Could Girder A2001 Possibly Have got Past the Side Plate on Column 79?

I c
On the far left of the drawing, at the bottom of the column, they call out three different finished floor elevations, 11th floor, 13th, floor, and 15th floor. In the middle of the columns that call out 12th floor, 14th floor, and 16th floor. At the top of the column they call out 13th floor, 15th, floor, and 17th floor.

7th tier looks to be the 11th, 12th, and 13th floors. 8th tier looks to be comprised of the 13th, 14th, and 15th floors. The 9th tier looks to be comprised of the 15th, 16th, and 17th floors.

Trying to confirm if this is the case.
Can't remember what drawing number is the column schedule, but I know I have it somewhere. I have a feeling that they are where the relationship between floors and tiers is set out.
 
I c

Can't remember what drawing number is the column schedule, but I know I have it somewhere. I have a feeling that they are where the relationship between floors and tiers is set out.
Drawing 1092 references tiers 10 and 11 in the title block. On the left of the drawing, at the bottom of the column, they call out both floors 17 and 19. in the middle of the column they call out both floors 18 and 20. Towards the top of the column, they call out both 19 and 21 floors. Looks like tier 10 is referring to floors 17, 18, and 19. Tier 11 is floors 19, 20, and 21.
 
Last edited:
so.. a tier is where they splice girders together to make one really tall column? or rather a tier is individual steel pieces that make up a column?
upload_2017-9-22_11-1-56.png
https://teklastructures.support.tekla.com/en/cchelp_usa_column_tiers
I believe in the case of WTC7's drawings, a "tier" refers to the location (which three floors) a particular (approximately 25'-6" in length) a column section (or piece) is to be set and "spliced" to other pieces located one "tier" below and one "tier" above.

So a column 79 section/piece within tier 7 would be spliced to a column 79 section/piece from tier 6 below and tier 8 above.

All tiers, when spliced/connected together, equal the entire column itself. So column 79 consists of about 25 different "tier" pieces.

All these callouts for columns", "zones" and "tiers" seem to have been to make erection easier in giving location for each piece.

That's what I think at least.
 
I believe in the case of WTC7's drawings, a "tier" refers to the location (which three floors) a particular (approximately 25'-6" in length) a column section (or piece) is to be set and "spliced" to other pieces located one "tier" below and one "tier" above.

So a column 79 section/piece within tier 7 would be spliced to a column 79 section/piece from tier 6 below and tier 8 above.

All tiers, when spliced/connected together, equal the entire column itself. So column 79 consists of about 25 different "tier" pieces.

All these callouts for columns", "zones" and "tiers" seem to have been to make erection easier in giving location for each piece.

That's what I think at least.
yea this paper says upload_2017-9-22_11-21-41.png

so that jives with the 3 floors each, I think. http://ws680.nist.gov/publication/get_pdf.cfm?pub_id=910105
 
Drawing 1092 references tiers 10 and 11 in the title block. On the left of the drawing, at the bottom of the column, they call out both floors 17 and 19. in the middle of the column they call out both floors 18 and 20. Towards the top of the column, they call out both 19 and 21 floors. Looks like tier 10 11 is referring to floors 17, 18, and 19. Tier 10 is floors 19, 20, and 21.

Ok, similarly for 1091,
Tier 7 = (11),12,13
Tier 8 = 13,14,15
Tier 9 = 15,16,17

So,
20170922-082313-cskjy.jpg

20170922-082640-3iqpx.jpg

20170922-083757-5tve8.jpg

So the differences here are just that the higher floors have 1" side plates, not 2" (makes sense, higher sections carry less load).
And then seat is shown at right angles, not at the (7/8 in 12) angle of the girder. But the holes are still angled.

Why might this be? It's the same girder. The plate (pf) is also specified as 8 1/4 x 12 in both Bills of Materials.

So why the difference? And is it significant at all? it would make very minor differences to the walk-off distance, and would perhaps make a slightly more significant difference in the pull-off (sagging beams) scenario. I don't really see it being a huge deal, but accuracy is important when we are dealing with conditions that come close to failure. What if the floor 13 seat was installed like this?

And why wouldn't it be? The seat is just specified as 8 1/4 x 12.
 
Clearly the splices are not every floor and some are what I would regards as traditional yet others, as in the first picture, appear unusual.
Of course this unusual splice may be perfectly normal in the States?


 
Clearly the splices are not every floor and some are what I would regards as traditional yet others, as in the first picture, appear unusual.
Of course this unusual splice may be perfectly normal in the States?


The splice plates are not really the issue here, it's the side plates we are interested in, and if they always stop the movement of A2001 off the C79 seat

Your pics show lower floors, and exterior columns, WTC7 was not a "perfectly normal" building, in large part due to the Con Ed substation deirdre mentioned.
 
But with construction photos, I've been wondering if there's way we could "ground truth" verify how generally accurate the drawing are vs. the field connections.
 
But with construction photos, I've been wondering if there's way we could "ground truth" verify how generally accurate the drawing are vs. the field connections.

Is there anyone in the "Big Building" construction industry that can speak to this fact? There is probably some amount of "slop" when cutting and fitting girders and such together. If a girder is cut a hair too short, it would probably be within the spec to perform the same under normal circumstances. This would then be different than what the plans call for. The question is, if this is true, how many hairs short is ok. The was a post earlier (not sure who from) that said something to the point of "how do we know that the girder was the correct length" when everyone was discussing expansion and contraction. If we are talking inches of movement, and the girder in question is shorter than the plan says it is... It would need to move and deform less to fall off of the seat. No?
 
I'm not very familiar with the structural drawings, but here is how Colin Bailey describes the connections to column 79 (with reference to certain specific drawings) in his expert report:



upload_2017-9-22_18-34-27.png

And here is the table of sources he uses to describe how they used the various plans to make their model:

upload_2017-9-22_18-46-33.png
 
This is C79. The upturned T pieces are additional reinforcement so gives an idea of the floor.
c79 t pcs.jpg

Also re the tier issue, E12/13 states that the tier level is 7 for 12 and 13 as seen on the bottom right in the pdf.

[ off topic text removed. ]
 

Attachments

  • WTCI-000272-I.PDF
    7.3 MB · Views: 836
Last edited by a moderator:
Not that it matters... but 6WTC is behind in the image and it was 8 stories (upper 5+ are visible and it was 102 feet tall)... this "level" is the deck over the Con Ed substation. The column in the fore ground appears to be col 73 or perhaps E3.
 
Mick understood that he was only showing a simplified view of the net lateral displacement. The elongation and other movements within the system are really impossible to show accurately without actually having the full output from NIST's ANSYS model. This is because the girder and column would have been moving together in a complex way as the fires progressed around them over time, as I noted in more detail here. As an example of how the elongation/trapping scenario is too simple, girder A2001 was cooling at the point in time when NIST determined it became unseated. Arup also found that it could fail in the cooling phase (Arup found it could fall off the seat to the north as it cooled, in fact). If (1) girder A2001 expanded and pushed column 79 south in its heated phase and then (2) girder A2001 sagged in its heating phase and began contracting in its cooling phase, it is quite plausible that it would no longer be within the bounds of the side plates of column 79 as column 79 was pushed east by the heating phase of the girder to its west. This is why it is such an incredible shortcoming in Hulsey's work to date that Hulsey did not test NIST's actual heating model (or any actual fire progression model).

Interesting. Was there any evidence of buckling of the bottom flange. I would have thought that the flange bolts provide some degree of fixity and that temperatures were high enough to cause buckling due to dl + ll + thermal effects. What would be the limiting factor, bolt shear or bottom flange buckling? And as you say the beam is much more vulnerable in the cooling phase
 
It has to get stuck behind the side plate.

The beams are heated to 600 C and the girder they frame into is heated to about 500 C according to NIST. At 500 C girder A2001 (which is 531 inches long) will expanded by about 3.6 inches. That is 1.8 inches on each side. It was only about that distance from column 79's flange at room temperature, so it can't get past the side plate.


Tony makes the same mistake as Hulsey does by not taking into account that the fires move around and heat beams at different rates in different places.

He makes the unrealistic assumption that bang, they all equally go from room temp to his stated temp evenly and smoothly.

All one needs to think about is the fire codes and the requirements of beams vs girders. Floor "beams" have a shorter fire resistance requirement than do floor "girders". Therefore, beams would heat quicker.

That alone should tell you just how complex an issue it is. And that simplifying it in the manner that Tony and Hulsey have is pure rubbish
 
Is there anyone in the "Big Building" construction industry that can speak to this fact? There is probably some amount of "slop" when cutting and fitting girders and such together. If a girder is cut a hair too short, it would probably be within the spec to perform the same under normal circumstances. This would then be different than what the plans call for. The question is, if this is true, how many hairs short is ok. The was a post earlier (not sure who from) that said something to the point of "how do we know that the girder was the correct length" when everyone was discussing expansion and contraction. If we are talking inches of movement, and the girder in question is shorter than the plan says it is... It would need to move and deform less to fall off of the seat. No?

ASTM A6 gives allowances from purchased lengths. Typically, pieces can be slightly longer than ordered, but not shorter, I think. I'd have to read it again when I'm in the office.

Typical tolerances are something like +/- 1/4" for dimensions in projects where I dealt with that. Pieces showing up either longer or shorter than required is maddeningly common. Inches of difference is pretty uncommon, but for the amount of steel in a building that size, it's pretty likely something showed up either too long or too short. There's no way to know unless there are pictures/measurements of that exact connection somewhere, which is pretty unlikely unless there was something wrong with it. Minor elevation issues are fairly common too. Not every project, but enough that it's not a surprise when it happens. Stuff will be off the thickness of the clip vertically or something like that.

http://www.steelemotion.com/dropbox/astm.a6.pdf
 
Last edited by a moderator:
Were the column and girder in question found and identified in the wreckage, or is this based solely on simulations and indirect evidence from other members that were identified?
 
Were the column and girder in question found and identified in the wreckage, or is this based solely on simulations and indirect evidence from other members that were identified?
None of the WTC7 steel was examined, so this is all based on the construction drawings, a few "before" photos, and simulations/calculations.
 
Why might this be? It's the same girder. The plate (pf) is also specified as 8 1/4 x 12 in both Bills of Materials.

So why the difference? And is it significant at all?

Different draughtsman, different day? The critical dimensions are the bolt holes, plate thickness and width.
An orthogonal plate with angled bolt holes (1092) would be more usual than the odd shaped plate with holes parallel to the edge (1091).

Why both options have been detailed, seeing as both were clearly drawn by the same person, is however rather interesting.
FWIW dwg 1091 will have a slightly longer weld length on the plate.
 
The A2001 construction drawing would be important to accurately figure out the position of the end of the girder, which would be relative to the bolt holes. Like here there 6 6/16 from the West hole to the face (left is north, down is west)



We don't have A2001, but Tony tells me this is similar (Sheet 4279)
upload_2017-9-23_11-10-55.png

That shows 3 3/4 from bolt to end of girder, (and 11 7/8 from bolt to what?)

However this image (9114) Shows 4 1/4 at the actual floor 13 C79 A2001 connection




So that's half an inch difference, so presumably that girder draw is not actually the same as A2001?

So how far ARE the holes from the end of A2001? And if we don't have a drawing for A2001, then how can an accurate model be made of that connection?
 
Who says none of it was examined?
I see pictures of John Gross next to WTC 7 steel.
Did he just happen to be passing by it and get photographed ???
[...]

NIST says:
https://www.nist.gov/el/faqs-nist-wtc-7-investigation

27. Why didn't the investigators look at actual steel samples from WTC 7?

Steel samples were removed from the site before the NIST investigation began. In the immediate aftermath of Sept. 11, 2001, debris was removed rapidly from the site to aid in recovery efforts and to facilitate emergency responders' efforts to work around the site. Once it was removed from the scene, the steel from WTC 7 could not be clearly identified. Unlike the pieces of steel from WTC 1 and WTC 2, which were painted red and contained distinguishing markings, WTC 7 steel did not contain such identifying characteristics.
Content from External Source
But the relevant point here that nobody found and identified the specific pieces of steel we are talking about here.
 
The A2001 construction drawing would be important to accurately figure out the position of the end of the girder, which would be relative to the bolt holes. Like here there 6 6/16 from the West hole to the face (left is north, down is west)



We don't have A2001, but Tony tells me this is similar (Sheet 4279)
upload_2017-9-23_11-10-55.png

That shows 3 3/4 from bolt to end of girder, (and 11 7/8 from bolt to what?)

However this image (9114) Shows 4 1/4 at the actual floor 13 C79 A2001 connection




So that's half an inch difference, so presumably that girder draw is not actually the same as A2001?

So how far ARE the holes from the end of A2001? And if we don't have a drawing for A2001, then how can an accurate model be made of that connection?
Mick, I think you have to go with what drawing 9114 tells you about the bolt hole location for A2001.

I did not say or mean that everything about the similar girders between columns 44 and 79, that we do have drawings for high in the building, is exactly the same as A2001.

They are all W33 x 130 structural shapes and have the partial height stiffeners (that the NIST WTC 7 report omitted on girder A2001 but are observed on drawing 9114 and in photos).

However, the girders in that position (between columns 44 and 79) higher in the building are a little longer, as these columns there are smaller cross sections due to lower vertical loads. It is not hard to understand why A2001 might have a slightly greater dimension from the end of the girder to the bolt holes. The sections of column 79 at the top of the building also did not have protruding side plates.
 
Last edited:
Mick, I think you have to go with what drawing 9114 tells you about the bolt hole location for A2001.

I did not say or mean that everything, between the other similar girders between columns 44 and 79, that we do have drawings for high in the building, is exact. They are all W33 x 130 structural shapes and have the partial height stiffeners NIST omitted on girder A2001 that is seen on 9114. However, those higher in the building are a little longer, as the columns are smaller there due to lower vertical loads. It is not hard to understand they might have slightly different dimensions from the end of the girder to the bolt holes. The sections of column 79 at the top of the building did not have side plates.

Agreed. The seat connection drawing for the 44th floor for A4279 to C79 is on 9179
20170923-115034-xaens.jpg

So that matches that girder 4279.

Still, the fact remains we don't have a drawing of A2001.
 
The fact that the drawing 9179 and 4279 girder bolt hole locations at the column 79 side match lends credence to the location that 9114 gives for those holes in A2001. I also think the protruding side plates on column 79 lower in the building is the reason for the slightly greater distance on A2001.

I don't think there is much of an argument there. However, asking why we don't have a fabrication drawing for A2001 is a good question.
 
This is C79. The upturned T pieces are additional reinforcement so gives an idea of the floor.
c79 t pcs.jpg

Also re the tier issue, E12/13 states that the tier level is 7 for 12 and 13 as seen on the bottom right in the pdf.

[ off topic text removed. ]

Here's a closeup
20170923-122446-38a4v.jpg

And a shot from another angle.
Mitchell-7-29-85-p3b copy 2.jpg


20170923-122737-2i1al.jpg
Of course this is just a splice connection on a much lower floor, however it gives a bit of a visual on the side plates. Looks like they are 3" thick here, but that makes no difference to the inside dimensions.

Notice the numbers on the side. 79, 84-15, and A107:
20170923-123102-xepjs.jpg

20170923-123154-txioo.jpg

One minor difference there is that the rectangular splice plates have huge holes in them which are not in the drawing. Presumable these were for the crane connection. Seems odd they are not in the drawing though.

I'm mentioning it just as an example of things not always being the same as in the drawings. I doubt it would make any structural difference.
 

Attachments

  • 0107.TIF.zip
    309.6 KB · Views: 579
came across this pic.. is this a side plate? does anyone know, if it is a side plate, how much force it would take to bend it like that?
side plate.JPG
 
Unfortunately @gerrycan has been banned for repeated and egregious violations of the politeness policy , as well as an accumulation of other posting guidelines violations.

Gerry was banned three years ago for similar reasons, and I lifted his ban to let him join the discussion on Hulsey's new study. So his position here was somewhat probationary. It has not worked out.

It's unfortunate because I think Gerry does mean well, however I have to enforce the rules.
 
It looks like a side plate. It would take a lot tip do that. Just breaking the weld(s) would be a lot, let alone buckling it and pulling it away from the flange. There's nothing there for scale, so it's hard to guess it's thickness.
 
It seems impossible that A2001 could have got past the 2" side plate on C79 by actually pushing it out of the way. I think everyone agrees that it would have to actually miss it, or get distorted itself, for the connection to fail.
 
It seems impossible that A2001 could have got past the 2" side plate on C79 by actually pushing it out of the way. I think everyone agrees that it would have to actually miss it, or get distorted itself, for the connection to fail.
that's what I thought. was just surprised when I saw that pic, maybe there was something attached to the side plate that pulled it in a twisty way as the building collapsed.
 
There are a couple of questions I have.

First the girder comes in at an angle to Col 79. This would put a large asymmetric bending on the bottom flange, which would force any buckling to start on the edge of the flange that Tony says would hold the beams in place

Interesting this tends to twist the column away from the beam supporting it. So general column roTation in fire could be an issue.

Secondly could tolerances be important.. Yes there is only a theoretical gap of what looks like half an inch. This could get used up in construction and the steel fixers would be happy to gouge hole to get beam to fit. The bolts are only erection bolts and this would be normal construction practice.

In addition the beam would try and rotate when concrete was added So it would be reasonable to assume a worst case that the beam was in contact at initiating stage. I don't think Arup or Wedlinger looked at this but unlike Hulsey they never claimed to have considered all possibilities And if the beam expands by 1.8in relative to the column restrained then it would cause immediate damage to the girder
 
If it took 180 kips to shear the bolts, could that be enough to damage the side plate and/or girder flanges and push past? I thing perhaps, but it depends on some things we don't know, or at least haven't been discussing here.
 
Back
Top