AE911 Truth's WTC7 Evaluation Computer Modelling Project

Status
Not open for further replies.
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
Relative to what exactly Mick. What fixed point did NIST have them moving relative to ?
 
Center of rigidity is the stiffness centroid within a floor-diaphragm plan. When the center of rigidity is subjected to lateral loading, the floor diaphragm will experience only translational displacement. Other levels are free to translate and rotate since behavior is coupled both in plan and along height. As a function of structural properties, center of rigidity is independent of loading.
Content from External Source
Not a concept that I was familiar with relevant to this either. Heard of it, as centre of rigidity before.
Assuming that "rigidity" and "stiffness" mean the same her (if they don't, then the quotation defines something that is not "center of stiffness"), then the definition is pretty worthless, as it translates to "Center of Stiffness is the Stiffness Center". D'uh.

(above) + Lots of published papers showing examples of this. Centre of rigidity also refers to it.

http://debug.pi.gr/Default.aspx?ch=726
This demonstrates that the CoS is a point around which all points in a rigid (!) plane rotate, once you subtract translations. It is essentially just a method to shift the coordinate system such that only rotations need to be described. This does not tell me much about the stresses, forces and deformations at any remote connection. In other words: I do not see how this concept applies to the problem at hand - how the girder end moves across its seat. Hulsey is not discussing that, or how, the WTC7 floor system rotates.

I don't see an issue with performing sub simulations on particular areas.
So expressing motion relative to CoS is ... what? Optional? Or 100% the right thing to do? What did you really mean by "the right place" and "it is [the right place] 100%" when you asked and answered: "Is the centre of stiffness the right place to state the amount of movement due to expansion from ? I say it is. 100%." That sounded like "the one and only right place to state the amount of movement". Now you say that not using CoS is okeydokey when focussing on particular areas such as the situation around column 79.

Confusing.
But it would come down to how these are applied, and I see the CoS issue as relevant to that. IOW the building will expand wrt the CoS.
It's great that you "see" that. Can you show it also? I.e. explain why the CoS issue is relevant?

Well they didn't mention their inputs either, but they surely had them.
They surely had inputs? Uhm yeah, d'uh. This doesn't address the question I asked, which was whether NIST made a conceptional mistake by not even mentioning CoS.
Do you claim that NIST did state, or compute, movement relative to CoS, but were silent about it, such that they did something right, only didn't talk about it?
Or do you assume they did not state anything relative to CoS? Then the question remains: Is that a fundamental mistake - or is CoS optional?

I suppose it depends how much you want to trust an agency that gets 11" confused with 1ft when it's written on a structural drawing bill of material right in front of them.
No, not at all. You are poisoning the well here without addressing the question.

No it wouldn't. We were checking the plausibility of NIST's hypothesis. Like with like. They were plain wrong. This has been proven for quite some time, but good to have that confirmed by the UAF study.
If their approach was wrong from the get go for not assessing movements relative to CoS, then repeating the same conceptual mistake is not likely to validly confirm or refute the result. It's a case of "Wrong with wrong" more than "Like with like" - and I guess my point was: Do you think today that your 2013ish analyses were even a valid approach at all, given that you had no idea that stating movements relative to CoS would have been the 100% right thing to do?

What I am getting at of course is for you to realize that stating movements relative to CoS is not just only optional, it is indeed entirely irrelevant, unnecessary. It's a completely arbitrary point. It is just a matter of which orthogonal coordinate system you use - but all orthogonal coordinate systems are completely equivalent to one another. Some are more practical than others, depending on the specific use they are put to. For example, if you want to know how a floor slab moves as a whole in a reaction to a lateral seismic force, CoS seems to be a useful coordinate system, as it can be shown that, as long as all connections hold and the slab does not deform laterally, all movements can be described as simple rotations. But we are in a different scenario, where in fact the slab deforms and we at least allow for connections to fail (that is indeed what we are after!) due to lateral forces of various magnitude going in all sorts of directions - it seems that the CoS coordinate is very ill defined in our case, and unpractical to describe what's going on at some remote connection. It seems a coordinate system local to Columns 79 is best suited to describe the motion of beams and girders near it.

The fact that a beam pinned at one end will expand away from the pinned end. Pin the opposite end instead and the direction in which it will expand axially reverses. Always expanding with respect to the stiffest point.
This explains why the expansion of a beam is best measured relative to a point a hundred yards away that may een be situated in the middle of a floor segment, away from any and all steel member? I think you just explained why Cos is a bad choice of coordinate system!

A beam is connected on both of its end to other members. These other members, due to their own material and geometric properties, to being assembled into a larger framework (which may change over time as other connections fail), and due to current conditions (temperature!), at any one point in time have each a certain stiffness. All you need to describe the movement of that particular beam is to look at what it is connected to, locally, and go from there.

ADD - how i am tending to think of it personally is this ( and this is my own personal way of thinking it out, so worth checking for error)
A 6 pointed star made up of I beams attached to a central hex shape. Heat it and the beams expand. They expand wrt the very centre of the hex shape.
If it is floating in free space. If the hexagon sits on the ground by one of its sides, the center will move up. That's called "boundary conditions", and I wonder at this point if we even know what Hulsey's boundary conditions are.

But that distiction, whether or not the center moves relative to the ground or any other [0,0] point is rather irrelevant if you want to figure out the stresses that arise in the connections in any one corner.

If you measure that from anywhere else you are on a bit steel that is moving, so dealing with 2 changes of position and not just one.
Yes. But remember: All orthogonal coordinate systems are completely equivalent to each other.

Now apply that to our column 79: In a local coordinate system pinned to column 79, only the girder moves. In a CoS-centered coordinate system, the column moves, and the girder moves.
Which one is better?

Apologies if that is maybe not the best way to think of it, or not the best way to explain it. But as I said, it's how I thought it through.
I should have added the "star" shape is lying flat on the ground and is entirely free.
Ah ok, there is your boundary condition.
Was WTC7 lying flat on the ground and being "entirely free"? I don't think so. The bottom ends of the columns were pinned to the foundations. I assume Hulsey modelled at least that correctly. With differential heating, the floor slabs on all floors will experience differential expansion, and with this, their CoS will move relative to the ground, and relative to one another. They will also move relative to pretty much all columns and all beam ends. It is indeed very difficult to see what advantage such an unpredictably moving and almost impossibly to identify point would have as origin of your coordinate system.
 
Now apply that to our column 79: In a local coordinate system pinned to column 79, only the girder moves. In a CoS-centered coordinate system, the column moves, and the girder moves.
Which one is better?
You can't choose the centre of stiffness. And the column moves with the girder. To the East.
I would advise the video with the Indian lady on it for CoS info.
I will address the rest of your post at some point.
 
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.
[GALLERY=media, 73]Firedamage modeling by Ph_ posted Sep 18, 2017 at 7:06 PM[/GALLERY]
 
[GALLERY=media, 73]Firedamage modeling by Ph_ posted Sep 18, 2017 at 7:06 PM[/GALLERY]
Are you saying that they did not build a 47 storey SAP 2000 model ?
Look apologies but I cannot debate 4 people at once. Plain to see the tactic though.
 
Mick, can we return to my post #905 now please ?

ADD or are we just going to bury that in huge long posts and forget about it ?
 
Are you saying that they did not build a 47 storey SAP 2000 model ?
Look apologies but I cannot debate 4 people at once. Plain to see the tactic though.
Just reply when you have a bit more time, i can wait patiently.
No tactics, just looking for information regarding this point, no hurries.
 
Don't insult my intelligence Mick. Please.
How far does the column move and in what direction ?

I'm not insulting you. It just seems like you are missing a point here. You keep talking about absolute motion, when relative motion is the only thing relative motion. The motion of the end of the girder relative to the seat which is welded to the column.

The motion of the column in a fixed frame of reference is irrelevant to the relative displacement. It contributes to it. But it's not an important number in itself. The important number is the relative change in position of the girder and the column/seat.
 
Here is the CoS part from the presentation

He says:

56:07 So it's [Col79] moving with respect to this over here, this area (points to CoS)
56:10 It's not moving with respect to this (points along the east wall)
56:15 It's not moving with respect to that (points along the west wall)
56:18 It's moving this way (an arrow points west from CoS)
56:19 or that way (an arrow points east from CoS)
56:20 with respect to what we call the thermal centroid.
Content from External Source
But it's plain nonsense to say Col 79 does not move with respect to the east or west wall - of course it does move with respect to either!
I understand that Hulsey merely informs us what his origin is - that it is neither that wall not that but - an arbitrary point somewhere quite far away and only loosly connected to col 79. That's a valid coordinate system, but he doesn't even begin to attempt to explain why it is a good choice.
And of course, as Mick has pointed out so often, he then compares his values relative to that point to NIST, who used a coordinate system fixed on column 79, which has moved relative to his.

And this is a also a great explanation of CoS and also CoM(ass)

The 2nd video is from NPTEL. Their channel is full of great lecture series.
Their channel is here https://www.youtube.com/channel/UCAVi5Zg6zSoyZUyKBtCJfmg

Ok, this confirms what I wrote earlier: CoS is a useful concept when you want to analyse torsion within a structure.

Hulsey does not discuss torsion. So CoS does not appear to be a particularly useful reference point.
 
I have a few papers on it here written about real world buildings and mostly I am noting that - the CoS is not always the same for each floor, but it seems to be a good idea to try to get it close to the CoM.
I studied the links you provided, and yes, what you say there is correct: It is a good idea to design a structure such that CoS is close to CoM, because a lateral force acting on the structure in its entirety, such as seismic forces or wind, acts on the CoM, but the building responds by rotating about its CoS (in addition to a translational response), and that means that torsion gets larger with increasing distance between CoS and CoM (unless the force vector, originating at CoM, points directly at or directly away from CoS).

Again, CoS matters when you analyse torsion in your structure.
This is not what Hulsey does.

Do you agree then that the CoS is the correct place to express movement in WTC7 elements relative to, when accounting for the whole building ?
I disagree completely - CoS is an arbitrary and impractical origin of coordinates for the purposes at hand.
 
@gerrycan By how many inches would column 79 need to move relative to the initial center of stiffness of floor 13 as calculated by Hulsey for A2001 to fail?

(PS: The reason you cannot answer this question is because your premise makes no sense.)
 
Last edited:
I understand that Hulsey merely informs us what his origin is - that it is neither that wall not that but - an arbitrary point somewhere quite far away and only loosly connected to col 79. That's a valid coordinate system, but he doesn't even begin to attempt to explain why it is a good choice.

I'm not sure he's even using the center of stiffness. At first he say "thermal centroid", which I've not found a good definition of.
 
Sorry for butchering your post Oystein, but can you explain this point to Mick for me please.

The column moved relative to the outer walls in Hulseys model (the one on the slide at that point). He says it does not. So Hulsey misspoke, at best.
 
The column moved relative to the outer walls in Hulseys model (the one on the slide at that point). He says it does not. So Hulsey misspoke, at best.

No. It moved relative to the CoS.
ADD - What does Dr Hulsey say it moved relative to Mick ?
 
He does in this bridge studies. CoS vs. CoM seems important concept to bridge analysis, of which Hulsey is an expert.
What does Dr Hulsey say the c79 moves with respect to in his presentation Mick ?

You have stated it yourself for goodness sake. Come on..
 
It's in this thread. the ANSYS outputs being applied to the LS-DYNA model. Wasn't that actually a reply to you already ?
It's the Chapter 11 ANSYS model, and the Chapter 11 ANSYS model only, that determined whether or not connections failed due to heating. The Chapter 12 LS-DYNA model had absolutely precisely ZERO to do with that part of the analysis. The Chapter 12 LS-DYNA model starts after all connections that ANSYS found to fail failed.

So you need to look at the heating regime in the Chapter 11 ANSYS model only to assess how floor 13, or any of the other 16 floors, moved as a response to the fires, and how connections responded to those motions.

That heating regime of course was highly dynamic and detailed, not even, as Hulsey did it.
 
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 above is EXACTLY what you said Mick.
How many times have I asked you in the last page or so what column 79 was moving RELATIVE to in Dr Hulsey's model ?
 
The above is EXACTLY what you said Mick.
How many times have I asked you in the last page or so what column 79 was moving RELATIVE to in Dr Hulsey's model ?
? pretty sure you only asked that recently. previously you asked about NIST.
Relative to what exactly Mick. What fixed point did NIST have them moving relative to ?


take a breather Gerry, and respond more slowly. a lot of posts are coming in fast at you, you don't have to respond in 30 secs... you can take a few minutes/hours to gather your thoughts.
 
? pretty sure you only asked that recently. previously you asked about NIST.



take a breather Gerry, and respond more slowly. a lot of posts are coming in fast at you, you don't have to respond in 30 secs... you can take a few minutes/hours to gather your thoughts.
Well why don't you guys take a breather and allow Mick and I to return to my post #905 ?
 
What does Dr Hulsey say the c79 moves with respect to in his presentation Mick ?

You have stated it yourself for goodness sake. Come on..
He says it moves with respect to CoS - which is true.
He says it does not move with respect to the east wall - which is FALSE. Of course it moves with respect to the east wall.
He says it does not move with respect to the west wall - which is FALSE. Of course it moves with respect to the west wall.
He doesn't say it, but Col 79 also moves with respect to orthogonal coordinate systems centered on column 76, column 69, the ground beneath the building, the Staten Island Ferry and the ISS. And all these coordinate systems are completely equivalent to one another. There is no practical reason that you or Hulsey have explained yet why CoS ought to be prefered to analyse the situation around C79.

And of course it is outright FALSE to compare numbers derived in his CoS-centered system with numbers derived in NIST's C79-centered system.
 
Well why don't you guys take a breather and allow Mick and I to return to my post #905 ?
all members of the public can post. just ignore those you want to.

when you write something like "lets go back to post #905. please link to it. no one wants to wade through several pages to figure out where post #905 is. or just repaste the questions you have.

thanks.
 
The above is EXACTLY what you said Mick.
How many times have I asked you in the last page or so what column 79 was moving RELATIVE to in Dr Hulsey's model ?

I thought that was a rhetorical question. Obviously he's measuring it relative to an absolute point with zero motion, somewhere around where he's waving his mouse after he says "where's the zero at?".
20170918-105036-cmzs1.jpg

Again it's irrelevant when considering connection failures, as only the relative motion of the two parts being connected is important to the connection failure.
 
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.
Exerts more ... what? Force?
Capacity and stiffness are not the same, nor are they proportional. Apples and oranges.

And also, this is irrelevant given that the discussion currently is about coordinate systems, and what the best choice of origin is to discuss and analyse the problem of that C79 connection. Coordinate systems are just a mathematical convention to give meaning to tuples or triples of numbers.
 
Look I will get back to you ok. I want to get through this point with Mick, as stated above. It's important, as he said.
 
What invalidates the comparison is the absence of a given fixed point in NIST's analysis.

I disagree entirely. Since the relative motion is what causes failure, then NIST is entirely correct in saying what the relative motion of the girder to seat was, and not the relative motion of the girder or the seat relative to something else.
What invalidates the comparison is that they are different things. Hulsey made the comparison. NIST did not. The comparison is invalid. Hulsey made a mistake.
 
I disagree entirely. Since the relative motion is what causes failure, then NIST is entirely correct in saying what the relative motion of the girder to seat was, and not the relative motion of the girder or the seat relative to something else.
What invalidates the comparison is that they are different things. Hulsey made the comparison. NIST did not. The comparison is invalid. Hulsey made a mistake.
So how far does the c79 move? You said that the seat plate was the point of reference for NIST but it cannot be. The column moving will either add or subtract to the movement of the girder relative to it. This is actually quite an important point.
You cannot tell the movement of either, but remember NIST said how far the girder moved explicitly when they claimed 6.25" west. The didn't mention the column. So are you going with the column did not move at all ? Or did it move ?
 
You said that the seat plate was the point of reference for NIST but it cannot be. The column moving will either add or subtract to the movement of the girder relative to it. This is actually quite an important point.
the seat plate is connected to the column. the column movement (if any) is calculated IN the relative motion number.

The didn't mention the column.
then why do you keep asking for a number?
 
So how far does the c79 move? You said that the seat plate was the point of reference for NIST but it cannot be. The column moving will either add or subtract to the movement of the girder relative to it. This is actually quite an important point.
You cannot tell the movement of either, but remember NIST said how far the girder moved explicitly when they claimed 6.25" west. The didn't mention the column. So are you going with the column did not move at all ? Or did it move ?

Relative motion incorporates the motion of both A and B. See my earlier diagram.


Let's say A is the column, B is the girder. They both move. There's a resultant relative motion.

The relative motion is the important quantity.
 
Nope. These are 4 different things. Stiffness is one of them.
...
And yet, Hulsey says in the bit that you linked earlier today:
56:07 So it's [Col79] moving with respect to this over here, this area (points to CoS)
56:10 It's not moving with respect to this (points along the east wall)
56:15 It's not moving with respect to that (points along the west wall)
56:18 It's moving this way (an arrow points west from CoS)
56:19 or that way (an arrow points east from CoS)
56:20 with respect to what we call the thermal centroid.
Content from External Source
That's a different thing from the stiffness centroid, isn't it?

So is Hulsey wrong to express movement relative to the thermal centroid?

Or have you been wrong all this time to stress that stating relative to stiffness centroid is 100% the right way to go?

Or were you wrong when you just claimed that they are different things?
 
Relative motion incorporates the motion of both A and B. See my earlier diagram.


Let's say A is the column, B is the girder. They both move. There's a resultant relative motion.

The relative motion is the important quantity.

Yes we have established this Mick. And we have also established that you are well aware that Dr Hulsey is expressing movement relative to the CoS.
Is that not the correct way to do this ? Be clear now please.

ADD the net movement of the girder will change depending on the motion of the column. You know this.
And you know that is why i am asking you .
 
Status
Not open for further replies.
Back
Top