Structural Analysis of 100 Year Old Structure 10

[pittguy12]

I'm running into an issue on an analysis that I am doing for a 100 year old 4 story building....and I'm wondering what I could be missing that's causing it to blow up on me.

General description of the building:
- 24' x 90' rough dimensions for first two levels. Then upper levels step in on one side and building is 20' x 90'
- Steel Lumber Floor system (basically cold-formed steel back to back channels) with 6" concrete slabs on each floor
- Upper floors span the narrow direction across the building and bear on exterior masonry walls
- Exterior walls are 12" - 16" brick
- Where the building steps in, the exterior wall (and thus half of the floor load from upper floors) is supported by a pair of steel I-Beams which, in turn, frame into 4 18" deep floor girders which also span across the narrow width of the building.
- No obvious signs of structural problems...bowed floors, cracked plaster, etc.

Issue:
I've taken into account the soft steel of the 1920's. I've found good design data on the steel lumber product and deduced it was designed for the typical residential 40 psf live loads plus the weight of the concrete floor. And I've done select demolition in different parts of the building enough that I have a pretty reasonable feel for how the building was framed and how it all works.

The problem is that when I run the analysis of the steel beams which are supporting the stepped in wall, it is blowing up on me. Stress ratios 2-3 times what the allowable was for 1920 steel (16,000 psi according to old literature). So much so, that even just running a dead load case it is still 1-2 times allowable.

I refuse to believe that the building has been standing 100 years and could be this far off. Even if you include the possibility that they ignored live load, it still seems unreasonable to assume they didn't account for the building materials they used! So I believe the issue is on my end.

I also am not eager to be the engineer who cries wolf that the building is unstable when it has stood for 100 years without any visible signs of structural distress.

Anyone with experience in similarly aged structures able to lend some expertise?

 

[r13]

Have you checked the possibility of a higher grade steel?

 

[pittguy12]

Good thought...I considered it. But were such grades available in 1920's? The Carnegie Steel Pocket Reference doesn't make any mention of them.

 

[SlideRuleEra]

But were such grades <structural steel> available in 1920's?

No.

The 16,000 psi is allowable bending stress, used with Allowable Stress Design (ASD).
Are you performing your checks, using 16,000 psi, per modern techniques such as Allowable Strength Design (ASD)?

http://www.SlideRuleEra.net

 

[r13]

I found something from AISC Steelwise, it shows steel grade from 1901 to 1968, however, it wouldn't help you much since the discrepancy is huge. The paper includes a table that shows tensile stress/yield stress/working stress. it indicates from 1923 on, the working stress is 18 ksi, but the working stress is not listed for 1901 to 1922. The paper is linked for interested party use, in case dealing with old steel structures.

Link

 

[kipfoot]

I wouldn't say that old steel was soft. The Carnegie Pocket Companion uses 16 ksi as an allowable bending stress in its examples, but 18 ksi was also common at the time and is what is in the ASTM design specification.

I don't mind using modern design equations on old structures, using the rationalization that the subsequent increase in allowable bending, from 0.5 to 0.66, for example, is based on research and improved understanding rather than on improvement of the fundamental methods of rolled steel production. You could also run coupon tests, but, none of this gets you to showing that a 3x overstressed girder is okay. I don't think that magic steel is your answer.

I think there's more information that we need here. How do the deflections work out for the pieces you're concerned about? There are more 'asking the obvious' kinds of questions like 'are you sure you're using the right girder size' and 'did you subtract the area of the windows from your wall load?' 'did you get the spacing on the girders right?" I'm only throwing these out there because it seems like we're missing something.

 

[SlideRuleEra]

Good article, retired13. The "AISC Iron & Steel Beams 1873 to 1952" is mentioned in it. The 1873-1952 book does not have the useful detail background info included in the article you posted.

I scanned that (public domain) 1873 - 1952 book in 2009 and posted it on my website. Later, AISC borrowed my scan and posted it themselves.
Here is a link to their posting, which downloads faster than the same scan from my site: Link

Interesting thing is that AISC claims "Design Guide 15 expands on this (1873-1952) publication and includes a summary of documents through 2000"... which is true. However, Design Guide 15 edits out much of the specific info such as which manufacturer produced a certain shape and the time period when that manufacturer had that shape in production. I prefer the older book for shapes thru 1952, Design Guide 15 for 1953 and later shapes.

http://www.SlideRuleEra.net

 

[Guest090822]

1920’s steel was mainly A7 with a yield stress of 33,000 psi and an ultimate tensile stress of 60,000 psi.

 

[r13]

Another bold suggestion if every check/effort turns no result, try "composite beam". This is not a joke, the linked paper indicates successful cases since 1902. But by doing so, you need to use ASD method though.

Link

 

[Blackstar123]

I think there's more information that we need here. How do the deflections work out for the pieces you're concerned about? There are more 'asking the obvious' kinds of questions like 'are you sure you're using the right girder size' and 'did you subtract the area of the windows from your wall load?' 'did you get the spacing on the girders right?" I'm only throwing these out there because it seems like we're missing something.
.

I second this. Are you absolutely sure that all the inputs such as material properties, sections, loads, units, unbraced lengths etc are right?
Most of the time wrong input data is the reason for odd results.
I ususally check displacements if I'm not sure that the computer results are right.
Most of the time doing this points me to my error.
Also which software are you using? I've noticed this in sap2000, if the model is giving you very unlikely results, importing the .sdb file again in a fresh model usually correct those inaccuracies.

Euphoria is when you learn something new.

 

[bridgebuster]

Here's a quick sanity check. look at The Carnegie Pocket Companion, what is the published safe load for the beams in question? The building is still standing, someone must have done something right. My other thought is the designer did a grillage analysis assuming some degree of continuity where the two beams frame into the four.

 

[bookowski]

You can have a few pieces of the steel tested for relatively cheap. The contractor nips off a few pieces of flange at the beam ends and you mail them off.

 

[miningman]

I have always suspected that todays computer programs leave a lot to be desired. What factors of Safety are involved here??? The old timers knew a thing or two and today's computer jockeys have lost the ability to run calculations manually , while applying a critical eye on the assumptions.

On more than one occasion during my career, I have been told by consulting geotechnical engineers, " I hope you realise that this excavation , which is obviously stable , is theoretically impossible" Mind you, there were a few that were theoretically stable , which failed catastrophically.

 

[hokie66]

"Cold-formed steel back to back channels"? 100 years ago?

 

[wiktor]

I think that there is something basically wrong in your approach - the first thing being that instead of providing a sketch, or drawing, you are trying to describe the structural system - "Steel Lumber Floor system (basically cold-formed steel back to back channels) with 6" concrete slabs on each floor" which is confusing to start with - what is steel lumber??? Is it steel or timber? It reminds me of some of my clients, which are doing exactly the same - describing the system which they do not fully understand, or computer modelling dudes, which could not solve by hand a simple supported truss - recent Miami Bridge collapse is a classical example.
I'm almost sure that there is nothing wrong with the building structure for the vertical loads - seismic and wind could be an issue, but this will depend on which code you are using.

 

[robyengIT]

one more doc

 

[Settingsun]

Is it the pair of I-beams or the four 18" girders that are failing? I don't think you said.

Did you remember to double the capacity of those I-beams since they're a pair?

The National Pressed Steel Company's 1921 handbook gives 2" as the standard slab thickness for steel lumber floors. Is yours really 6" throughout the building? That's triple the load if not actually the case which is the magnitude of your problem.

If indeed 6", is it doing something special? Maybe composite with steel beams as suggested earlier. Or carrying the upstairs wall to some degree.

 

[r13]

What is the spacing between the 18" floor beams? The pair of I beams only carry the wall weigh and distributed to the floor beams at where they framed into. The I beam shares a little portion of the floor load, the floor beams carry the majority load and the wall weight. Are they failing?

 

[pittguy12]

Lots of great responses....I'll try to address the most pertinent of them.

Am I sure of the sizes and measurements:
Beam sizes were all verified by digital calipers in multiple locations thus I'm confident I have the proper beam sizes. Overall building dimensions were also likewise checked multiple times using a laser distance finder. I'd comfortably say they are within +/- 1". The 18" beams are spaced at 16-6" OC. There are two 10" beams spaced at 6" OC which run directly under the stepped in wall and connect to the 18" girders...thus they span 16'-6".
* Also, floor thickness was confirmed in multiple test holes. Each floor is 6" concrete.

Am I applying modern ASD to working stress:
Yes...while taking into account a yield stress of 33 ksi. I also verified with the Carnegie pocket companion the safe allowable working loads and find a proportional discrepancy so I don't think the error is in my analysis technique. For instance, one of the beams in question is a B10X25.4 from Carnegie book (this is one of the two beams under the stepped in wall). This has an allowable lbs/ft load at 16 ft span of 1,024 lbs/ft...on these beams rests the 30' tall exterior 12" brick wall and 9' of trib floor for the 3rd, 4th, and roof levels. The 3rd and 4th floor each have a 6" concrete floor...so just taking a 40 psf apartment load (actually is correct for the period also) plus a 75 psf floor weight, each of these B10's would be seeing 1,035 lbs/ft. And then I still need to account for where the weight of the roof, snow, and wall load are going.

Magic steel:
I agree, this seems unreasonable. My research shows that some higher strength steel was available in 1920, but mostly used for ship building. Hardly seems reasonably for it to end up in a apartment building in West Virginia

Steel Lumber:
It is actually quite interesting. Steel lumber was developed in the late teens as a replacement for wood floor and wall framing in Massillon, OH. Some research indicates that it was to be the next big thing in construction. Less than a decade after its development though, open web trusses took hold and steel lumber was mostly forgotten. I did manage to find a design reference in the National Archives which is the only thing I could find proving its existence:

https://archive.org/details/NationalPressedSteelCompany/mode/2up.

It is essentially a cold formed steel system.

Is anything failing:
No. Nothing I can see or measure. The beams themselves look good. Plus the building is all plaster...any reasonable movement would show.

Composite beams:
I think this is interesting and I had not considered it. I do have a portion of the floor above the 18" girder open and I don't see studs or coils, but that doesn't mean they are there. Admittedly, I have do not know much about the development of composite beams and whether they were in use in the 1920's. What is interesting though, the 6" concrete floor does seem thick even when comparing some of the details in the steel lumber and Carnegie companion books. Does anyone know if composite beams were used? If so, did they utilize something other than studs or coils to transfer the load?

 

[JLNJ]

I wouldn't grasp onto the possibility of the steel being of higher strength. I'd be looking to my analysis for errors in load path or logic

For example, if you have a beam under a masonry wall, the wall is probably arching and the beam doesn't see but a fraction of the load. The wall might even be holding up the beam.