Oz,
Howdy & welcome. Excuse me for missing it, if you mentioned it before, but what is your background. You sound like an engineer to me. I'm mechanical.
Exactly leftysergeant.
This is the bit that failed:
http://conleys.com.au/webstuff/003.jpg
And (Courtesy Heiwa for the base diagram) this is how the first few floors of the outer office space started the global collapse.
(So leftysergeant's words and my picture.)
http://conleys.com.au/webstuff/globalstart.jpg
NOTE the accumulating floors. No "one for one" BUT the Top Block would progressively yield to other forces and mostly collapse.
And, disclaimer, this only explains the outer office space. See earlier post for the core.
A couple of points, but ones that are significant.
In your rework of Heiwa's sketches, I believe that it is important to NOT include the damaged 6 floors or so in either section A or C. A & C are basically intact & not weakened by damage, distortion or fire. Those weakened floors (93 - 98 in WTC1) are likely the ones that collapsed first. By the time that any collapse got to an intact section of the towers (A or C), it had already gathered 6 stories worth of momentum.
Second in your last sketch, you show the debris collecting predominantly in the part A, the lower section. This is wrong. Think about any floor that is hit by debris and begins to fall. It is starting from a standing start, as a fast descending hammer comes down from above. For the first 2-5 stories (my SWAG), the faster descending upper block is going to sweep the debris into its open structure and pack in there. After a couple of stories, the bottom surface of the upper block C will be so impacted with debris, that it'll approximate a solid surface.
Thrid point. Think about the effect that the stagger of the peripheral & core columns had on the destruction of any give floor. By the time that the descending upper block arrives at some given floor (say the 80th), fully 2/3rds of the columns supporting the 80th floor have already been destroyed. This is because 1/3 of all those columns reached up to the 82nd floor and another 1/3 reached up to the 81st floor. These have been ripped away BEFORE the descending block even arrives at the 80th floor. There ain't that much more work to do to snap the 1/3rd connections that are left supporting that floor.
And I feel strongly that Bazant goes off course in calculating work required to buckle columns. This is an important calculation in the collapse initiation, but not after that point.
AFTER the collapse had initiated, the failure mode was snapped connections, gross breakup of the concrete floors and deformation of the cross trusses. This it the calculation that should have been used in calculating the work of disassembly.
Your correction to Heiwa's sketch, that the unsupported columns become immediately unstable & snap off, is right on the money. I would tend to think, tho, that the impacting of debris at the bottom of the upper block will give those bottom columns of the upper block some lateral support. The top "stubs" of broken off columns in the lower section have approximately zero ability to resist side loads, however.
This is one (of many) of Heiwa's fundamental errors. He distributes the strain energy of the of the lower (larger) section uniformly over its entire structure. And he ignores the stress concentrations and side loads that are applied to the non-braced column in the top floor that is being crushed.
Just a couple of points.
tom