Hi
Something that I had been wondering about is the crush up. I've been following the thread and very much enjoy your posts (Im not an Engineer, just an IT proffessional). The bit I cant quite figure out, and its the only bit in Heiwas argument I cant quite resolve, is why the upper part C doesn't break apart as much as lower part A when collape occurs. I though perhaps it does, but that the general mass of the collapsed material keeps the progression going.
But (unless I have missunderstood) Bazant says theres a crush down first, followed by a crush up when C finally reaches whats left of A. Now I understand that its a simplified model, and I think I intuitively understand that this could happen, but Im not sure why.
Lastely I wish Heiwa would stop pratting about and get into the discussion properly so we can get to the end of your thought experiment. Im interested in the conclusions.
Cheers
Tim
Hey Tim,
Sorry, took a bit to spring free the time.
Your instincts are right on the money. There is an asymmetry to the damage that needs to be explained.
If the collision between the upper & lower parts had been horizonal, then the damage would have been symmetric into both pieces.
So what is the source of the asymmetry? Exactly what your instincts tell you: Gravity.
In a sense, it's the difference between jumping on somebody & having somebody jump on you. But with an extra twist...
The best way to view this is as the side of an incredibly steep slope subject to a rockslide. All the various rocks have had braces carefully designed & placed under them. Unfortunately, all those little braces are "mutually supporting". They cannot withstand their own load if too many of the adjacent braces are gone. And now, when the avalanche starts, even tho every individual collision does obey Newton's 3rd law (action = reaction), the whole landslide moves inexorably downhill once it gets going.
The twist is that you're going to ride down the avalanche in a shed above the debris. You'll see what I mean at the end of this post.
Again, the asymmetry is provided by gravity.
So, I think the question question of "why it progressed to the street" is simple. Just like an avalanche, it gathers momentum & mass as it moves down.
The separate (& independent) question is "why did it not move up into Part A?"
Unless you're pretty good with math, it's tough to get a sense of this from Bazant's paper (IMO). His is not a model of what really happened. It is idealized to favor an arrested collapse. And his idealizations are tough to translate into "what does this say about the real building as it came down?"
To get a gut feel, imagine that you are inside the upper block of the building, at about the 102nd floor, and everything below you, down to the crush floor, is transparent.
Before the collapse began, several of the floors on the 97th thru 99th floors had collapsed. This was seen from outside the building. Others were sagging and were pulling in on the external columns (& out on the core columns). Suddenly the connections failed, and the peripheral columns flung outward, becoming massively weaker (because of unsupported length & bowing).
As the collapse begins, the columns of the 98th floor buckle. They do NOT fail as Bazant describes them in any of his papers. Bazant examines them as if they buckled over a length of one floor, with 3 (or 4) knuckles. [See any of Bazant's papers that describe the columns becoming unstable.] This makes any of the individual straight segments between knuckle points equal to about 3' to 4' long. [Again, this is one of Bazant's "conservative", arrest favoring, idealizations.] Remember, short beam are strong, long ones are weak.
An examination of any of the collapse videos shows that the external columns did not fail like this. They failed at the column to column joints. That means that the lengths of the buckling segments was not 3' to 4', but rather about 36' long. This means that the real failures of virtually complete floor destruction extended up about 3 stories and down 3 stories from the 98th floor. Plus partial destruction up 2 additional & down 2 additional stories.
Lacking lateral support, the core columns buckled too, probably mirroring the outer columns (i.e., 6 floors total collapse, plus 2 up & 2 down partial.)
And when you pull the supports out from under a concrete floor that weight about 2.5 million Kg each, what happens? They all start to fall, together.
But there are still partial connections to the standing lower columns, so they don't free fall. The upper block does. The upper stories (3 stories of rubble first) then overtake the 3 stories of dragging rubble. The point is that the collision velocity between something that is free falling (upper block) and something that is ALMOST free falling (rubble) is much less than between something that is almost free falling (rubble) and something that is stationary (lower floors).
The impacts between the upper block and the rubble is a much less forceful impact than between the rubble and the lower stories precisely because of the lower relative velocities.
Meanwhile, a second effect occurs. The rubble is "packing into the lower 2 - 3 stories of the upper block. It's packing in at a low relative velocity, and it is being tamped in by air pressure and many, many collisions.
Now, we're ready for your observations from the 102nd floor. The debris has been packing in in the stories below you. At fairly low relative velocities. By the time the upper block has fallen 6, 7, or 8 stories, there is a LOT of debris packed into those lower couple of stories of the upper block. This debris became a barrier between the impacts and the upper block. The impacts were happening on the BOTTOM side of the debris. The upper Part A was riding down on the UPPER side of the debris. The growing debris layer PROTECTED the upper block from the destruction.
I tried to describe this in another posting here:
http://www.internationalskeptics.com/forums/showthread.php?postid=4743226#post4743226
I hope this helps.
Tom