Rivets and Analysis of Historic Structures 2

[HeavyCivil]

My firm has been tasked with replacing the bearings on a through-truss railroad bridge. The bridge was built in 1927 and was originally property of the Delaware and Hudson. The span is long so the jacking loads are quite high - the bridge weighs over 700 Tons. Analyzing the connections is fairly simple, but using the lowest allowable stress value for rivets listed in AREMA produces unsatisfactory capacity in the floorbearm-bottom chord connection to jack under the floor beam. A502 Grade 1 Hand Driven rivets give you 11ksi allowable shear according to AREMA. AISC lets you use a higher value.

Analyzing the plans shows lots of information about the geometry but nothing about the material properties. It does reference a standard "D+H Co 1914 Spec". Does anyone know where one could find this standard spec?
Does anyone have any information on the historic use of Grade 2 and Grade 1 rivets in Railroad structures? Grade 2 would work, but that's not an assumption I can make without a high degree of probability.

 

[kipfoot]

Any chance you're close to St. Louis? It looks like there's a copy at the Univ. of Missouri St. Louis library, where they apparently still have paper books. link

 

[bridgebuster]

Two ideas if you can't locate the railroad specs:

1. Remove a rivet and have it tested by a metallurgical lab
2. Remove the rivets in the connection and replace with high strength bolts.

Regarding the allowable stress: You could go to google books and see what was published at that time (couldn't find the D&H specs). In one book I have, it says 24 ksi for shop rivets and 20 ksi for field driven.

 

[SlideRuleEra]

If vintage AREMA is not available, another approach is to visit this page of my website and download early 20^th century steel manufactures' handbooks:
Assorted Structural Steel Handbooks

Maybe some detective work going though those handbooks will help. I'd suggest starting with the "Cambria Steel" series.

http://www.SlideRuleEra.net

http://www.VacuumTubeEra.net

 

[HeavyCivil]

kipfoot - I'm not in St. Louis. Thanks anyway.

bridgebuster - We removed a couple rivets last week (exploratory) and replaced them with bolts. Good news is that one is 1" while the other is 7/8" as expected. The 1" works even at the more conservative allowable stress in AREMA. We may end up removing all the 7/8" rivets and replacing them with bolts, but I'll look into the metallurgical study.

SlideRuleEra - thanks for the link. I'll take a look.

 

[Lomarandil]

I deal with a lot of historic riveted bridges -- typically not for use with AREMA though, so my advice will have to be more general.

I don't think A502 came around until the 1960s, so it's unlikely that your rivets match that spec. (As bridgebuster mentioned, you can always send one to a lab).

For 1927, the material properties were usually based on a project/company specific set of specifications (your "D+H Co 1914") -- so there's no way to be sure without that original copy. If you called the University of St Louis that kipfoot posted and asked nicely, they'd probably scan that document in for you.

Without that information, rivets from that time were typically similar in practice to early A7/A9 rivets. Pre-1928, you're looking at 10ksi allowable shear through the shank Quick reference (I'd argue that with a spec written in 1914, this may be slightly unconservative). I wouldn't assume any Fu over 50ksi for that time period, and more likely around 45-46ksi.

The good news is, replacing old 7/8" rivets with A307 or A325 bolts is pretty inexpensive if you find a contractor with experience doing it.

I'd be careful assuming many of the rivets were nominal 1" (although it would cause me to take a second look at the drawings, sometimes sizes were mixed in critical areas). More likely than not, that particular hole was slightly misaligned and had to be made fair to drive the rivet -- and it was just a case of good workmanship that let the rivet fill the entire diameter.

 

[CELinOttawa]

You may also want to be quite careful about what privets are used where.... In that era you frequently find that there can be more than one grade of rivet used in the same bridge. Hate to be the bearer of very frustrating news, but I have seen bridges for which this was the case. If you take out a rivet and have it tested you effectively know about that one connection.

An excellent recent paper:

http://www.sparksengineering.com/wp...on-and-Steel-in-Historic-Bridges-P-Sparks.pdf

 

[tolchijb]

I am curious how easy it is to remove the rivets. I have always been told it is a nightmare because the rivets deform over time.

 

[Guastavino]

Does this help? It's from the "Structural Engineer's Handbook" by ketchum. Date of book is 1918

 

[bridgebuster]

Tolichijb - it depends. How many plies of steel, how much clearance there is for the helldog, position of the rivet - horizontal or vertical. Driving a rivet horizontally through 3 tplies of plate isn't that bad, 5 or more plies is very difficult.

 

[HeavyCivil]

Lomarandil: I think the test rivet was representative since the heads on all of those rivets are quite large. I am the contractor - I will remove another when we're back to strengthen the floor beam prior to jacking because I agree that a 1" rivet is unusual. AREMA does say 11 ksi which is in the middle of the range of values in the Cambria document shared by SlideRuleEra. I may just call UMSL for a copy of the D+H spec.

CELinOttowa - I'll check out the paper.

Tolchijb- Removing rivets takes skill - as does everything we do that is often discounted as "unskilled labor". You need an expert with a torch and a good helper. You use a gouging tip to burn the head off off the rivet, then pierce the shank itself to get some heat into it. The helper immediately takes his pneumatic hammer (rivet buster) and pounds it out. On a good day my guys can do 200. On an average day they can do 100.

njlutzwe- That is helpful. That and SlideRuleEra's source both reinforce that using the 502 Grade 1 Hand Driven Rivet spec in AREMA is not far off (11ksi vs 10ksi). Moot point here because they're all getting removed and replaced with bolts unless I could confirm a higher grade, but that seams unlikely given the historical sources provided here. In the grand scheme of the reinforcing that the floor beam needs, removing a few dozen rivets isn't that big a deal. And as cool as it would be to find out what they're really worth it's probably easier for me to do than commissioning a metallurgical study.