The 9/11 Forum

femr2 wrote:Well the very simple answer is, yes, you are misinterpreting the question.
It does not refer to the free-falling debris, it refers to the debris ejecta on the right hand side of the tower:

Yes, that is a different question altogether. And a bit more difficult to answer.

The differential appears nearly from the get-go, as if the initial failure actually occurs well below where the simulation places it. Very nice work, femr2.

femr2 wrote:..the initial failure position is known and pretty clear.

I think there may be room for debate, here. Little about this collapse is "known and pretty clear," in my opinion. You are beginning to sound like our friend Dr. Benson.

The collapse front does not appear to have been a horizontal surface. The initiation was likely internal, and could have involved only part of the core, for example, giving one side a head start. Is this not possible?

Femr2, from the image and video a few posts ago I cannot be convinced that the dust from the south face is of the floor-by-floor ejection type. I suspect it is also from large freefalling perimeter sections which you cannot see in the clip.

Video from the south doesn't allow us to see ejections directly out of the building (on south face, west side) at that elevation because large perimeter chunks leaving dust trails have already obscured the view.


Trails from two huge perimeter chunks (very interesting in themselves) from the south face cover the west side of the S face.

You'll have to show that the dust you refer to is actually from the collapse front and not from some other obscured source. I've never been able to get a glimpse of collapse front ejections in that region, though I can clearly see them on the east side of the south face.

These 2 photos show them

http://www.sharpprintinginc.com/911/index.php?module=photoalbum&PHPWS_Album_id=9&PHPWS_Photo_op=view&PHPWS_Photo_id=1512

http://www.sharpprintinginc.com/911/index.php?module=photoalbum&PHPWS_Album_id=9&PHPWS_Photo_op=view&PHPWS_Photo_id=1511

Wild. The largest chunk in each photo is actually falling next to the other. The right edge of the chunk in the second photo is the edge of the largest chunk in the first photo. They fall together, possibly slightly attached.

Thanks for the new view (edit: I meant the still image from a few posts back... the latest gif is great, too), very helpful. Based on that image, the crush front is at ~story 60; about 11 floors ahead, as you say. I've tried to construct a plausible parametric adjustment to achieve the desired displacement. I'm not saying this is how it went, just throwing out ideas.


1) Initial elevation

Your upper block is composed of 12 stories, 99 through 110. On the south side, bowing is occurring as low as floor 94. This is external visible evidence strongly suggesting initiation occurred on the south side at or below this level. Even on the north side, there are perimeter sections moving down with the upper block from the lowest levels of the impact area. Choosing story 98 as the missing story for this model is arguably high in the general sense and almost assuredly too high for the south side. Given the geometrical characteristic of bowing, regardless of root cause, it's reasonable to infer deflection of floor 94 to a point well into story 93, pre-initiation. There was initial tipping towards the south immediately preceding the break which would further lower the starting elevation for an independent progression on the south side.

Just based on the above, I would offset the missing floor down to 93.


2) Timing of regionally specific catastrophic initiation

Continuing with the idea of independent progression on the south side, there's the matter of whether an internal failure could precede the externally visible global failure, and what magnitude of floor mass displacement could occur before becoming apparent from the external view. This is obviously a fuzzy area. I'm going to ask for a little under a second here, about 0.8s. That's not too much, really. The movement before total global initiation is visible on video for about 4 seconds. The scenario I envision is local floor collapse. This will produce a belch of smoke, yes, but only towards the latter half, a period under a half second - and it's on the other side.

Rather than change the timing, I'll just offset the missing floor one down to 92.


3) Mass distribution

Short of having something very solid and heavy punching through floor diaphragms (the upper S wall?), pancaking floors would be the obvious choice for an internal failure in the floor area. Should conform well to this model of inelastic slab collisions. Ignoring for a moment the detailed work of you and Hambone and others in collecting mass info, a reasonable approximation for the mass of each story is all equal. This from the impacts involving floors and contents; however, the actual mass of the upper block is another thing. DBB has claimed that the hat truss assembly and so forth made the cap heavier than it would be if one just counted actual stories. I don't know about that, but I'm going to invoke it anyway.

As I recall, he used 8 extra stories. I'll take 6.


4) Stretch/compaction

No quibble with your default slab thickness, but it's reasonable to say that it could be larger overall, and particularly in the first half of collapse. It's debatable. It doesn't seem like much of a stretch (pun intended) to go with a stretch of 0.2, effective at the 5 second mark, giving a slab thickness of about 3/4 meter.



The snippet of code from before is adequate for this task, if the overall story count is set to 116 and the missing story at 92, for an actual Zone C height of 66.6m. Note that these adjustments double the upper block mass. Using these values with a stretch of 0.2, I get 118.4 meters / 32 stories for displacement of the crush front at 5 seconds.

In other words, story 60.


Please check my work within your own framework, using corresponding parameter changes.

Of the four adjustments listed above, the biggest gain came from #1. This one is the least questionable, I think. No doubt a drop through story 98 is taking it up as high as it can go, realistically it needs to be lower even for the general, global case, where 96 might represent a better average location. There's evidence available in one of your frames to bolster the adjustments from item #1 and possibly #2. Definitely expulsion on the far side occuring a full five floors below your crushing story depicted at the 0.63 second mark:



The stretch may be too large, and the additional six floors of mass is questionable, but this is an example of tweaks to fit a crush front at story 60.

Edit: catching up with the posts, I see Un-hyphenated has also mentioned the timing and location, and I agree, as noted.

Edit2: I guess it's not fair to invoke a localized pancake collapse AND have a rigid block the mass of 24 stories. Is it? No. I was trying to stay within the slab model. The best quick fix I can come up with is to claim the wedge: upper block sliding down inside the south wall shearing it away from the floor and shoving it outward, sort of like WTC 2. Accumulating debris in front of it.

I should add that's not the impression I get from the expulsions in the animation you posted above, but impressions are not always reliable. This was just an attempt to tackle the apparent anomaly with givens in a simple model.

Cool. I agree with the WTC1 images and I agree with the possibility of applying it to WTC2. Just not from the angle you posted before.

It is nice having you around. Quite a show.

OneWhiteEye wrote:

upper block sliding down inside the south wall shearing it away from the floor and shoving it outward, sort of like WTC 2. Accumulating debris in front of it.

And what would the lower part north wall do to the upper part at the same time? Shearing it away accumulating debris on top on it?

Heiwa wrote:

OneWhiteEye wrote:

upper block sliding down inside the south wall shearing it away from the floor and shoving it outward, sort of like WTC 2. Accumulating debris in front of it.

And what would the lower part north wall do to the upper part at the same time? Shearing it away accumulating debris on top on it?

Not exactly. There's absolutely no reason whatever to expect such symmetry between the two cases, but it still may not be too far off the mark. The NW corner around the 97th floor hangs on quite a while, still there at 3 sec into femr2's video. I wouldn't describe it as 'accumulating on top' but it does seem to clog up in that region for a bit and lag behind the advance of the south side.

Edit: The lower part of the upper block north wall, I believe, was sliced off by and fell outside its counterpart below. Leaving a bunch of rubble to drop inside. Would you expect debris to do a balancing act perched up there on the north perimeter? Or that somehow the north side floor diaphragm could support a third to a half the upper block piled up on it as debris? Arrested? Seriously, there are questions, and then there are...

OneWhiteEye wrote:

Heiwa wrote:

OneWhiteEye wrote:

upper block sliding down inside the south wall shearing it away from the floor and shoving it outward, sort of like WTC 2. Accumulating debris in front of it.

And what would the lower part north wall do to the upper part at the same time? Shearing it away accumulating debris on top on it?

Not exactly. There's absolutely no reason whatever to expect such symmetry between the two cases, but it still may not be too far off the mark. The NW corner around the 97th floor hangs on quite a while, still there at 3 sec into femr2's video. I wouldn't describe it as 'accumulating on top' but it does seem to clog up in that region for a bit and lag behind the advance of the south side.

Edit: The lower part of the upper block north wall, I believe, was sliced off by and fell outside its counterpart below. Leaving a bunch of rubble to drop inside. Would you expect debris to do a balancing act perched up there on the north perimeter? Or that somehow the north side floor diaphragm could support a third to a half the upper block piled up on it as debris? Arrested? Seriously, there are questions, and then there are...

Well, if the upper part slides inside the south wall below, it is the upper south wall that shears off the floors below. This I can accept. But what happens one the north side? You suggest that the lower part north wall of the upper part was sliced off and it must then be by the lower part north wall slicing the floors off the wall of the upper part on the inside.
This scenario I have already described in one of my papers - if the upper part actually drops. The upper part weak elements/floors are sliced into several parts by the lower part vertical strong elements, i.e. the upper part cannot one-way crush down the lower part as described in the crush down model under discussion.

Heiwa wrote:Well, if the upper part slides inside the south wall below, it is the upper south wall that shears off the floors below. This I can accept. But what happens one the north side? You suggest that the lower part north wall of the upper part was sliced off and it must then be by the lower part north wall slicing the floors off the wall of the upper part on the inside.

Seems we are in basic agreement on what can shear off what.

femr2 wrote:I'm cautious to offer 'agreement' to such a large movement of the initiation level, but the model is fully capable of change in the required parameters so I have no problem setting such a run up and rendering the result.

Remember this only applies to the southern end, it's application of the model to a more specific vertical slice in an attempt to match a later ejection in that horizontal region. It would represent the extreme lowest point, but that's precisely the case. IF the model can work under these circumstances (i.e., there really is flat-on-flat inelastic impact as opposed to wedging which conserves momentum but doesn't fully entrain material and isn't necessarily fully inelastic), it can only be expected to work in a 1D slice representative of the conditions at that location. We're only trying to fit the very fastest region. What if the problem were reversed?



It is my opinion that the northernmost, uppermost slice of the building is still above the top of the green region. Can you tell me what tweaks need to be applied to the spreadsheet model to match that? (in case the tweaks should result in arrest, you'll be obligated to provide a mechanism for continuation! haha)

I suggest some additional inspection of the relevant video footage, to which end I've uploaded a 1/3 speed best resolution video of the initial segment of the descent:

I can't dispute that, from this angle, a lot of the top is dropping along your default elevation. Moreover, I haven't examined this issue at the level of detail that others have, likely including yourself. If I had come to some conclusion based on careful research, the tone would be different. Naturally, there's a difference between asserting that the phenomena in question is justified based on an analysis and playing around with uncertainties. As example, if I say that this view most accurately represents the maximum split elevation, it's just an impression. If I'm wrong, oh well.

The mere presence of the ejecta at the level indicated by the lower line indicates this is a more accurate position for an extreme front, especially towards the south. As you note, the video also depicts the progression south to north. In discriminating conditions along the north-south axis, this is worth - on the outside - a second of head start or offset down one floor. It does not necessarily get added to the above but, failing a detailed temporal examination proving otherwise, it would be a prudent assumption. This takes it nominally down to the 94th floor.

Visible evidence from that video must be reconciled with other knowns. One of which is the bowing extent I mentioned earlier. This is strong, independent corroboration for choosing a value no higher than 94. All was I was asking for was two stories beyond that!

The other known is the crop of one of your frames I posted above. Going back to those calculations, and judging crush front location by exactly the same standard as the five second mark, it's already 5 stories ahead. So, either this gets applied to the initial conditions for the model or the model tries to fit this early, huge discrepancy, and I don't think it would succeed. It would surely be more interesting than trying to fit the discrepancy at over 4 seconds later in the descent. That's why I opted to consider this early 5 floor discrepancy better incorporated as an initial offset. I tried to grab one more story with caternary action, but OK, that's cheating. I'll take it on the basis of uncertainty as a reasonable limit, though.

At least I've made a decent case that floors 93-94 are actually a better choice for a southern slice than 98, in absence of more compelling evidence otherwise. Can we agree on 93-94 as nominal, and 95 to 92 as extremes, for the purposes of this context? If so, then it's 92 we must consider, and so you have.

Shall do.
...

Reduces the 11 storey deficit to ~3.5 storeys.

Not sure why we have a three and half floor difference, but it doesn't seem that important to hash out at this moment. The mass distribution may be quite different and I'm always willing to toss in an off-by-one error times two on my part for anything I hack out in 20 minutes - like the previous code. It more than accounts for the difference.

My preference will be to estimate the actual mass of the hat truss and include it within the model. (It already includes addition for the aircraft (fully loaded) and the antenna)

Has anyone produced an estimation of the hat truss mass ?

I don't know of any, but I'll bet it's out there. Best figures are always best(!) but there has to be caution in applying precision inappropriately. Once again, the issue is matching the conditions of a limited region, so using the best figures for even the slender region - in a static layout - is not necessarily the most accurate picture of what mass distributions may have developed dynamically. One does not have to believe that there was necessarily an ongoing 'piling on' to this specific location to admit that there is uncertainty of distribution. Therefore, in order to account for the most extreme visible occurrence of ejecta, it is necessary to adjust to extrema of reasonable potential distribution, and understand that is the real centerline, and there are tighter plus/minus bands around that. Other, prior observations are employed to set constraints, as done above. And, again, the only reason it has to be done is the objective of matching later knowns also involving specific rather than aggregate location. If the purpose is to match the average behavior, then the opposite extreme would be the NW corner between your red lines, still near that elevation a couple of seconds into it.

In this matter, the known tilt must be considered. Even if one allows for total cap destruction within a few seconds (which I would argue against), the last verified position and orientation of the upper block puts the diagonal of the rectangle over the south half. It is a different cross section, too, and not necessarily as dense because of core inclusion. This leads to further speculation. If a larger amount of the block remains intact through the five seconds, then shearing and/or wedging could occur as with the WTC2 east side and the model cannot be expected to be applicable as is.

The other choice is that the upper block partially or largely disintegrates over this period and some amount of rubble is contained. The known existence of the spire not only ensures there was some containment on the other side of the floor span, but also strongly suggests that some of the mass, which would stay in the core region in an unadjusted model, has to make its way downward through the floor region. Containment is guesswork, but skewing of the total impinging mass upward from the static distribution is not to be ignored and, if it fits it almost has to be worn.

In light of that, and knowing the hat truss had to add some mass, additional floors have to be added to my simple uniform mass calculation just to expect proper behavior. I did, six may have been too high, but probably not by much, and may have actually been too little. However, this says nothing about your mass distribution, the two are not comparable. Do you have a handy graph of the upper floor masses contained in your sheet?

Regardless of the progress going forward, I can say this has already proven to be a rewarding discussion as far as I'm concerned. I'll cut to the chase now, since I'm not going to defend to death a 1D simple model in an attempt to match reality. If it comes up short, I will fail the model, simple as that. The next step would be the wedge and, while it's not too hard to support existence, it's a devil to refute because WTC2 did it like nobody's business. Arguing with certainty to the contrary, just on the surface of having WTC2 as an example, and without video of the WTC1 S face, is not a tenable position. We are currently examining a much more restrictive model which can only be shown to succeed or not, it wouldn't be fair to refuse the other side of the coin when it comes up.