it even compatible with what we know actually happened on 9/11?

The collapse simulation including the initiation... requires understanding what happened to the frame from and including the plane strikes, through the destruction caused by the fires. NIST presented a scenario of heated floor trusses. So I would expect that there were enough floor trusses heated to "do whatever" to the facade columns to cause it to buckle. They point to on example on the south facade of 1WTC. Yes when 1WTC top was "released" there was apparently a break around the entire perimeter (buckling?) I am not sure I understand how that worked... If you do you could explain it. Your load transfer cartoons look more like a description of inside the core failure (progression).

Regardless, I find the NIST explanation unsatisfying though it can explain what happens to one truss. I doubt it occurred to enough trusses. (My guess). So I look for a explanation that involves the core losing capacity... And it supported approximated 50% of the floor loads outside the core. What would happen if the core collapsed at the plane strike levels? I suspect if that happened it might explain the observed movement of the top.

In the case of 1WTC the jet plowed into the center of the long side of the core taking out multiple columns and it's likely that momentum carried the fuel largely into the building.... in its path. Of course the fuel would spread go down elevator shafts and so on. And yes some would spread to the entire footprint... at least on the south side. I have no clear picture of the fuel spread not the flame spread. I think that the core experienced more fire than the areas outside the core. But maybe as the trusses were light steel they were more vulnerable to heat??

However if the core steel was being "attacked" by fire its failure might not be a 1+ hr progression of failures... but a 1+ hr weakening of the entire core... (most of it) leading to a failure of all (many) core columns at once. With that what they supported would drop down and this might pull the core side of the outside the core slabs... doing two things... starting the ROOSD collapse and buckling the perimeter.

Regardless whatever the "initiation" process was... it would / should be expressed in the collapse simulation. It was complex and involved 47 columns of the core...and all the perimeter...

 

[investigating911]

Secondly, I think this may be a flaw with the software, but the sudden impact of the top section on the lower section overloads the columns near ground level. I think this is not realistic, but I guess in real life the whole lower section was obscured by dust so it was difficult to see what exactly was going on down there.

Other than the lack of peeling away of exterior columns, this is another feature I noticed that doesn't seem like the actual collapses. In the actual collapse, even as the first exterior columns and dust was reaching the ground, the lobby level still looked intact and did not appear to be coming apart, nor did the exterior columns above the lobby level but below the present region of the downward cascading floors (as the still from 1:14 to 1:21 in the following video of the south tower collapse from Liberty Street shows):

Source: https://www.youtube.com/watch?v=qLShZOvxVe4&list=UUdNkRyt3S0oTQQdzlKJtXyg&t=74s

Similarly see the still at 2:19 in the same video for the north tower, where the lower section is still relatively intact as the first debris and dust is reaching the ground

Source: https://www.youtube.com/watch?v=qLShZOvxVe4&list=UUdNkRyt3S0oTQQdzlKJtXyg&t=139s

From the eventual WTC debris pile, notable areas of exterior perimeter columns at the lobby level were left standing (for the north and northeast side of the north tower and for the west and southwest side of the south tower). Had the lower exterior columns begun to collapse before the collapse front reached this region, I don't see how significants sections of above ground exterior perimeter columns near the lobby level can remain standing on certain sections of the building. Yet from the eventual WTC debris pile, this was what was seen.

There didn't appear to be any breakup of the exterior columns (other than emission of dust from broken windows) until the region of downward cascading floors had reached the floor height of those exterior columns.

We're any simulations done by NIST of the fires? And in particular, the affect of wind on the fires at 1000ft?

There were extensive simulations down by NIST of the fire spread and the effect of this on the expected temperature of the steel columns and trusses in the region of fire. If I'm remembering right, this was the part of the research that proved that the steel in the region of fire was heated sufficiently enough to soften and buckle, although not melt.

See Section 6.10 (starting on page 121), Section 6.12 (starting on page 131) and Section 6.14.5 (page 146) specifically in the paper for information on the fire simulations carried out.

A summary of the research done for the reports of both the causes of the collapse of the twin towers and WTC 7 was released shortly before the 20th anniversary of 9/11:
https://www.nist.gov/blogs/taking-m...rld-trade-center-investigation-and-its-legacy


I don't think the effect of wind on the fires was examined in detail other than the researchers at NIST saying that the broken windows (from the plane impact and explosion) helped continue to fuel the fires that were burning because oxygen from outside was able to reach the fires that way.

You can clearly see there must have been a reasonably strong North Westerly wind by the smoke pattern. This would have created a furnace type effect and would have caused the fires to be more severe in the South and East sides. In fact I think this is backed up by the images of 1WTC South facade just before collapse which appear deformed, and also by the people who were seeking refuge in the North Western corner which must have been the coolest place.

I think part of the reason the fires were most severe on the south and east sides at the time of collapse of the north tower is because the fires started more on the north face of the building (the area of the floor where the airplane had first made impact and the fuel in the wings had spilled and ignited) and moved over time to the opposite face on the south side of the building.

Similarly, at the time of collapse of the south tower, the flames from the fires were most visible and intense on the east and north faces of the building, even though the airplane and fires had initially started more on the south face of the building where the airplane had first made impact and the fuel in the wings had first spilled and ignited.

It appears that the fires moved over the floor area to the side opposite of the impact hole as the supply of combustible fuel in a particular floor section of the building that was ignited earlier became exhausted. This would have probably happened first for the area of the floor closest to the initial airplane impact.

Breaking complex things down into separate building blocks is a sound, proven teaching technique. I'm aware that it may not suit some members. If other members also show no interest in comprehending the physics which is the foundation of the OP proposed simulation - I won't press the matter further.

I doubt there is much difference in the practical aspects. The big and unresolved issue with physical models or simulations is in the presumption that they will benefit a large number of persons who would not be sufficiently helped by other methods of teaching/learning. The actual mechanism of the initiation stage of Twin Towers collapses cannot be accurately modeled or simulated in detail. So, whether sim or model, it would have to be a simplified demo of the main principle.

BTW you don't need 22,000 objects to comprehend the real event. Sub-set the tower into a handful of sub-systems. maybe a dozen parts in a sub-system to illustrate principles. The reality of 22,000 objects is actually one of the main reasons FEA analysis of the whole system is impractical.

So from what you are saying, even a physical model (such as pasta columns and wooden blocks or wax columns and wooden blocks) still wouldn't capture the intricacies of the collapse of the towers (presumably because of the necessary small scale of those physical models if designed by a layperson without significant financing) and so a more simple computer model with a sub-system of blocks would have to do, since FEA cannot be run on a large number of blocks without very high computing power?

Why is it that the impact of a plane on a building can't be easily simulated in a physics simulation from Besiege (per discussion in https://www.metabunk.org/threads/ho...ngs-penetrate-the-wtc.3326/page-5#post-259628) but a collapse of a WTC-like building can?

 

[econ41]

So from what you are saying, even a physical model (such as pasta columns and wooden blocks or wax columns and wooden blocks) still wouldn't capture the intricacies of the collapse of the towers (presumably because of the necessary small scale of those physical models if designed by a layperson without significant financing)

The problem is not scale size per se. It is complexity. Far too many details to model physically. And whilst modelling the number of components may be possible in a computer sim we cannot know enough to input the data to specify all the motions of the many parts. We don't know and cannot access that data from evidence of the real extent.

But I suggest that there are a couple of more fundamental issues:
(a) Adequate and accurate models can be made for a sufficient selection of the sub-components of mechanism to demonstrate the overall mechanism AND
(b) (Remember the real purpose is teaching someone) Explaining bit by bit on a "building block" approach is almost certainly a more effective teaching and learning technique.

and so a computer model with a number of sub-systems of blocks would still be the approach, since FEA cannot be run on a large number of blocks without very high computing power?

Computer or physical model - "horses for courses". Some bits lend more to physical models. Others are easier to achieve by computer.

 

[investigating911]

The problem is not scale size per se. It is complexity. Far too many details to model physically. And whilst modelling the number of components my be possible in a computer sim we cannot know enough to input the data to specify all the motions of the many parts. We don't know and cannot access that data from evidence of the real extent.

Yes that makes sense. I'd also expect that even given the same starting conditions for collapse, hypothetically repeating such a collapse numerous times (whether in a physical or computer model) would yield slightly different results each time in terms of debris trajectory, exact time of full collapse and the layout of the debris pile. This would probably be because of slightly different conditions each time around (especially ambient condition differences in the case of the physical model or initial condition differences in the computer model) but also because of a certain amount of randomness involved. In any case I wouldn't expect that an exact model (considering where each individual part of the tower fell and punched through before settling) would ever be able to be made that exactly matches what was seen in the collapse of the WTC towers down to every last detail.

But I suggest that there are a couple of more fundamental issues:
(a) Adequate and accurate models can be made for a sufficient selection of the sub-components of mechnaism to demonstrate the overall mechanism

Is the modeling of fire spread on the affected WTC floors by NIST an example of such sub-component modeling? From the questions and discussions I've had here concerning modeling the collapses, I think I am beginning to understand why NIST never released a physical or computer model of the full collapse of the WTC twin towers: it was just too complex to model every stage of the collapse, so it was simply concluded that once the collapse began it was inevitable and the details were left at that instead of going into detail about all the individual stages of collapse.

(b) (Remember the real purpose is teaching someone) Explaining bit by bit on a "building block" approach is almost certainly a more effective teaching and learning technique.

Yes, I think the most important elements to have in modeling the collapse of the WTC twin towers would be: correctly modeling the degree of tilt of the upper section prior to it descending and beginning the collapse, the lower sections of the exterior columns remaining intact (i.e. not peeling away or coming loose) until the region of descending, cascading floors had reached the floor of those exterior columns, and a portion of the core columns (about halfway up the height of where the building once stood, such as the 60th floor) standing for at least 10 seconds after most of the exterior columns had collapsed, before the core columns finally fall too.

These are three elements that were common in visuals from both collapses.

Computer or physical model - "horses for courses". Some bits lend more to physical models. Others are easier to achieve by computer.

Could a physical model potentially be more useful than a computer model for modeling the stages of destruction at individual floors or a set number of floors if such a physical model of the towers was ever able to be constructed and slow motion photography and analysis could be taken of the various stages of the collapse once the progressive collapse had initiated?

 

[econ41]

@investigating911 I'll respond section by section.

Yes that makes sense. I'd also expect that even given the same starting conditions for collapse, hypothetically repeating such a collapse numerous times (whether in a physical or computer model) would yield slightly different results each time in terms of debris trajectory, exact time of full collapse and the layout of the debris pile. This would probably be because of slightly different conditions each time around (especially ambient condition differences in the case of the physical model or initial condition differences in the computer model) but also because of a certain amount of randomness involved. In any case I wouldn't expect that an exact model (considering where each individual part of the tower fell and punched through before settling) would ever be able to be made that exactly matches what was seen in the collapse of the WTC towers down to every last detail.

All agreed. Then remember the goal - help some other person understand the real collapses. And you don't need to be accurate in all the details to achieve that goal. If the "trainee" or "learner" comprehends all the key parts of the mechanism AND can see how they fit together >> goal achieved. Remember the model or computer sim is simply a means to an end. The end is that the other person learns. NOT that the model or sim be 100% accurate.

 

[econ41]

Is the modeling of fire spread on the affected WTC floors by NIST an example of such sub-component modeling? From the questions and discussions I've had here concerning modeling the collapses, I think I am beginning to understand why NIST never released a physical or computer model of the full collapse of the WTC twin towers: it was just too complex to model every stage of the collapse, so it was simply concluded that once the collapse began it was inevitable and the details were left at that instead of going into detail about all the individual stages of collapse.

Both those are true but they are actually separate issues. Yes, overall simulation was too complicated. And - for the same reasons we have been discussing - it was not necessary when explaining for a genuine, reasonably informed and honest "audience". It was never NIST's goal - never could have been - to present a bullet proof explanation that no conspiracy theorist could ever fault.

And you don't even need either a physical model or a computer sim. I've used this simple graphic many times. It is part of a sequence of about 1 dozen "step by step" explanations. IF you can understand how load-redistribution can overload column "C" in this model you are well on the way to comprehending even the complexities of Twin Towers Initiation Stage. And I invented the model - simply to explain a principle.

 

[econ41]

Yes, I think the most important elements to have in modeling the collapse of the WTC twin towers would be: correctly modeling the degree of tilt of the upper section prior to it descending and beginning the collapse, the lower sections of the exterior columns remaining intact (i.e. not peeling away or coming loose) until the region of descending, cascading floors had reached the floor of those exterior columns, and a portion of the core columns (about halfway up the height of where the building once stood, such as the 60th floor) standing for at least 10 seconds after most of the exterior columns had collapsed, before the core columns finally fall too.

Actually the questions about tilt are not very important. BUT tilt was such a prominent feature that I think it should be modelled. people expect to see "tilt" and removing it in a teaching situation would unnecessarily confuse the "student". For those of us who are on the 'teaching side," it is important that we do understand tilt - it has caused a lot of confusion. The main problem is that a lot of people think that "tilt" is what caused or prevented the falling column ends landing on the lower columns. So it is a subset of the wrong understandings about the1D approximations. "Tilt" was NOT the cause of failure. It was a consequence. Columns had already failed and their top ends were bypassing - missing - their bottom parts. And tilt occurred because more columns failed one side of the building than the other. So tilt followed column failure. Didn't precede or "cause" it. << That explanation is a bit cryptic -- we may need more details.

 

[econ41]

Could a physical model potentially be more useful than a computer model for modeling the stages of destruction at individual floors or a set number of floors if such a physical model of the towers was ever able to be constructed and slow motion photography and analysis could be taken of the various stages of the collapse once the progressive collapse had initiated?

I doubt it is necessary. Remember that most of the floors - over 80 of them - involved in the "progression stage collapse" were identical. If you understand this:

.. you should understand "this" times another 80 identical replicas.

And the remaining 20 or so floors were involved either in the events of the ignition stage involving the impact and fire damaged zones OR in the "Top Block" which shouldn't be too hard to comprehend once the "student" understand the main features of the other 80 floors. Simply turn the process upside down. << Again something you may wish to discuss in more detail. Your call.

 

[Jeffrey Orling]

I doubt it is necessary. Remember that most of the floors - over 80 of them - involved in the "progression stage collapse" were identical. If you understand this:

.. you should understand "this" times another 80 identical replicas.

And the remaining 20 or so floors were involved either in the events of the ignition stage involving the impact and fire damaged zones OR in the "Top Block" which shouldn't be too hard to comprehend once the "student" understand the main features of the other 80 floors. Simply turn the process upside down. << Again something you may wish to discuss in more detail. Your call.

minor point.... the floors were not identical... Mech floors were different. However faced with the dynamic loads it hardly mattered.

 

[investigating911]

For those of us who are on the 'teaching side," it is important that we do understand tilt - it has caused a lot of confusion. The main problem is that a lot of people think that "tilt" is what caused or prevented the falling column ends landing on the lower columns. So it is a subset of the wrong understandings about the1D approximations. "Tilt" was NOT the cause of failure. It was a consequence. Columns had already failed and their top ends were bypassing - missing - their bottom parts. And tilt occurred because more columns failed one side of the building than the other. So tilt followed column failure. Didn't precede or "cause" it. << That explanation is a bit cryptic -- we may need more details.

Were the top ends of the columns bypassing the bottom parts because of inward bowing of those columns, such that the downward trajectory of these columns would cause them to impact the floor area and not the lower columns?

Also I recall that one of the lead investigators on the NIST report,Shyam Sunder, had said in an online interview (part of the series here https://www.pbs.org/wgbh/nova/wtc/, see transcripts for each section for example the transcript for https://www.pbs.org/wgbh/nova/tech/debunking-9-11-bomb-theories.html) that just prior to the collapse, some of the exterior perimeter columns in the impact zone that had bowed inwards had snapped back outwards as the connection of the floor trusses with the exterior columns had been severed after sagging of the trusses and thermal expansion from the heat of the fires had put enough stress on the connections between the inward bowing exterior columns and the floor trusses. Just like inward bowing columns would mean the falling columns would impact the floor area and not the lower columns, wouldn't this mean that part of the reason that the falling column ends did not land on the lower columns was also because those falling columns that snapped back outwards fell down outside of the region of the lower exterior perimeter columns, bypassing the tower perimeter and interior floor area entirely?

minor point.... the floors were not identical... Mech floors were different. However faced with the dynamic loads it hardly mattered.

I remember reading that the mechanical floors were designed to carry more weight than normal floors because of the heavy mechanical equipment that had to be stored on these floors and that the mechanical floors were taller than normal floors, with open air vents instead of windows. The compressed air and dust escaping outwards from the collapsing towers had an easy passage through these open air vents on the mechanical floors compared to on normal floors where it would escape through broken windows or breaking through previously intact windows, which was another reason NIST had given for why puffs of smoke (suggested to be demolition squibs by 9/11 conspiracy theorists) were seen just below the collapse front, especially on the darker mechanical floor sections of the towers.

 

[Jeffrey Orling]

Were the top ends of the columns bypassing the bottom parts because of inward bowing of those columns, such that the downward trajectory of these columns would cause them to impact the floor area and not the lower columns?

Also I recall that one of the lead investigators on the NIST report,Shyam Sunder, had said in an online interview (part of the series here https://www.pbs.org/wgbh/nova/wtc/, see transcripts for each section for example the transcript for https://www.pbs.org/wgbh/nova/tech/debunking-9-11-bomb-theories.html) that just prior to the collapse, some of the exterior perimeter columns in the impact zone that had bowed inwards had snapped back outwards as the connection of the floor trusses with the exterior columns had been severed after sagging of the trusses and thermal expansion from the heat of the fires had put enough stress on the connections between the inward bowing exterior columns and the floor trusses. Just like inward bowing columns would mean the falling columns would impact the floor area and not the lower columns, wouldn't this mean that part of the reason that the falling column ends did not land on the lower columns was also because those falling columns that snapped back outwards fell down outside of the region of the lower exterior perimeter columns, bypassing the tower perimeter and interior floor area entirely?

There was no "separation" of columns except for those severed by the planes.
Column / axial failures occurred from INLINE buckling or lateral displacement of the column ends leading to inadequate bearing area and web and flange crippling.

It is impossible for all of the perimeter columns to pass outside or inside. If one side... passed outside... the opposite face would have to pass inside. It's more likely that there was lateral displacement (1WTC) in both long and short axis of the core.... leading to measurable displacement in both axes leading to buckling on all sides.

2WTC's top tilted Southeast as it dropped. It looks like it also was displaced southward leading to facade columns breaking and passing on the south and north faces. The upper block appears to move east so the west columns were broken and no longer in axial alignment. Intact slabs were impacted by the broken column ends. But the dropping mass of the upper floors was likely what kicked off the ROOSD.
Hard to know.

 

[Jeffrey Orling]

How was the huge mass of the top blocks made to move laterally... for surely they had to because at some point all columns were no longer axially aligned?

 

[Marc Powell]

How was the huge mass of the top blocks made to move laterally... for surely they had to because at some point all columns were no longer axially aligned?

At collapse initiation, each top block rotated about its center of mass as it descended. Inertia prevented it from moving laterally.

 

[Jeffrey Orling]

At collapse initiation, each top block rotated about its center of mass as it descended. Inertia prevented it from moving laterally.

explain

 

[Marc Powell]

explain

In videos, upon the initiation of collapse, the bottom of the top block can be seen to move laterally in a direction opposite the direction of tilt. Absent the application of external horizontal forces, the tremendous inertia of the top block constrained it to only rotate about its center of mass as it fell.

 

[econ41]

explain

Try this:

Original source not known. The clip is taken from a longer version posted on The911Forum by truther researcher "achimspok" who applied the floor and column grid overlay. I further shortened the length of the clip. I have also used the following single frame clip from the motion *.gif to demonstrate the mechanism which started "ROOSD" - note how the moving perimeter walls (yellow arrows - my additions) apply a massive linear concentrated overload onto the office space floors (blue lines - also added by me):

Note also - that single frame clip also serves to rebut Bazant & Verdure's "Crush Down/Crush Up" hypothesis. It falsifies the B&V premise that the falling "Top Block" crushed the columns of the lower tower. One of four fatal errors in "CD/CU". The "Top Block" did NOT wait to start "Crush Up" until it reached ground level AFTER "Crush Down". Top Block broke up at the start of progression concurrently with break up of the upper levels of the lower tower. And neither "Down" nor "Up" were by "Crushing"...

 

[econ41]

Were the top ends of the columns bypassing the bottom parts because of inward bowing of those columns, such that the downward trajectory of these columns would cause them to impact the floor area and not the lower columns?

How the column ends "missed" or "bypassed" is a topic that has seen a lot of confused thinking.

Let's put it in context. The defining characteristic of WTC Twin Towers collapses is the 2/3rds "G" "speed" of descent of the collapse progression. The reason for that was "ROOSD" - i.e. the columns were mostly ineffective because the falling debris missed the columns and landed on the floors. So the key question is the lead-up - "Why did the top parts of the falling columns miss their bottom parts?"

There are two reasons which are:
(a) The micro-level reality than as any single column fails in axial overload whatever the combination of heat weakening or load re-distribution it get shorter. The load applied to the top of the failing column pushes it down and the top end gets closer to the bottom end. It can do that in several ways. (i) localised buckling - see the following graphic; (ii) bowing of a full length between horizontal braces and/or (iii) physically breaking into two parts. This is one example:

Note it is from WTC 6? It shows that the "top" is already starting to bypass the "bottom" and the buckling has halted for reasons not relevant at this stage of discussion.

(b) Macro level gross motions as er this graphic:

(See fuller explanation in preceding post #118)

Type "a" - micro-level buckling of every failing column ensured that the "missing" or "bypassing" scenario was set-up in preparation for ROOSD. Type "b" - macro-level motion reinforced the amount of "miss" but was strictly redundant. (Micro effects mean the column ends are already bypassing. It doesn't change anything if more macro or gross motions cause them to miss by a greater distance.)

The other type of failure was "clear-cutting" as resulted from the aircraft initial impact. The same situation would result from explosive cutting if there had been any - ironically the truther claims for CD do not change the explanation of what actually happened. It merely changes the balance of contributing factors. >> take a rain check on more detailed explanatory proof.

So - back to your question about perimeter inward bowing and "elastic spring back". Put simply that motion does not change the explanation. All it does is change the timing of the failure of those affected columns. If they had not failed when they sprang back they would simply fail later. Probably by gross full-length buckling. Remember that the "spring back" resulted from the failure of the bracing effect of floor joists. So the effective length of the "springing back" column would be at least two stories - maybe three. Also heat-driven "micro" buckling was less likely for perimeter columns - there would probably be insufficient temperature rise at those exposed locations

 

[Jeffrey Orling]

In the vid of the top of 2WTC dropping and tipping it is clear that what led to this motion was most likely a failure of the SE core. It appears that there is a fairly clean break of the facade columns... not a staggered jig saw which might be a tell that when the top "dropped" it had enough interior structure... slabs? core? Hat truss? to hold it as a solid block. It seems also to be displaced so the facade columns are not "hitting" but passing. The bottom floors of the top crushed the top floors of the bottom... the columns punched through slabs. There appears to be a mutual destruction as they "come together"... not unlike what happens when two cars collide. I suspect in 1WTC the top moved "en mass" laterally and dropped with columns sheared and no longer aligning... and the same crushing in the "zone of impact".