ASAP, we are editing part 3 now. We originally intended on outlining all the 'errors' that NIST had made in their analysis but came to the realisation that such a video would be several weeks in length, and we'd prefer to keep these short. The intention is to pin them together with some extra material and make this into a sort of mini documentary.

Yes, it would make sense that the total expansion would be divided, measured both ends equally--- unless there was some reason for one end to stop, say contacting an immovable surface.

The focus for NIST has always been on the beam expanding against the movable girder, but did they consider the girder expansion? The girder did indeed have limited expansion room--on the order of about 1.2 inches both ends. In the case of the northern end connected at column 44, it was literally buried in the channel there and could not move sideways very far.

In the case of column 79 side to the south, the girder end was in a channel 1.8" deep, which was formed by the addition of two 26" wide by 2" thick side plates welded to the already extremely beefy column. Column 79 was a W14 x 730, which had a 3" thick web and 5" thick flanges. For all practical purposes immovable.

According to the spreadsheet, at 600C, the girder would have expanded 4.88" using NIST's incorrect formula and 3.99" using the right formula. Both are beyond what the columns would allow. Even at 400C the numbers would be 3" and 2.48" respectively.

I guess one would have to take many factors into account and run simulations. There are so many variables.

If the beam heated up first and the girder somehow stayed cool I suppose you could imagine a scenario where the bolts at column 79 sheared and the girder flew off to the west. But that is way too much of a stretch of my imagination, knowing what I do about the fire being out on floor 12 before 5PM, the presence of shear studs, the five beams still attached to the girder, they in turn still attached to the exterior columns.

The whole thing gets pretty silly.

and applied much more of a horizontal restraint to the beams than did the girder between columns 44 and 79 and that the expansion would have been towards the interior girder.

To know if this is true one would need to have both in a model or do the hand calculations knowing the girder strength of the exterior.

I don't believe we have been given the girder sizes and connections of the exterior. From photos we know the connections were substantial moment connections. However, if the girder had the same moment of inertia as the one on the interior there certainly would have been some deflection.

Even so, NIST's claim is pure nonsense as they can't get the expansion they need to push that girder off its seat even when it all goes one way.

Yes. But that doesn't suit NISTs conclusions, so we are going to ignore it for now. The aim of these videos is to prove NIST wrong by their own standards of investigation, within which, they seem to be weighting every variable in their favour. I would guess that steel will expand in all directions to some extent when subjected to the conditions that NIST cite, but looking at the detail for this girder/column79 connection it appears that NIST need more than a mere 2.5" head start to make their hypothesis work. http://img521.imageshack.us/img521/3176/1091topzoom.jpg
We will continue to analyse their explanation on the basis that if there is a doubt about any variable, that variable will be increased fully in the direction that favours their theory. For example, the outside wall to the east of column 79 will be taken to be infinitely strong. It looks to me as if there will be no need to dispute any of the many 'oversights' that NIST have made, yours being a good example of one. It should make for an interesting conclusion to see what happens when we apply NISTs story to the reality of the construction, even in the temperatures/conditions that they suppose existed.