Mineapolis bridge - yes again

[csd72]

I noticed the following news thread on the aol site:

http://news.aol.com/story/_a/2003-p...e/20080323153909990001?ncid=NWS00010000000001

I just love how they quote lawyers as if they are expert witnesses.

 

[HgTX]

I was at TRB when a bit of a report was given by FHWA and one of the consultants who worked on the investigation. The gussets were 1/2" when similar gussets elsewhere in the bridge were 1", which is what got people looking at their design to begin with.

Because the original design calcs are gone, there is no way to know if it was a design error or a drafting error. The investigator said because of the particular distribution of the construction loads, that node in the bridge was seeing the highest load it had ever been under before. That's what was different.

Everything was a factor; if the bridge hadn't been redone over the years with more concrete barriers and more traffic, the undersized gussets might have survived (then again, maybe not, considering the particulars of the construction load distribution), but if the gussets hadn't been undersized, all the concrete barriers, increased traffic, and construction loads would have been fine.

Hg

Eng-Tips policies: faq731-376

 

[Qshake]

HgTX - I'm not sure I agree with your last statement "if the gussets hadn't been undersized, all the concrete barriers, increased traffic, and construction loads would have been fine."

Those or other gussets may be overloaded as more load is added and no analysis done to show what is the capacity.

It is normal procedure for us and our clients to submit calcualtions showing that for additions such as bulk materials and batch plants etc will not overstress a structure. If anything this horrific failure has shown us that procedure isn't adhered to as it should be.



Regards,
Qshake

Eng-Tips Forums:Real Solutions for Real Problems Really Quick.

 

[msquared48]

I find it hard to imagine a bridge, subject to the lane loading live load of 300 psf, failing due to the imposition of construction loading. Obviously there was no traffic in the area of the construction....???

If the committee is pointing there as the reason for the failure, as I gather from the posts above, I don't think they have looked far enough. It just doesn't add up to me.

Mike McCann
MMC Engineering

 

[HgTX]

They were saying it was the particular pattern of where the construction materials & equipment were placed on the bridge. A concentrated load here, an absence of load there, etc.

Either that or it was a big conspiracy.

Qshake: "if the gussets hadn't been undersized, all the concrete barriers, increased traffic, and construction loads would have been fine." An earlier version of that sentence, before I posted, included something about analyses for increased load that assumed the connections were stronger than the members framing into them--which they weren't, because of the undersized gussets.

Hg

Eng-Tips policies: faq731-376

 

[Qshake]

But HgTX, the AASHTO code only requries gusset plate connections to be designed for the average of the actual stress (or factored loads) and member design strength but not less than 75% of the member (deisgn) strength....hence connections are not designed over and above members which the NTSB chairman has stated.

As such, gusset plates, like all structural elements should be checked and not ruled out just because engineers believe them to be stronger than any other elements....

I hope I'm not missing your point.

Regards,
Qshake

Eng-Tips Forums:Real Solutions for Real Problems Really Quick.

 

[civilperson]

Where did the live load value of 300 psf come from? Not 550 plf/lane?

 

[HgTX]

Even with the 75% rather than 100% requirement, there's still an assumption that can be made about the connection without directly analyzing it. IF one knows which code the connection was designed to...IF the connection was designed correctly...

I'm not sure what my point is. I think it's that analyses were probably done for the various load increases using assumptions reasonable at the time. The collapse shows that those assumptions aren't reliable.

Hg

Eng-Tips policies: faq731-376

 

[electricpete]

I'm not particularly mnechanically oriented and haven't followed the discsussion. Just looked at the link and have an observation and a question.

Observation: Photo's 1 and 2 have different date stamps... 2 days apart. Seems like someone saw something unusual and went back for another look?

Question: What's a gusset and where is the supposed bent gusset in those photos?

Thanks

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[Qshake]

electricpete - Those are two different joints that are photographed, one on 6/10 and the other on 6/12.

For a symmetrical truss, that is symmetrical about centerline of mid span and centerline of structure, there should be 4 joints that are nearly identical except for location. for that joint in question, it is U10. U10 has four duplicate locations. For example, U10W, U10E and U10'W and U10'E. W and E designate the west and east trusses and in this case, the prime (') and non-prime represent the nort and south locations about centerline of mid span. clear as mud?!

Regards,
Qshake

Eng-Tips Forums:Real Solutions for Real Problems Really Quick.

 

[jmw]

Actually, I'm with Electricpete (not just in knowing nowt about bridges) in not really seeing what these photos are supposed to show.

Qshake, for us ignoramuses, is it the plates on the sides of the joints that connect across from one structural member to another?

Are these the sort of view an engineer would take or a tourist? Perhaps to the trained eye the problem is obvious but I'd have thought a view that is exactly end on would be better to show the warping to better advantage.

JMW

http://www.ViscoAnalyser.com

 

[GregLocock]

Photo 1 and 2 shows a clear kink in the vertical edge of the gusset plates. That would make them useless in compression.

the gusset plate is the thin vertical squarish plate reinfrcing the joint between the main members.

Although the kink looks small, it would be enough to destroy the stiffness of the plate.









Cheers

Greg Locock

SIGlease see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[Qshake]

Greg - I differ from you stance on the matter of the kink and it's effect on the plate.

Because of the location of the kink, I believe it reduces the capacity of the plate to transfer signicant and maybe even the design compressive load but I believe that if the buckle or deformation where at the end of the member framing in, then it would indeed zero out the plate as a means of load transfer and any transfer after buckling would be left to binding between members. But the deformation at the edge away from the fastener pattern where the load is being developed from member to plate is likely to continue to have some capacity.

JMW - The gusset plates are those that sandwich the members (verticals, diagonals and horizontals ) together.

Regards,
Qshake

Eng-Tips Forums:Real Solutions for Real Problems Really Quick.

 

[jmw]

Thanks,
I think that's what I thought.

JMW

http://www.ViscoAnalyser.com

 

[msquared48]

civilperson:

Sorry that I pulled the rabbit out of the hat here... I got the300 psf value from an old 1963 California Bridge Design manual we used in school. This "lane loading" controlled in short span situations where the H20 and others controlled in longer spans.

Mike McCann
MMC Engineering

 

[bridgebuster]

lane load is 640 plf plus a concentrated load of 18k for moment and 24k for shear