Continue to Site

Eng-Tips is the largest engineering community on the Internet

Intelligent Work Forums for Engineering Professionals

  • Congratulations waross on being selected by the Eng-Tips community for having the most helpful posts in the forums last week. Way to Go!

I-35W Redux

Status
Not open for further replies.

bridgebuster

Active member
Jun 27, 1999
3,969
This article was sent to me by a colleague:

I-35W Suit To Target Engineer, Contractor

Independent investigation pinpoints rusty bearings and contradicts federal report

04/01/2009

By Aileen Cho and Tudor Hampton, with Tom Ichniowski



Contradicting the National Transportation Safety Board’s report that blamed too-thin gusset plates for the Aug. 1, 2007, collapse of the Minneapolis Interstate 35W bridge, an independent analysis has concluded that rusted, frozen roller bearings prevented thermal expansion and caused a truss chord to fail, triggering the gusset-plate failure. The analysis by Thornton Tomasetti, New York City, using forensic bridge information modeling, was presented to victims’ families and will be deployed in an anticipated lawsuit against URS Corp., San Francisco, and Progressive Contractors Inc., Minneapolis, in the next few months, according to Chris Messerly, one of the pro bono lawyers with Robins Kaplan Miller & Ciresi LLP, Minneapolis.



Slide Show



Photo: Associated Press

“Our experts’ findings confirm that URS, the engineering company hired by the state to ensure the safety of the bridge, and PCI, the construction company who chose to pile its materials on the bridge rather than keeping them just off the bridge, are responsible for this catastrophe,” says Messerly.

Lawsuits on behalf of 117 families of the 13 people killed and 145 injured will be filed in state court within the next few months, he says. “This is a negligence claim. URS knew these things: The design was obsolete; there was no redundancy in the bridge; some gusset plates were bent; the roller bearings were corroded and would not move with broad temperature changes....As they were hired to assess the safety of the bridge, they had a duty to see that [it] was safe. They failed in this duty.”

URS declined to comment. A lawyer for PCI was unavailable to comment due to a medical emergency. Thornton Tom-asetti (TT), which was not allowed to comment, consolidated more than 50,000 paper documents on the collapsed bridge into a forensics information modeling program made available by NTSB. The team created an interactive computer graphic model of the bridge using shop drawings, photos, videos and maintenance records.

Messerly says there is no written report to issue. “We have not asked TT for a report, and I am not sure we will, as they have already given us a very comprehensive presentation of why the bridge fell.”

The issue is essentially one of maintenance versus design. NTSB’s report had blamed the quality-assurance procedures of original bridge designer Sverdrup & Parcel and Associates Inc., St. Louis, for the gusset plates on the main truss. Sixteen of the gusset plates were 1/2 in. thick rather than the required 1 in., according to NTSB Chairman Mark Rosenker.

“We stand by our investigation,” says Terry Williams, NTSB spokesman, adding the safety board has no comment on the TT presentation. Bruce Magladry, the NTSB investigator who was instrumental in alerting state departments of transportation about possibly underdesigned gussets before the NTSB report was released, said in a recent speech, “Forty years ago a design mistake was made that led to construction of a bridge that was in danger of collapsing the day it was completed....It took 40 years for the right circumstances to come about.”

Officials with Northbrook, Ill.-based Wiss, Janney, Elstner Associates Inc., the forensic engineer that discovered the gusset-plate problem, declined to comment on the new findings, citing the ongoing litigation. But its engineers considered the effects of bridge bearings, thermal expansion and construction loads in great detail in their investigation.

Its 164-page report, commissioned just after the collapse by Minnesota Gov. Tim Pawlenty (R) and released along with the NTSB findings, allows that previous URS inspections indicated that the roller bearings at Pier 6 were locked up. However, Wiss, Janney notes that it was unable to determine whether or not they were in motion at the time of collapse.

It speculates, however, that locked bearings may have actually increased the bridge’s capacity, while operational ones would have forced an earlier collapse as the bridge heated up from summer temperatures. “Compromised roller bearings and the associated restraint forces could have resulted in a substantial increase in the bridge’s ability to carry superimposed loads at the time the construction material was being placed on the deck,” it says.

The report continues, “If the bridge structure was still realizing a net heat gain at the time of collapse, the buildup of restraint forces in the Pier 6 bearings could have caused them to slip.” Either way, it dismisses the bearings and points to the gussets as the trigger.

One source close to the NTSB investigation says about the new theory, “I don’t think it’s well-founded,” adding that personal-injury attorneys are “getting a little bit nervous” and pointing to new defendants in an effort to discourage the victims’ families from taking state settlement money, the offer for which expires this month. “I don’t think that things are always driven by facts,” the source adds. “They are driven by dollars.”

State lawmakers last year passed a $38-million bill to compensate victims, up to $400,000 each. Messerly testified in favor of creating the fund. “By taking the money from the fund, they do waive any claim against the state,” Messerly says, adding , “The statute specifically allows the victims to bring claims against all other responsible parties—including URS and PCI.”

 
It is a difficult issue Bridgebuster. I was very surprised at the report by the NTSB, I find hard to believe that a design error on the gussets would allow the bridge to operate for 40 years and then make it collapse under the 'right' circunstances of loading. It does not seem very logical to me.

On the other hand,they carried out a complete review and analysis of all the information available for months, so I have no reason to believe that they reached the wrong conclusion. They certainly know more than me!

You could have 5 consulting companies analysing the data and have 6 failure modes. I guess we will never know for sure

What is you oppinion?
 
At this stage, a 164 comprehensive report by Wiss Janney Elstner certainly bears more credence than a statement by a lawyer that he has an unwritten opinion by Thornton Tomasetti saying that WJE is wrong. The gusset plates were definitely grossly underdesigned. It is not clear if other factors contributed.

PCI has also filed suit against the State and Jacobs.
 
I am always amazed by those structural engineers questioning the strength of a ill-designed gusset plate.

There should really be no surprise that this gusset plate performed as it should have for 40 years. The plate, according to the NTSB, did not meet the allowable stress limitations imposed by the code. This hasn't anything to do with a limit state for failure. As structural engineers we should expect our works to perform when an allowable stress is exceeded so long as the ultimate capacity is not exceeded or a specific limit state exceeded.

And while the gusset plate is identified as the location wherein damage manifested I don't believe the NTSB reports exactly what caused the plate to fail the way it did.

And yes, it is likely a myriad of circumstances that happended in concert.

IMHO, I don't think the NTSB report goes far enough and this new work by Tomisetti is not well founded. Pictures from the NTSB website show the chord in question, presented by the Minneapolis Star Tribune article on this Tomisetti analysis do not show a buckling failure as Tomisetti claims.

So, if TT really believes a lower chord buckled at least show the people a photo that supports the analysis.

Regards,
Qshake
[pipe]
Eng-Tips Forums:Real Solutions for Real Problems Really Quick.
 
Last year my initial thought was a bearing failure; that was based on published photos. I recall there was one span that appear twisted whereas the others looked as if they fell straight down.

I thought the NTSB was a little weak and even though TT is a good firm they're not a bridge outfit - at least on the east coast.

I agree with kelowna's assessment: 5 firms; 6 answers.
 
Several important contributing factors were omitted during the investigation. One was improper inspection of the bridge, which, in my opinion, was never inspected by a bridge engineer. Comparison of some photographs, presenting PP #13, bottom chord in the West Truss (2004 report p.37, 2005, p.37 and 2006 p.34 reports) leads to very interesting conclusions sustantiating my opinion regarding the inspection of the bridge. Also, even one look by experienced bridge engineer at the 2" thick plates framing into 1/2" gussets could lead to extra investigation.
As to the condition of the expansion bearings, and their influence on the collapse of the bridge, my initial and current opinion is that sudden release of one side bearing could trigger the collapse. URS reports lists the thermal reaction at pier 6E at 184 kips for 20F temperature differential. For 60F differential and sudden release of one of the bearings, the extra force in the diagonal will be very high. Also, the URS report is not very specific regarding the way the RC deck was modeled and included in the analysis. As the original expansion joints were likely paved over, or completly filled with debris, the continuous for compression deck could be the big contributor to the magnitude of the thermal forces in the structure. So, TT report could be correct.
On the other hand, locked bearings increased the force in L9-U10 diagonal for temperature rise and even speculation, that their frozen state increased the bridge capacity is a pure nonsens. But everybody can have different opinion.
 
One matter that is overlooked with respect to the bearings being locked up is that a purely locked situation is impossible due to the inherent stiffness of the columns supporting the bearings. Had the bearings been locked, the displacment would have simply been manifest in the colummns.

Regards,
Qshake
[pipe]
Eng-Tips Forums:Real Solutions for Real Problems Really Quick.
 
Qshake,
In a regular bridge that would be the case. The problem with this structure was that the piers were relatively short and stubby, providing good horizontal resistance. The URS's structural model took this all into consideration, that's why the thermal forces were so high.
Wiktor
 
Just a couple of thoughts:

Originally, my first thought was poor maintenance would be found as the cause of the collapse. After hearing some of the investigation details, things centered on the under-designed gussets. It does seem unusual that in the 40 year life of the bridge, this is the first time the gussets were stressed high enough to fail - possible, but unusual.

Also, what conclusion has been reached about a local overload due to the construction operations that were underway in the area of the collapse?

Finally, the most unusual aspect of bridgbuster's OP is "...one of the pro bono lawyers". I've NEVER heard of pro bono lawyers on a catastrophe of this magnitude.
 
wiktor - I submit that is the case for any bridge due to the matter of proportions. For example, you say it happens in "regular" bridges and I agree with that and note that the substructure is designed to account for the level of force for the bridge size. In the case of I-35, I note that this bridge is much, much larger than "regular" and hence the loads are higher which when acting on the substructure will produce a proportional displacement. A displacement that we know was not intended since the bearings were supposed to work.

Regards,
Qshake
[pipe]
Eng-Tips Forums:Real Solutions for Real Problems Really Quick.
 
Actually jheidt2543 there are quite a few firms doing pro-bono work on this case. See link below.
It reinforces my opinion about some lawyers.

The case is still not clear on my mind. No reasons to doubt the NTSB, and not happy with a cause that does not seem logic to me.

One question for consideration. The video of the collapse seemed to indicate a symmetric failure. Wouldn’t the failure of one of the gussets cause an asymmetric failure?

 
kelowna:

Thanks for the link, it does change my opinion of lawyers - at least some of them.

I have a daughter and son-in-law that travel that bridge once or twice a week. Luckily, that was not one of the days.

I could be quite some time before the entire picture becomes clear on the causes of the failure. I have an inclination to favor WJE, but the first "reporter" doesn't always get the full story. Also, the proximate cause is not always the ultimate cause of a failure. As you all know, most of the time these things are caused by a confluence of events.
 
kelowna - Having viewed the video several times (still no expert and have no theories, i.e. no monday morning quarterbacking!!) I can say the south end of the bridge failed first while the north end was a resultant failure of the south end (i.e. too much weight for the newly formed cantilever span from the north river pier) to handle. Again, I can reason that the loads went right to the weak link in the chain, which was the U10 gusset plates. I do not believe that the U10 gusset plates caused the failure but were clearly victims of overload. I further believe that the north gusset plates, not having been loaded as the southern counterparts were did not fail in the same way as the south plates. All of that is conjecture on my part but for me makes logic sense.

So while the video shows a very near simultaneous failure of the river span it was, in my opinion, started on the south end.

Keep looking at the video and let us know if you see this distinction.

Regards,
Qshake
[pipe]
Eng-Tips Forums:Real Solutions for Real Problems Really Quick.
 
Thanks Qshake

If I understand correctly, the longitudinal axis of the bridge runs north/south. In that case what I was trying to say is that the failure was symetrical, meaning both trusses, the east and the west trusses, seemed to have failed at the same time (again, with the poor quality of the video, that is just speculation, as you well said)

The point is that the loading of the bridge was not symetric, one lane was closed and there was construction loading, the others were open and with traffic. If the loading was not symetrical, it would seem unreasonable to assume that the gussets on both trussess failed at the same time. I would have expected to see a truss going down first, dragging the other behind but with some delay.

Maybe that is what happened anyway, or maybe the bulk of the loading on those gussets was not from live load but from dead and environmental.

Sorry, I know I am beating a dead horse, I should know when to stop. This will be my last post on the matter
 
Well, Kelowna maybe that's a good point. we know the south end was loaded most...can see that from the released photos of the material on the bridge at the time of collapse. And the material was loaded in the area near U10 on the south. So perhaps the additional dead load of material and equipment and perhaps even the lack of decking in some areas nearby finally pushed the plates into overload...not just overstress but real plastic deformation. Hence the bridge trusses east and west failed nearly simultaneously.

It's logical but the NTSB didn't say it was due to material and equipment weight. Perhaps they dont' want to say that. What should be said then is a combination of thin plates and too much construction loading.

Regards,
Qshake
[pipe]
Eng-Tips Forums:Real Solutions for Real Problems Really Quick.
 
And the statistics. I would say every 9 in 10 collapses of buildings happen when either works are ongoing, besides, within or under them. Left alone, they use to stay as they were, "I do not know why".

It is the incompetence of those acting, sometimes unfortunately (I muse, improperly) licensed people.

We had here town works with say 15 to 20 m tall unstabilized pairs of masonry walls (of some building to be revamped, really rebuilt) that collapsed when some trusses were being placed atop. What a surprise!

And thes guys are those that decide our buildings may have license or not! They live happily, more budget kills any worries.

The failed moment connections of San Francisco by man(mad?)datory details may be another case; sorry guys, it seems there is always someone that knows better. Yet they also live happily.
 
Status
Not open for further replies.

Part and Inventory Search

Sponsor