Challenge: Demonstrate Sagging floor Trusses Pulling in Perimeter Columns

[jaydeehess]

@ozeco
well post 1 and post 4 assume that posters here be able to illustrate that they have the same competance as NIST researchers or Umani

 

[ergo]

Done. Usmani (Scottish structural engineer) shows that floor membrane tension for a three story WTC1,2 fire, even at relatively low temperatures (400C) and without plane damage is one cause of the perimeter columns failure.
Report done in 2003, before the NIST reports.

http://www.era.lib.ed.ac.uk/bitstream/1842/1216/1/WTCpaper.pdf

For the benefit of layfolk here, can you or anyone else cite very quickly where in this paper Usmani talks about the columns being pulled in by sagging trusses? I am looking specifically at the Discussion section of the paper, from pdf pages 26 to 31. Thanks.

 

[LSSBB]

For the benefit of layfolk here, can you or anyone else cite very quickly where in this paper Usmani talks about the columns being pulled in by sagging trusses? I am looking specifically at the Discussion section of the paper, from pdf pages 26 to 31. Thanks.

Third to last page, point 4:

As the membrane stiffness reduces (through either geometric or material effects), the outward movement of the column is arrested and the stored strain energy in the column makes it recoil with an increasing rate of inward displacement pushing the softened floor system back in.

 

[ozeco41]

For the benefit of layfolk here, can you or anyone else cite very quickly where in this paper Usmani talks about the columns being pulled in by sagging trusses? I am looking specifically at the Discussion section of the paper, from pdf pages 26 to 31. Thanks.

I had a very quick scan through the paper and I cannot find a reference to "pulled in".

The only reference I found in the range of pages you suggest and which seems to be relevant is this from page 31:

4. It is clear that the thermally expanding truss floor system initially pushes the column out (while it is sufficiently stiff in membrane behaviour). As the membrane stiffness reduces (through either geometric or material effects), the outward movement of the column is arrested and the stored strain energy in the column makes it recoil with an increasing rate of inward displacement pushing the softened floor system back in.

So the authors see it as columns pushing the floor membrane inwards rather than columns being pulled in by catenary sag of the floor.

E&OE as I said - I only took a quick scan through and could easily have missed something. However the quoted bit is diametrically opposed to the "catenary pull in" concept.

 

[ozeco41]

Third to last page, point 4:

Ooops.

You posted whilst I was reading the paper and writing my response.

One day I will learn to check for thread updates before I post.

 

[NoahFence]

Jesus then all this thread is, is some silly attempt at reversing the burden of proof?

You were expecting intelligence and honesty?

 

[ergo]

Ooops.

You posted whilst I was reading the paper and writing my response.

One day I will learn to check for thread updates before I post.

Are you aware that you quoted the same paragraph as LSSBB?

And you were right the first time:

So the authors see it as columns pushing the floor membrane inwards rather than columns being pulled in by catenary sag of the floor.

.... the quoted bit is diametrically opposed to the "catenary pull in" concept.

 

[ozeco41]

Are you aware that you quoted the same paragraph as LSSBB?..

Yes - after I posted. Hence my semi apology. But I did go a bit further with interpretation.

...And you were right the first time:

So the authors see it as columns pushing the floor membrane inwards rather than columns being pulled in by catenary sag of the floor.

.... the quoted bit is diametrically opposed to the "catenary pull in" concept.

Yes - I still have not read the paper in any detail but it looks like they come from the academic club where FEA results rank higher than observation and analysis of what really happened.

As you should be aware I belong to that minority group who prefer to explain what really happened rather than rely on academic abstractions.

I may read the paper more carefully but it is a long read.

 

[Oystein]

I had a very quick scan through the paper and I cannot find a reference to "pulled in".

The only reference I found in the range of pages you suggest and which seems to be relevant is this from page 31:
So the authors see it as columns pushing the floor membrane inwards rather than columns being pulled in by catenary sag of the floor.

E&OE as I said - I only took a quick scan through and could easily have missed something. However the quoted bit is diametrically opposed to the "catenary pull in" concept.

Number 4. on page 30 continues:

... A point to note here is that there is potentially a dynamic magnification effect here (not considered in this analysis) which could be another possible factor contributing to the instability. The
column eventually overshoots the original position significantly and the eccentricity thus caused is quite different from the relatively stable configuration when the floors were pushing the columns out. The floor meanwhile has been pushed beyond its original position and deflections therefore increase to the extent that the membrane compression now changes to membrane tension, adding further to the destabilising force on the column, potentially triggering the instability and progressive collapse (as seen in Figure 28)

So Usmani et al. describe the tension (pull) of downwardly deflected (sagged) floors as aggravating the inward bow and potentially being the decisive factor that triggers collapse. This however they could not establish as this would require a dynamic model; theirs was static.


This is how I understand their collapse initiation scenario:

1. As multiple floors burn and heat up, those in the middle expand much, while those above and below expand to a lesser degree.
2. This causes the middle floor trusses to push out against the exterior columns (the core columns are assumed fixed in the model!), resulting in lateral comnpression of the floor trusses.
3. Of course, Newton's 3rd, the force with which the truss pushes out is equal to the force with which the column pulls in
4. Since the floor trusses are mainly designed to carry vertical loads (office contents) and less so for lateral loads, and as they are very slender, their stiffness, even while relatively cool, is low, such that at some point they buckle.
5. This buckling reduces drastically their lateral capacity for compression, hence the force with which they push the column outward, and hence the column "wins": It accelerates inwards, initially with the reverse of the force that the floor truss previously pushed out.
6. Buckled floor no longer can provide lateral stiffness and now sags, exerting tension on the column. Column now bows inward, but is laterally unrestrained by floor truss.
7. With laterally unrestrained column length doubled, load capacity falls to 25%
8. Tension by sagging truss may likely aggravate this situation
9. Collapse ensues.

Makes some sense.
But doesn't take into account the 3D-situation; for example, when a floor truss buckles in compression, but its neighbors on the same floor do not, then would its column be free to move inward beyond plumb, or would the spandrels resist this? If the spandrels resist this, failure of one truss increases compression of neighboring trusses - setting the stage for failure progression.
Etc.

 

[ozeco41]

I would need to read it more carefully and thoroughly.

The countervailing comment which you took into account included this bit of disclaimer:

...potentially a dynamic magnification effect here (not considered in this analysis) which could be another...

I regarded the "(not considered in this analysis)" as negating or relegating to speculation what followed.

I was (and still am) confused by their multiple references to dynamic aspects.
[DIVERSION]
The visible evidence is clear - columns on one face at least bowed inwards to an extent sufficient to reduce their axial strength significantly. AND - subject to any contradictory argument - I would expect such inward bowing to be the result of self propagation of the bowing once it had been initiated. It wouldn't need the sagging floors to pull it in the full amount. That is my long held view - I have always regarded with suspicion any analysis either way which is based on catenary sag being the sole cause of inwards bowing.
[/DIVERSION.....

After that lengthy now somewhat shorter derail o) I am still uncomfortable that the Usmani et al paper is yet another academic abstraction which seems to totally ignore observations of reality.

You know where I will come from in such a setting. Reality wins over academic abstraction. Esp if reality negates academic. Sorry Bazant et al.

However thanks for the comments - you did manage to get their paper back towards "floor pull in" as a real possibility. I will need to give it more thought.

I also note your comments about three dimensions - it is years since I tried to explain possible cascade failure propagation . but it has to be explained in 3D - you don't get load shedding from column/joist to adjacent/nearby column/joist in a 2D simplification. And load shedding is the key factor in a cascade. So what price academic 2d abstractions dealing with an event which is essentially 3D? You can explain an individual mechanism. But you cannot also address propagation. Cascade in this case being multiple mechanisms (at least two) plus propagation of failure and load shedding. I won't even attempt to comment further at this stage.

 

[jaydeehess]

This non-engineer reads it as there first being outward push by heated floor membrane, on the columns. Then the floor membrane heats to the point where it loses its strength enough to begin sagging at which point the columns rebound inward exacerbating the sagging which in turn now pulls on the already inwardly deforming columns.

IMHO, from there we have a few choices;
-the perimeter coluimns fail first, collapsing the floors and dropping the upper section, loss of lateral support and extreme buffeting from heavy debris, as floors progresively collapse, causes core columns to fail.
-floor seats fail or trusses break away from seats, collapsing floors, loss of lateral support to both perimeter and core etc.
-core columns fail, floors collapse, perimeter columns fail, etc.

Its a dynamic event in which forces change over time.

 

[Oystein]

Well, ozeco, Usmani et al. don't descibe the "inward pull by sagging truss" as THE reason for inward boing, but merely as a possible contributing factor. It could, as I read them, turn out to be decisive in a more complex and realistic scenario (3D, dynamic...). At the least they do find that, after the trussed had buckled in compression and the columns sprang back from having been pushed outward, that the floor trusses provide increasing tensile forces. So those have to be reckoned with.

Their model predicts a modest outward bowing of perimeter columns by something like 15 mm. Would that be observable? Has it been observed? Has it been observed that this did NOT happen?

 

[Oystein]

This non-engineer reads it as there first being outward push by heated floor membrane, on the columns. Then the floor membrane heats to the point where it loses its strength enough to begin sagging at which point the columns rebound inward exacerbating the sagging which in turn now pulls on the already inwardly deforming columns.
...

Uhm no. This occurs usually at temperatures so low that the steel retains most of its strength. The more important factor in their model is the increasing compression of the trusses which are caught between columns that don't yield. The more the truss expands, the greater the force it exerts to the columns, that of course means the greater the force that the columns exert to compress the truss. Eventually, that force would exceed the capacity of even the cold truss. So, with increasing temperature of the truss steel, geometric change (expansion) beats material change (softening).

 

[jaydeehess]

Uhm no. This occurs usually at temperatures so low that the steel retains most of its strength. The more important factor in their model is the increasing compression of the trusses which are caught between columns that don't yield. The more the truss expands, the greater the force it exerts to the columns, that of course means the greater the force that the columns exert to compress the truss. Eventually, that force would exceed the capacity of even the cold truss. So, with increasing temperature of the truss steel, geometric change (expansion) beats material change (softening).

Ok, its a minor point really but I do want to try and understand in layman's language the order of events as Usmani sees it.
How's this?

-Membrane heats and expands
-Thus pushing on the core and perimeter columns
-Columns yeild very little giving rise to compressive force on constrained membrane
-Force increases to the point where compression causes floors to deform
-columns rebound

They are saying that the columns rebound well past their original vertical position and are kept there by the deformed/sagged trusses.

Am I correct now?


Why is this considered a more probable scenario than that the floor trusses sagged due to heat as is seen in numerous other office fires? Or is this simply the results shown by the FEA? The NYPD pictures show floors that are a red inferno, certainly temperatures well in excess of 500-600 degrees F and I can get a steel poker glowing red hot in a wood fueled campfire. Is it the concrete that remains much cooler due to it being less of a conductor than the steel, and thus the concrete that keeps the floor intact for a while? IIRC the top chord of the trusses were embedded in the concrete and only the lower chord was exposed.

Not sure that its very intuitive that the columns would rebound that far if they only deformed outward by 15 mm. ( though I realize that is calling in personal incredulity ) I suppose that the floors deforming downwards would have a certain momentum to them that would add to movement back inwards, of the columns and of course the truss span from core to perimeter is now shorter with trusses bent. Why they do not characterise this as a pulling inwards though is puzzling, or is that the tensile force in the membrane that they refer to?.

Truss span (that is the straight line distance between truss ends) gets shorter due to sag/deformation and once this exceeds the extra length of truss due to heat expansion, the columns are being pulled inwards. It seems that switching frames of reference to the floor membrane is much of what they did and refering to the tension in the deformed downwards membrane is pulling on the columns.

Holding a rope between two people and putting a weight in the centre of the rope, the people feel a pulling in, but the rope 'feels' it is being pulled apart(tensile force)

Then again I see that you say that they are simply saying that it was the geometric deformation of the membrane that allowed the columns to rebound and deform inwards and that pull-in was merely a possible contributor to that.

 

[Oystein]

Ok, its a minor point really but I do want to try and understand in layman's language the order of events as Usmani sees it.
How's this?

-Membrane heats and expands
-Thus pushing on the core and perimeter columns
-Columns yeild very little giving rise to compressive force on constrained membrane
-Force increases to the point where compression causes floors to deform
-columns rebound

Yes, so far this is pretty much how I understand this

They are saying that the columns rebound well past their original vertical position and are kept there by the deformed/sagged trusses.

Am I correct now?

The first part yes, that's what they say (although, like you, I don't quite understand why they rebound past the original position; I don't see that the velocities they give in the paper make this an issue of momentum or string-like vibration).
The second part I think no: I don't recall they say that the columns are kept in place at an inwardly bowed position by the pull of sagged trusses.

Why is this considered a more probable scenario than that the floor trusses sagged due to heat as is seen in numerous other office fires? Or is this simply the results shown by the FEA?

The latter.

The NYPD pictures show floors that are a red inferno, certainly temperatures well in excess of 500-600 degrees F and I can get a steel poker glowing red hot in a wood fueled campfire.

It is difficult to gauge actual temperatures within a building from just images of the fire. Usmani et al. ran their simulations for a wide range of (maximum) temperatures and found that the sequence as described above eventuated at lower as well as higher temperatures, so even if the temperature eventually increased enough to make the steel sag due to changed material properties, the inward bowing through geometric effects and buckling came earlier than sagging due to softening.

Is it the concrete that remains much cooler due to it being less of a conductor than the steel, and thus the concrete that keeps the floor intact for a while?

Usmani et al. did include the concrete flooring in their simulations. They considered a simple 1D modelling of heat flow, i.e that the surfaces heated before the inside of the concrete. In contrast, the truss steel, being rather flimsy and out in the open, was assumed to attain rising ambient temperature immediately.

IIRC the top chord of the trusses were embedded in the concrete and only the lower chord was exposed.

No. The concrete was on top of the metal decking, and the metal decking was on top of the top chord, so the top chord was out in the open. Only the "knuckles" (top bends) of the diagonal web bars extended a couple of inches or so into the concrete.

Not sure that its very intuitive that the columns would rebound that far if they only deformed outward by 15 mm. ( though I realize that is calling in personal incredulity )

I suppose that the floors deforming downwards would have a certain momentum to them that would add to movement back inwards, of the columns and of course the truss span from core to perimeter is now shorter with trusses bent. Why they do not characterise this as a pulling inwards though is puzzling, or is that the tensile force in the membrane that they refer to?.

According to some of the Figures in the paper, the main fire floor pushes outwards in their sims with a force on the order of 50,000 N - that's "5 tons" for laypeople
At the same time, the non-fire floors above and below experience tensile forces - going the other way - of a similar magnitude.
This reverses after the fire floor buckles and the column bows in.
The way I read those graphs, the pulling force eventually exceeds the earlier pushing force.

Truss span (that is the straight line distance between truss ends) gets shorter due to sag/deformation and once this exceeds the extra length of truss due to heat expansion, the columns are being pulled inwards. It seems that switching frames of reference to the floor membrane is much of what they did and refering to the tension in the deformed downwards membrane is pulling on the columns.

Yes, there is a pull on the columns, but, as the earlier push only resulted in a lateral deflection of the column by only 15 mm or so, you wouldn't expect them to bow inwards by more than that from an equal pull - all else being equal.
But all else is not equal: The buckled (failed) floor provides no (or insufficient) lateral bracing for the column in the inward direction (there is still, through tensile strength, bracing wrt outward bowing), and thus the vertical capacity drops sharply. The column is now free to bow inward even further due to its vertical load. To which the sagging truss contributes, but it's not the main factor, and perhaps not even necessary.

Holding a rope between two people and putting a weight in the centre of the rope, the people feel a pulling in, but the rope 'feels' it is being pulled apart(tensile force)

Then again I see that you say that they are simply saying that it was the geometric deformation of the membrane that allowed the columns to rebound and deform inwards and that pull-in was merely a possible contributor to that.

Right.



(Disclaimer: I am certainly as much a layman as you are - just trying to explain things the way I understand them. Don't trust me!)

 

[ozeco41]

Well, ozeco, Usmani et al. don't descibe the "inward pull by sagging truss" as THE reason for inward boing, but merely as a possible contributing factor. It could, as I read them, turn out to be decisive in a more complex and realistic scenario (3D, dynamic...). At the least they do find that, after the trussed had buckled in compression and the columns sprang back from having been pushed outward, that the floor trusses provide increasing tensile forces. So those have to be reckoned with.

Sure. It is a two different range of magnitudes issue. All the stuff they mention is in the few milimetres range - the fatal inward bowing in the 1.5 metre range. So my "engineer's gut feeling" automatically consigns all the lesser stuff to "second order" and not significant. I may be wrong. I've still only skip read that bit of the paper. Last night's effort with the assistance of a bottle of wine which may account for what I thought at the time was brilliant lucidity.

...Their model predicts a modest outward bowing of perimeter columns by something like 15 mm. Would that be observable? Has it been observed? Has it been observed that this did NOT happen?

Dunno either way. I doubt it.

And let's not forget we are on a side track to the OP.

 

[jaydeehess]

Yes, so far this is pretty much how I understand this


The first part yes, that's what they say (although, like you, I don't quite understand why they rebound past the original position; I don't see that the velocities they give in the paper make this an issue of momentum or string-like vibration).
The second part I think no: I don't recall they say that the columns are kept in place at an inwardly bowed position by the pull of sagged trusses.

I guess not, after all there is no force at all that is going to push the inward bowing columns back the other way and their load is now far from axial. They will continue to bow simply due to gravitational loading from above.

The latter.

Which I do understand can and does often demonstrate that forces that dominate are not neccessarily the ones that a person would assmue do.

It is difficult to gauge actual temperatures within a building from just images of the fire. Usmani et al. ran their simulations for a wide range of (maximum) temperatures and found that the sequence as described above eventuated at lower as well as higher temperatures, so even if the temperature eventually increased enough to make the steel sag due to changed material properties, the inward bowing through geometric effects and buckling came earlier than sagging due to softening.

Okay I'm getting this now. As I said its the FEA showing that the expansion of the membrane is the force that dominates early and does the damage before temps are that which would severely soften the steel.

Usmani et al. did include the concrete flooring in their simulations. They considered a simple 1D modelling of heat flow, i.e that the surfaces heated before the inside of the concrete. In contrast, the truss steel, being rather flimsy and out in the open, was assumed to attain rising ambient temperature immediately.

The concrete held itself together and the trusses still had enough strength to support the foor through the expansion period before compressive failure.

Interesting, as of course one expects that loss of strength would happen first.There's that personal incredulity aka 'common sense' that leads so many astray.

No. The concrete was on top of the metal decking, and the metal decking was on top of the top chord, so the top chord was out in the open. Only the "knuckles" (top bends) of the diagonal web bars extended a couple of inches or so into the concrete.

No I recall, yes. I knew I remembered that something extended into the concrete.

According to some of the Figures in the paper, the main fire floor pushes outwards in their sims with a force on the order of 50,000 N - that's "5 tons" for laypeople
At the same time, the non-fire floors above and below experience tensile forces - going the other way - of a similar magnitude.
This reverses after the fire floor buckles and the column bows in.
The way I read those graphs, the pulling force eventually exceeds the earlier pushing force.

Holy Carp that sounds like the world's biggest bow and arrow.

Yes, there is a pull on the columns, but, as the earlier push only resulted in a lateral deflection of the column by only 15 mm or so, you wouldn't expect them to bow inwards by more than that from an equal pull - all else being equal.
But all else is not equal: The buckled (failed) floor provides no (or insufficient) lateral bracing for the column in the inward direction (there is still, through tensile strength, bracing wrt outward bowing), and thus the vertical capacity drops sharply. The column is now free to bow inward even further due to its vertical load. To which the sagging truss contributes, but it's not the main factor, and perhaps not even necessary.

Not dissimilar to what occured to column 79 of WTC 7 when it loses lateral bracing on one side and would be free to bow/buckle in that direction.

So once the rebound began the columns bowed inward but as you say, unlike while bowing outward they now have no opposing bracing but still have the same vertical load and they buckle under load in the direction they have already moved.

Yes, off topic and perhaps feeding a troll (though he did say he would discuss such matters only with original authors ).

,,,,,,,,,,,,and still we wait for a description of what occurred in the towers wrt to therm?te and/or explosives with the same attention to detail as we see from NIST or, as we see above, Usmani.

and wait

and wait

and wait

 

[ergo]

It is difficult to gauge actual temperatures within a building from just images of the fire. Usmani et al. ran their simulations for a wide range of (maximum) temperatures and found that the sequence as described above eventuated at lower as well as higher temperatures, so even if the temperature eventually increased enough to make the steel sag due to changed material properties, the inward bowing through geometric effects and buckling came earlier than sagging due to softening.

Not only are they not saying that the columns were pulled in by sagging trusses, they also confirm what many observers have stated, which is:

It is very unlikely that the temperatures could have been that high in the interior of compartments of such a large aspect ratio (height to width) and the consequent resistance to ventilation it would create ... The video footage of the events shows dark smoke spewing from the openings caused by the initial impacts for the whole duration while the buildings remained erect. Thus, the authors favour the hypothesis that such enhanced ventilation did not exist.

The external columns had three of their faces open to atmosphere that in the absence of external flaming is inconsistent with high temperatures.

Hmm. But what about the Very Hot Fires and the Hurricane Force Winds and the argument that "Lots of smoke means lots of FIRE"? Could 9/11 bedunkers be wrong??


What else do they say?

The FEMA report also indicates that the despite the structural damage, the remaining columns retained significant load carrying capacity (as most of the external structure was designed to resist very high wind loading and in the absence of wind, would have provided considerable redundancy for gravity loads). The utilisation factor of for gravity loading is reported to be a very low 20% for exterior columns and a moderate 60% for the core columns. Considering these arguments (which clearly were not available to the authors at the time of the publication of their paper), the scenario of a large number of columns suffering creep buckling almost simultaneously is not credible.

Oops! Is Usmani bedunking Bazant?? Uh oh!

Furthermore, no gross assumptions about structural behaviour have been made (such as the persistent claims about connection failures and resultant ‘pancaking’ of floors, without any supporting quantitative analysis).

Ouch. So many bedunker arguments trashed in one fell swoop. No wonder you guys get so confused about your WTC collapse-and-grind models!

 

[Newtons Bit]

I see Gamolon has issued a challenge in another thread. I will offer an even simpler one. Show with your own FEA or calculations how the perimeter columns of WTC 1 were pulled in by up to 54” by sagging floor trusses.

Bonus, demonstrate how this instability caused the initiation sequence for collapse of WTC 1.

I created a thread about it back in 2008. Calculations are done. What's my prize?

 

[ergo]

Does that thread answer anywhere the question of how the truss-to-column connections can be so robust that the floor trusses, while sagging something like 40" (a feat in itself, without breaking) can pull in 14" steel box columns, causing them to break, but simultaneously be so flimsy as to then be unable to prevent rapid progressive floor collapse? Just curious.