One interest point--while there is no singular diagram of the unseating of girder a2001 in the NIST reports, NIST's technical briefing from 11/19/08
contains a simple drawing of the unseating on pg. 33):
While certainly not proof that NIST properly modeled this connection correctly in its detailed sim, this drawing does include side plates and shows the girder clearing them. Whether this was possible (strictly from a side plate trapping sense) in the real world seems to me to be entirely a product of two factors (assuming the eastmost beams were expanding and pushing the girder to the west): (1) what was the average temperature of girder a2009 at the time of alleged dislodgement, and (2) regardless of its average temperature at such time, had girder a2009 previously suffered damage from the fire (e.g., had it already buckled or substantially sagged and then started cooling). We don't know the answers re these considerations in respect of Hulsey's study yet. I don't think we really know them in real exacting detail in respect of NIST's study, either, but I made be mistaken in my recollection on that point.
Re the issue summary thread--
Mick--I would also note that the slides Hulsey used to dismiss Arup and WAI's studies out of hand lack support. For Arup, Hulsey relied on the same error Tony has made here re dismissing Arup's collapse initiation model due to a flaw in Nordenson's subsequent hand calculations. For WAI, he dismisses the study on the bare assertion that the fire temperatures used in their primary model were too hot (even though he didn't do any independent fire modeling). He also does not address the fact that WAI received from Arup copies of iArup's girder failure models and confirmed them.