1. Mick West

    Mick West Administrator Staff Member

    Fire and debris damage was the cause of the collapse. C79 buckling was the initiating event. So you'd need the same fire damage and debris damage - both of which are inexactly known.

    It would be an interesting experiment though. NIST did it and the building collapsed in their model.
     
  2. gerrycan

    gerrycan Banned Banned

    So we are "stuck" if you like, with column 79, because of the demise of the penthouse and the lack of substantial external deformation to the building at that point.
    So on the basis of real world observation, we reasonably can assume that the events at column 79 (whatever they were) were the cause and likely initiation point for the first signs of real destruction that we see in the building real world, externally ?
    We still agree?
     
  3. Mick West

    Mick West Administrator Staff Member

    Well the first sign was the penthouse, so it was probably something under that.

    There's always going to be unknowns. Part of the problem with making definitive statements.
     
  4. Mick West

    Mick West Administrator Staff Member

  5. Mick West

    Mick West Administrator Staff Member

    Please avoid insulting each other. I will time-out people from the thread if they continue to snipe.
     
  6. gerrycan

    gerrycan Banned Banned

    Do you realise how far the penthouse actually is from C79 ? It's a fair bit, I will try to dig out the drawing of it.

    Getting back to the model, would a unidirectional element be appropriate to model walk off in the girder, do you think ?

    Or would you object to that ?
     
  7. Mick West

    Mick West Administrator Staff Member

    Can you explain how your line of thought is related to Hulsey's study?
     
  8. NoParty

    NoParty Senior Member

    As I read, I'm learning that every time two posters contort to defend this new semi-transparent
    (now "interim"?) powerpoint, the game is to yell "...but NIST..." to distract from the topic at hand.

    For what it's worth gentlemen, I can't be the only observer who now interprets every one of your
    invocations of "NIST" (on a thread that's supposed to be about the Hulsey project) as little more than a
    a thinly veiled attempt to distract from the obvious shortcomings of what Hulsey has delivered,
    compared to what was promised.
     
    Last edited: Sep 14, 2017
    • Agree Agree x 5
  9. gerrycan

    gerrycan Banned Banned

  10. Mick West

    Mick West Administrator Staff Member

    It depends on prior analysis. Could you make your actual point please. Then we can just work backwards if I don't understand it.
     
  11. gerrycan

    gerrycan Banned Banned

    NIST decided the net direction of travel in the girder would be West, before they even started, by allowing for movement in only one direction.
     
  12. Mick West

    Mick West Administrator Staff Member

    Can you give a reference for that? Sorry, didn't see the pic.

    What control element did Hulsey use for that node? That should be a pretty easy bit of data to share, even if we don't have the software to use it.
     
  13. gerrycan

    gerrycan Banned Banned

    I really don't know. You should ask that question.
    Would you recommend something unidirectional? Doesn't sound right to me.
     
  14. Mick West

    Mick West Administrator Staff Member

    Much as I like to delve into technical detail, you are trying to critique elements of the NIST model. But this thread is about the Hulsey model. It's about NIST in the sense that he claims to have done it better. But since we don't know anything at all about the the type of elements Hulsey used, then it's a bit pointless to compare them to NIST.
     
  15. gerrycan

    gerrycan Banned Banned

    We surely have to check that the same shortcomings observed in NIST's analysis have not been repeated in the UAF one. We know that the elements are correct as per UAF because we looked at NISt and realised just how much had been omitted.
    The real point here is that we don't know what values were applied to this COMBIN37 in NIST's model.
    No inputs known = fog = NIST.
    UAF is stating that they will be open with their inputs and model. I have no reason to think that they will not be.
     
  16. Mick West

    Mick West Administrator Staff Member

    Well they are not right now, hence the claim that their study shows that fire didn't do it is debunked.
     
  17. gerrycan

    gerrycan Banned Banned

    The model evidence that you base that on has how much input data released ? ZERO
    NIST made the extraordinary claim that they had proven that fire COULD lead to a progressive collapse.
    There's no evidence to support that claim. No input data.

    Yet, when faced with 2 separate models from UAF, which you know to be more structurally representative of the building you choose to put your faith in the NISt model, which is less accurate, and you choose to support their extraordinary claim that fire could do this, on nothing more than faith. Because you have no evidence for it, but rather lots to the contrary.
    Telling. Especially at the interim stage of a project.
     
    • Disagree Disagree x 2
  18. Mick West

    Mick West Administrator Staff Member

    What? How do I know this? All I've seen so far from Hulsey is this odd claim that both of his models make everything around C79 move east together with zero relative motion. Can you explain how this indicates things are more accurate?
     
  19. gerrycan

    gerrycan Banned Banned

    The inclusion by UAF of the elements that you know NIST left out.
    I do agree that there needs to be more meat on the bones from UAF, but this presentation was an update, and we can already see that the connection in question has been modelled in a way that truly represents the connection as per the available drawings.
    Surely you can't be saying that you believe NIST did a better job with the elements at the C79 connection ?
     
  20. Mick West

    Mick West Administrator Staff Member

    Gerry, can you explain what you think this diagram represents in terms of what is happening to

    A) The floor slab
    B) C79
    C) A2001

    [​IMG]

    I'm quite open to the idea that I'm missing something. Tony said I was talking gibberish, so perhaps I just don't understand what the 1.85 is here, and how it's comparable to the 5.5"/6.2" figure from NIST.
     
  21. gerrycan

    gerrycan Banned Banned

    Not a lot of detail given about what is moving with respect to what there, but I think this is a separate simulation of the whole building reaction to stated temperatures, and that this movement would wrt the centre of stiffness.
    Floor slab expansion and the modelling of the exterior to a non fixed, or pinned state would be what I would suspect may account for the difference between this and NIST.
    Also open to the idea that I am missing something though.
     
    Last edited: Sep 13, 2017
  22. Mick West

    Mick West Administrator Staff Member

    Can you give more details on that? A link?
     
  23. Oystein

    Oystein Active Member

    Why do you think that "unidirectional" is a problem? What difference does it make compared to a similar element that would be multidirectional?

    The way I understand it is that this element monitored whether or not the girder end satisfied the "walk-off" criterion, which was defined as "x inches relative to the seat towards relative west". This opens two possible outcomes at runtime:
    • Either the girder does not travel x inches west relative to the seat; then girder stays in place; no failure
    • Or it does travel x inches west relative to the seat; then girder loses vertical support; failure.
    As it turned out, the criterion was satisfied, and failure was observed. Now you: Would the outcome have been different if the element had been modeled monitoring more than one direction?

    The worst that could have happened, if this "unidirectional" is interpreted as an actual direction, namely "west", is that the girder walks off towards north (e.g. girder sags) or east (e.g. girder is pulled upon cooling of beams), in which case ANSYS would have missed the walk-off despite it actually occurring. Since this case did not, in fact, occur, no harm was done. Right?

    You haven't even begun to explain why this unidirectional element is a "shortcoming".

    No, we cannot possibly know at all that UAF has even one single element "correct" for two reasons:
    1. UAF has not actually shown us the model (broken promise that data would be fully available)
    2. There is no absolute value for "correct" in modelling such things. This reason breaks down into two again:
    2a. We cannot possibly know "correctly" the actual situation of any connection as per 2001/09/11 8:45 a.m.
    2b. No FEA model can and ever will be a perfect representation of reality.

    You are dealing in false dychotomies

    But you are already convinced that UAF has it all correct. Congratulations to your prescience!

    Since Hulsey and AE911Truth already announced in public headlines the final top level conclusion, not having revealed ALREADY the data is a breach of promises, and THAT is a good reason not to trust them blindly. Promise breakers WILL break promises.

    You cannot with a straight face claim that you have "No inputs" from NIST when in fact they devote three lengthy paragraphs of detailed explanations to their ANSYS modelling of this connection. What you ought to have written, since this thread is about Hulsey:

    No inputs known = fog = promises broken = Hulsey/AE911Truth.

    We haven't seen their analytical model AT ALL, or have we?
     
  24. Mick West

    Mick West Administrator Staff Member

    20170914-083240-at1x9.
    This graph (Hulsey, page 71) shows the Load Capacity of Column 79. That's basically how much weight (in kips, where 1 kip = 1,000 pounds) a given unbraced length of column can hold up before it fails. It's actually lots of different graphs for different temperatures. The top one 68°F being room temperature. Higher temperature and longer lengths of column have reduced load capacities.

    Hulsey has marked off two lengths, 12'9" and 25'6". These represent the column when it either spans a single floor or two floors (for example if one floor had collapsed, or had been removed).

    He then has added a short blue line on the 212°F curve marked "Nist Col 79 temp (392°F) All floors". Presumably this was intended to be on the lower 392°F curve, not the 212°F curve, but it makes little practical difference. This line is showing the difference in carrying capacity between a single floor and a double floor section of unbraced column. You can see it's not that different to the room temperature (68°F) curve, which shows (assuming the graph is correct) that heating of the column to 392°F is not going to make it fail.

    Two slides earlier he shows this:
    20170914-084244-oa233.

    So this is showing column 79's Dead Load (the static load, i.e. the weight of the building held up by that column) of 5142 kips.

    We can add this value to Hulsey's graph, as a thick blue horizontal line.
    20170914-090205-pgdt8.
    Now we can see where it intersects a floor marker (the red vertical lines) to see what temperature it would fail at for a given number of floors (1 floor being the normal case, nothing removed). So we can see that for 392°F (or 212°F) the column exceeds its load bearing capacity at 5 floors in length, which would be the same as the collapse of 4 floors.

    In the NIST model, 9 floors collapse away from C79, as seen in Hulsey's slide 18. This is an effective length of 10 floors, or 127.5 feet, which is off the chart, reducing the load capacity to less than half the dead load, hence causing collapse.
    20170914-091004-8yddw.

    However on Hulsey's slide 82:
    20170914-091112-6lq8e.

    He says "UAF:Based on NIST Column Temperatures; col 79 did not buckle under gravity loading."

    This is presented as a comparison with NIST's results. We have already seen how the other comparison in red text is meaningless (5.5" is a local differential movement, 2" is a global structural expansion). But is this a valid point?

    The problem here is that the temperature is irrelevant. NIST has never claimed that the column buckled due to temperature, and specifically said that the heating of the column was not a factor.
    (NCSTAR 1A, Page 53, Pdf page 95)
    What they said, and Hulsey's graph confirms, is that removal of lateral bracing causes the column's load bearing capacity to be reduced to less than the dead load, and so it would fail. Here Hulsey and NIST seem to be in rough agreement. Since they agree on this point it cannot be used as a comparative point.

    The point should read something like "UAF: Col 79 did not lose lateral support, and hence did not fail" which is the actual substantive difference between Hulsey's study and the NIST report.

    Hence the comparison is invalid and misleading.

    References:
     
    • Informative Informative x 1
  25. benthamitemetric

    benthamitemetric Active Member

    Yes, I'll try to work up a fulsome post with both a concise summary of these issues and full documentation in reference to Hulsey's slides, the Arup Report, Norden's expert report, and the WAI report. If I can't get to it today/tomorrow, I'll post it over the weekend.
     
  26. gerrycan

    gerrycan Banned Banned

    Well, we have my best guess and your presumption that he is expressing this movement at C79 wrt the stiff point. but ok.

    I wouldn't say irrelevant.
    If for example though, the girder fails. That does not mean that the column is entirely withough support. You could argue that C79 would still retain orthogonal support structurally having suffered the loss of the C79-44 girder.

    Again, you would need to state whether the column was totally unbraced or not. I would argue that given the loss of the girder it remains supported. But you still need an initiation. But more importantly for this issue, and to your question to Bentham above, you need progression.

    I think you should be clear about what exactly you are saying here.
    You are over reaching with your values. And you still do not have collapse progression, let alone initiation.
    There are issues with the retention of support in C79 also.
    And again, you seem to be just disagreeing with the model outputs and findings that Dr Hulsey is stating. Do you see any potential problems in the modelling ?
    NIST did everything they possibly could to "encourage" the girder at C79 to fail in their model. They left off key elements and ignored others, while exaggerating walk distances for the girder.
    Dr Hulsey on the other hand has modelled the whole building with the correct elements applied, unlike NIST.

    Why would you NOT expect their results to be different.

    BTW Bentham was trying to talk about the stiffness calculation (K=1), and progression. Be interesting to see his reply to you though.

    ETA - I don't have the data to hand, but I will later.
    If you know off hand, in which model did NIST substitute C80 for 79 ? Think it was 80. Just a thought.
     
    Last edited: Sep 14, 2017
  27. benthamitemetric

    benthamitemetric Active Member

    What are "fire simulation results"? The fire simulation is a time-dependent function. The "results", by which I think you mean certain temperature values produced by the fire simulation at some arbitrary set time, are not an accurate reflection of the actual full simulation. That Hulsey failed to actually run a full fire simulation is a failing of his report that introduces error into his result; that he specifically failed to actually run NIST's full fire simulation is a failing of his report that prevents him from concluding other independent variables in his study (e.g., added structural elements) were the reason for the difference between the outcome of his model and NIST's. Hulsey failed to control for the single most important independent variable shared by his model and NIST's.
     
    • Agree Agree x 1
  28. Mick West

    Mick West Administrator Staff Member

    What's not clear? I think my post explains it quite well. Hulsey says under "Comparative Studies": "UAF:Based on NIST Column Temperatures; col 79 did not buckle under gravity loading.", since this is part of a table of comparisons between UAF and NIST it hardly seems likely that he means: "we agree with NIST".

    He's making a comparison, but the comparison is not valid. At best it's highly misleading because it misses the WHY part.
     
    • Agree Agree x 1
  29. gerrycan

    gerrycan Banned Banned

    Not going to go round the houses like the peer review issue again here, but....
    Dr Hulsey has used the same criteria as NIST with fire, so we now know what the correct elements and the whole building being modelled with proper spring elements makes.
    Also, you should look at how these dynamic values are applied to an end output simulation ie from a dynamic fire simulator into ANSYS.
     
    • Disagree Disagree x 1
  30. benthamitemetric

    benthamitemetric Active Member

    The bolded is completely false with respect to the fire simulation.

    NIST ran a full, time dependent simulation of fires across 16 floors of the building, which simulations included a detailed progression of the fires across all 16 floors over time, and then, from that simulation, NIST modeled fire damage in a time dependent simulation across the east side of those 16 floors.

    Hulsey heated up certain elements on smaller portions of the east side of 2 floors to maximum temperatures that he pulled from some point in time in NIST's simulations. It is not clear what sort of heating curve he used to apply those temperatures, but it is completely clear that he did not actually model the progression of the fires, which caused some elements to heat up/cool down/otherwise become damaged/etc. before other elements.

    You never answered from many pages back whether you think it is accurate to model all elements of the building except for elements contained in small portions of two floors floors of the building as fixed, by the way.
     
  31. gerrycan

    gerrycan Banned Banned

    Listen to what Dr Hulsey says re the column temperature here.

    Source: https://youtu.be/3iVm94V0lLE?t=20m7s

    A little further on, when talking about potential column failure due to temperature, "nobody seems to think that".
    That's fairly clear.
     
  32. gerrycan

    gerrycan Banned Banned

    With all respect Bentham. You need to go and take another look at what NIST did in their models before we can discuss this properly. Especially as far as modelling elements as fixed or pinned is concerned.
     
    • Disagree Disagree x 2
  33. Mick West

    Mick West Administrator Staff Member

    Column 79 is directly underneath the penthouse. A bit north of center. So what does "a fair bit" refer to?
    20170914-111615-vjxiq.
     
  34. Oystein

    Oystein Active Member

    This is a terrible move of goal posts and thus totally fails to address the actual argument benthamitemetric provides you with.

    But no, actually YOU need to revisit what NIST did as far as modelling elements as fixed or pinned is concerned. You have several times expressed the idea that NIST modeled the walls as fixed and rigid. I showed you already, twice, that this is FALSE. You did reply to one instance of my showing you this, but your reply had nothing at all to do with what I wrote.

    Please revisit Post #490: https://www.metabunk.org/ae911-trut...r-modelling-project.t5627/page-13#post-210990

    This posts also has another case of Hulsey completely getting wrong what NIST modelled. You seem to have missed this, too.
     
    • Agree Agree x 1
  35. deirdre

    deirdre Moderator Staff Member

    is it because Quan was testing something different for his phd work? I initially thought it odd they didn't model the fire progression (or full fire(s) ), because I thought Quan specifically was wanting to model fire progression and it's effects on buildings.. but relistening to where Husley discusses this.. am I misunderstanding? Is "fire test" something different?

    re: zhili quan phd student 2:40 Husley says "doing his phd work on fires, on fire modeling in buildings. and fire modeling not only on buildings but on fire tests, on how you
    actually conduct a fire test and how reliable those are, so we're looking at alot of that kind of stuff. "

    Source: https://www.youtube.com/watch?v=EKN4qilUOfs
     
  36. Mick West

    Mick West Administrator Staff Member

    So it makes it doubly odd that he included it as a comparative point on the later slide.

    It seems that from his reading of the slide in this version of the presentation he intended it as a single point, not two:

    Sept 2017, @1:03:54


    However it's different from the previous version of his talk 11 months ago, where he lists it as a point of AGREEMENT, and not a consequence of the previous comparision:
    20170914-113209-0key9.

    Seems like a bit of a flip-flop?
     
  37. gerrycan

    gerrycan Banned Banned

    My point was how far back from the front penthouse wall C79 is. I think about 35-40 ft.
    The penthouse falls NE side first, then NW. The kink in the EPH lines up more or less with C44, but the corner goes all the way to C47.
    My point with the distance from C79 was that despite being relatively far from it, the thin walls of the penthouse still fall right in and back at C47. For a failure initiated at floor 13 to progress to that not only to that height, but to that extent, so quickly is unimaginable.
    There's a good view of the EPH failure here
     
  38. gerrycan

    gerrycan Banned Banned

    No, what we are looking at are notes used to guide a presentation, sparse and to be elaborated on. The tone and inference is quite clear in the presentation. Perhaps a more purpose specific PDF will accompany the final report from this project.
    In the lecture Dr Hulsey is very clear that there's no difference with NIST here, although he does earlier allude to column temperature maybe being a potential suspect issue for him earlier on in the project.
     
  39. gerrycan

    gerrycan Banned Banned

    Which model are you talking about here ? And can you be specific about which floors if possible.
    It looks to me as if NIST have pinned most of the building.
     
  40. Mick West

    Mick West Administrator Staff Member

    What do you mean so quickly? What two events are you measuring?
    How do you know that? It could have sunk in the middle, under C79 and C80.
     
    • Agree Agree x 1