Summary of WTC tower collapses

jaydeehess

Penultimate Amazing
Joined
Nov 15, 2006
Messages
20,849
Location
40 miles north of the border
Given that there is great debate about using an acronym to summarize the collapses, ROOSD for eg., I wrote out a summary of my own. I invite correction and suggestion.
http://www.internationalskeptics.com/forums/showpost.php?p=11207104&postcount=2065

Generic Description of the Collapse Progression of the Two WTC Towers:

Video evidence demonstrates aircraft impact destroyed or bent a number of perimeter columns. A lesser amount of core column severance or damage is expected. Some floor spans also directly damaged.

Aircraft fuel spread through several vertically adjacent floors of the structure and ignited. This created a condition of very large area office contents fire on those floors with various levels of rubblization of the contents. This is a fire condition rarely if ever, reached within seconds of fire initiation in large office structures.

Aircraft impact debrided spray on fire insulation in varying degree, on a significant number of structural components on impact floors. This reduced fire resistance of those structural steel components.

Structural response to impact was a redistribution of loads previously carried by missing or foreshortened ( due to being bent) columns. Load redistribution was not even across all remaining columns.

Fire spread and movement was similar, but not identical, on each involved level and also spread to higher levels. Heat began affecting steel structural components, both vertical and horizontal, and the connecting components between them on all fire involved levels.

Floor spans had long trusses with their lower chords exposed to the heat and the trusses in several areas began sagging. A few areas had already seen truss or truss to perimeter connection failure at impact.

Perimeter columns and truss to perimeter connecting components heated less than the core columns and truss to core structure components. The interface between long span trusses and core was a belt truss rather than direct connection to core columns.

Core columns are subject to heating basically from all sides, over the several floors fully involved in the fires, and over the course of fire spread around each level.
( that is to say a column at level x may heat more on one face than that same column at level x-1 or x+1 but that over the several levels involved some columns are heated on all sides and taking into account the changes through time)
The capability of the structure to resist any collapse was being eroded. One very obvious indication of this is the inward bowing of part of the perimeter. Each tier also twisted and tilted noticeably up to collapse. These demonstrate that the structure was responding to further weakening due to the heat. Loads were continually changing and redistributing during this time.
At some point a redistribution caused one or more columns to exceed its capacity. Column capacity in this case does not necessarily mean it original engineered load bearing ability. Loads were not necessarily on axis and columns were not necessarily at room temperature, both of which would tend to lower any individual column's strength.

At this point a rapid redistribution the remains structure occurred, rapidly causing other columns to exceed capacity, causing another rapid load shift and column failures progressed to all columns over one or more levels.

A large section of the upper building then falls, accelerating to next contact. A combination of factors result in few if any buckled columns impact on lower parts of themselves. Upper falling section column ends are the first contacts with lower structure. That contact would be proximate to the lower section columns and thus the greatest impact occurs near or directly on lower floor truss to column connecting components.

This is occurring both to long span truss to column connecting components, at perimeter and belt truss, as well as with the core's beams running between core columns.

As upper mass continues its lowest floor begins contacting lower floors. This first contact is between floor systems that are already severely compromised. They are among those that have been affected by impact and heat and one has suffered loss of connections to columns as per last paragraph.

Both upper structure and lower structure now have major connection failures, both in the vertical/horizontal connections as well as connections between column sections. The later is due to both the destruction of Lateral Support and the buffeting collisions from the falling mass.
Perimeter column "trees" peel away, the first to do so are on the order of 1000 feet above ground as they pivot away.

Collapse continues with a relatively small amount of falling mass being ejected at the collision interface. Since initial contact was an order of magnitude greater than that which the long span open floors could absorb, little effect is seen on the falling mass. This means that the still intact portion of upper mass will continue to accelerate.

Again first contact is the lower column ends, though some movement of those column ends might not see them directly affecting column\truss(or beam) connections , they will spear through that next lower floor.

Next long span floors will now be impacted by more mass , albeit more rubblized, that is necessarily moving faster than first contact. Those floor spans must also be hit by an order of magnitude greater load thanthey can withstand, and indeed greater than at that earlier, first contact impact.

There existed no mechanism by which this floor truss and perimeter peel destruction could be arrested as this continues.

The core area beam destruction lags that of the long span destruction since those beams are heavier and shorter. Perimeter peel necessarily lags long span destruction as it is a direct result of that long span truss destruction.

Core columns stripped of inter-core Lateral Support succumb to a combination of slender column buckling and heavy buffeting from falling debris. A relatively small, heavily damaged, portion of the core remains after long span floor and perimeter destruction ( no longer being buffeted by debris impact) but cannot remain intact and it comes apart as well
 
Last edited:
I'll never understand why these need to be so wordy.

20 floors fell on one floor. One floor couldn't hold the weight, so 21 floors fell on one floor, which couldn't hold the weight. So 22 floors....

repeat until completed.
 
I'll never understand why these need to be so wordy.
.

Well it depends on the audience. For the general public your two liner is sufficient.

For forensic engineering one needs a full engineering report along the lines of the NIST reports.

For the in-between, those interested in the mechanics but not versed in engineering, one needs an explanation that is understandable yet not too simplistic.

Much discussion recently centered on how the acronym ROOSD did not describe the events well enough. I pointed out that once you begin adding to an acronym it becomes unwieldy while subtracting can make it nigh on useless.

Seems few are still interested.

To that end I produced the above narrative summary.
 
Last edited:
For the in-between, those interested in the mechanics but not versed in engineering, one needs an explanation that is understandable yet not too simplistic.
That's me and it was very clear on first read.:thumbsup:

Don't know if you care but there are a few spelling and grammar errors (I only mention as it seems more like a polished piece of writing than a normal post).
 
That's me and it was very clear on first read.:thumbsup:

Don't know if you care but there are a few spelling and grammar errors (I only mention as it seems more like a polished piece of writing than a normal post).

Unlike many, I invite correction. For my illumination point out the errors plaese. :blush:
 
Unlike many, I invite correction. For my illumination point out the errors plaese. :blush:
No problem! :) I've put the corrections in bold.

- Column capacity in this case does not necessarily mean it [its] original engineered load bearing ability.

- The later [latter] is due to both the destruction of Lateral Support

- Those floor spans must also be hit by an order of magnitude greater load thanthey [than they] can withstand

- A relatively small, heavily damaged, portion of the core remains after long span floor and perimeter destruction ( no longer [(no longer] being buffeted by debris impact)

These two may be wrong on my part:

- At this point a rapid redistribution the remains structure [of the remaining structure?] occurred

- A combination of factors result [not sure if this should be 'results', referring to singular 'combination' or plural 'factors'] in [as?] few if any buckled columns impact on lower parts of themselves
 
I don't agree with some of the concept expressed such as:

Aircraft impact debrided spray on fire insulation.

I don't for a moment doubt that some SRFM was scraped off of parts of the steel that came contact with parts of the plane which might have hit the steel.

Frankly I don't think much SRFM was removed by direct plane impact.

Second

I don't think there was as much load re distribution in the core region with the exception of the columns adjacent to severed core columns and those columns loads moved to adjacent columns. The face lost a significant number of adjacent columns and there was a redirection of the loads around the missing columns to those adjacent. The large spandrel design facilitated this... but the core was quite different.

third

The loss of axial capacity was at first likely from core columns being forced out of axial alignment. This was functionally equivalent to severing them. Columns above still loaded would then "drop" and become hangers and probably separate from at/ the hat truss truss level... slabs and contents would then drop down inside the core.

fourth

The OOS failures and collapse had several modes but the main one was the fracture of the slabs, with the fractured chunks falling and driven down by the collapsing mass falling on it. A secondary failure mode was the breaking/failure of the truss connections to the spandrels and the core belt girder... A third failure mode was the separation/failure of the connections of the belt girder to the perimeter core columns

The bits of floors and structure still connected would represent loads which were redistributed and these eventually exceeded the capacity of the increasing few connected aligned core columns. They buckled... the core collaped...the OOS floors feel inward without core support. Slabs fractured and dropped... the while frame which was barely connect moved laterally... destroying all axial alignments and the everything above the strike some where the fires raged... dropped with no resistance.

The unstoppable floor destruction which raced to the ground left the facade columns unstable, caused the facade to bulge out and connections between panels began to break and the facade panels either dropped or wre pushed away by the debris mass in large assemblies and this was the so called "peel"

The core lost lateral bracing in a similar floor collapse leaving the core columns to tall, too slender for their height and they either self buckled or fracture their connections from swaying and jostling caused by the floor debris mass growing in volume.

and so on...
 
:thumbsup: Good summary! :thumbsup:


Aircraft fuel spread through several vertically adjacent floors of the structure and ignited. This created a condition of very large area office contents fire on those floors with various levels of rubblization of the contents. This is a fire condition rarely if ever, reached within seconds of fire initiation in large office structures.


This is a very important point! "Truthers" like to say that the jet fuel "burned off quickly" but then ignore that that means a LOT of heat ENERGY was released quickly, which means that more
steel got hotter, faster than any "normal office fire." NIST (or was it FEMA?) estimated that about 3000 gallons of fuel burned inside each building (not counting the fireball outside and run-off). That's enough energy to lift a 150-ton 767 to over 30,000 feet and then push it at 500 miles an hour for two hours -- released within minutes, and then the office contents were burning over the same large area and spreading quickly.

Much discussion recently centered on how the acronym ROOSD did not describe the events well enough. I pointed out that once you begin adding to an acronym it becomes unwieldy while subtracting can make it nigh on useless.

Agreed, but what did you think of my compromise suggestion of Runaway Floor Destruction for the progression phase?
 
Last edited:
Brilliant stuff JDH - Even in "pedantic engineer mode" I would not fault it as a good summary.


Overall most of your paragraphs identify and concisely catalogue multiple subtler points of engineering. Many of which have attracted lengthy debate.

A lot of details summarised in few words.

And I don't see any point that is in need of correction or even "polishing" for greater clarity. (I'll leave the typo/spelling/textual proof reading to you and Georgio.)

Do we have an "I'm jealous" smilie?

:rolleyes:

PS
One - only one - minor "nit".

Possibly a reference to "well outside design envelope" in your second paragraph - the key issue being it was way beyond what would reasonably be designed for rather than "rarely reached" - it has AFAIK only been reached on two occasions. WTC1 and WTC2.
 
Last edited:
No problem! :) I've put the corrections in bold.

- Column capacity in this case does not necessarily mean it [its] original engineered load bearing ability.

- The later [latter] is due to both the destruction of Lateral Support

- Those floor spans must also be hit by an order of magnitude greater load thanthey [than they] can withstand

- A relatively small, heavily damaged, portion of the core remains after long span floor and perimeter destruction ( no longer [(no longer] being buffeted by debris impact)
All noted.
These two may be wrong on my part:

- At this point a rapid redistribution the remains structure [of the remaining structure?] occurred

- A combination of factors result [not sure if this should be 'results', referring to singular 'combination' or plural 'factors'] in [as?] few if any buckled columns impact on lower parts of themselves
arrgh. Sometimes I revise something but inadvertently leave in words from the original, other times I miss an auto-correct that buggers up meaning. I'll try to determine just what exactly how I wanted it to read.
 
These two may be wrong on my part:

- At this point a rapid redistribution the remains structure [of the remaining structure?] occurred

- A combination of factors result [not sure if this should be 'results', referring to singular 'combination' or plural 'factors'] in [as?] few if any buckled columns impact on lower parts of themselves

How about:

"At this point a rapid progression of column failures occurs. A column or columns fail and in redistributing load cause other columns to exceed capacity, causing another rapid load shift. Column failures progressed to all columns over one or more levels."


-"A combination of factors result in few, if any, buckled columns impacting on lower parts of themselves"

I see another now too:
"Each tiertower also twisted and tilted noticeably up to collapse"
 
Last edited:
I don't agree with some of the concept expressed such as:
Sander why recycle objections which have been responded to multiple times - without you taking on board the suggested corrections that have been put to you?

EDIT:
(On second thought I removed the detailed critiques of JSanderO's points most of which actually accept the JDH summary but are posted as if they were disagreement.
Let's not divert attention from the "Quality Improvement Process" :) )
 
How about:

"At this point a rapid progression of column failures occurs. A column or columns fail and in redistributing load cause other columns to exceed capacity, causing another rapid load shift. Column failures progressed to all columns over one or more levels."


-"A combination of factors result in few, if any, buckled columns impacting on lower parts of themselves"
:thumbsup:
 
Thanks JSO
I don't agree with some of the concept expressed such as:
Difference of opinion noted
I don't for a moment doubt that some SRFM was scraped off of parts of the steel that came contact with parts of the plane which might have hit the steel.

Frankly I don't think much SRFM was removed by direct plane impact.
While perhaps not a lot, I believe it was significant enough to compromise the resistance to heating. IOW, the fuel was an accelerant in fire spread, loss of fire retardant is an accelerant in steel heating. One more reason the building succumbed so soon.

I don't think there was as much load re distribution in the core region with the exception of the columns adjacent to severed core columns and those columns loads moved to adjacent columns. The face lost a significant number of adjacent columns and there was a redirection of the loads around the missing columns to those adjacent. The large spandrel design facilitated this... but the core was quite different.
I see the core as being impacted by a huge heating effect, basically from all sides. The perimeter was necessarily cooler due to the open air. As core columns suffered load creep due to increasing loss of capacity from heating, loads would slowly transfer to other columns.

The loss of axial capacity was at first likely from core columns being forced out of axial alignment.
From what mechanism?
This was functionally equivalent to severing them. Columns above still loaded would then "drop" and become hangers and probably separate from at/ the hat truss truss level... slabs and contents would then drop down inside the core.
Slow creep driven loss would simply see the core shortening at those columns affected. This would cause the tower to move and twist. It did.


The OOS failures and collapse had several modes but the main one was the fracture of the slabs, with the fractured chunks falling and driven down by the collapsing mass falling on it.
The very first fracturing would be due to upper column sections spearing the floor. Those column sections are right next to their counterparts below, which is also where truss seats are. Breaking a seat is going to fail that truss's usefulness immediately.
A secondary failure mode was the breaking/failure of the truss connections to the spandrels and the core belt girder... A third failure mode was the separation/failure of the connections of the belt girder to the perimeter core columns
Belt girder would be pretty insignificant in holding the core together. Tearing away the truss connections to that girder would weaken the girder itself(its a violent action)

The unstoppable floor destruction which raced to the ground left the facade columns unstable, caused the facade to bulge out and connections between panels began to break and the facade panels either dropped or wre pushed away by the debris mass in large assemblies and this was the so called "peel"

I wanted to avoid anthropomorphism such as "raced".

The core lost lateral bracing in a similar floor collapse leaving the core columns to tall, too slender for their height and they either self buckled or fracture their connections from swaying and jostling caused by the floor debris mass growing in volume.
.
No significantly different from what I said.
 
My issues are sort of "details" / mechanisms... not much to argue about and ultimately not provable or unprovable. My biggest "beef" is stating the connection failures are what drove the floor slab collapse. What drove it was the driving crushing mass which destroyed what it fell upon and the "crush front" was far from uniform.
 
My issues are sort of "details" / mechanisms... not much to argue about and ultimately not provable or unprovable. My biggest "beef" is stating the connection failures are what drove the floor slab collapse. What drove it was the driving crushing mass which destroyed what it fell upon and the "crush front" was far from uniform.

Got it. Mostly agree. For me I cannot get the image of those column sections being the first impact on the lower floor slab. They either (1)just punch through the concrete and miss the nearby truss, (2)they land on the slab directly above the truss and snap off its seat, and of course this is not a binary situation, it could impact concrete directly above, (3)but not fully in line with, the truss. A glancing blow. Its a spectrum. You believe its dominated by (2). Probably, but the damage wrought by (1) is more significant individually as it fails lateral support immediately.
 
Last edited:
Perhaps tomorrow I will do a rewrite, taking all the above, and any new, suggestions into mind.

BTW, JSO, the inclusion of the belt girder is directly attributable to you having pointed it out on many an occasion.
 

ISF - Join now!

Every member here is approved by hand. No bots, no spam, just people who care about evidence and honest debate.

Membership is free!

Create your free account

Back
Top Bottom