An old structural paper / question 7

[youngstructural]

Hello All;

Having contributed to another thread has made me think of a rather old engineering mystery. I'm hoping someone out there *cough* insert standard regulars here *cough* might be able to help:

I don't suppose anyone knows the "mythical" paper from the 1910s or 20s that discusses how much fixity a truss gusset connection needs to actually affect the system sufficiently to make a pin-pin analytical solution inappropriate? I've always wanted to get my hands on a copy, but have never found out even the paper or author name. The paper became a bit of a legend at Public Works Canada (now PWGSC, long story), but no one could find the thing for me... The department was being wrapped up when I was brought on board as a bit of "young blood", but all of the old training materials had disappeared in a recent move. The remaining engineers were wonderful and very giving of their time, however it would have been good to get a bit more documentation for some of the tips and tricks I learnt.

The evidence of the paper's effect is clear: Most bridge structures before the 1920s had enormous pins, then after they had fixed-style gussets. None of the analysis methods changed, engineers just learnt to recognize that a fairly innocuous redistribution of forces occured and the structure's true behaviour "came out in the wash". My mentor once told me that he felt the work likely inspired, or potentially even underlaid, Hardy Cross' pioneering methods.

Knowing it exists, and never finding it, always makes me wonder just how much we must repeat simply by being unable to locate original papers!

Hope someone can help,
Cheers,

YS

B.Eng (Carleton), P.Eng (Ontario), MIPENZ (Structural-New Zealand)
Working in Canada, and missing my adoptive New Zealand family... at least I brought the little Kiwi with me!

 

[DRC1]

Contact the Linda Hall Library. They are on the web. I would call, as the people that work there are very helpful. Cost once they find the papers are very reasonable. I would call rather than try to contact by e-mail as there will be several questions I am sure.

Good Luck!

 

[BigH]

SRE - can't you help???

 

[msquared48]

Wasn't there a lot of discussion in the forum here on this subject relative to the Minneapolis Bridge failure recently?

Mike McCann
MMC Engineering

 

[BigH]

youngstructural - have no clue to the paper you are referring. see links below that might be of some interest. However, the one article suggests that you might try contacting some professor emeriti of various universities and they might be able to help. My original advisor at school worked on the Hoover Dam. They might be a start.

http://www.jflf.org/pdfs/wi301/historicbridges.pdf

http://www.marygrove.edu/ids/papers/Dry_Dock_Engine_Works.doc

http://www.in.gov/dnr/historic/files/bridge28.pdf

 

[hokie66]

YS...I don't know either, but a paper of that vintage would have dealt with riveted connections rather than bolted. The leap to assume pinned action in riveted joints is probably not as big as for bolted joints, as rivets are much softer.

 

[youngstructural]

Gentlemen: Some interesting leads, and I'll do a bit more digging. I think I'm finally motivated enough to use what little spare time I have to run this mystery to ground. I'll let you know if I find out more.

Cheers,

YS

B.Eng (Carleton), P.Eng (Ontario), MIPENZ (Structural-New Zealand)
Working in Canada, and missing my adoptive New Zealand family... at least I brought the little Kiwi with me!

 

[SlideRuleEra]

YS, BigH - This sounds interesting, will try to turn something up on this - may take some time.

http://www.SlideRuleEra.net

http://www.VacuumTubeEra.net

 

[SlideRuleEra]

Good News... the paper does exist. It is titled "Secondary Stresses in Framed Structures" and was written by E. W. Pittman and presented, by him, in 1909.

Mr. Pittman addresses several potential improvements in riveted structures, such as proper alignment of members at the joints. The ground-breaking part is to INCREASE the thickness of gusset plates so that each rivet is fully loaded (bearing) in double shear. This will decrease the number of rivets needed in any particular joint. Therefore, the gusset plate will be much smaller. His insight is that the smaller gusset plate will DECREASE the fixity of a member, allowing a reduction in secondary stresses. The net result was longer riveted trusses.

Many other engineers quickly accepted this improvement, including E.S. Andrews in his 1914, "Further Problems in the Theory & Design of Structures" (see page 219)

http://books.google.com/books?id=KA...e=gbs_v2_summary_r&cad=0#v=onepage&q=&f=false

Another was E. McCullough in his 1921, "Practical Structural Design" (see page 205)

http://books.google.com/books?id=iJ...e=gbs_v2_summary_r&cad=0#v=onepage&q=&f=false

Even one of the most respected bridge engineers, J. A. L. Waddell, seems to have liked the concept.

The only "big-name" holdout appears to have been Milo S. Ketchum, who still preferred the pinned trusses, as late as 1921.

The complete 1909 paper, "Secondary Stresses in Framed Structures", is in the file below:

http://www.SlideRuleEra.net

http://www.VacuumTubeEra.net

 

[Ron]

Nice Job SRE...came through as usual.

Ron

 

[youngstructural]

SlideRuleEra: You are the structural man. I am so very, VERY impressed!!! I can tell you that there are several (retired) engineers in Canada who are going to be super keen to hear that!

Well impressed am I; Yet another star for you!

Sorry for the delay; I'm out in the field at the moment, so this is the first time I've been on the net in four days. Please don't let the delay be interpreted as a reduced gratitude, I greatly appreciate you effort.

Now if you'll excuse me I believe I have a couple of papers to read. *smiles*

Regards,

YS

B.Eng (Carleton), P.Eng (Ontario), MIPENZ (Structural-New Zealand)
Working in Canada, and missing my adoptive New Zealand family... at least I brought the little Kiwi with me!

 

[Qshake]

YS - The method, by the way, did change drastically from that used with pin joints over those with gusset plates.

Most older structural books or texts on bridge design will demonstrate the differences in methods.

The problem, in my humble opinion, is that AASHTO has for many decades insisted on determining the maximum direct, bending and shearing stress. It does not say how. Most engineering examples show a calculation of bending stresses in the plane of the plate due to the unbalanced member loads which places the maximum stress on the very edges of the gusset plates. This is a well documented folly as the actual tensile and compressive stress on the plate edges is practically zero and most load transfer occurs in the plate directly at the terminus of the member. Fisher, Kulick et al have demonstrated this in the book on connections. The only real good approximation is shearing stress across the plane parallel to the lower or upper chords.

Whitmore also documented the poor design procedures for gusset plates and did so in 1952! It's surprising that the methods have persisted as long as they have.



Regards,
Qshake

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

 

[youngstructural]

QShake: Interesting point; I'll have to try to get a copy of Fisher, Kulick, et al. and have a read of their procedure.

This actually reminds me of the shear in concrete issue from the 1950s and 60s. It was known for fifty years before large structures started to fail in shear (US Airforce base Wilkins, etc) when Talbot stated that the shear resistance of a concrete beam would be found to be dependent upon depth, longitudinal reinforcement, length over depth, etc.

Thankfully it seems obvious that the current procedure is not unconservative, given that we don't see failures due to the design against these fictitious forces.

Cheers,

YS

P.S. I'm really enjoying reading the discussion to the Pittman paper: I suppose the internet has now caused these paper discussions to occur on forums like this one, or between engineers by e-mail. Sadly I think we've lost something by not having it all together in this convenient format. You can clearly see that the immediate reaction to the paper was quite favourable, and it is obvious that many of these engineers were putting the ideas into practice right away. I suppose I'll have to add Pittman to the list of engineers who I believe caused fundamental shifts in our profession...

B.Eng (Carleton), P.Eng (Ontario), MIPENZ (Structural-New Zealand)
Working in Canada, and missing my adoptive New Zealand family... at least I brought the little Kiwi with me!

 

[slickdeals]

YS:
The Fisher/Kulak book which Qshake is referring to maybe available for free download at

http://www.boltcouncil.org/files/2ndEditionGuide.pdf

QShake: Is this the material that you are referring to?

 

[slickdeals]

YS/Qshake,
I found this resource after the failure of the I-35 bridge. Take a look at their procedure for checking gusset plates. I am curious as to what your thoughts are with regards to the issue you raised with stresses.

http://www.ntsb.gov/Recs/letters/2008/H08_1_Design_Adequacy_Report.pdf

 

[Qshake]

slickdeals - that is the reference I was referring too.

Also I should note that due to non-symmetry in connection plates there may very well be a eccentricity in the connection that causes out of plane bending on the joint connection. But my previous discussion concerns the usual practice of calculating an inplane bending stress on the gusset plates due to differences in member forces.

Chapter 15 of the Fischer reference is all you need to wonder why we continued to design plates this way...that and the Whitmore paper of 1952.

My guess is that engineers just said, nothing has failed yet...so we keep doing it....

slickdeals - In response to the NTSB report, I have read that and the direct quote from AASHTO is exactly what I noted in my earlier post in this thread. Moreover, it is the so called bending stress in the plane of the plate that draws my ire. I agree with the shearing stress calculations from the procedure and experimental results have at least borne those stresses out. But where a flexural formula applied to the plate is supposed to yield high tension and compressive stress at the edges of the plate it actually yields none and this is shown with a very crude FEA model in the Fischer reference. I have seen many many more such models since that support the Whitmore stress distribution as more accurate and more critical.

My guidance would be to determine the shear on the weakest plane and then focus on the Whitmore stresses. And where applicable check the out-of-plane bending stress due to unsymmetrical plate arrangements.

Regards,
Qshake

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

 

[BigH]

Rob - great job!