Exactly. Reproducing fire and damage effects on one floor wouldn't lead to collapse in any of the models I've seen.
That's right: they were establishing baselines, which had never been done by fire testing of intact truss samples before the towers were built. BTW, the restrained sample that was insulated most like the tower trusses failed to achieve its fire rating.
Anyway, the problem may be much more complex than simply heating trusses until they sag. Bazant
says,
NIST's global models did take creep buckling into account, but apparently predicted less than Bazant believes existed. Of course, this was not tested when the truss samples were cooked.
So, to build a full-scale, furnished, multistory reconstruction of the tower impact and main fire floors, with proper live and dead loads, with structural damage, fireproofing damage, ventilation, and fuel and debris distribution simulated, fully sensored, and to have to rebuild the structure and re-furnish and re-damage it for each scenario run, as was done with the computer models, all to simulate events that are, to say the least, extremely rare? No.
What would be more interesting, more feasible, and more relevant to other buildings, would be to build a reduced-scale multistory model to see if claims of the
intact buildings' vulnerability to fire, as made by Arup, U. Edinburgh, and Dr. Quintiere, might be correct.
