AE911 Truth's WTC7 Evaluation Computer Modelling Project

Status
Not open for further replies.
Depends on the connections. But what matters is what actually happens at the ends. Stiffness is just an abstract measurement.
No it isn't. If I have a beam and one end is stiffer and I heat it, which end exerts more on it's connection as a % of it's capacity ?
Not abstract at all. Would be handy for you if it was though.
 
You can't compare that way. Just because they both possibly suck, in different ways, doesn't mean one is better than the other.
Husley completely neglected to model the actual events. The event is important, ergo: not good
Nist omitted some elements for whatever reason. IF these elements are important: not good.

You haven't even seen anything like or near the progressive collapse model from Dr Hulsey.
It isn't done yet.
So which model would you trust. The one that sucks more, or the one that sucks less ?
 
It's in this thread. the ANSYS outputs being applied to the LS-DYNA model. Wasn't that actually a reply to you already ?


So we agree that UAF did a much better job of modelling the NE of the building, including the elements that were there on the drawings for both parties to see. Elements that NIST omitted. Thanks.

It's better to include those elements as a general matter, but I don't think adding elements that were immaterial makes one model "much better" than the other, especially when the the model that includes those elements is far less accurate on actual material issues (such as the heating progression). You are still assuming facts not yet established re the materiality of those added elements. And you have no response re Arup, which is noted.

Re NIST's heat model [...]. NIST generated its thermal load data from FDS and then applied the temperatures over time to its ANSYS model using FSI, from which it calculated damage to the floors. Once it was determined that enough damage to the floors had occurred for the collapse to progress globally, then NIST output the temperature data from the ANSYS model at that point in time into the LS-DYNA model so that the LS-DYNA model could take over the event simulation where the ANSYS model left off. NIST's floor damage calculations were absolutely done in accordance with a varied temperature progression applied from its FDS data, and even the temperature output to the LS-DYNA model was not "even heating"--it was just the output of the FDS-derived variable temperatures at the time of global failure.

This is all described in incredible detail in NIST NCSTAR 1-9, chapters 10, 11 and 12. How many years have you been arguing about the NIST report? How can you honestly not know how their models worked? [...]
 
Last edited by a moderator:
I should have added the "star" shape is lying flat on the ground and is entirely free.

Which brings us to another question about Hulsey's models. Were the columns just floating in space, or sliding around on some frictionless ground?

In the real world, and in the global collapse ANSYS model the columns were fixed to the ground. In the 16-floor NIST ANSYS model they were fixed at the top and bottom. I'm not sure about the sub-assembly.

But what about Hulsey. Did the columns resist movement?
 
No it isn't. If I have a beam and one end is stiffer and I heat it, which end exerts more on it's connection as a % of it's capacity ?
Not abstract at all. Would be handy for you if it was though.

Of course, but that's just describing what happens. It's a result of the structure, or of the model. It's not a variable, it's an emergent property of the connections and the structure.
 
Of course, but that's just describing what happens. It's a result of the structure, or of the model. It's not a variable, it's an emergent property of the connections and the structure.
Hang on Mick. You said it was just an abstract value. Yet on the previous page it was oh so important because it was moving the centroid on a single element ?
You need to clarify this. Is stiffness important. Forget connections and steel shapes for now. As a value applied to or used in a model, is stiffness important or not.
It cannot be both.
 
Why do I have to trust either model? if they both suck, they both suck. degrees of suckness don't matter.
People on here, yourself included seemed happy enough to take NIST at their word to an extent.

Degrees of SUCTION surely do matter. Inversely less suction can keep something from falling in this instance.
 
Hang on Mick. You said it was just an abstract value. Yet on the previous page it was oh so important because it was moving the centroid on a single element ?
You need to clarify this. Is stiffness important. Forget connections and steel shapes for now. As a value applied to or used in a model, is stiffness important or not.
It cannot be both.

I'm not sure you are using terms here consistently. Nor am I sure what you are paraphrasing there.

Of course "stiffness" is important, if that's an actual parameter to your model.

But if it's just the result of other parameters, then no, it's not important, it's just a derived value, or an emergent value.

Compare it with the center of gravity (or center of mass of a spaceship). Is that an important position, sure. It's the point about which the system will tend to rotate. But it's just an emergent position, the average of the mass of the full system.

The "center of stiffness" of the entire building is just a calculated point in space. Since you've been Googling this I'm sure you've seen this slide:
20170918-083846-h28d1.jpg

It's perhaps a useful point for doing calculations on individual members, or an an elastic system. But if the system is deforming inelastically it's not really relevant.
 
If you heat a metal structure evenly and slowly, it expands evenly, about the center
No it doesn't. The exception being if it is equally stiff at either end.

( i just edited a reply to an earlier post you made by mistake on the last page and i cannot undo it, sorry if it breaks the context)
 
I'm not sure you are using terms here consistently. Nor am I sure what you are paraphrasing there.

Of course "stiffness" is important, if that's an actual parameter to your model.

But if it's just the result of other parameters, then no, it's not important, it's just a derived value, or an emergent value.

Compare it with the center of gravity (or center of mass of a spaceship). Is that an important position, sure. It's the point about which the system will tend to rotate. But it's just an emergent position, the average of the mass of the full system.

The "center of stiffness" of the entire building is just a calculated point in space. Since you've been Googling this I'm sure you've seen this slide:
20170918-083846-h28d1.jpg

It's perhaps a useful point for doing calculations on individual members, or an an elastic system. But if the system is deforming inelastically it's not really relevant.

Nope. These are 4 different things. Stiffness is one of them.
And if a system has undergone plastic deformation, then what does that do to the centre of stiffness?
I could always direct you to the previous page.....
OF COURSE it is relevant. And remains so.

ADD And that POOR horse in the photo there, on it's way to Sunder's shredder no doubt.
 
Abstractly, yes. But what if the outer connections were badass fully moment resisting connections, and the center ones were just glued to a bit of wood?
But they're not are they because we specified that it was I beams and did not say any part was of a different grade or material.
 
Gerry, you keep saying the center of stiffness is important. But I'm really not getting why it's important? I mean the strong nuclear force is important, gravity is important, sure. But what exactly are you disputing here?

It seems like you just want to give some importance to the 2" movement?

Surely we can both agree that the important result of the model is if the connections fail? So while the center of stiffness is certainly a position that can be calculated from the model, I don't really see what you are arguing for or against?
 
Gerry, you keep saying the center of stiffness is important. But I'm really not getting why it's important? I mean the strong nuclear force is important, gravity is important, sure. But what exactly are you disputing here?

It seems like you just want to give some importance to the 2" movement?

Surely we can both agree that the important result of the model is if the connections fail? So while the center of stiffness is certainly a position that can be calculated from the model, I don't really see what you are arguing for or against?
You claimed that Dr Hulsey was in error using the centre of stiffness as a point to express movement in elements with respect to.
Do you still think that ?
 
But they're not are they because we specified that it was I beams and did not say any part was of a different grade or material.
I didn't specify how they were connected either, and you just said "attached". This seems like a trivial and moot point. Unless you can give a really clear explanation of what your problem is, then perhaps we can move on?

You claimed that Dr Hulsey was in error using the centre of stiffness as a point to express movement in elements with respect to.
Do you still think that ?
Did I say that? I think I would have said it's an irrelevance. The important movements are the relative motions of connected members. It's not an error, it's just not useful in respect to connection failures.
 
Did I say that? I think I would have said it's an irrelevance. The important movements are the relative motions of connected members. It's not an error, it's just not useful in respect to connection failures.

Yes, you state

"The study makes incorrect displacement comparisons.
In both 2016 and 2017 Dr. Hulsey made much of a difference in the displacement at column 79 (5.5" west vs. 2" east). But he appears to be comparing the wrong values — global instead of local displacements. https://www.metabunk.org/posts/210992/ "

First of all, you do not know what NIST are moving ANYTHING at that connection relative to. And that is NIST's error in not specifying a datum point.
Secondly, Dr Hulsey actually did replicate NIST's simulation and got 5.1" movement to the West in the girder. You ignore that above.(just 0.4" less than NIST)
Thirdly. If Dr Hulsey had attempted to make a direct comparison to NIST again looking at the whole floor, he would merely have been repeating NIST's error.

Given that we reasonably can say that C79 moved to some degree, tell me what NIST have it and the girder moving with respect to. But please don't just ignore the above 3 points.
Scrub that point of your list please Mick. It's not just a weak point, it's an own goal.
 
Yes, you state

"The study makes incorrect displacement comparisons.
In both 2016 and 2017 Dr. Hulsey made much of a difference in the displacement at column 79 (5.5" west vs. 2" east). But he appears to be comparing the wrong values — global instead of local displacements. https://www.metabunk.org/posts/210992/ "

First of all, you do not know what NIST are moving ANYTHING at that connection relative to. And that is NIST's error in not specifying a datum point.

The datum point is the seat. The 5.5" is the displacement of south end of the girder relative to the seat.

And I fully stand by my point, in fact it's one of the most important points. He is comparing two different things.
 
I didn't specify how they were connected either, and you just said "attached". This seems like a trivial and moot point. Unless you can give a really clear explanation of what your problem is, then perhaps we can move on?

What I actually said I was envisaging was this
"A 6 pointed star made up of I beams attached to a central hex shape. Heat it and the beams expand"
Different isn't it. Because the I beams are equally stiff in the centre in my albeit crude imagined model.
What's in the centre of yours ?
 
The datum point is the seat. The 5.5" is the displacement of south end of the girder relative to the seat.
No. The seat would move with the column. Unless you are saying that the column did not move at all. You need to clarify that. Also you need 6.25". Not 5.5.

And I fully stand by my point, in fact it's one of the most important points. He is comparing two different things.
Who is in error though. Who states a clear reference point ?
you stated that Dr Hulsey was in error when in fact the onus would be on NIST to clarify their fixed point. Which is certainly not the seat.
 
No. The seat would move with the column. Unless you are saying that the column did not move at all. You need to clarify that. Also you need 6.25". Not 5.5.


Who is in error though. Who states a clear reference point ?
you stated that Dr Hulsey was in error when in fact the onus would be on NIST to clarify their fixed point. Which is certainly not the seat.

You are conflating issues here. Hulsey compares his 2" to NISTs 5.5" (or 6.25") which was the distance need for the girder to fail. Since they said it failed in the probably collapse sequence, then obviously it must (in their model) have moved 5.5" or more relative to the seat.

So he is comparing a relative movement to an absolute one. Hence it's not a valid comparison.
 
What I actually said I was envisaging was this
"A 6 pointed star made up of I beams attached to a central hex shape. Heat it and the beams expand"
Different isn't it. Because the I beams are equally stiff in the centre in my albeit crude imagined model.
What's in the centre of yours ?

I did not specify. Like you did not specify how your beams were "attached". And it's not really relevant to the point under discussion is it? The validity of Hulsey's 2" vs. 5.5" comparison.
 
Would you agree this is the correct comparison?

NIST: The south end of A2001 moved 6.25" west, relative to C79
Hulsey: The south end of A2001 did not move relative to C79, but they both moved 2" east due to overall expansion of the building in that general region.
 
You are conflating issues here. Hulsey compares his 2" to NISTs 5.5" (or 6.25")
No I am not. You are seeing a difference in the 2 analyses amd correctly so. There is a difference. One gives a fixed reference point and NIST's doesn't. So your default erroneous position is to presume fault in Dr Hulsey wherever a difference occurs, clearly. The problem here is that NIST give no fixed reference point. You though the seat was a fixed point 5 minutes ago. Do you think it won't move with the column, or that the column won't move. Has to be one or the other, or you are wrong. Pick one.
which was the distance need for the girder to fail. Since they said it failed in the probably collapse sequence, then obviously it must (in their model) have moved 5.5" or more relative to the seat.
West by 6.25" does not fail the girder, not only because of the side plate, but also because of the stiffener plates that you already admitted NIST omitted from the end of the girder. again Dr Hulsey includes them because they are right there in front of his eyes, and yours. And NIST's.

So he is comparing a relative movement to an absolute one. Hence it's not a valid comparison.
What invalidates the comparison is the absence of a given fixed point in NIST's analysis. NIST are in error and you are compounding that error. Willingly and knowingly and also publicly.
 
Would you agree this is the correct comparison?

NIST: The south end of A2001 moved 6.25" west, relative to C79
Hulsey: The south end of A2001 did not move relative to C79, but they both moved 2" east due to overall expansion of the building in that general region.
Yes. You can take the above statement to the bank. That's what happens when the floor and it's composite nature along with the shear studs(which NIST omitted) that you admit were on the C79-44 girder are considered.
It moved together I believe the man said.
 
What invalidates the comparison is the absence of a given fixed point in NIST's analysis. NIST are in error and you are compounding that error. Willingly and knowingly and also publicly.

Again, only relative motion is relevant to connection failure.
 
Again, only relative motion is relevant to connection failure.
Relative to WHAT, according to NIST, be specific please. You said the seat plate 10 minutes ago, What is their fixed point now ?
And you totally avoided answering the rest of that so far.

ADD By "that" I mean post #905 just above.
 
Relative to WHAT, according to NIST, be specific please. You said the seat plate 10 minutes ago, What is their fixed point now ?
And you totally avoided answering the rest of that so far.

Relative to the thing it is connected to.
 
Relative to the thing it is connected to.

And the girder is connected to the seat plate. So is the seat plate not moving with the column, or is the column not moving. Pick one.
Maybe re-read NIST's report first though.
So you are forgetting about the rest of post #905. Fair enough, I am comfortable that anyone reading this will see that for what it is.
 
And the girder is connected to the seat plate. So is the seat plate not moving with the column, or is the column not moving. Pick one.

The seat plate is welded to the column. While it's possible that that might fail that's not a result anyone found. So the column and the seat plate can be considered the same frame of reference for the relative movement of A2001.
 
The seat plate is welded to the column. While it's possible that that might fail that's not a result anyone found. So the column and the seat plate can be considered the same frame of reference for the relative movement of A2001.
So you are supposing that the column 79 DID NOT MOVE ? That's a YES/NO one.

And eh post #905. What about the rest of that. You addressed the last sentence of it only.
 
...
Maybe re-read NIST's report first though. ...

Did you re-read the NIST report re temperatures? Do you want to update your false claim that NIST evenly heated elements in its models? Hard to take your bare assertions about what NIST says seriously after you have demonstrated such poor fidelity with the facts. Going forward, you should cite to exactly what in the NIST report you are referencing each time you make a reference. The exercise should, at least, force you to actually read the material you are commenting on.
 
Did you re-read NIST report re temperatures? Do you want to update your false claim that NIST evenly heated elements in its models? Hard to take your bare assertions about what NIST says seriously after you have demonstrated such poor fidelity with the facts. Going forward, you should cite to exactly what in the NIST report you are referencing each time you make a reference. The exercise should, at least, force you to actually read the material you are commenting on.
I stated that they transferred their temps over to LS-DYNA evenly. It tool 2s to apply the temps.
 
Does anyone know why only floor 12 and 13 were modeled for firedamage? (Methodology slide in Hulsey's presentation)
When you think of the center of stiffness in a vertical plane of the building it surely matters that there were fires on other floors (6 and 7) as well.
 
Does anyone know why only floor 12 and 13 were modeled for firedamage? (Methodology slide in Hulsey's presentation)
When you think of the center of stiffness in a vertical plane of the building it surely matters that there were fires on other floors (6 and 7) as well.
They modelled all 47 storeys in SAP 2000. I think that was an early slide.
 
I stated that they transferred their temps over to LS-DYNA evenly. It tool 2s to apply the temps.

Nope. You stated that NIST heated the elements in its models evenly. By that you meant that NIST "evenly" applied variable temperatures? What does that even mean?

Did you forget what you said?

NIST heated them evenly in their models.

NIST analyzed damage to the floors using variable temperatures over time derived from its FDS model, and NIST applied variable temperatures based on the same model to its global collapse model. There was no "even" heating in NIST's model, not even with respect to a single element: all temperatures applied to all elements flowed from a dynamic simulation of gas temperatures around those elements over time.
 
So you are supposing that the column 79 DID NOT MOVE ? That's a YES/NO one.
It's irrelevant, because it's the relative motion that's important.

Do you understand what I mean by relative motion? For two points A and B it's the change in A-B. (i.e. (A'-B') - (A-B) )
So if the column is A and the end of the girder is B, then any absolute movement of A and/or B is incorporated into the relative motion of A-B
 
Status
Not open for further replies.
Back
Top