WTC7 and the girder walk-off between column 79 and 44

[LSSBB]

Guys.........great discussion!

It's nice to watch people who are knowledgable in a field discuss details and have a back and forth discussion on the specifics.

Too bad the truthers can't contribute much..........but no matter..........for those of us in other Engineering fields.....it's nice to watch.

Keep up the great work!

+1 (from a Mechanical Engineer in a different field than Structural Engineering).

 

[femr2]

Can't wait to hear the excuses you're going to use to not say anything...

I take it that you're fine with my previous response ?

Perhaps you could also explain why you felt it necessary to make the statement above.

 

[ozeco41]

Who did you have in mind.

There has only been one member who erected a false claim on an unproven/unsupported assumption. Plus his tag team partner who merely echoed the same errors.

In summary this is the range of discussions in the thread:

• Gamolon posted the OP requesting help - he seems to have been satisfied with the help he got;
• The discussion morphed into a claim by Tony Sz that the NIST explanation was impossible and discussion of that claim;
• Those discussions took two main branches with a third area in overlap;
• The "details" branch where several members led IMO by tfk took Tony on at the detail level he presented his argument. Prima facie the status of that debate reached the stage where Tony had been shown to be wrong at the level of details.
• I then posted at the level of overall argument and identified three problems with Tony's claim, each of which was fatal to his claim of "impossiblity of the NIST explantion. My three objections were - in order of importance not chronology:
  1. He was not considering all the factors;
  2. He was making an assumption that the end conditions for the subject girder were not changed by the effects of heat through the building; AND
  3. Because he had not proven or supported his assumption he had not met his 'burden of proof' needed to 'make out' the claim.
  ...therefore his claim failed. Prima facie the status of "my" branch of discussion is "Tony shown to be wrong at three levels and he has failed to respond to those claims of error."
• Meanwhile other members addressing the "detail" level of argument also reinforce my Point 1 - that Tony was not considering all the factors. Some of those members may have beaten me to identify the error - I cannot remember. So that aspect - "not considering all factors - was common ground between the two approaches to discussion with Tony;
• After failing to adequately address the identified errors at any of the levels Tony and C7 "bailed out" of discussion;
• This recent series of discussions has progressed the reality that Tony is:
  A) wrong for not considering all factors; AND
  B) Has identified - and I suggest "proved" - some factors that Tony did not consider.
• From my perspective the status of debate is that Tony's claim "the NIST explanation is impossible" is falsified - several times over and at each of the three levels of debate. So that is "dead in the water" by my pedantic semi-legal standards.
• HOWEVER we have not proved "NIST was right". My own view "NIST explanation is plausible and the best we have". Others with better expertise than mine at the technical detail level may be more assured that NIST was right. I don't need that level of assurance.

The interplay between femr2 and tfk, once you filter out the personal aspects, doesn't IMNSHO change the overall status of the main discussion in the thread.

And the OP, Gamolon, still appears satisfied with the answers to the OP.

Guys.........great discussion!

It's nice to watch people who are knowledgable in a field discuss details and have a back and forth discussion on the specifics.

Too bad the truthers can't contribute much..........but no matter..........for those of us in other Engineering fields.....it's nice to watch.

Keep up the great work!

Definitely agreed. Top work folks.

 

[tfk]

Detail at Dwg E118A bottom right 2nd from bottom shows (13th floor floor)
"Core column grid 79 at ....13th... flrs" shows the col. flange side plates pa/paa welded at the splice and pd welded at top col.
Also I believe pb and pc were temporary lifting plates removed after installation.
The columns were two stories tall. Their splices at 11th , 13th, 15th floor floors.

OK, that would be something completely different.

Thanks for the info.

I'll add in those plates, unless anyone knows anyplace else that someone else has already provided these assembly drawings.

Anyone know of any other pertinent drawings (other than E118A & E118)?

 

[Tony Szamboti]

OK, that would be something completely different.

Thanks for the info.

I'll add in those plates, unless anyone knows anyplace else that someone else has already provided these assembly drawings.

Anyone know of any other pertinent drawings (other than E118A & E118)?

Don't forget the welds at the column section interface, which are in addition to the welded on plates.

 

[BasqueArch]

.............

Those are pretty good drawings, where did they come from? They would have helped me when I initially built my model. Even though these drawings show the stiffener plates, did the NIST include the stiffener plates in their analysis? And if I do move my stiffener plates over the bearing plate, would you expect a change in the results?

The drawings were done by my draftsman in ACAD from the shop drawings using TS's displacement numbers.
NIST did not include the web stiffener plates in their analysis. They included a stiffener plate under the seat which wasn't there.
The girder would still slide off the seat because after the girder sagged, its load would be placed at the tip of the seat. This torque would have bent or sheared the seat plate and the caused the girder to slide off.

Detail B below shows the girder displacement as calculated by TS without the column 76 girder push. The girder load is at the tip of the cantilevered seat.

Detail B1 below shows the girder displacement with the 1” column 76 girder push. The girder load is at the tip of the 4 7/16” cantilevered seat.
“The temperature of the girder between Columns 76 and 79 on Floor 13 was sufficient to displace Column 76 to the west and Column 79 to the east.” (NCSTAR 1-9 p527).

Detail C shows the sagging girder with a 20” deflection, similar to the beams deflection modeled by NIST. Any girder deflection would place the girder load at the tip of the cantilevered seat.

Detail D shows the condition at 13th floor Column 79 seat. The 83.7k load was calculated at 100 psf for 837 sf of area at the tip of the 4 7/16” cantilevered seat. It looks like torque would fail the seat in bending or shear and the girder would slide off, but I can’t do the math.

NIST didn’t consider the vertical failure of the seat and its model included a seat stiffener at Col 79 not shown in the plans.
“Since vertical failure of the seat was not considered (Section 11.2.5), the connections at Columns 79 and 81 were both modeled as stiffened seats.” (NCSTAR 1-9 p.558) See “Figure 12-25 Seat connection in global model Column 79. (p. 559)

Column 79 under the EPH failed first due to its unbraced length ~ 8 stories. Loss of column buckling strength is exponentially proportionate to its unbraced length. At low temperatures the beams shear bolts failed and the concrete cracked due to this failure, restrained expansion and beam sagging. At east-west Col 79 axis at low temperatures the west and south girder knife connections (they were not the seat type) failed due to bolts shearing or the angle welds failing. The east had no girder connection and the concrete had failed to provide stiffening to this column on different floors. At east-west axis Column 79 had lost bracing over multiple floors, making it unstable regardless of whether the Cols 44-79 girder at floor 13 had failed last.

The collapse of WTC7 was due to the failure of the sufficiently unbraced Col 79 caused by fire damage.

 

[Tony Szamboti]

The drawings were done by my draftsman in ACAD from the shop drawings using TS's displacement numbers.
NIST did not include the web stiffener plates in their analysis. They included a stiffener plate under the seat which wasn't there.
The girder would still slide off the seat because after the girder sagged, its load would be placed at the tip of the seat. This torque would have bent or sheared the seat plate and the caused the girder to slide off.

It is hard to understand how the girder would deflect downward and sag by 8 degrees as you show in Detail D of your drawing. This would require the downward deflection to be about 37". Do you have calculations to show this would occur? Is this a result of applying the 83,700 lb. load to the end of the seat?

NIST claims the girder between columns 44 and 79 was at 500 degrees C. At that temperature, the AISC says the modulus of elasticity retention factor is 0.63. In my calculations I used the reduced modulus of elasticity and girder loading that would produce reaction loads on each side that are 1.3 times the 83,700 lbs reaction load that you show, and they only show about 1.5" of downward deflection and 0.31 degrees of sag.

The girder would be expanding by about 2" per side at 500 degrees C and the horizontal deflection and sag combined wouldn't cause more than 1" of total shortening of the girder, or about 0.5" per side. Thus, the web and stiffeners would remain well over the 14" deep vertical plate beneath the seat at column 79, making the failure mode you are proposing seem impossible.

 

[BasqueArch]

It is hard to understand how the girder would deflect downward and sag by 8 degrees as you show in Detail D of your drawing. This would require the downward deflection to be about 37". Do you have calculations to show this would occur? Is this a result of applying the 83,700 lb. load to the end of the seat?

NIST claims the girder between columns 44 and 79 was at 500 degrees C. At that temperature, the AISC says the modulus of elasticity retention factor is 0.63. In my calculations I used the reduced modulus of elasticity and girder loading that would produce reaction loads on each side that are 1.3 times the 83,700 lbs reaction load that you show, and they only show about 1.5" of downward deflection and 0.31 degrees of sag.

The girder would be expanding by about 2" per side at 500 degrees C and the horizontal deflection and sag combined wouldn't cause more than 1" of total shortening of the girder, or about 0.5" per side. Thus, the web and stiffeners would remain well over the 14" deep vertical plate beneath the seat at column 79, making the failure mode you are proposing seem impossible.

I used your previously calculated numbers which you now have heavily revised.
NIST calculated a 20" sag for the heated beams. I assumed the same heated stress/strain and same 20" deflection would apply to the girder.
I used ACAD to generate a circle where the circle intersected the two end points at the length of the girder and a tangent point at -20" at the center. ACAD measured this angle at 8 deg. at the seat tip.

Your calculation chart for deflection is wrong because it assumes a uniform load along the girder and that the ends were fixed (bolted).
You should use one with beam point loads along the girder and not fixed ends (hold down bolts were sheared). The difference would be substantial. Any deflection of the girder away from level would place the load at the tip of the seat and bend or shear it. Will you provide this revised deflection chart.

The 2" expansion at each end would cause further westward lateral buckling since the gap to the column was about and inch, placing the girder farther away from the seat.

Re calculations: I'm not an ME or SE. I asked a busy SE consultant/friend of mine to do some calculations and he said he would (or rather he tepidly said "ok"), about a week ago. After explaining to him the CD vs.fire claims for WTC1,2,7 he thought the CD claim "is stupid." I've bugged him a couple of times for the calcs. It don't look like he's going to spend his time on this.

 

[Animal]

I used your previously calculated numbers which you now have heavily revised.
NIST calculated a 20" sag for the heated beams. I assumed the same heated stress/strain and same 20" deflection would apply to the girder.
I used ACAD to generate a circle where the circle intersected the two end points at the length of the girder and a tangent point at -20" at the center. ACAD measured this angle at 8 deg. at the seat tip.

Your calculation chart for deflection is wrong because it assumes a uniform load along the girder and that the ends were fixed (bolted).
You should use one with beam point loads along the girder and not fixed ends (hold down bolts were sheared). The difference would be substantial. Any deflection of the girder away from level would place the load at the tip of the seat and bend or shear it. Will you provide this revised deflection chart.

The 2" expansion at each end would cause further westward lateral buckling since the gap to the column was about and inch, placing the girder farther away from the seat.

Re calculations: I'm not an ME or SE. I asked a busy SE consultant/friend of mine to do some calculations and he said he would (or rather he tepidly said "ok"), about a week ago. After explaining to him the CD vs.fire claims for WTC1,2,7 he thought the CD claim "is stupid." I've bugged him a couple of times for the calcs. It don't look like he's going to spend his time on this.

NIST clearly states that the girder was free to rotate at both ends because the bolts had failed (Pg 353) Another point is that the floor beams supported by the girder failed and buckled. If those floor beams were no longer supported by the seats at the exterior, most of the floor load would have been transferred to the girder.

 

[enik]

NIST didn’t consider the vertical failure of the seat and its model included a seat stiffener at Col 79 not shown in the plans.
“Since vertical failure of the seat was not considered (Section 11.2.5), the connections at Columns 79 and 81 were both modeled as stiffened seats.” (NCSTAR 1-9 p.558) See “Figure 12-25 Seat connection in global model Column 79. (p. 559)

It sounds like we have a different scenario not considered in the original post? A failure mode in which a torque would have bent or sheared the seat plate and then caused the girder to slide off [in a vertical direction]. And this was not considered by the NIST since they included a stiffener plate in their model that was NOT actually on Column 79.

Is this correct?

 

[Newtons Bit]

It is hard to understand how the girder would deflect downward and sag by 8 degrees as you show in Detail D of your drawing. This would require the downward deflection to be about 37". Do you have calculations to show this would occur? Is this a result of applying the 83,700 lb. load to the end of the seat?

NIST claims the girder between columns 44 and 79 was at 500 degrees C. At that temperature, the AISC says the modulus of elasticity retention factor is 0.63. In my calculations I used the reduced modulus of elasticity and girder loading that would produce reaction loads on each side that are 1.3 times the 83,700 lbs reaction load that you show, and they only show about 1.5" of downward deflection and 0.31 degrees of sag.

The girder would be expanding by about 2" per side at 500 degrees C and the horizontal deflection and sag combined wouldn't cause more than 1" of total shortening of the girder, or about 0.5" per side. Thus, the web and stiffeners would remain well over the 14" deep vertical plate beneath the seat at column 79, making the failure mode you are proposing seem impossible.

Your calculations show that the girder has buckled, which results in your deflection calculations being completely wrong. Now I haven't been following this sub-forum that closely. Maybe you've run some numbers that take this into consideration. Perhaps you'd like to show this revised calculation?

 

[Tony Szamboti]

Your calculations show that the girder has buckled, which results in your deflection calculations being completely wrong. Now I haven't been following this sub-forum that closely. Maybe you've run some numbers that take this into consideration. Perhaps you'd like to show this revised calculation?

I have shown that the west side girder framing into column 79 at the 13th floor would have buckled before it could deflect column 79 enough to the east to make a difference.

However, that isn't the girder between columns 44 and 79, which is the one I was discussing above, and which Basque Arch was referring to as part of the walk-off claim/counter claim controversy. I don't see a way for it to have buckled, as there is no compressive load on it.

 

[Newtons Bit]

I have shown that the west side girder framing into column 79 at the 13th floor would have buckled before it could deflect column 79 enough to the east to make a difference.

However, that isn't the girder between columns 44 and 79, which is the one I was discussing above, and which Basque Arch was referring to as part of the walk-off claim/counter claim controversy. I don't see a way for it to have buckled, as there is no compressive load on it.

1) Where are your column 79 calculations?
2) Lateral-torsional buckling does not involve axial loads.

 

[Tony Szamboti]

1) Where are your column 79 calculations?
2) Lateral-torsional buckling does not involve axial loads.

I discussed the west side girder buckling earlier in this thread. In that case there would be a compressive load.

Are you saying the girder between column 44 and 79 would fail due to lateral-torsional buckling with the same vertical loads it always had on it? Remember, it didn't have shear studs according to NIST and it handled those loads fine for years. The beams would have provided lateral restraint so you can't use the full length of the girder in the calculation.

I'll take a look at it with the 500 degree C temperature on it and reduced modulus if that is what you think would cause it.

 

[Newtons Bit]

Err...

I don't see a way for it to have buckled, as there is no compressive load on it.

2) Lateral-torsional buckling does not involve axial loads.

I discussed the west side girder buckling earlier in this thread. In that case there would be a compressive load.

And this is why I hate discussing engineering with truthers.

 

[Newtons Bit]

Are you saying the girder between column 44 and 79 would fail due to lateral-torsional buckling with the same vertical loads it always had on it? Remember, it didn't have shear studs according to NIST and it handled those loads fine for years. The beams would have provided lateral restraint so you can't use the full length of the girder in the calculation

Yes, it failed in lateral-torsional buckling. This occurred when the beams that normally braced the girder failed. This is all detailed in the NIST report. See Section 8.8.

 

[Tony Szamboti]

Err...






And this is why I hate discussing engineering with truthers.

This is an immature comment. Nobody is twisting your arm to discuss anything.

 

[Newtons Bit]

This is an immature comment. Nobody is twisting your arm to discuss anything.

No. Pointing out sliding arguments is not immature. Pretending to be an expert when you lack a complete understanding of the basic principals of structural engineering, on the other hand, is immature.

eta: added the new edit to the quote.

 

[carlitos]

eta: added the new edit to the quote.

You have to wait a certain period of time before quoting Mr. Szamboti. It's a nebulous and subjective period of time, like when should a widow start seeing gentleman callers.

 

[Tony Szamboti]

Yes, it failed in lateral-torsional buckling. This occurred when the beams that normally braced the girder failed. This is all detailed in the NIST report. See Section 8.8.

Section 8.8 does not say the girder failed due to lateral-torsional buckling. It says the northmost girder failed due to lateral-torsional buckling and that the remaining beams did also and that they rocked the girder off it's seats to the east.

Of course, they only show the north most two beams and the model is missing the three beam stubs from the north exterior to the northmost beam.

I did a finite element analysis with and without those beam stubs. Without them the northmost beam did fail in lateral-torsional buckling and the second beam was deformed some sideways but hadn't rotated. With the stubs the northmost beam did not fail in lateral-torsional buckling.