A virtual model illustrating some aspects of the collapse of the WTC Towers

[Mick West]

And keep in mind the current model has no vertical connections between columns in different assemblies. Earlier more "2d" models had progressively heavier and stronger Colin's as you went down. It's just al little more work needed there. A slow process of gradual improvement.

 

[Mick West]

I agree--having an undamaged version of the building that stands indefinitely should satisfy the current criticism re how the version that is designed to fail from the outset does, in fact, fail.

And it should support some dynamic load.

 

[Cube Radio]

And it should support some dynamic load.

Aren't we now approaching a tower that isn't "designed to fail" at all, then? The twins themselves can hardly be said to have been designed to fail, and yet both failed with spectacular speed and destructive power from two very different initiation points.

I don't feel as though the objective of Mick's modelling effort would be fairly described as a Rube-Goldberg machine that is deliberately and precariously balanced in order to collapse in a manner that bears useful resemblance to the actual events. If it can support itself, some dynamic load and some lateral force, it is already a long way from such a contrived scenario.

However, I don't won't to be accused of attempting to move the goalposts, so invite and naturally take into account the views of all forum members -- and a Rube-Goldberg machine would be interesting in itself.

 

[Oystein]

Aren't we now approaching a tower that isn't "designed to fail" at all, then? The twins themselves can hardly be said to have been designed to fail, and yet both failed with spectacular speed and destructive power from two very different initiation points. ...

The difference is quantitative, not qualitative.

It is often said that the twins, like other highrises, were designed with some factor of safety - or their structural elements. The idea being that the elements could bear, say, 2-3 times their maximum load; or the assembly 2 times the weight on top (I am making these number up, but you recognize the concept).
(Reality is more complex than this, but you get the point)

You could turn this on its head and say that "the twins were designed to fail at 2 times their max load", or "the twins were designed to fail at or below 50% of capacity". Not literally, not consciously, but in effect.

This "factor of safety" is maintained only for the as-designed assembly.
Once the assembly has taken wide-spread fire damage, once it has suffered structural damage from plane crashes or debris smashing into the walls, they are no longer "as-designed", and no factor of safety can be assumed any longer.

Once the top part has started to move downward more than seismic stress could induce, you are outside the envelop - no factor of safety can be designed into the structure to accomodate this mode of being. It's a case that a building simply not "designed for". The progressive failure and dynamic loading both increase local loads and decrease local capacity.

 

[Keith Beachy]

Aren't we now approaching a tower that isn't "designed to fail" at all, then? The twins themselves can hardly be said to have been designed to fail, and yet both failed with spectacular speed and destructive power from two very different initiation points. ...

Ironically, or unfortunately,
WTC towers were designed to fail as seen, unintentionally, but as designed. Floors held up by by core and shell, the collapse seen was due to the design of the WTC. We can see the overwhelming mass destroying floor by floor, and that same mass rips apart the shell in large sections, and the core that remains can no longer stand because the shell which was the lateral support is gone. The towers design drove the failure as seen. A gravity collapse started by massive office fires, because the aircraft impacts 7 and 11 times the actual design point to stop a 187 pound of TNT KE impact destroyed fire systems and dislodged insulation accelerating the failure due to fire.

Not saying the WTC was designed to fail, I am saying the designed failed as it would due to the design. Evidence, collapse videos, full up models of the WTC. As for the spectacular speed, it is exactly the speed of a momentum model for the WTC floors failing, and the destructive power is spectacular, because the E=mgh is released and is equal in energy to more than 100 2,000 pounds bombs; thus the speed is physics, and the energy is physics, the speed is normal, the energy (destructive power) is spectacular, but explained by physics and expected.

Most, if not all, modeling of WTC towers fail to include floors held up by core and shell fail above a certain weight. WTC floors fail at connections to the core and shell at weights above 29,000,000 pounds. Has anyone seem models including this feature.

 

[Mick West]

Aren't we now approaching a tower that isn't "designed to fail" at all, then? The twins themselves can hardly be said to have been designed to fail, and yet both failed with spectacular speed and destructive power from two very different initiation points.

I don't feel as though the objective of Mick's modelling effort would be fairly described as a Rube-Goldberg machine that is deliberately and precariously balanced in order to collapse in a manner that bears useful resemblance to the actual events. If it can support itself, some dynamic load and some lateral force, it is already a long way from such a contrived scenario.

I'm a little confused as to your point here. I'm trying to create models that support themselves, and similar loads to that which the WTC could support. At the very simplest I'm only trying to illustrate why the cardboard box model is wrong. More accuracy and robustness would be great though.

Building are designed not to fail within a certain envelope of potential conditions, and generally with some safety factor tacked on. When you go outside the envelope then you are outside the design.

 

[benthamitemetric]

Aren't we now approaching a tower that isn't "designed to fail" at all, then? The twins themselves can hardly be said to have been designed to fail, and yet both failed with spectacular speed and destructive power from two very different initiation points.

...

Mick's model is "designed to fail" in the sense that the failure is the initiation of the drop of the top block over the distance of approximately one story. What happens after the top block drops is the essence of the tower challenge as @aka hypothesizes that, for any tower that meets his stated criteria (and Mick's will), the result of the dropping of such a top block would not result in a collapse that is similar the collapses witnessed for WTC1 & WTC2 (i.e., a top-down total collapse wherein the floors are shorn from the columns in relatively rapid succession on a floor-by-floor basis).

 

[Mick West]

Source: https://www.youtube.com/watch?v=oswEUUhdLCQ


This is more of an experiment with rendering to video than anything. Previous videos had been screen captures. This one is actually rendered properly frame by frame (using OpenGL Rendering, so simplified lights), so should be time-scale correct.

It's a 12 assembly three floor 3x3 structure, so 36 floors. I'd set it up with no damage to see if it collapsed with 2 second ramped gravity. (it did not). Then I manually did a bit of damage and reran the physics. The result was quite asymmetric as I'd just remove columns on one side. But I think that speaks quite well to the general stability of the structure, in that there was a lot of destruction on one side before the other side started to collapse. The main collapse wave does not lag much behind free fall.

 

[Bruno D.]

Source: https://www.youtube.com/watch?v=oswEUUhdLCQ

That's interesting. This collapse resembles WTC7's one

 

[Mick West]

That's interesting. This collapse resembles WTC7's one

Yes, it made me think I could try that maybe a 4x2 grid (this is 3x3). More irregular shaped.

Probably needs wall panels too. I might stat looking at other engines.

 

[Cube Radio]

That's interesting. This collapse resembles WTC7's one

I can't agree, except insofar as they were both collapses. Sudden onset observed across the entire roofline, which began to move rapidly "as a piece" with a high degree of symmetry (I'm not going to get drawn into a semantic discussion of what degree of symmetry can be reasonably be called "symmetric") is just one critical factor defining WTC7's destruction, and it's obviously not what we have here.

But of course Mick isn't seeking to model WTC7, and even NIST couldn't achieve anything approaching a representation of the collapse that reproduced the features I mention.

Is the source available for this latest model, please, Mick? Which engines are you considering instead of Blender? It's obviously important that this all remains open source and easily tested by a wider community, which is why Blender is the engine I've got most interest in.

 

[Mick West]

I just did a quick try of a WTC7-style collapse, 60 floors, removed some interior columns.
Source: https://www.youtube.com/watch?v=BHCg19XwHEw


On of the most significant factors in the way WTC7 collapse was that the outer walls had connections that were moment resisting (i.e. they resisted twisting). The inner connections were largely just seated connections - so much easier to break with lateral forces. This mean that the outer walls held together as a skin for several seconds after the interior had collapsed.

 

[Mick West]

Here's the 60 floor model prebuilt to show it does not collapse with gravity ramping (so a 15MB zip). You can removed some columns, or do something like grab a section and move or shrink it

 

[Jeffrey Orling]

The form of collapse is driven by the structural design. Collapses are gravity driven and we saw top down driven collapses in the twin towers and a lwd down failure undermining axial load paths in 7 wtc.

To simulate the FORM of the 7wtc collapse you need to include the unusual transfer structures in the lower floors of the design. From floor 8 to the roof the design was not unusual except perhaps that the plan was a trapezoid shape. The tower was built spanning a huge 3 story power station. The fires which were the proximate cause of structural failures were fed by unusual quantities of diesel fuel. This is not typically present in office towers.

 

[Mick West]

Regarding the reddit thread on this topic. Discussing the same thing in two places is not going to work. It's effectively a violation of the no-click policy to have partial excerpts of an ongoing external discussion as context is generally required.

If @aka wants to make suggestions, then he can do it here.

 

[Mick West]

Which engines are you considering instead of Blender? It's obviously important that this all remains open source and easily tested by a wider community, which is why Blender is the engine I've got most interest in.

I was thinking possibly Unity. The biggest problem I have with Blender is the time it takes to construct a model in script. I like to iterate over different models rapidly, and having to wait three or four hours to make a 60 floor model is no fun.

Unity is a game engine, and uses C# for scripting (nicer than Python). It's more focussed on real-time simulation, but there are plug-ins for things like Bullet Physics. I think it would be much faster for procedural mesh/objects/connections/scene generation. There's also been some work on destroyable buildings (2013):
https://forum.unity3d.com/threads/destroyable-buildings-v4.169718/


I'd also look at Javascript wrappers to physics engines. ammo.js, Cannon.js, Physijs, JigLibJS, etc. But none of them leaps out.

 

[benthamitemetric]

Mick--have you considered using a render farm? You might be able to significantly cut render time down at a relatively low cost. Regardless of which program you end up using, it may make sense from a time/money perspective as you iterate.

 

[Mick West]

Mick--have you considered using a render farm? You might be able to significantly cut render time down at a relatively low cost. Regardless of which program you end up using, it may make sense from a time/money perspective as you iterate.

Rendering isn't really the issue. Blender has problems with scripted model creation (at least the way I'm doing it).

 

[benthamitemetric]

Rendering isn't really the issue. Blender has problems with scripted model creation (at least the way I'm doing it).

Ok, gotcha.

The models are looking interesting so far and the WTC7-esque progressive collapse model is the most interesting of all. If that video output a real time output of the simulation? It seems to me some objects, such as the red beams that fall out of the structure at various places, that should be in free fall (because they are detached from the structure during the collapse) fall a bit sluggishly, but not sure if that is a result of the video being slowed down, an illusion of some kind given the perspective, or something else.

 

[Mick West]

If that video output a real time output of the simulation?

Yes it's real time, in the sense that the video should reflect the real-world physics of something that size. "Real-time" can have a bit of a different meaning in games and simulation - meaning that the physics and rendering run fast enough to display everything accurately. Here the video runs at real-world speed but was pre-calculated and pre-rendered.

The model is set up with approximate WTC Tower dimensions, so it's 200 feet wide. (actually set with a 20m gap between 1m columns, so 64m, so 209 feet. ) Floors are 3.77m apart, so this is 226 meter, 742 foot high tower.

The time to fall 742 feet is about 6.8 seconds. The beams look like they are moving slowly because they are so big. Compare with video of debris falling from the towers as they collapse.

 

[Cube Radio]

@aka has conducted a comprehensive analysis of Mick's models so far. Perhaps the most interesting (in terms of visual resemblance at least) is this one:


Further details in the Reddit thread:
https://www.reddit.com/r/towerchall...adebunk_a_review_of_mick_wests_virtual_model/

 

[Landru]

@aka has conducted a comprehensive analysis of Mick's models so far. Perhaps the most interesting (in terms of visual resemblance at least) is this one:


Further details in the Reddit thread:
https://www.reddit.com/r/towerchall...adebunk_a_review_of_mick_wests_virtual_model/

Why don't you explain what it says? This is close to a no click violation.

 

[Cube Radio]

It's a version of Mick's model with the strength of the connections increased to strengthfactor 7.5. Top-down collapse is initiated and the video shows the result, which I think is interesting because it bears some resemblance to the actual event.

 

[Mick West]

It's a version of Mick's model with the strength of the connections increased to strengthfactor 7.5. Top-down collapse is initiated and the video shows the result, which I think is interesting because it bears some resemblance to the actual event.

I've not looked at the model since my last post. Was there any conclusion (or just discussion) reached regarding this particular run? Can you quote it here please?

One thing to consider with my 3x3 model is that I never got around to adding column-column connections, which would change both the other strengths needed for static/wind/earthquake stability, and the visual appearance of collapse (more "spire" type structures). It would also need something like stronger and heavier columns in the center.

 

[qed]

@aka has conducted a comprehensive analysis of Mick's models so far. Perhaps the most interesting (in terms of visual resemblance at least) is this one:


Further details in the Reddit thread:
https://www.reddit.com/r/towerchall...adebunk_a_review_of_mick_wests_virtual_model/

Why don't you explain what it says? This is close to a no click violation.

It's a version of Mick's model with the strength of the connections increased to strengthfactor 7.5. Top-down collapse is initiated and the video shows the result, which I think is interesting because it bears some resemblance to the actual event.

Dude, given your rep, not cool. Be more careful please.

 

[Cube Radio]

Dude, given your rep, not cool. Be more careful please.

I didn't realise I had a reputation to uphold around here -- except as an object of ridicule, that is. But thank you for that. I am in a difficult position really as @aka's analysis is as comprehensive and succinct as possible, but posting his comments from the external source is also frowned upon here. However, I will try again in the following post.

 

[Cube Radio]

The most interesting point about the model I've posted here first is that, with strengthfactor 7.5, we have -- as you see -- a collapse that, when initiated, does resemble the actual events to some extent. It takes much longer than was observed in the actual events, and the model has some jolts and asymmetries, but leaving that to one side it is not unlike the reality.

The problem with this structure at strengthfactor 7.5 is that it does not stand up on its own: it cannot support itself for very long, and falls apart.

[Video description: although there is no indication how long the Blender software was running before filming started, in this demo we see at 1 min 23 secs we see 2 non-consecutive floors about 3/4ths up the building on right side collapse for no apparent reason.]

Was there any conclusion (or just discussion) reached regarding this particular run? Can you quote it here please?

Observations

Unlike Bazant asserted repeatedly, the upper block does not remain rigid, it is clearly and without exception the first to fall apart, unless in virtual freefall through towers so weak that they offer no resistance at all.

Without exception, the towers with West's strength preset were unable to stand up against their own static weight and fell apart immediately, even without an artificially induced collapse. Making the connections stronger and stronger resulted in stiffer and stiffer buildings (clearly, even building a completely safe tower is not out of the question), but yielded longer and longer collapse times, jolts – and even collapse arrests, just as predicted by many "demolitionists".

Content from external source

tl;dr

Mick West builds extra flimsy model to demonstrate a possible mechanism for the collapse of the Twin Towers. Upon analysis, model proves to be "demolished" already, being too weak by a huge margin to even stand up its own weight. Stronger connections yield increasingly stable structures, but greater probability of collapse arrest. A tower that stands up safely when intact but collapses smoothly, rapidly, symmetrically and completely once initiated by dropping the smaller top on the larger bottom has not been presented. The model serves to demonstrate the difficulties any towerchallenger will encounter. Bazants and NISTs claim of "inevitability" can be regarded as thoroughly debunked. An intentional or accidental design error does evidently also not serve as excuse. A premeditated sudden or sequential weakening of many connections, columns and girders within a short time frame due to additional energy sources is the more plausible explanation for the anomalous phenomenon observed on September 11th, 2001.

Content from external source

https://www.reddit.com/r/towerchall...adebunk_a_review_of_mick_wests_virtual_model/

 

[Mick West]

The problem with this structure at strengthfactor 7.5 is that it does not stand up on its own: it cannot support itself for very long, and falls apart.

That's more of a function of the physics engine though. There's no reason at all that the models as build would collapse one their own if the physics was being calculated correctly. The problem is just tiny impulses getting magnified as they instantly propagate though the structure - something that should not happen with proper physics or reality. Hence you have to make the building stronger to correct for this.

It would probably be greatly improved if the columns were connected, and even more improved if the proper connections were used. That just needs a few hours work by someone.

 

[Cube Radio]

That's more of a function of the physics engine though. There's no reason at all that the models as build would collapse one their own if the physics was being calculated correctly.

To clarify: you believe that all of your model builds so far have been strong enough to stand up on their own and support their own weight, and in all cases it is entirely the fault of the Blender physics engine that they are observed to collapse when they are simply set up to stand on their own?

 

[Mick West]

To clarify: you believe that all of your model builds so far have been strong enough to stand up on their own and support their own weight, and in all cases it is entirely the fault of the Blender physics engine that they are observed to collapse when they are simply set up to stand on their own?

Yes.

 

[Cube Radio]

Presumably then you regret using Blender in the first place, as you cannot expect me to give you $100 for a virtual model that cannot be shown to support its own weight before collapse is initiated. Or do you feel that such an expectation or condition is unreasonable on my part? Do you in fact feel that I owe you $100 at this stage? That was certainly your assertion upthread.

 

[Mick West]

Presumably then you regret using Blender in the first place, as you cannot expect me to give you $100 for a virtual model that cannot be shown to support its own weight before collapse is initiated. Or do you feel that such an expectation or condition is unreasonable on my part? Do you in fact feel that I owe you $100 at this stage? That was certainly your assertion upthread.

It's entirely up to you. I'm not doing this for the money, so the precise $100 criteria is irrelevant.

My aim was to make a model that illustrates some aspects of the WTC collapse. It does that, but obviously there's loads of room for improvement.

I think at the very least it illustrates that it's possible for something that looks like the WTC towers to collapse top down. Can you see that?

 

[Cube Radio]

It's entirely up to you. I'm not doing this for the money, so the precise $100 criteria is irrelevant.

My aim was to make a model that illustrates some aspects of the WTC collapse. It does that, but obviously there's loads of room for improvement.

I think at the very least it illustrates that it's possible for something that looks like the WTC towers to collapse top down. Can you see that?

As deirdre observed earlier, computers can be made to represent anything.

The only thing we can all see clearly here is you can't make a tower in Blender that stands up under its own weight and also collapses like the towers did in any meaningfully representative way.

What I didn't expect is that you'd blame your tools.

 

[deirdre]

The only thing we can all see clearly here is you can't make a tower in Blender that stands up under its own weight

how come @aka 's model has a big gap (location A). ? and why after 1 min 20 would those two floors (B) 'fall' first?

 

[Mick West]

The only thing we can all see clearly here is you can't make a tower in Blender that stands up under its own weight and also collapses like the towers did in any meaningfully representative way.

What I didn't expect is that you'd blame your tools.

I'm just explaining why THAT model falls down.

Consider if you simply modeled (in Blender) 100 blocks stacked one upon the other. Would that tower fall if properly modeled? No - without any external impulse the tower should in theory be perfect stable.

My tower is no different, however the minute differences in floating point positions mean it's not exactly perfect, and the way impulses are propagated, and the lack of damping, and the way joint failures are implemented, all that adds up to what we see here.

Different approaches, an different tools, would yield different results.

 

[Cube Radio]

I'm just explaining why THAT model falls down.

All your Blender models fall down. Not just THAT one -- all of them are observably incapable of standing up.

Different approaches, an different tools, would yield different results.

The result we're looking for in the first instance is a model that can stand up on its own. I didn't realise that was going to be such a big problem for you.

 

[Mick West]

The result we're looking for in the first instance is a model that can stand up on its own. I didn't realise that was going to be such a big problem for you.

Well I would't exactly call it a "big problem". I put a few hours into building a model, that's how far I got, and I moved on to other things. I might get around to doing more work on it in the future.

Do you actually think it's impossible to model a building that both stands up, and then falls down when damaged?

 

[Cube Radio]

Do you actually think it's impossible to model a building that both stands up, and then falls down when damaged?

All the evidence in this thread is that it's quite impossible for you to build such a model using Blender. Of course you could build a different model that stands up and then falls down, but I put it to you that such a model would not meaningfully illustrate any aspects of the mechanism that you believe caused the twin towers to collapse so swiftly, smoothly and totally, with such destructive force.

If your models had been able to stand up, and then made to collapse in a manner that successfully illustrated the collapse mechanism you believe in to any reasonable degree, I would expect you and other members of Metabunk to heap ridicule upon me if I then refused to pay you your $100 on the basis that Blender's physics engine is so inadequate to the task at hand that its output is effectively meaningless.

 

[Nada Truther]

As deirdre observed earlier, computers can be made to represent anything.

The only thing we can all see clearly here is you can't make a tower in Blender that stands up under its own weight and also collapses like the towers did in any meaningfully representative way.

What I didn't expect is that you'd blame your tools.

Are you saying that Blender is no good for this, or Mick can't do it? If you are knocking Blender, why did you suggest it in the first place?

Blender is free, and the last time Mick said he could build in tower in less than a month (working part time) he insisted he'd do it for free when I suggested a fundraising effort.

But perhaps you didn't see that.

Maybe I am wrong, but it seems like you brought it up.

This thread, and really, your part in it... and most of what you say, Cube, seem like Typical Truther formula..... "That's not good enough.... You should do this... You should do that.... I won't do anything myself, or talk any of the over 1000 Architects and Engineers who allegedly support this claim, but I will demand that someone from the other side of it try to provide proof, so I can sit back and tear down every part as insufficient, so I can further "prove" my theory."

 

[Mick West]

All the evidence in this thread is that it's quite impossible for you to build such a model using Blender.

A few hours work on a model is hardly a lot of evidence. If evidence of anything, it's that I'm not incredibly interested in this topic. I may return to it at some later date though.