Modeling a progresive collapse. Is it possible?

[Quad4_72]

One main argument of the CFs is that NIST did not model the collapse after initiation. My question is, how would one go about modeling such an incredibly complex and unpredictable event? This question is mainly directed at the nutters, but if someone in touch with reality had a reasonable theory on how to do this I would be very interested in hearing your ideas.
(My apologies on the spelling error in the title.)

[Earthborn]

Well, a scientific model is by definition a simplified representation of reality. And scientists do model things that are even more incredibly complex. The Earth's climate for example. And they do it highly successfully. How hard it is to model it depends on how much you simplify it.

The complexity of the WTC collapse is therefore not the main reason why NIST didn't attempt to build a computer model of it. They probably figured with much simpler calculations (also a model in the scientific sense of the word!) that a collapse would have been highly probably given the condition the buildings were in and didn't see the need to explore the hypothesis that it might not have been possible without added explosives.

[leftysergeant]

It is feasible, but it would be horrendously expensive, I should think. There are too many variables to consider, some of them unknowable, such as the degree of damage done in the impact of the planes, the subsequent fires and the collision of structural elements with each other and surrounding structure. Then there is the variable loading of each floor, random distribution of weight pre-and-post-collapse. One could probably expect a floor to fall differently if it is loaded only with somewhat evenly-distributed office furnitiure than it would with a nasty great bank of lead-acid batteries in one corner.

[Quad4_72]

It seems that the twoofers want some sort of model that demonstrates what happened down to every single detail. This just seems to difficult to accomplish because there are just too many unknowns. Like I said though, I would love to hear a response from any of the nutters in this forum.

[Sizzler]

I would be happy if someone could create a tube-in-tube structure, made of any material, in which the top 15%, when raised the height of one or two scaled floors and dropped, would cause "global" collapse of the remaining 85% of the structure.

Any material would satisfy me.

I have even considered creating a prize of $1000 to be donated to a charity selected by the winner. That is, if anyone can make such a model.

[GregoryUrich]

I would be happy if someone could create a tube-in-tube structure, made of any material, in which the top 15%, when raised the height of one or two scaled floors and dropped, would cause "global" collapse of the remaining 85% of the structure.

Any material would satisfy me.

I have even considered creating a prize of $1000 to be donated to a charity selected by the winner. That is, if anyone can make such a model.

Lego.

[GregoryUrich]

It seems that the twoofers want some sort of model that demonstrates what happened down to every single detail. This just seems to difficult to accomplish because there are just too many unknowns. Like I said though, I would love to hear a response from any of the nutters in this forum.

Call me a nutter, just don't call me late for dinner.

I'm doing my best to try to educate people on a couple of truth sites.

The people who understand the physics (very few, but more than you would imagine) usually acknowledge that after initiation, complete collapse is inevitable. Unfortunately, the visual impressions and superficial CD arguments are very effective on people who only have high school physics. It's hard to get past peoples' intuition if they don't possess the tools to challenge it themselves.

Many still question the initiation though and many don't believe the fires could have heated the steel to the required 600 degrees C. Personally, I don't that temperature was necessary. I think almost no one is considering that the towers were most likely bent by the airplane impacts. Although I have not seen forensic evidence, the NIST impact model gives this result. I suspect that even the slightest bend would greatly reduce the load capacity. Newtons Bit of Rwguinn could probably speak to this point better than I.

[rwguinn]

Call me a nutter, just don't call me late for dinner.

I'm doing my best to try to educate people on a couple of truth sites.

The people who understand the physics (very few, but more than you would imagine) usually acknowledge that after initiation, complete collapse is inevitable. Unfortunately, the visual impressions and superficial CD arguments are very effective on people who only have high school physics. It's hard to get past peoples' intuition if they don't possess the tools to challenge it themselves.

Many still question the initiation though and many don't believe the fires could have heated the steel to the required 600 degrees C. Personally, I don't that temperature was necessary. I think almost no one is considering that the towers were most likely bent by the airplane impacts. Although I have not seen forensic evidence, the NIST impact model gives this result. I suspect that even the slightest bend would greatly reduce the load capacity. Newtons Bit of Rwguinn could probably speak to this point better than I.
Newton's Bit is the buckling expert, but yes--it takes very little distortion in any part of a beam (column) to cause loss of strength.

[Sizzler]

Lego.

You figure you could do it with lego? The upper 15% cannot be denser than the lower 85%.

How large of a base do you think it would need? How many units by how many units?

[e^n]

You figure you could do it with lego? The upper 15% cannot be denser than the lower 85%.

How large of a base do you think it would need? How many units by how many units?

What's important is you limit the slenderness of columns and the lateral connections with those columns. I think a bit of weak glue holding long and thing bits togther with belts of adhesive tape around the structure would probably work quite well, although the tape would radically exceed steel's ability in tension.

It is an interesting topic, I have enough space but not enough lego to give it a go :)

[ellindsey]

I have been experimenting with modeling progressive collapse using some highly simplified computer models. The physical modeling package I'm using is based on some open source collision and graphics code, along with some scripting I put in to simulate strength and failure modes of various materials. Most of this is actually aimed at physics simulation for computer gaming, so the code is optimized for speed rather than accuracy, but I figured it should be able to demonstrate how a failure in a small part of a building's structure can lead to a progressive collapse of the entire building.

Unfortunately, I've been running into hard limitations on the number of interactions and moving objects my code can simulate. I'm starting to think that an actually realistically modeled collapse of a WTC-sized building would be staggeringly difficult if not completely impossible. Furthermore being such a chaotic event even if fully and accurately modeled the simulated collapse still might not resemble the actual collapse closely enough to satisfy truthers.

[Myriad]

I have been experimenting with modeling progressive collapse using some highly simplified computer models. The physical modeling package I'm using is based on some open source collision and graphics code, along with some scripting I put in to simulate strength and failure modes of various materials. Most of this is actually aimed at physics simulation for computer gaming, so the code is optimized for speed rather than accuracy, but I figured it should be able to demonstrate how a failure in a small part of a building's structure can lead to a progressive collapse of the entire building.

Unfortunately, I've been running into hard limitations on the number of interactions and moving objects my code can simulate. I'm starting to think that an actually realistically modeled collapse of a WTC-sized building would be staggeringly difficult if not completely impossible. Furthermore being such a chaotic event even if fully and accurately modeled the simulated collapse still might not resemble the actual collapse closely enough to satisfy truthers.


Ellindsay, what happens if you drop it to two-dimensional cross section? It would become more of a conceptual illustration than an actual model, but some of the phenomena observed or hypothesized (such as peeling outward of the wall on the downtilt side, formation of a debris zone, and crush-down running ahead of crush-up) should still be replicable, and there might be some value in being able to see it happen in slow-motion detail.

Respectfully,
Myriad

[Myriad]

You figure you could do it with lego? The upper 15% cannot be denser than the lower 85%.

How large of a base do you think it would need? How many units by how many units?


Sizzler and I had been discussing similar ideas in the "Ice Model" thread (http://forums.randi.org/showthread.php?t=114132), when I think our exchange got interrupted by the forum outage.

Sizzler, you never answered my question about whether you'd permit a fixed guide structure (like a rod or taut cord) running vertically through the model unattached to it, to compensate for the lower ratio of angular momentum to torque on the upper block at reduced scale, by preventing the upper block from tumbling off in the model.

Also if you'd like to continue the old thread, I think it's your turn.

Respectfully,
Myriad

[ellindsey]

Ellindsay, what happens if you drop it to two-dimensional cross section? It would become more of a conceptual illustration than an actual model, but some of the phenomena observed or hypothesized (such as peeling outward of the wall on the downtilt side, formation of a debris zone, and crush-down running ahead of crush-up) should still be replicable, and there might be some value in being able to see it happen in slow-motion detail.

I've considered it. I don't know that the physics libraries I'm using can be made to operate in two dimensions. I'll look into it.

[Sizzler]

Sizzler and I had been discussing similar ideas in the "Ice Model" thread (http://forums.randi.org/showthread.php?t=114132), when I think our exchange got interrupted by the forum outage.

Sizzler, you never answered my question about whether you'd permit a fixed guide structure (like a rod or taut cord) running vertically through the model unattached to it, to compensate for the lower ratio of angular momentum to torque on the upper block at reduced scale, by preventing the upper block from tumbling off in the model.

Also if you'd like to continue the old thread, I think it's your turn.

Respectfully,
Myriad

yes that is right. i will respond to the old thread after i take the pooch for a walk:)

[Mancman]

Lego.

Yes.

http://www.brickshelf.com/gallery/srezkall/skyscrapers/wtc/31.jpg

http://www.brickshelf.com/cgi-bin/gallery.cgi?f=74081

[Profanz]

One main argument of the CFs is that NIST did not model the collapse after initiation. My question is, how would one go about modeling such an incredibly complex and unpredictable event? This question is mainly directed at the nutters, but if someone in touch with reality had a reasonable theory on how to do this I would be very interested in hearing your ideas.
(My apologies on the spelling error in the title.)

Sure it would work. By why would say NIST spend the money or time on yet another test that does absolutely nothing to support their hypothesis? Don't they have enough of those already? Some might say NIST quit while they were ahead. I say they quit before they got any further behind.

[WildCat]

The complexity of the WTC collapse is therefore not the main reason why NIST didn't attempt to build a computer model of it. They probably figured with much simpler calculations (also a model in the scientific sense of the word!) that a collapse would have been highly probably given the condition the buildings were in and didn't see the need to explore the hypothesis that it might not have been possible without added explosives.
You're right about that, but the truthers are demanding a model that shows every last detail of the collapse. They actually believe that the WTC should have arrested the collapse once initiated, or that it should have taken several minutes to collapse (the "clunkety clunk" model some truther idiot thought up).

[GregoryUrich]

You figure you could do it with lego? The upper 15% cannot be denser than the lower 85%.

How large of a base do you think it would need? How many units by how many units?

I would use 10 block high stacks of the 1x1, 1x2, 1x3 pieces to model the columns for each floor because they will have eccentricities and bend somewhat before coming apart in failure. I think it would be sufficient to model the core. The main issue is creating the core such that slightly more that 2x the load causes faliure. The complete structure should sway fairly easily. Remember that you still have to drop the top from the real height (3.7m) unless the strength is scaled to relative the drop height.

[Quad4_72]

Sure it would work. By why would say NIST spend the money or time on yet another test that does absolutely nothing to support their hypothesis? Don't they have enough of those already? Some might say NIST quit while they were ahead. I say they quit before they got any further behind.

Ok...this is the part where you give me your evidence that creating such a model would be possible.

[Profanz]

Ok...this is the part where you give me your evidence that creating such a model would be possible.

The fact that they modeled the plane crash is evidense they could model the collapse. Now if it's your claim that something like a plane crash and collapse initiation is less complex then the collapse itself then you provide evidence of that. Either they know for a fact that the global collapse is possible and can support it with facts and figures that can be applied to a model or they can't. And right now NIST claims they can't so how do you know any better?

[ktesibios]

The impact simulations done by NIST modeled only 8 floors in the impact areas. The building models for these eight floors contained over 1,300,000 nodes each with another 740,000 nodes for the aircraft models. To simulate only about 715 mS of each impact took approximately two weeks of computer time. (This information is found in NIST NCSTAR 1-2.)

Anyone claiming that NIST can and should model the collapses in their entirety ought to support their claim by estimating how many individual elements and nodes would be required to create a sufficiently fine-grained model of each tower, how long the computer run would take, how long it would take to develop and debug the models and what all that would cost.

They could go on to explain exactly what public good would be served by such an expenditure of the taxpayers' money. In the case of the impact simulations, gaining insight into how the crashing airplanes affected the towers' structures pretty clearly serves the purpose of gaining insight into how the towers ultimately failed, which in turn serves NIST's mandate to improve the survivability of high-rise buildings by developing information and recommendations for engineers and regulatory authorities.

What comparable public good can be achieved by a project which basically amounts to catering to a load of paranoid idiots who have already demonstrated that they will reject out of hand any conclusion which doesn't include George Bush with one hand on an old-fashioned rack-and-pinion detonator and the other twirling a Snidely Whiplash mustache?

[Profanz]

The impact simulations done by NIST modeled only 8 floors in the impact areas. The building models for these eight floors contained over 1,300,000 nodes each with another 740,000 nodes for the aircraft models. To simulate only about 715 mS of each impact took approximately two weeks of computer time. (This information is found in NIST NCSTAR 1-2.)

Anyone claiming that NIST can and should model the collapses in their entirety ought to support their claim by estimating how many individual elements and nodes would be required to create a sufficiently fine-grained model of each tower, how long the computer run would take, how long it would take to develop and debug the models and what all that would cost.

They could go on to explain exactly what public good would be served by such an expenditure of the taxpayers' money. In the case of the impact simulations, gaining insight into how the crashing airplanes affected the towers' structures pretty clearly serves the purpose of gaining insight into how the towers ultimately failed, which in turn serves NIST's mandate to improve the survivability of high-rise buildings by developing information and recommendations for engineers and regulatory authorities.

What comparable public good can be achieved by a project which basically amounts to catering to a load of paranoid idiots who have already demonstrated that they will reject out of hand any conclusion which doesn't include George Bush with one hand on an old-fashioned rack-and-pinion detonator and the other twirling a Snidely Whiplash mustache?

Well if they can model a plane crash and collapse initiation involving eight floors of a tower in two weeks then they should be able to model the rest of the collapse in about two months at the most. I mean if this was a progressive collapse then how different is the collapse of say floor 40 once it's modeled to say floor 39? And just how unique is a progressive collapse that happened twice in one morning? And as for what good would it do what good was modeling the plane crash and the collapse initiation? How about trying to get an idea on just how long firefighters and people might have to get out of a burning skyscraper before it falls on top of them for one? Would that cost too much tax payer money that could be better invested in something like this... http://usinfo.state.gov/media/misinformation.html?

[leftysergeant]

The amount of energy available to drive the collapse changed with each floor, at an unknown rate, because each floor was loaded differently with furniture and such. Any model would have to take into account how much debris, blocked from downward travel and spreading outward, was pushing against the perimeter columns. As the collapse progressed, those columns were forced outward with more force, amplified by leverage due to their height and weight. Again, because we do not know how much the debris weighed, it is almost impossible to calculate an energy budget from one floor to the next.

As for materials with which to build a scale model, Legos would work only if they were taller than long and were stacked in multiple spirals as were the perimeter columns.

Putting tape around the model would spoil the effect, because it would force the walls to bend outward uniformly on each floor. I don't see that happeneing in real life.

[e^n]

Well if they can model a plane crash and collapse initiation involving eight floors of a tower in two weeks then they should be able to model the rest of the collapse in about two months at the most.
No, the modelling done was largely in the non-linear region of deformation. Once the collapse progresses into plastic sections with impacts occuring the process is much slower. Bear in mind many of their computer models were simplified as best as possible to attain their speed. Truthers reject outright any simplifications it seems. It's simply not possible to plausibly model this collapse.

I mean if this was a progressive collapse then how different is the collapse of say floor 40 once it's modeled to say floor 39?
Floor 40 collapsed before floor 39, so as a result it requires the cumulative simulation results and the error margins from every floor above. Not to mention the building was modified a reasonable amount and the precise distribution of internal walls, partitions, office equipment etc is not well known.

And just how unique is a progressive collapse that happened twice in one morning?
Models which simplify out the unique elements have been created and predict collapse, presumably you're asking for an FEA simulation which includes a lot of unique factors per building.

And as for what good would it do what good was modeling the plane crash and the collapse initiation? How about trying to get an idea on just how long firefighters and people might have to get out of a burning skyscraper before it falls on top of them for one?
The point of the NIST report was to provide suggestions for engineering standards to prevent the collapse occuring. If they could not, a building such as the WTC would not be built. I don't think you really know enough about the topic to comment if this is your attitude.

[Profanz]

No, the modelling done was largely in the non-linear region of deformation. Once the collapse progresses into plastic sections with impacts occuring the process is much slower. Bear in mind many of their computer models were simplified as best as possible to attain their speed. Truthers reject outright any simplifications it seems. It's simply not possible to plausibly model this collapse.


Floor 40 collapsed before floor 39, so as a result it requires the cumulative simulation results and the error margins from every floor above. Not to mention the building was modified a reasonable amount and the precise distribution of internal walls, partitions, office equipment etc is not well known.


Models which simplify out the unique elements have been created and predict collapse, presumably you're asking for an FEA simulation which includes a lot of unique factors per building.


The point of the NIST report was to provide suggestions for engineering standards to prevent the collapse occuring. If they could not, a building such as the WTC would not be built. I don't think you really know enough about the topic to comment if this is your attitude.

Either you buy into the modeling that NIST did attempt and accept it as relevant enough that they could continue with it or you don’t. NIST claims they can’t do the collapse progression and I say the part they did do is irrelevant for the same reasons.

[e^n]

Either you buy into the modeling that NIST did attempt and accept it as relevant enough that they could continue with it or you don’t. NIST claims they can’t do the collapse progression and I say the part they did do is irrelevant for the same reasons.

False Dichotomy. I am aware of the problems with FEA modelling and the likelyhood of obtaining an accurate result. You are dismissing their initial modelling without considering the intent of the NIST report.

Tell me, how would modelling the collapse help improve the protection of life safety systems in the future?