How Buckling Led to "Free Fall" acceleration for part of WTC7's Collapse.

Thomas B

Active Member
Okay, I think I understand.
a) That the loss of support occurred below the Red Line and
Agreed.
(b) involved three (3) factors, two of which are partially overlapping.
Three important things for me are that

(1) the loss of support propagated laterally across the entire facade before any of the columns above the red line began to move downwards, and
(2) the columns above the red line and up to the roof moved downwards together, in parallel, which is why this region can be considered a black box in the load path diagram, and
(3) the mechanism of this propagation was load redistribution after the failure of either a column or a floor connection.
 

econ41

Senior Member
Okay, I think I understand.
Great.
Agreed. Three important things for me are that
The "three factors" I was referring to are (i) buckling; (ii) cantilever failure and (iii) Transfer tRuss failure (and the last two are related AFAIK.)

Now, addressing your "Three important things":
Three important things for me are that

(1) the loss of support propagated laterally across the entire facade before any of the columns above the red line began to move downwards,
Yes. That is the sequential cascading process that is driven by load redistribution. It was very fast but still was - actually "had to be" - sequential. That is the only way it can progress for reasons commented on in several recent posts.
and
(2) the columns above the red line and up to the roof moved downwards together, in parallel, which is why this region can be considered a black box in the load path diagram, and
Yes
(3) the mechanism of this propagation was load redistribution
Yes
after the failure of either a column
Yes
or a floor connection.
Maybe. I've been referring to "cantilever" and/or "transfer truss" failures. We will need to work out the details - @Jeffrey Orling understands the details better than I do - but I've been agreeing with him on TTF as ONE plausible sub-mechanism for probably 10 years.
 
Last edited:

Jeffrey Orling

Senior Member
So it might be the time to discuss how the load transfers on floors 5-7 might have or likely failed.

The starting assumption is that there were "failures" at / around column 79, 80, & 81 on floors 12-14 (actual floors not critical). The failures involved:
over heated beams and girders framed into these columns
those beams and girders came expanded and were pushed off their beat seats
this caused all the floors they supported to drop
it also left these columns unbraced and unstable/weaker
The columns likely bucked and this led to the drop of the East penthouse
all the floors slabs and contents on the floors above 12-14 came crashing down
this mass crashed down onto floors 5-7 containing mech equipment and a number of transfer structures:
TT1, TT2
TT1 and TT2 were dislodged by the onslaught of 40 floors of debris - slabs and steel beams and floor contents
TT1 and TT2 were framed into the East side of the core
The core had "belt girders" around it which braced the core and supported the core side of the floor slabs
The belt girders were dislodged when TT1 and TT2 collapses
The north belt girder supported 8 - MG27 which were spanning over the Con Ed and supported the most of the north side of the moment frame on the cantilever
the building"s perimeter contained a belt truss from floors 5-7 (rigid) which supported the moment frame
collapse of the MG27, "disturbed" the belt truss
the cores belt truss collapses/disturbances led to the core's disintegration, the disturbance of TT3 and that led to all floors losing core side support. - they dropped
the entire interior of the building has lost support and dropped down inside the intact perimeter and moment frame
the debris could not be contained by the steel structure below the red line and especially at the base of the tower.
The lateral force of the debris cause the structures below the moment frame to buckle which was loaded by the 40 story moment frame
the moment frame collapse at almost FFA

The steps proceeded one to the next very rapidly and were run away and unstoppable.
 

Thomas B

Active Member
The steps proceeded one to the next very rapidly and were run away and unstoppable.
On page 64 of the NIST report:

Had contemporaneous standards and practices been available to expressly
design WTC 7 for prevention of fire-induced progressive collapse, it would have been sufficiently
robust to withstand local failure due to the fires without suffering total collapse.
Content from External Source
And page 65:
Possible options for developing cost-effective fixes include:
• More robust connections and framing systems to better resist the effects of thermal expansion
on the structural system.
• Structural systems expressly designed to prevent progressive collapse.
Content from External Source
https://nvlpubs.nist.gov/nistpubs/Legacy/NCSTAR/ncstar1a.pdf

My hope is that we can identify the (theoretical) connections and systems that could have been made more robust and therefore prevented the collapse. That, after all, will give us a model of the building as truthers imagine it actually was. According to NIST, while it wasn't that way in reality, it may have been a "cost-effective" possibility.

That's why I'd like to see some load path diagrams with the connection strengths noted. How much stronger, and where, would they have had to be?
 

econ41

Senior Member
Take care @Thomas B - you are drifting into losing the focus:
My hope is that we can identify the (theoretical) connections and systems that could have been made more robust and therefore prevented the collapse.
Which "collapse"? We are discussing the cause of the FFA portion of the collapse of the perimeter shell of WTC7. NOT the overall collapse of WTC7. The "...connections and systems that could have been made more robust.." are the ones that could have "...prevented the collapse.." of the whole building. Probably by making the bracing of Col 79 more robust. Then, if the building did not collapse, there is no point in making the buckling columns, cantilevers or transfer trusses stronger to resist the FFA of that final stage of collapse. Because the situation would not arise.
That, after all, will give us a model of the building as truthers imagine it actually was.
I like your generosity to "Truthers". It is misplaced. The genuine truth-seeking truthers needed assurance that the collapses were caused by fires and not by CD. That truth was explained in layman's language by about 2009-10. From that time genuine truthers gradually lost interest and there are few genuine truthers still active. The deliberately dishonest "truthers" won't be satisfied by any reasoned argument. And those who are genuinely CT obsessed won't be released from their need to be obsessed. Facts and truth are irrelevant to them - their problem is a psychological pathology that is not resolved by arguments of technical truth.
According to NIST, while it wasn't that way in reality, it may have been a "cost-effective" possibility.
Yes. But don't interpret NIST out of context. They were referring to preventing the collapse of the building - not preventing FFA of part of the building part way through a collapse.
That's why I'd like to see some load path diagrams with the connection strengths noted. How much stronger, and where, would they have had to be?
What do you want them to achieve? If you refer to the points of vulnerability of WTC7 it is a separate topic - off-topic for this thread. And, if I understand you, your goal is even larger. Viz "To build buildings that cannot collapse due to fires". And that raises that other topic where our discussions are unresolved. Also not a topic for this thread.

@Jeffrey Orling is "on-topic" with his suggestions as to what caused the FFA of the perimeter shell of wTC7.
 
Last edited:

Thomas B

Active Member
Take care @Thomas B - you are drifting into losing the focus:
Sorry, Econ. Please don't let me distract you. You have a plan for how to address the OP in a step-by-step way, and I'm looking forward to seeing how you proceed.

The "...connections and systems that could have been made more robust.." are the ones that could have "...prevented the collapse.." of the whole building. Probably by making the bracing of Col 79 more robust.
My understanding of NIST's proposal is that an improved design would have slightly (i.e, cost-effectively) stonger connections at all points. The assumption would be that one would not be able to predict which part of the structure might fail, and one would want the spread of failures to be arrested from any point of origin.

I like your generosity to "Truthers". It is misplaced. The genuine truth-seeking truthers needed assurance that the collapses were caused by fires and not by CD.
I'm primarily intersted in people who find the truthers' argument that steel-framed buildings can't collapse like WTC7 (unless most of their columns are cut simutaneously) compelling. Showing them a load path diagram of a building that indeed wouldn't collapse from a local failure, and then weakening it slightly (without making it into a house of cards with respect to, say, wind loads), resulting in a building that would totally collapse from local failure, would help such people.
@Jeffrey Orling is "on-topic" with his suggestions as to what caused the FFA of the perimeter shell of wTC7.
Agreed. I take the last stage of the collapse (the free fall of the perimeter shell) as problem that can indeed be isolated (as in this thread). We have a structure (the perimeter) that is stable (i.e., the roofline hasn't moved) after the core has failed and is then damaged (by the ongoing destruction inside the building) which iniates a rapid sequence of failures that causes the "red line" to descend uniformly across the whole facade.

It's that (theoretical) sequence of failures, driven by load redistribution, that I want to understand.
 

econ41

Senior Member
Hi @Thomas B - I'll take them one at a time for the three themes. So this one first:
Sorry, Econ. Please don't let me distract you. You have a plan for how to address the OP in a step-by-step way, and I'm looking forward to seeing how you proceed.
At this stage we have identified three plausible contributions to "removing the support to allow FFA":
(i) Column buckling which is the NIST explanation presumed by the OP. I haven't checked out exactly what that means or to what degree it is true.

(ii) Cantilever failures. The one that has always been the prima facie most obvious to me. Because it near-instantly creates the necessary 7-8 storey "gap" to allow the free fall. BUT I have never looked into thedetails esp how many columns it would "release to fall"; and

(iii) Transfer Truss failures which @Jeffrey Orling has for many years been identifying as a plausible cause. We have agreed on "plausible" possibly for as long as 10 years in discussions on several forums. Jeffrey tends to see it as a sole cause. I've never been assured of how it related to column buckling. And I have never checked out the relationship with cantilever failure - since I have seen cantilevers as part of the transfer bridging structures over the Con Ed Sub.

So, IF we want to go further. we need to get clear what is meant by all three and how many columns each one affects. Most of my interest is resolved by the recognition that column buckling was probably not the sole cause of the FFA. And I doubt we can ever know all the details. Maybe we could clear up our understanding of the balance of contributions from the two or three sub-mechanisms. I doubt that column buckling was the sole cause. I cannot prove that at this stage.

SO - There are three further steps needed (If we have sufficient reason to keep going. And I don't):
(1) Identify which columns are alleged to have buckled;
(2) Clear up our shared understanding of the relationship between TT failure and cantilever failure. Identify how many columns failed from either as the sole cause or from a combination of the two.
(3) Decide whether we need to disagree with NIST. << and that is often not a popular choice. ;)
Agreed. I take the last stage of the collapse (the free fall of the perimeter shell) as problem that can indeed be isolated (as in this thread). We have a structure (the perimeter) that is stable (i.e., the roofline hasn't moved) after the core has failed and is then damaged (by the ongoing destruction inside the building) which iniates a rapid sequence of failures that causes the "red line" to descend uniformly across the whole facade.
That ^^^ paragraph restates what you and I, probably Jeffrey, have agreed on.
It's that (theoretical) sequence of failures, driven by load redistribution, that I want to understand.
Understanding that it was some combination of three sub-mechanisms contributing to cascading failure by load re-distribution is probably as far as I am interested for a pragmatic reason. I doubt that we can definitively say much more other than broadly quantify the proportions of contributions. (If any - we still haven't "proved" the principle.)
 

econ41

Senior Member
@Thomas B - the second "theme":
My understanding of NIST's proposal is that an improved design would have slightly (i.e, cost-effectively) stonger connections at all points. The assumption would be that one would not be able to predict which part of the structure might fail, and one would want the spread of failures to be arrested from any point of origin.
That opens up a can of undefined worms. Including the key points of the other contentious discussion, we have had elsewhere. Including the "New Paradigm" vision put forward by Prof Torero including both what he says and what he doesn't say. So it is a discussion for another thread if you ever want it.

The goal of "..the spread of failures to be arrested from any point of origin." is "pie in the sky". There will never be a building design process that can guarantee a building to withstand anything that can be thrown at it. That is the factor that Torero does not mention in his address and for valid reasons. The bottom-line - buildings will always need to be designed to withstand a target scenario. One legitimate target which the WTC Towers were not designed for could be "Should withstand burnout under any predictable office fire situations". That is central to Torero's suggestions and near enough what he means, Discussion of that topic should be in an appropriate thread as I have commented several times.

Two more points: I doubt that making all connections stronger would ever be cost-effective. And it is a post hoc solution for WTC7 - not globally applicable. Making other buildings "burn out proof for office fires" may require different solutions - other than stronger connections.
 

econ41

Senior Member
@Thomas B - and the third theme:
I'm primarily intersted in people who find the truthers' argument that steel-framed buildings can't collapse like WTC7 (unless most of their columns are cut simutaneously) compelling. Showing them a load path diagram of a building that indeed wouldn't collapse from a local failure, and then weakening it slightly (without making it into a house of cards with respect to, say, wind loads), resulting in a building that would totally collapse from local failure, would help such people.
You need to be pragmatic when addressing: "people who find the truthers' argument...compelling". They will either be open to reasoned explanation or already conspiracy-obsessed. For those who are CT obsessed no technical argument is likely to help. And certainly, not one that is technically complex.

Keep the explanations simple: They will be interested in (or pretend interest in) the WTC 9/11 collapses. Explain that the WTC Towers were steel framed buildings which are known to be vulnerable to fire. They are designed with a fire rating (so explain that also). All three WTC towers had fires far more extensive than they were designed for and those fires were not actively fought.

Any rational non CD obsessed person should accept that explanation - possibly with a few more details. You won't persuade the pathological psychology of a truly obsessed Conspiracy THeorist. Read Mick West's book for a good easy-to-read explanation. And Mick is slightly more optimistic than I am.
 

Thomas B

Active Member
Understanding that it was some combination of three sub-mechanisms contributing to cascading failure by load re-distribution is probably as far as I am interested for a pragmatic reason. I doubt that we can definitively say much more other than broadly quantify the proportions of contributions.
Okay, it sounds like I'm not going to get my load path diagrams. I'll keep looking elsewhere.

There will never be a building design process that can guarantee a building to withstand anything that can be thrown at it.
I'm talking about a design (described in a load path diagram) that would survive the failure of any one element (or a few elements in a localised area).

Any rational non CD obsessed person should accept that explanation - possibly with a few more details.
My view is that any rational person who believes the collapses are explicable (without CD), would imagine that the load path diagram I'm looking for is possible to draw. The inability to find such a diagram as part of the discussion is unnecessarily odd.
 

econ41

Senior Member
Okay, it sounds like I'm not going to get my load path diagrams. I'll keep looking elsewhere.
Best of luck. Why do you think it is plausible? (Or that you can find someone who understands the topic and is still prepared to provide a diagram?)
I'm talking about a design (described in a load path diagram) that would survive the failure of any one element (or a few elements in a localised area).
That is your immediate goal and specific to this discussion. My comment referred to your bigger picture goal as I clearly stated.
My view is that any rational person who believes the collapses are explicable (without CD), would imagine that the load path diagram I'm looking for is possible to draw.
I'm aware of your view. Your global claim "any rational person" is unsupportable. Not every rational person thinks that way. Some may. And any "rational person" who adequately understands the topic would know the true status with possible "load paths"
The inability to find such a diagram as part of the discussion is unnecessarily odd.
Why? I've explained. If you disagree - stay so and be prepared to show why your position is more realistic than mine.
 

Thomas B

Active Member
Not every rational person thinks that way. Some may. And any "rational person" who adequately understands the topic would know the true status with possible "load paths"
Are you saying* that a rational person who understands what a load path diagram is would not find it odd to hear that it is impossible to draw one for WTC7? In any case, I'm trying to reach an "adequate understanding" so that I can draw the load the paths. I think such a diagram would put at least some minds at ease.

*Edit: For example when you say this:
Why do you think it is plausible? (Or that you can find someone who understands the topic and is still prepared to provide a diagram?)
I think it would be odd to meet someone who claims to understand the subject but is not prepared to draw a simplified load path diagram of a purported progressive collapse whose main mechanism is load redistribution after local failure.
 
Last edited:

econ41

Senior Member
Are you saying that a rational person who understands what a load path diagram is
Take care - you have already changed the target group towards what I"m advising.
would not find it odd to hear that it is impossible to draw one for WTC7?
You can draw one. Will it correctly explain whatever you want to show? BTW - I'm suspecting that you have again lost focus and are trying to draw the wrong load path. Let's check. Respond by telling me what the load path would show - what failure???
In any case, I'm trying to reach an "adequate understanding" so that I can draw the load the paths. I think such a diagram would put at least some minds at ease.
Could it be "correct"? And again you shift the target to "some". How many? What proportion of the demographic you have specified?
 

econ41

Senior Member
Well, @Thomas B, it looks like all the recent off-topic meanderings have been removed together with a couple of my attempts to close off the rambling and focus back on the topic.

Probably the most relevant on topic concern you are expressing is your preference for a "load path diagram" to help move the discussion forward by your preferred method. I've given some reasons why it is not practical. I'll make this one more attempt at showing why.

The status of the "on-topic" debate is that you and I are agreed that there are three plausible contributing sub-mechanisms causing the FFA of the perimeter shell identified in the OP. That means we potentially disagree with the OP inference that column buckling was the sole cause of FFA.

We also agree that, however we more clearly define those three causes and their interrelationships, the driving process was one of load re-distribution which must be a sequential process.

You want a load diagram which is a means of defining a specific situation as the (singular) process the event would follow. Load redistribution is generic. For events such as the OP event of FFA (and also the example we have quoted - the initiation stage of Twin Towers collapses) we cannot - never will be able to - specify the specific sequence of failures.

We can rigorously explain "How the first column failed" and "How that led to second column failure" and "Third" etc etc. BUT we will never know which specific column was first. So we cannot draw a specific load diagram. We can draw a generic one. I've already presented a couple of generic diagrams to explain load re-distribution.

And this is where the overriding barrier appears. Some people can readily relate a generic explanation to a specific situation or event. And some people cannot. I have zero difficulties arguing the reasons why the first columns failed. BUT I know from multiple experiences of explaining WTC collapses that as soon as I say "the first columns" many people will ask "Which column was first" because they simply cannot process a generic argument. They have to know that it was "Column Q14" or "The third column in the second row." And given the chaos of both examples of WTC events, we will never know"Which was first". << So read that and understand but don't risk going off-topic please.

So there is the bind. We will not be able to produce a valid load path diagram that is more specific than the style of generic diagram which I have already presented. And a lot of people cannot process generic answers or relate generic models to an unpredictable situation.

We can take one more step in this discussion - the one I have already identified. viz define the details of how the three mechanisms operated and how many columns each mechanism could have made vulnerable to failure. THEN, if we can establish that column buckling was NOT the sole cause - we may be able to ballpark estimate the proportion that each contributed to the collapse.
 
Last edited:

Thomas B

Active Member
We also agree that, however we more clearly define those three causes and their interrelationships, the driving process was one of load re-distribution which must be a sequential process.

You want a load diagram
I'm happy to leave it there. I'm going to try to draw a plausible load path diagram of the facade, after the core had collapsed but before the roofline began to fall. It's going to take a little while to learn how. See you then.
 

econ41

Senior Member
I'm happy to leave it there.
No problem - it's your call. It might be appropriate if you read the full thread and put our recent narrowly focused discussion in perspective.

I've already noted that the OP seems to assume that the NIST explanation of "column buckling" is correct. As you may remember my SOP for over 15 years has been agnostic towards NIST's (or Bazant et al's) explanatory reasoning. I also agree with @Jeffrey Orling that the 7-8 storeys of FFA automatically suggests that the bridging arrangement over the Con Ed substation should be one of the first places to look. Hence the focus of our discussions.

But the NIST explanation is the "on-topic" focus of the thread. And, AFAICS, NIST relies totally on its simulation which shows column buckling in the floors above the sub-station bridging - floors 7 to 14 or thereabouts. IF failure originated in those levels then column buckling as a single cause is feasible. Whilst my SOP is never to treat NIST, Bazant or other "authorities" as infallible NIST's hypothesis, in this case, looks both plausible and probable.

So that complicated your bid for a "load path diagram" because you will need to EITHER prepare one for each of the two competing scenarios OR decide which scenario you want to prove by your load-path diagram.

Whichever you pick I still hold to my reservations about the challenges you face in attempting to prepare such a diagram.
 
Last edited:

Jeffrey Orling

Senior Member
I suggest that interested members agree on a possible/probable sequence of failures.

When complex systems fail it is often difficult how the failure will progress. Column failures progress as Econ's diagram shows. However it should be noted... that in addition to column buckling/failure there was beam and girder failure which resulted in local floor collapse... (pancaking) and that would unload (bypass) columns. But it also likely leads to Euler buckling.

So the sequence might have been:
Heating on floors 12-14 (or thereabouts) on the East side of the building
Cols 79, 80 & 81 lose their bracing.. and buckle when beams and girders expand and move from heating
EPH loses support and drops
Floors on the east side lose connection to columns and drop
Massive east side floor collapse crashes down of TT1 & TT2
TT1 & TT2 involve the girders surrounding the core (magenta)
TT3 is disturbed and collapses
West side floors collapse with TT3
MG27s north of the core collapse
North side of moment frame has no support as it was on the end of the MG27a
Core columns lose bracing and buckle
Beams and girder framing from the core to the moment frame lose core side support and collapse
Floors south, west, north and east collapse.
Building is hollowed out
Debris pile from the entire building builds at the base and exerts lateral force on the columns of the perimeter.
Columns of the perimeter at the base lose lateral support and buckle - Moment frame has no support
Moment frame and attached curtain wall decent as a hollow rigid tube at FFA for 7 stories (2.25 sec)
 

Thomas B

Active Member
So that complicated your bid for a "load path diagram" because you will need to EITHER prepare one for each of the two competing scenarios OR decide which scenario you want to prove by your load-path diagram.
I may not have made myself entirely clear. The diagram I'm imagining is going to be "theoretical".

It's going to resemble the real WTC7 about as much as your A-G diagram. Like your diagram, it will be 2D. It's just going to have 10-story columns, and floor trusses between the columns rather than a rigid beam across the top.

With 7 columns, there'll be 60 floor trusses, each connected to two columns.

If we call the load on each floor truss L, the load on each floor connection will be .5L and the load on the top column sections .5L for the perimeter columns and 1L for each of the internal columns, 6L altogether spread over 7 columns (I don't know, maybe the load would be spread more evenly on all the columns?).

The next floor down in its 12L, then 18L, then 24L, etc. (again distributed in some manner over 7 columns).

We then introduce some damage. Like, we remove a floor truss on each side of column C, which doubles its length and therefore reduces its capacity while removing only 2L from its lower portion.

Presumably the loads and safety factors can be determined under which that column section would fail. And the theoretical result of this failure on the rest of the structure can be traced. That's all I'm doing.

The kind of connection (especially its lateral strength) is something I haven't got a handle on yet.

In any case, I'm not trying to prove that any particular scenario happened. Only what such a progression might look like "in theory".
 

Jeffrey Orling

Senior Member
I may not have made myself entirely clear. The diagram I'm imagining is going to be "theoretical".

It's going to resemble the real WTC7 about as much as your A-G diagram. Like your diagram, it will be 2D. It's just going to have 10-story columns, and floor trusses between the columns rather than a rigid beam across the top.

With 7 columns, there'll be 60 floor trusses, each connected to two columns.

If we call the load on each floor truss L, the load on each floor connection will be .5L and the load on the top column sections .5L for the perimeter columns and 1L for each of the internal columns, 6L altogether spread over 7 columns (I don't know, maybe the load would be spread more evenly on all the columns?).

The next floor down in its 12L, then 18L, then 24L, etc. (again distributed in some manner over 7 columns).

We then introduce some damage. Like, we remove a floor truss on each side of column C, which doubles its length and therefore reduces its capacity while removing only 2L from its lower portion.

Presumably the loads and safety factors can be determined under which that column section would fail. And the theoretical result of this failure on the rest of the structure can be traced. That's all I'm doing.

The kind of connection (especially its lateral strength) is something I haven't got a handle on yet.

In any case, I'm not trying to prove that any particular scenario happened. Only what such a progression might look like "in theory".
All beams, trusses and girders are framed to the SIDE of columns... not the tops of columns.
 
Top