AE911 Truth's WTC7 Evaluation Computer Modelling Project

[Mick West]

Not at this juncture no. I would rather explore the potential of not considering the outside walls to be infinitely strong. Perhaps it is that difference in the models that accounts for the disparity in the outputs, in part at least ?

So now you think that when the fire heated the beams and the girders they all expanded at more or less the same rate, and they pushed the columns around, so there was no displacement at any of the connections? What's the largest displacement between adjacent parts in this diagram

C79 is in the middle of the 1.47 to 2.03 band, and says it moved 1.85 east, 0.94 North (or maybe sourth).

But it ALSO means that the beams and the girders and the slab all moved smoothly with it. That makes zero sense.

(EDIT) and means the comparison to the NIST 5.5" figure is meaningless.

 

[gerrycan]

So now you think that when the fire heated the beams and the girders they all expanded at more or less the same rate, and they pushed the columns around, so there was no displacement at any of the connections? What's the largest displacement between adjacent parts in this diagram

View attachment 28845
C79 is in the middle of the 1.47 to 2.03 band, and says it moved 1.85 east, 0.94 North (or maybe sourth).

But it ALSO means that the beams and the girders and the slab all moved smoothly with it. That makes zero sense.

(EDIT) and means the comparison to the NIST 5.5" figure is meaningless.

At this point, Dr Hulsey explains what the movement is relative to in the model.

 

[benthamitemetric]

NIST had these connections fixed. Not realistic. UAF considered them.

Perhaps not as realistic as they could have been. Materially unrealistic? Remains to be seen. Were all of the elements on all of the floors except for 12 and 13 fixed in reality?

I believe the project replicated NIST's stated conditions at some point.

You believe in error. They did not model NIST's fire scenario at all. They didn't even model an actual fire scenario. They applied certain max temperatures from the NIST studies to certain elements (beams/girders/columns) according to their structural element type. There was no actual fire progression model used. I've pointed this out several times, so perhaps you should take a few moments to review the thread's contents.

UAF modelled the whole thing. NIST did not.

No. They modeled fire damage on only two portions of the east sides of floors 12 and 13. They did not model any other fire damage in the building.

Once an element in NIST's model was "deemed to have failed" it was removed, along with any other element attached to it. NIST's walk off theory is beyond the "in trouble" point. It can be discarded. This is a far more accurate model from UAF that contradicts NIST's findings directly and consistently across 2 programs.

How were connection failures or buckled elements treated in Hulsey's model, again? As far as I can tell, we don't actually know, so its really putting the cart before the horse for you to claim Hulsey's treatment of them was more accurate than NIST's.

Right.

Glad we agree.

 

[Mick West]

At this point, Dr Hulsey explains what the movement is relative to in the model.

But that's irrelevant. As the collapse criteria is the movement of the connection point of one member relative to its connection point on the other member. That's what the 5.5" refers to. So the comparison is invalid. And this displacement diagram is meaningless because it's not going to cause any failures, or even put any strain on anything, because there's no discontinuities in the displacement.

 

[Mick West]

The columns (inner and outer) must have some resistance to moving. They are not floating in space - which is how it looks like they are modelled here.

 

[Oystein]

Sure, yes. But the connection has been modelled and we have been shown it, along with the drawings.
UAF are way more accurate than NIST. ...

...one one (1) out of thousands of connections and other elements.

They are FAR less accurate on most of everything else, such as the entire business of assessing the fire situatiuon, or the floors below and above the two they considered.

So which was, overall, the more accurate model?

Again? Such question (yours primarily; mine obviously, too, as I am mocking yours) are silly and a detraction. Your focus on one (1) element to the exclusion of literally everything else is [...].

 

[Oystein]

Not at this juncture no. I would rather explore the potential of not considering the outside walls to be infinitely strong. Perhaps it is that difference in the models that accounts for the disparity in the outputs, in part at least ?

This is a misunderstanding of the NIST model!

NIST did not model the walls infinitely strong. They merely did not give the connections within the wall any specified failure modes.
I detailed that in a post earlier tonight, which got duly ignored.

Hulsey does not understand the NIST modelling in a couple of aspects. That is worrisome.

 

[Oystein]

I don't believe that Hulsey has proven anything or can conclude that fire could not destroy a steel framed high rise... a core theme of Gage's CD belief.

Both gerrycan and kawika have already agreed that Hulsey's conclusion is invalid. Tony Szamboti already agreed to this tacitly the other day.

 

[Oystein]

NIST had these connections fixed. Not realistic. UAF considered them....

More precisely: NIST did not consider any connection failure modes within the walls, having determined from real-world observation and from preliminary models that this would be unlikely anyway.

Now: Did Hulsey's modelling observe and failure modes occurring within the wall? Or did his model show that no connection with in the walls failed?
If the latter, then Hulsey has validated NIST's judgement to not consider the potential of connection failure within the walls.

 

[Oystein]

Since the issue of allegedly "infinitely stiff exterior walls" has been brought up despite it having been debunkted two pages ago, allow me to repost:


On pages 19 and 20 of his presentation, Hulsey shows the floor area inside which NIST modeled connection failures, and outside which they did not. This is from the 16-story ANSYS model, Hulsey shows NIST's "Figure 11–9. Area of the floor where connection failures were modeled" on page 476 of NCSTAR 1-9 (Although, weirdly, his presentation actually references NIST 2004 - why would they do that? I imagine that he took recourse to a very old AE911Truth presentation given to him). Note the bolded word: failures. It's the operative word here! NIST provides an explanation for their choice:

The extent of the area with detailed connection models was based on the results of single floor fire simulations, where connection damage west of Columns 73 through 76 were not found to contribute to an initial failure event on the east side of the structure. The area where break elements were modeled was selected to reduce the model size without biasing the results for simulating the initial failure event.

Hulsey does not address this. Instead, he claims:

1. Outside the selected area, connection failures were not modeled (NIST used fixed or pinned connections).

This omits what NIST actually did to model damage to the western part (which is not entirely irrelevant, IMO):

Outside the selected area, structural damage–such as buckling of the steel frame and crushing and cracking of the concrete slab–was modeled over the entire floor, but connection failures were not modeled.

More importantly, Hulsey claims that "2. Connections were not modeled for the exterior moment frame". I think he (and now @gerrycan !) misconstrues this as "exterior moment frame was totally rigid". But read the NIST report in context: You find this all in section "11.2.5 Modeling Connections", which starts on page 473. On the same page, they start discussing "Modeling Failure with Break Elements". The critical sentence is on page 475:

The floor area where failure of floor framing connections and shear studs was modeled with break elements on Floors 8 to 14 is shown in Figure 11–9. This area is east of the north-south line passing through Column 76 and the core area east of Column 73.

Bolding mine: It's not that they didn't model the connections at all - they probably did! -, they just didn't allow for the connections in the wall or west of the selected area to fail under the modelled loads. I thus assume that particularly the east wall was allowed to respond laterally to the pressure of expanding floor beams. I have not found explicit confirmation of this, but hints such as on page 490 (my bolding):

Interior columns were not thermally restrained and were free to expand. Exterior columns had some thermal restraint due to the moment framing, but their temperatures remained low. Axial column stresses remained essentially constant because (1) interior column temperatures were less than 200 ºC and exterior column temperatures were less than 150 ºC, except for a few columns in the 12th and 13th floors which were less than 300 ºC; and (2) the failure of girders in a floor did not noticeably decrease the column load.

Also, Section "11.2.7 Boundary Conditions and Loads" does mention how the top and bottom of the 16-floor model were fixed, but do not mention that the walls or the western part are fixed, so I assume they were not.

(Page 21 of the presentation presents a quote attributed to "NCSTAR 1-9 Page 525, 2008" - but it's neither on the actual page 525, nor on page 525 of NIST's PDF. I again wonder if he lifted this from some very old AE911Truth presentation).

To summarize: Hulsey tries to convince us that the 16-floor ANSYS model of Chapter 11 was modelled with the exterior and the western part made rigid, when in fact NIST had only decided not to model connection failures outside the east floor framing. They did allow the exterior and the western part to respond laterally to what the beams did when heated.


Now on to a big blunder in Hulsey's presentation, which he hasn't corrected since a year ago - and it is a pity that AE911truth as well as other Truther sites actively censored me:

Please turn to page 24:

• On the left, he again shows Figure 11–9. Area of the floor where connection failures were modeled, only upside down now. Notice: This is a boundary within the ANSYS model.
• On the right, he shows a snapshot from LS-DYNA animation, a model introduced in Chapter 12.
• The page is captured with this statement: "Connections were not modeled; outside selected blue space."

That statement is UNTRUE for both the ANSYS and the LS-DYNA model!!

• It misconstrues what Figure 11–9 actually shows - "Area of the floor where connection failures were modeled"
• More importantly: This distinction simply does not apply to the LS-DYNA model! Hulsey conflates two distinct models!

In other words: Dr. Leroy Hulsey, as of September 06, 2017, has not understood the NIST models!

 

[gerrycan]

Since the issue of allegedly "infinitely stiff exterior walls" has been brought up despite it having been debunkted two pages ago, allow me to repost:


On pages 19 and 20 of his presentation, Hulsey shows the floor area inside which NIST modeled connection failures, and outside which they did not. This is from the 16-story ANSYS model, Hulsey shows NIST's "Figure 11–9. Area of the floor where connection failures were modeled" on page 476 of NCSTAR 1-9 (Although, weirdly, his presentation actually references NIST 2004 - why would they do that? I imagine that he took recourse to a very old AE911Truth presentation given to him). Note the bolded word: failures. It's the operative word here! NIST provides an explanation for their choice:

The extent of the area with detailed connection models was based on the results of single floor fire simulations, where connection damage west of Columns 73 through 76 were not found to contribute to an initial failure event on the east side of the structure. The area where break elements were modeled was selected to reduce the model size without biasing the results for simulating the initial failure event.

Hulsey does not address this. Instead, he claims:

1. Outside the selected area, connection failures were not modeled (NIST used fixed or pinned connections).

This omits what NIST actually did to model damage to the western part (which is not entirely irrelevant, IMO):

Outside the selected area, structural damage–such as buckling of the steel frame and crushing and cracking of the concrete slab–was modeled over the entire floor, but connection failures were not modeled.

More importantly, Hulsey claims that "2. Connections were not modeled for the exterior moment frame". I think he (and now @gerrycan !) misconstrues this as "exterior moment frame was totally rigid". But read the NIST report in context: You find this all in section "11.2.5 Modeling Connections", which starts on page 473. On the same page, they start discussing "Modeling Failure with Break Elements". The critical sentence is on page 475:

The floor area where failure of floor framing connections and shear studs was modeled with break elements on Floors 8 to 14 is shown in Figure 11–9. This area is east of the north-south line passing through Column 76 and the core area east of Column 73.

Bolding mine: It's not that they didn't model the connections at all - they probably did! -, they just didn't allow for the connections in the wall or west of the selected area to fail under the modelled loads. I thus assume that particularly the east wall was allowed to respond laterally to the pressure of expanding floor beams. I have not found explicit confirmation of this, but hints such as on page 490 (my bolding):

Interior columns were not thermally restrained and were free to expand. Exterior columns had some thermal restraint due to the moment framing, but their temperatures remained low. Axial column stresses remained essentially constant because (1) interior column temperatures were less than 200 ºC and exterior column temperatures were less than 150 ºC, except for a few columns in the 12th and 13th floors which were less than 300 ºC; and (2) the failure of girders in a floor did not noticeably decrease the column load.

Also, Section "11.2.7 Boundary Conditions and Loads" does mention how the top and bottom of the 16-floor model were fixed, but do not mention that the walls or the western part are fixed, so I assume they were not.

(Page 21 of the presentation presents a quote attributed to "NCSTAR 1-9 Page 525, 2008" - but it's neither on the actual page 525, nor on page 525 of NIST's PDF. I again wonder if he lifted this from some very old AE911Truth presentation).

To summarize: Hulsey tries to convince us that the 16-floor ANSYS model of Chapter 11 was modelled with the exterior and the western part made rigid, when in fact NIST had only decided not to model connection failures outside the east floor framing. They did allow the exterior and the western part to respond laterally to what the beams did when heated.


Now on to a big blunder in Hulsey's presentation, which he hasn't corrected since a year ago - and it is a pity that AE911truth as well as other Truther sites actively censored me:

Please turn to page 24:
View attachment 28846

• On the left, he again shows Figure 11–9. Area of the floor where connection failures were modeled, only upside down now. Notice: This is a boundary within the ANSYS model.
• On the right, he shows a snapshot from LS-DYNA animation, a model introduced in Chapter 12.
• The page is captured with this statement: "Connections were not modeled; outside selected blue space."

That statement is UNTRUE for both the ANSYS and the LS-DYNA model!!

• It misconstrues what Figure 11–9 actually shows - "Area of the floor where connection failures were modeled"
• More importantly: This distinction simply does not apply to the LS-DYNA model! Hulsey conflates two distinct models!

In other words: Dr. Leroy Hulsey, as of September 06, 2017, has not understood the NIST models!

If that were the case for NIST then they would have mentioned the K3004 beam connection to column 39 and how that would result in less expansion to the West by K3004. The do not account for it.

 

[Mick West]

Displacement map again. Hulsey says earlier in the presentation that in his simulation expansion of the beams pushes girder A2001 against the flange of C79, making it get stuck there and not fail.

That distance is 4" inches. NIST says the girder would have to move 5.5" (or later 6.25") in order to fall off the seat. Hence Hulsey, et al, consider the NIST report to be invalid.

But later in the presentation we have what is the result of the SAP model.

DL: 5,142k
x-1.92"; y=0.73"
The floor moved
East @ column 79.

(east is to the right in this view)

What does this mean? Is the girder not part of the floor? If the floor has moved right 2", and the girder has moved 4" to the left relative to the column (as in the animation above), then does that mean that the column has "moved east" 6"?

Hulsey says:

59:42 movement of that section right there of
59:47 almost two inches but not west, east!
59:54 totally different the floor moved east
59:57 at column 79 not west and by the way
60:01 they move together they didn't move
60:02 separately

Then we have this oddly smooth displacement map, supposed the output of some Abaqus model:

Is this indicating the "floor" moving? No, here Hulsey says column 79 is moving.

61:05 Floor 12 and 13 UAF Abaqus
61:10 structural expansive movements. Just kind
61:12 of want to share with you this is the
61:13 model we're talking about this is floor
61:16 13 this is the column 79 at 4:13 1.85
61:24 inches to the right. That means East that
61:28 doesn't mean West. So this whole thing
61:30 moved

61:31 I'll show you their different picture
61:33 here hopefully in a moment so this is
61:36 all this stuff to them to the east of
61:39 this so you can begin to see that the
61:43 displacements if you start looking at at
61:47 minus point 5 oh no where's the zero at
61:50 it's it's over here somewhere here's
61:52 column 79 so it's moving with respect to
61:55 this over
61:56 in this area it's not moving with
61:59 respect to this it's not moving with
62:03 respect to that it's moving this way in
62:05 that way with respect to what we call
62:07 the thermal centroid which is your
62:13 stiffer point so you so basically got a
62:16 center of stiffness by which this thing
62:18 is behaving and and and that's what's
62:21 happening here
62:22

The then compares these two figures of floor movement to NIST's 5.5"

The problem is they are measuring radically different things. The SAP and ABAQUS seem to be measurements of the region around Column 79 relative to a "thermal centroid" near the lower middle of the plan. The NIST figure, on the other hand is a measure of the relative movement of the 79-44 girder relative to column 79. That would be the same as the approximately 4" (west) movement that Hulsey himself found in the more detailed model.

It seems like in Hulsey's models the concrete, the beams, the girders, and the columns, have all been glued together in a homogenous slab where nothing moves relative to anything it is in contact with, so no connections are ever broken, things just expand and move out of the way. Columns bend (or move) several inches away from their positions. If that's the case, then what's the purpose of the simulation of the motion of the girder relative to the beam?

This seems impossible for a number of reasons - for one the steel and the concrete have very different coefficients of thermal expansion (7.78e-06 for steel, 5.9e-06 for concrete - dolomite aggregate, Hulsey slide 68). For another they heat at different rates, so the steel will heat and expand quicker than concrete, and smaller beams will heat quicker than thick girders and columns. This is seen in the temperature distribution that Hulsey says is being used:

So this homogenous slab approach does not seem plausible. Even if it was, the comparison of the "floor" movement with a relative girder/column movement is not one that makes any sense.

 

[gerrycan]

The first example that you show shows a forced condition (expansion +40%) to get a NIST type failure mechanism. The result is no failure because of the sideplate. On drawing 9114 you can see where the stiffener is and the 3/4" setback.
So the East bottom of the flange has about 1" or so to travel to the C79 flange face. The West flange has about 1.3/4" approx.
On the actual AFU analysis of this on page 43 of the pdf, you can see that the east side of the flange is contacting the C79 flange. It's stuck.

 

[Tony Szamboti]

Both gerrycan and kawika have already agreed that Hulsey's conclusion is invalid. Tony Szamboti already agreed to this tacitly the other day.

I most certainly did not agree that Leroy Hulsey's conclusion is invalid. In fact, I stated the opposite and explained how he would have come to that conclusion in post #323.

If the columns never got hot enough to fail then the only way fire could cause a problem is for dislodgement of horizontal members removing lateral support for a significant number of stories (at least five in most cases) to cause the column(s) to be slender enough to buckle and fail. Once you include the items they ignored and omitted the NIST analysis could not show it happened between columns 44 and 79, and every other situation in the building was even less likely to have it occur. It is clear that it is groping at straws trying to say the building collapsed due to fire once you understand these details.

The temperatures Hulsey used on the structural members involved were also worst case for the NIST simulation, so it is nonsense trying to claim any other combination of temperatures may cause it.

 

[Mick West]

and every other situation in the building was even less likely to have it occur

How do you figure that?

 

[Mick West]

@Tony Szamboti, can you address my post 492, above,regarding the homogenous slab. Is Hulsey actually saying there's no movement at any of the connections?

 

[gerrycan]

@Tony Szamboti, can you address my post 492, above,regarding the homogenous slab. Is Hulsey actually saying there's no movement at any of the connections?

Hope you're not attributing that statement to me. It's the centre of stiffness that doesn't move.

 

[Tony Szamboti]

How do you figure that?

The NIST WTC 7 report (chapter 8) explains that the reason the phenomena of a girder flange folding and allowing it to fall had never happened in previous building fires is that those buildings did not have asymmetric framing of floor beams into girders with seat connections. They said this situation (occurring at the northeast and southeast corners in WTC 7) would allow lateral travel by the girder, and with it being pushed by the asymmetrical framing during its thermal expansion, cause the girder to be vulnerable to unseating. However, we now see that couldn't actually happen if the structural items they ignored and omitted (column side plates and girder web stiffeners) had been considered and included in the analysis. All other situations in the fire affected areas (besides the northeast and southeast corners) had symmetric framing and/or knife connections, which would not allow the girder to travel laterally.

The opportunities for horizontal member dislodgement from the columns, and any subsequent loss of lateral support for enough stories to cause enough slenderness for column buckling, due to heating from fires, were essentially nil. Therefore, no vertical collapse could have occurred due to fire. Dr. Hulsey is right.

The only way vertical collapse could happen was for the columns to be directly weakened and the fires could not heat them enough since the columns had a heat transfer path to cooler areas. The fire simulations bear this out, as they show the columns never got hotter than 300 C. At 300 C structural steel still retains all of its yield strength.

 

[Jeffrey Orling]

OT... How odd that the scientific work was timed to coincide with the anniversary of 9/11. What's up with that?

 

[Mick West]

Hope you're not attributing that statement to me. It's the centre of stiffness that doesn't move.

No I'm not, I'm attributing it to Hulsey who said (#492),

here's
61:52 column 79 so it's moving with respect to
61:55 this over
61:56 in this area it's not moving with
61:59 respect to this it's not moving with
62:03 respect to that it's moving this way in
62:05 that way with respect to what we call
62:07 the thermal centroid which is your
62:13 stiffer point so you so basically got a
62:16 center of stiffness

So again, he shows a generic value of floor displacement relative to a center of stiffness, and compares it to a specific displacement of a girder relative to a column.

Would you agree that's an invalid comparison?

Would you also agree that the displace diagram appears to show ZERO local displacements, only the expansion of a homogeneous slab?

 

[Oystein]

If that were the case for NIST then they would have mentioned the K3004 beam connection to column 39 and how that would result in less expansion to the West by K3004. The do not account for it.

Huh?
If what were the case?
Did you quote the wrong post? Because your reply in no way whatsoever references the post you quoted.
The main parts were:

1. Hulsey misconstrues how NIST modeled connections inside and outside a certain area in the eastern part of the 16-story ANSYS model of Chapter 11. In particular, the claim that the east wall was modeled as infinitely stiff is plain wrong. NIST merely decided, with reasons stated, not to give the connections properties modeling various enumerated failure modes. In other words, they assumed that connections between wall elements would not fail. The elements, and wall as as such, however was still modelled with connections, and responding vertically and laterally to the stresses that came up in the simulation.
2. Hulsey conflates the 16-story ANSYS model (Chapter 11) and the 47-story LS-DYNA model (Chapter 12) when he juxtaposes Figure from either Chapter and writes about the full collapse model animation that NIST did not model connection failures west of a certain line. This is plain WRONG.

Both errors, which occupy a total of 4 or 5 slides, demonstrate that Hulsey to this date does not understood what, how and why NIST modelled what they modelled. This is cause for grave concern. Like you, like AE911Truth, he concentrates on one detail, while demonstrating incompetence on the global picture.

 

[Oystein]

I most certainly did not agree that Leroy Hulsey's conclusion is invalid. In fact, I stated the opposite and explained how he would have come to that conclusion in post #323. ...

As a matter of fact, you DID agree that Leroy Leroy Hulsey's conclusion is invalid in post #323 where you wrote:

I just want to explain why it is extremely difficult for a steel framed high-rise building to collapse due to fire...

"Difficult" <> "Impossible".
Post #323 goes on not talking about what Hulsey did in his study, but what you think about other studies, and what your base belief is about buildings (that highrises cannot collapse from fire, period). By inputting this FALSE premise you are certain to arrive at the FALSE conclusion - but you failed to show how Hulsey's study actually supports Hulsey's claim. Instead, you employ a subjunctive: "...how he would have come to that conclusion..."
"Would" <> "Did"

 

[Oystein]

The NIST WTC 7 report (chapter 8) explains that the reason the phenomena of a girder flange folding and allowing it to fall had never happened in previous building fires is that those buildings did not have asymmetric framing of floor beams into girders with seat connections. ...

"girder flange folding" is not the only failure mode that NIST considered.

In fact, NIST found that quite a number of girders and beams lost vertical support, on several floors and several columns as well as several beams framing into girders. The other serious engineering studies found further candidates of such connection failures, involving further failure modes.

Hulsey did not even begin to assess these multiples of multiples of floor frame failures when he refused to model fire progression entirely, and employed a heat distribution snapshot to only two floors.

Therefore, Hulsey's most definitely did not, could not, show that fire didn't bring down WTC7.

 

[Mick West]

Let me focus on this objection again, as I'd like @gerrycan and @Tony Szamboti to weigh in on this specifically.

It seems that the comparison Hulsey made on this slide is incorrect:

The NIST value (6.2" west), is the movement of the end of A2001 (the girder between Columns 79 and 44) relative to C79, due to the theorized thermal expansion of the beams perpendicular to A2001.

The ABAQUS value (1.85" east) appears to be the movement of a composite model and includes the movement of the column, the floor slab, and the beams and girders. Hulsey says in the presentation "they move together they didn't move separately". The ABAQUS movement is relative to a thermal centroid or center of stiffness. It does not appear in this ABAQUS model that there is any relative movement between A2001 and C79, as there are no discontinuities in the displacement map:

So if one number is a measure of local relative displacement between a girder and a column, and the other number is a measure of the movement of the system as a whole, where girder and column move together, then how can a comparison be valid?

 

[Oystein]

OT... How odd that the scientific work was timed to coincide with the anniversary of 9/11. What's up with that?

The calendar days Sept 11+12 are the two days in the year where reporting of 9/11 related stuff reaches the largest audience.

By presenting FALSE conclusions just days before those days, Gage makes sure that a measured and truthful response is not ready during the two days of maximum attention.

This is a Propaganda trick, plain and simple. It's understandable if you put yourself in Gage's shoes, but ethically questionable in Hulsey's case, particularly given that solicitation for MONEY was intertwined with the promise, completely broken as we see today, that Hulsey would keep his study transparent throughout the process, such that the public stayed informed of its progress the entire time; and the promise, also broken thus far, that all data would be made public.

 

[deirdre]

a read a bit of Chapter 8 since Tony brought it up, and it says there were no sheer studs, and they got this from structural drawings "Cantor 1985".. I don't want to watch an hour long video again, did Husley say why he thinks there are sheer studs?

 

[Mick West]

Partly answering my own question, back on October 2106 Hulsey gave a talk with the same displacement slides as above. At the end of the talk he was asked a question that gives some indication of the state of the research, at 46:05

Question: I guess the NIST assumption is that those bolts sheared on the girder seat and then pushed it off.

Hulsey: yeah, yeah ...

Question: Did you look at the shear strength of those bolts, along with the shear strength of whatever shear studs on top of that girder. I mean, the amount of force you must have to take to shear all that off and displace that girder has got to be pretty huge. [inaudible] have the capacity to displace it like that

Hulsey: We're looking at that very issue today. Feng who is my graduate with a PhD is looking at that today and in the last couple of days he has said those forces are huge, but in reality that can't be the case because that a most flexible piece. And so you've got to remember that it's also connected to that floor system and so if it can't, if it's going to deform, then, and it can't handle those forces, it's going to hand them off to something that can handle them and so that model is being looked at just to make sure that that's being handled correctly. So I don't know the answer to that yet pat, but I do know we're gonna check and see on how much movement actually occurred and where those forces get distributed to, who's carrying what. So we're going to completely dissect that whole area and see what's happening.

So basically at that point they have not even modelled the motion of the girder relative to the column. Yet they still made a huge point of comparison of composite free-floating slab expansion with the NIST relative motion of the girder to the column.

So his comparison (1.85" east vs. 6.2" west) was utterly meaningless in 2016, and still meaningless today because he's just saying the exact same thing, with the same slides, as he said in 2016.

 

[gerrycan]

@mick Do you think that UAF were right to include and account for the floor system in their model ?
NIST did not.

 

[Oystein]

@mick Do you think that UAF were right to include and account for the floor system in their model ?
NIST did not.

What the...!??

 

[Mick West]

@mick Do you think that UAF were right to include and account for the floor system in their model ?

NIST included the floors, Hulsey has included them in a different way.

Are you following my objection here? It seems like he included the floors so well that nothing is moving relative to them. You know that makes no sense. It also makes most of his presentation (and all your work) regarding the side plates, stiffener plates, and beam buckling irrelevant.

 

[gerrycan]

I see exactly what your objection is. If the floor system moves as a whole to the East at C79 relative to to centre stiff point, or indeed the column does then that has to be translated as, for example, movement of the girder relative to the column in order to be compared on a like for like basis.
I think the key phrase from Dr Hulsey is that we are looking at a week or so worth of examples from a 3 year game. And we are looking at them in what seems like a very mixed and matched way. It's difficult to make wholesale judgements on so little info, without context.

 

[Mick West]

Since a lot of information is getting buried in this thread, I'm writing a post including the major problems with the study.
https://www.metabunk.org/debunked-uaf-study-shows-wtc7-could-not-have-collapsed-from-fire.t9056/

It is intended to be a go-to reference for people who want to quickly understand the study and its problems. It's in skeleton form right now. I invite suggestions before I publicise it.

 

[gerrycan]

Since a lot of information is getting buried in this thread, I'm writing a post including the major problems with the study.
https://www.metabunk.org/debunked-uaf-study-shows-wtc7-could-not-have-collapsed-from-fire.t9056/

It is intended to be a go-to reference for people who want to quickly understand the study and its problems. It's in skeleton form right now. I invite suggestions before I publicise it.

So you are claiming to have debunked the UAF study ? This should be good. Fire away.

 

[Oystein]

So you are claiming to have debunked the UAF study ? This should be good. Fire away.

Rule of "So": No, you misconstrue.

We have debunked that the specific claim "Fire did not and could not have caused the failure of this building" follows from the study. You already agreed to this, remember?

This does not mean that the entire study is debunked, as it may or may not turn out that it raises valid objections to NIST's results re a certain single connection.

 

[James Ice]

Has the presentation PDF been linked to already? Anyway, here it is:

http://ine.uaf.edu/media/92216/wtc7-structural-reevaluation_progress-report_2017-9-7.pdf

I skip ahead to the conclusions:

Page 83:

Did Building 7 Collapse from Fires?
• No (this is based on our calculations)

Page 84:

UAF Conclusions:
• The concrete floor diaphragm stiffness is significant and even with no shear connectors, frictional resistance to thermal expansion is not trival.
• The thermal expansion of the concrete deck cannot be ignored and it is likely less than steel (the value is highly dependent upon the type of aggregate).
• The research team evaluated fire by considering the air space below the beams in the space between the drop ceiling and the structural steel framing. The result is that a fire underneath will likely burn through the drop ceiling quickly and its resistance to heat transfer is likely not available to help.
• The NIST vertical collapse was not consistent with that of the actual collapse. The difference was primarily influenced by not modeling a significant portion of the structural framing connection details.

The last bullet point on page 84 apparently refers exclusively to the C79-44 connection.

We see that indeed the top-most conclusion is this "Did Building 7 Collapse from Fires? No" thing, which Tony Szamboti, kawika, gerrican and John85 are so studiously avoiding, thereby giving the impression that all four of them agree with us that this conclusion cannot validly follow from the study Hulsey did.

Which at this point invalidates completely the presentation and the study.

Tony Szamboti wrote a nice thing about how Hulsey was justified in saying what he said, also going into the acknowledgement of the validity of the objection that strictly speaking fire might not be ruled out as a cause.

Perhaps you are aware of the philosophy thing where it is impossible to prove that you are not just a brain in a vat (as it often goes). You cannot prove that this world you are experiencing is real. Well, this is more or less on the same level as the objection that Hulsey has not proved that fire could not have caused the destruction of WTC7.

You may as well say, one cannot prove that I am a brain in a vat, and this invalidates completely the presentation and the study.

 

[Ph_]

You may as well say, one cannot prove that I am a brain in a vat, and this invalidates completely the presentation and the study.

This is called Russell's Teapot. Your reasoning is flawed.

He wrote that if he were to assert, without offering proof, that a teapot orbits the Sun somewhere in space between the Earth and Mars, he could not expect anyone to believe him solely because his assertion could not be proven wrong.
To illustrate that the philosophic burden of proof lies upon a person making unfalsifiable claims, rather than shifting the burden of disproof to others.

So Hulsey needs to prove his point, not shift the burden of disproof to us.

 

[gerrycan]

Since a lot of information is getting buried in this thread, I'm writing a post including the major problems with the study.
https://www.metabunk.org/debunked-uaf-study-shows-wtc7-could-not-have-collapsed-from-fire.t9056/

It is intended to be a go-to reference for people who want to quickly understand the study and its problems. It's in skeleton form right now. I invite suggestions before I publicise it.

Debunk page 48 please.

 

[Mick West]

Let me focus on this objection again, as I'd like @gerrycan and @Tony Szamboti to weigh in on this specifically.

It seems that the comparison Hulsey made on this slide is incorrect:

The NIST value (6.2" west), is the movement of the end of A2001 (the girder between Columns 79 and 44) relative to C79, due to the theorized thermal expansion of the beams perpendicular to A2001.

The ABAQUS value (1.85" east) appears to be the movement of a composite model and includes the movement of the column, the floor slab, and the beams and girders. Hulsey says in the presentation "they move together they didn't move separately". The ABAQUS movement is relative to a thermal centroid or center of stiffness. It does not appear in this ABAQUS model that there is any relative movement between A2001 and C79, as there are no discontinuities in the displacement map:

So if one number is a measure of local relative displacement between a girder and a column, and the other number is a measure of the movement of the system as a whole, where girder and column move together, then how can a comparison be valid?

This whole seemingly false comparison seems to go back even further, to October 2015:

Source: https://youtu.be/R-zPna-OI70?t=30m10s

It's supposed to move about 6 inches, it only moves about 2.9 inches, 1.6, not enough to move off the bearing

Again, this appears to be a global absolute movement, not a local relative movement. It's the relative movements that cause connection failure.

 

[Mick West]

Debunk page 48 please.

Do all the other problems mean nothing to you? You asked me earlier if I knew about the C79 change of direction. Is this no longer relevant?

 

[benthamitemetric]

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.