With the supposed upcoming release of NIST's draft report on WTC 7's collapse, I thought it would be useful to revisit NIST's original report for WTC 1 and 2. As many of you know, NIST found that the towers collapsed from aircraft damage and fire alone and that they found no corroborating evidence for alternate theories of the collapse. What I find interesting is that the same can be said for NIST's own hypothesis - namely that no corroborating evidence was found to support its fire theory.

In order to be exact, let us restate NIST's exact fire collapse hypothesis, specifically for WTC 1:

1) Aircraft impact cause damage to the North face and several core columns were severed. Aircraft debris also dislodged fireproofing that helped speed the later collapse.

2) The jet fuel ignited a large fire that originally started on the north side of the tower and over the course of 90 min worked its way to the south side.

3) The fire temperatures seen were in excess of 800C in some places, causing weakening of the core but specifically causing the long span floor truss assemblies to sag.

4) The sagging floor trusses exerted a pull in force on the external columns of only the south side of the tower. These forces eventually bowed the south wall inward sufficiently that it became unstable.

5) The instability of the south wall spread and could not be contained by the damaged core and remaining external columns - global collapse ensued.

Therefore the singular question to be asked and answered is this: "Is this hypothesis supported by the evidence?". Fortunately the NIST investigation answered this question and the answer is an emphatic NO. Let me therefore explain, with the appropriate findings from NIST's actual report to prove it.

Previous researchers have cited the following reasons why many of the points above cannot be true:

1) NIST generated various cases for both aircraft damage and fire temperatures but consistently chose the most extreme cases, with little to no justification other than the fact that it made their hypothesis slightly more likely.

2) Of the pitifully tiny 236 steel samples salvaged, only 2% saw temperatures greater than 250C. None saw temperatures above 600C. While the sample size is small, there is no evidence here to SUPPORT the high temperature hypothesis.

3) In only 2 out of 9 test cases was NIST able to generate sufficient bowing (~ 33 inch in the model vs 55 in observed) using an extremly simplified computer model with point forces. They had to assume all fireproofing was gone, maximal fire temperatures, and more time was required in the simulation than on 9/11.

4) NIST ignores the interior column connector when looking for pull in forces. This was the weak link and would have failed first.

See the Journal of 9/11 Studies for more articles and data on these points.

Of course this is a very small sample of the available criticisms, but it gives you an idea. Today I hope to contribute another, this time in NIST's own words. Many people have neglected to actually READ the report which is too bad - since the more one reads it, the more glaring the contradictions become and the less likely NIST's hypothesis is actually true. It's simple for the media to say the "10000 page report" or Ronald Wieck to say the "good people of NIST" but to actually SEE the contradictions in NIST's own report is sobering. What follows is a verbatim excerpt from NIST NCSTAR1-6D detailing how NIST actually arrived at its collapse hypothesis.

---

Pull-in Forces to Cause Observed Bowing

As discussed in Chapter 5, inward bowing of the exterior wall on the south face of WTC 1 can be seen in photographs of the event at 10:23 am (about 97 min after the aircraft impact), while no bowing of the south wall is evident at 9:55 am (about 68 min after the aircraft impact). The inward bowing at 97 min extended from Floors 95 to 99 between Columns 308 to 326 (possibly to 340); NIST estimated the maximum bowing to be 55 in. at Floor 97.

The isolated south wall model of WTC 1 did not bow inward under Case A temperature condition, and it bowed slightly over a very limited area under Case B temperature condition. The isolated wall model results did not capture the actual bowing.

The most plausible explanation for the discrepancy between the observed and modeled bowing of the exterior walls is that pull-in forces were not captured in the full floor model, and floors sagged to a greater extend than these models predicted. Following are reasons the floor models likely underestimated sagging and did not accurately calculate pull-in forces.

* The exterior wall boundary conditions used in the floor models were realistic only if a single floor was heated. The floor trusses were supported on exterior columns that extended one story above and one story below the floor modeled. The far ends of the columns were restrained against translation in the direction normal to the exterior wall and rotation about the axis parallel to the exterior wall. This boundary condition was much stiffer in translation normal to the exterior wall than actually occured for sequential floors heating simultaneously.

* The floor model did not include creep behavior in the steel or cracking of the concrete floors.

* The floor models did not include strap anchors or studs

* More thermal insulation may have been dislodged from the trusses than estimated from the impact analysis. The impact analysis did not account for the effect of impact or vibrations on dislodging insulation.

* The floor models assumed a uniform live load. Debris accumulated in large piles was observed in some floor areas.

To model the effect of pull-in forces on inward bowing of the columns, trial values of pull-in forces were applied to the exterior columns of the south wall over five floors from Floor 95 to Floor 99 where bowing was observed.

The magnitudes of the pull-in forces were determined by trial and error, matching the observed inward bowing of exterior walls for Case B temperature condition. The Case B temperature condition was used because temperatures of the south office floors and south wall columns were much higher than those of the Case A temperature condition, and because the full floor models with the Case Bi temperature condition showed much larger floor sagging in the south office area than did the floor models with the Case Ai temperature condition.

---

NIST NCSTAR1-6D pg. 96-97, author's emphasis


What is immediately striking is that NIST is basically saying that their prior studies on fireproofing dislodgement, temperatures, and floor response modeling must be thrown out the window! Thus the work of researchers in sections 1-6A (fireproofing) , 1-6B (floor tests), and 1-6C (floor modeling) of the report is meaningless! The data are simply thrown out since they do not predict large inward bowing! Instead, the exercise is turned on its head: The hypothesis that floor sagging cause the bowing is now ASSUMED to be true (contrary to the data), and the task now falls to simply guessing through trial and error how much force is required on each column to match the observed data! This is the exact point in the report where NIST totally sheds the scientific method in favor of quick results (this section was one of the last to be completed before the deadline). From here on out NIST's models are no longer predictive, but rather attempts to match the model to the observed data. So when you graphics of massively bowed columns from NIST, you know that it was produced by working backwards - why bother with models - they should have hired a graphic artist!

The "plausible... reasons" for throwing out the data are basically an attempt, pardon the language, to throw shit on the wall and see what sticks:

1) The magnitude of the "discrepancy" between NIST's data and the observed bowing. The "bowed slightly over a very limited area under Case B temperature condition" is actually a reference to:

"The maximum inward out-of-plane displacement was 0.66 in. at Column 336 at Floor 98 at 100 min, while the maximum outward out-of-plane displacement was 2.3 in. at Column 342 at Floor 97. At these locations, the floor was disconnected from the exterior wall so that the unsupported length of the columns was two to three stories." (NIST NCSTAR 1-6D pg. 81)

NIST was only able to achieve 0.66 in of inward bowing (vs observed of 55 in.) in a SINGLE column. Even more remarkable - they found a case of 2.3 in. of OUTWARD bowing as well. Clearly these data are totally inconsistent with large scale bowing over a full five floor area of WTC 1.


2) While it is true that the floor model was limited to a single floor, NIST cannot be saying that large swaths of the south wall were unrestrained since only 14 columns over two floors were observed to have disconnected (NIST NCSTAR 1-6D, pg. 49). Thus the floors both in and above the bowing zone were still available to provide either full or partial restraint.


3) The reason that the floor models did not include the strap anchors is because NIST found them too cumbersome to model and that they were irrelevant:

"In the actual buildings, the strap anchors and studs must have been capable of transmitting a significant amount of force between the floor system and exterior wall. However, the full floor analysis with the strap anchors and studs resulted in sequential failure of these components and an extremely slow convergence in the analyses. Because these components were found to fail at early stages of fire, these elements were then removed from all the full floor models to obtain solution within a reasonable period of time. Therefore, in these analyses, the only structural elements in the full floor model that could transfer horizontal interface forces between the floor system and the exterior walls were the gusset plates and seat bolts. This cause premature failure of the gusset plates and seat bolts in the analyses, which result in horizontal floor/wall disconnections. In addition, friction between bearing abgles and seats was not modeled in the full floor analyses. Therefore, the full-floor model did no show significant tension at the floor/exterior wall interface."
(NIST NCSTAR 1-6D pg. 37, author's emphasis)

Therefore how could components that failed at early stages of the fire be around later to exert increased pull-in forces on the south wall? Obviously they can't.


4) Since NIST cannot provide any evidence that "more thermal insulation may have been dislodged" above and beyond its already flawed computer models - how can we take this statement as anything but hand waving?


5) Same goes for the distribution of load - NIST has no evidence of where the load actually was - and more importantly, no data as to how this might have affected their results - more hand waving.


6) Finally, it is completely clear that NIST choose the Case B temperatures here not because there is any evidence to select it, but because it was hotter and makes NIST theory look slightly more plausible. Apparently Jones, et al. were right all along - NIST cherry picked its scenarios.

There is much more that could be said about NIST subsequent modeling after this point, but suffice it to say that NIST was able to show throughout most of its report that the steel temperatures were not extreme, fireproofing dislodgement were large but not complete, and that floor trusses subject to extreme temperatures do not sag to such a degree as to be able to generate sufficient pull in forces on the exterior columns. It was only in the last mile of the report, section 1-6D, that when push came to shove and deadlines fast approaching, ANY explanation (in the politically correct vein) was needed to explain the bowing, data be damned.