Bending Capacity and E of Existing Wood Joists in Historic Building 5

[mfstructural]

Hey everyone -
I'm interested in getting people's thoughts on the strength of existing wood joists in an existing building built in 1909. My scope was related to evaluating the roof framing for new RTUs and the 2nd and 3rd floors for some new mechanicals, but we came across an issue where I ended up evaluating the existing joists. The building is 32' wide x 155' long. The joists on the 2nd and 3rd floor span from front to back between 15'9" and 17' (span lengths), between steel beams. The building was renovated in 2007 and the drawings from that renovation list the LL as 100 psf and the joists as 1.75"x13". I performed a check on these joists and they were not working for a Fb of 1,000psi. I informed my client and they were not happy. In fact, they found another engineer to run some numbers and see if it works to verify my opinions. A calc I received from a supposedly licensed engineer uses a Fb of 1750psi for Douglas Fir Larch #1, with a repetitive use factor included. I feel like this is somewhat unconservative, but the drawings from 2008 also listed these joists for 100 psf LL capacity. I wanted to get people's thoughts on using 1500 psi for existing wood rough sawn joists. I've included some photos. They are not happy with me and my findings and I've been asked to sharpen my pencil, but I'm not sure how far I want to go. Reinforcing all the joists would be a serious undertaking of costs and time, but I can't let that dictate my decision.

 

[WesternJeb]

Where is the building roughly located? This will make a difference with the type of wood used in your locale.

 

[mfstructural]

The building is in Chicago. I can get the joists to work using a Fb of 1200psi and use of repetitive member factor.

 

[XR250]

1,200 PSI seems reasonable to me.

 

[Harbringer]

I'm kind of amazed when these questions come up. There's so much information available out on the internet. I personally have pdfs of the UBC going back from 1994 to 1927 where the max for Dense Select DF Coastal Region is 1750 psi so the "supposedly licensed" engineer may be right you will need to determine the grading yourself. However, you don't just blindly apply that allowable to the current code method for designing wood you will need to read the standard from the time and determine what reductions/increases you can take per that standard. Apples to apples.

Also, there are experts out there who can come in and give you a visual grading of in-situ lumber if it's critical.

TLDR do some research.

 

[WesternJeb]

Agree with those above. If you are that skeptical about it, require some NDT.

 

[jerseyshore]

I've talked to AWC directly about this years ago and they said very clearly that if you are checking something to modern standards you should be using modern design values. There are no adjustment factors in the NDS so the current standards must be followed. Unless you are getting the wood tested, you have to use the current code values.

 

[mfstructural]

That's interesting jerseyshore..I've mostly read about going back at determining design values based on the time it was installed. In this case for example, E is 1.8 in 2018 NDS for DFL, but it is 1.6 for select structural DFL in 1927 UBC. I have read through many codes and understand that 1750psi is the allowable for DFL, but this calculation was performed without a site visit and knowing the wood type. In our area DFL was prevalent in the early 1900s, so that's what I'm assuming it is. It's just a matter of choosing a design value. Even for DF, there are several different design values for select structural see table below from 1927 UBC, which I believe is a good representation of the 1909 wood building we have here. I did a lot of research and didn't find much regarding wood strengths in early 1900s. Since this building will be used as a public place (restaurant/night club), I would prefer to be conservative. I may go out to the building and ask the ceiling finish to be remove to inspect the ceiling joists. If I sand a bit off the surface and see the grain is tight and doesn't exhibit a lot of knots, I'd feel comfortable using a select structural. Select structural DFL in 2018 NDS is 1500 psi.

 

[Eng16080]

the joists as 1.75"x13"

A calc I received from a supposedly licensed engineer uses a Fb of 1750psi for Douglas Fir Larch #1

Using NDS 2018, the unadjusted Fb for DFL #1 is 1,000 psi. With a depth of 13", a size factor, CF, of 0.9 should probably be used (although this is perhaps debatable based on how the Table 4A adjustment factors are presented). The repetitive member factor, Cr, increase of 1.15 is applicable if the joists are spaced no greater than 24" o.c. I mention this because the joist spacing often exceeds this for older buildings. I'm guessing you already verified that though. In any case, I don't agree with the number from the other engineer.

I informed my client and they were not happy. In fact, they found another engineer to run some numbers and see if it works to verify my opinions.

I hate when clients pull this. I'm a little surprised they aren't using the other "engineer" for the project. (I've had this happen on far too many projects where they found some other engineer to come up with a simpler solution, often overlooking something that made my design more complex.)

 

[phamENG]

I do a lot of historic buildings, and I do not just use the modern design values (though I do use them as a starting point). A few reasons:

1) We're mostly southern pine around here, and that species group changed about 10 years ago and they downgraded the allowable stress values. Anything put in before 2013, I bump it up a grade and/or use the values from before the change (depends on whether I'm using software with preset materials or doing it by hand - the error is less than 2% if I remember correctly).

2) Applying current design values to existing displays a misunderstanding of where those values come from and why they are what they are. For some good reading on how these are developed, check out these sources: TFEC Tech Bulletin 2018-11, The Wood Handbook (Chapter 7), ASTM D2555, and ASTM D245. These values are conservatively low because wood is a terrible material for structures. I don't really mean that, but I also do. I guess I should clarify by saying it's terrible for engineered structures. It's not clean and tidy like steel. It can be manipulated and formed like concrete. It is a natural and highly variable material. Reference design values are set at the 5th percentile - roughly 95% of all visually graded lumber used will be slightly to significantly stronger than we give it credit for. But until it's used, we don't know. To quote a recent Structure Magazine article:

The National Design Specification for Wood Construction (NDS) is a reliable standard for the structural design of new timber structures but is not a good standard for predicting the actual behavior or adequacy of existing structures. Within timber grade classifications, there is a wide variation in strength properties. The published allowable stress values are calculated based on the weakest 5% of timbers of a given species and grade. Consequently, the published reference design values are very conservative for most of the timbers that are in service.

I've mentioned this before and maybe somebody has reposted it, but I can't ever find it: there's a Canadian code about evaluation of existing buildings that gives guidance for adjusting design values based on time in service. I think I first saw it from KootK. Or maybe jayrod? The idea is that if a structure has been in service for a long period of time and has not failed even though our calcs say it 'should have' it's probably stronger than those values give it credit for. In the commentary for ASCE Chapter 4, they give "Typical Live Load Statistics." If the building is older than the reference period for the class of structure, I feel comfortable saying it can support the previously applied load at least. The Existing Building Code bears this out as it allows you to continue using a structure for the same or even slightly higher loading as long as the structure is not damaged.

3) Trees today are not the same as trees growing 150 years ago. Today, we've exhausted the commercially available stocks of old growth trees. The ancient trees that stood for a couple hundred years, growing through the mini-ice age with super tight growth rings and the sort of density that make it impossible to reproduce a Stradivarius: we cut them down and built stuff. And that wood is pretty amazing. Today, we grow it on tree farms as quickly as possible in warmer conditions that result in less dense wood. The wood itself is the same, but the density is different and the grain spacing is different. The result is lumber that isn't as strong. So depending on the age of the structure, the current reference design values may be far enough off that they don't really apply.

That's not to say wood gets better with age. It doesn't. It typically becomes more brittle and less ductile. It'll be more prone to splitting. And if it is overloaded, it'll be more susceptible to a rupture type failure than a 'younger' piece of wood.

A word of caution: don't just look at older books to get design values, either. In some cases they may not be compatible with our current methods of developing allowable stresses (which are still generally valid, but need minor modification to capture actual behavior). They also may be wrong. For instance, there was a general misunderstanding about tension capacity of wood for a long time. Up until the 50s or 60s we just assumed allowable tension stress would equal flexural tension stress. That didn't change because wood changed, it changed because we (the engineering community) realized we were wrong. So look at them, but also do some research to understand why and how they arrived at those values. Using them may not meet the requirements of the code, but then again it might.

So this is my suggestion:
- Is the load increasing above what it's been for the past 50+ years? No? If it's not damaged, then move on. If it is increasing, proceed as follows:
- Figure out the allowable and reference design stress you need.
- Figure out what species you have. Don't guess at it. Identify it. If you don't know how, hire somebody who does.
- Grade the lumber. If you don't know how, there are certified graders that you can hire to do it for you.
- Check the reference design value in the NDS for the applicable species group - is it what you need? If so, stop and call it a day. If not, continue:
- Go to the ASTM references above and see if they list values for the specific species. If so, work through them to develop a reference design stress for that particular species. Does it work? If so, you're good. If not, then you need to reinforce.

 

[JAE]

I think most of the comments above are fine - don't disagree much at all - but the whole discussion about taking wood joists from 1909 and simply applying 2023 code provisions to it - or for that matter using 1909 code provisions, doesn't take into account the fact that these joists are 114 years old.

Older structural lumber like that typically came from more dense "old growth" timber so the density, and resulting strength, actually gets higher on average than current lumber.

But even so, over time, wood dries out, becomes embrittled, and a level of conservatism is called for.

1. Your idea of visually inspecting the joists is a very good idea - look for splits, irregularities, dry rot, water exposure, etc. to ascertain what possible damages may exist.
2. You could load test a portion of the floor (IBC has provisions for this) This costs time and money but is a good way to get a better understanding of what you have.
3. The for-hire grading inspectors will typically only get you a species, not an Fb.
4. You could conceivably have several joists removed and tested in a lab - In the past I tried to find an ASTM for this but none exist that I found - so you probably need to just test to loads and using your engineering judgement come up with an allowable stress from that (I think the prevailing safety factor for wood is around 2).

Having said all that I'd agree that 1,200 isn't that high to worry about in general with older wood.

 

[JAE]

It typically becomes more brittle and less ductile. It'll be more prone to splitting. And if it is overloaded, it'll be more susceptible to a rupture type failure than a 'younger' piece of wood.

Yes.

 

[lexpatrie]

Can we get a research article backing up this "more brittle" claim?

Chicago, so.... everybody in the evaluation is a licensed structural engineer in Illinois?

American Wood Council had a video presentation on existing structures you may find useful. It's more oriented to wood bridges and exposed wood, but the methods apply, probably.

Provided you look at this and determine it's structural engineering under the Illinois practice act, and have an S.E., then you are dealing with several separate issues -

a ) what was it originally "engineered" for, which I would suggest is it was "engineered" for nothing, it was 1909. It was built and hasn't fallen down yet. Loading history, while it sounds good, is unproven speculation.

b) what could it reasonably be expected to support at the time of original construction (so, rough sawn dimensions (or not), repetitive member factor, allowable stresses from that time and whatever code provisions you can find), I wouldn't make any increases in strength based on "successful service" as, unless it's some former warehouse building, it could have been very lightly loaded. I think there's a potential for higher bending stresses based on closeness of grain, and it's maybe old DFL wood, but I can't definitively say if the DFL values from say 1929 are larger than current (because there are factors here that changed over time, or were added, the code for wood has evolved since 1929.) My impression is they would be stronger on the whole than current production grade wood, but that needs to be documented. Not by testing, perhaps, but at least in some referenced design materials.

c) (the real issue) what decay or modifications have taken place that have compromised it (infilled deleted stair openings, joists that don't bear on walls, cut joists, "fire cuts", splits, sections missing, nailed "splices" between pieces, large knots, cut out holes, etc. What kind of site work did the last guy do, before they vouched for the floor at 100 psf? Site unseen? Ceilings left in place? Depth measurements taken at a stair and inferred based on 1" plank and lath and plaster ceiling? This is a very old building. What are the in place dead loads for that matter.

d) What is it going to resist now, per "current standards." And in the mix is the whole standard of care question, what is the current use compared to that 100 psf it was given a few years ago.

e) this guy diagnosing bending stress on paper is a problem. This is not the standard of care.

If you find it's structural engineering under the practice act and you don't have an S.E. you walk away. If others giving "advice" aren't S.E.s in that scenario, you point out the issue with the "advice" others are giving, and the owner can file a complaint about their work to the board.

 

[JAE]

lexpatrie - I'm an IL SE but I don't think posting advice on Eng-Tips violates the IL statutes.

But in any case, very good list you posted there. Appreciate your insight.

As to old wood embrittlement, I'd have to go dig a bit but my past work on older structures, and wood framing for the US National Park Service, provided that information years ago.
I'll see what I can find.

 

[phamENG]

Can we get a research article backing up this "more brittle" claim?

Here's one doctoral thesis: Link

Most of the strength properties of wood seem to change rather slowly with aging, or not at all. However, fracture behaviour of aged wood was often found to be different to that of recent wood. Wood might become more brittle with aging as mentioned by Attar-Hassan (1976) and Kawai et al. (2008). Many authors report a decrease in the absorbed energy in the impact bending test for aged wood (Kohara and Okamoto 1955; Schulz et al. 1984; Weimar 2000; Lang 2004; Baron 2009; Yokoyama et al. 2009). Samples investigated in these studies consist mainly of softwoods, the only exception being the study of Kohara and Okamoto (1955) that also deals with the hardwood Keyaki. All the wood samples investigated served in constructions for 300-1600 years. Inspections of the fracture surfaces reveal a higher proportion of brittle fractures for aged wood than for recent wood (Weimar 2000; Lang 2004) and more uneven and complex surfaces on the microscopic level (Ando et al. 2006). Breaking strain values determined in bending were found to decrease in Hinoki specimens aged up to 1580 years (Yokoyama et al. 2009).

This one indicates that in a test of ancient Chinese structures, they found a loss of cellulose and hemicellulose while the lignin remains. In the Wood Handbook, we read that:

In a general sense, cellulose can be understood as a long string-like molecule with high tensile strength; microfibrils are collections of cellulose molecules into even longer, stronger thread-like macromolecules. Lignin is a brittle matrix material.

 

[bookowski]

Data point fwiw - 1899 nyc building code, "Safe extreme fibre stress in bending" - Yellow pine = 1,200, oak = 1,000, spruce = 800

 

[JAE]

Thanks, phamENG!

 

[lexpatrie]

I didn't mean the folks here I meant the OP and the "alleged engineer" on the actual project, and the prior guy who vouched for the 100 psf live load.

Ok so the wood is more brittle versus impact loads, or are we making a leap to lower tensile strength/bending strength from that?

 

[JAE]

I think I would be a little more conservative with a brittle element vs. a ductile element - same as ACI using 0.9 for flexure (usually a ductile failure) and 0.85 for shear (brittle, abrupt failure).

 

[jerseyshore]

Just to follow up regarding what I said above, here was the quote I got back from AWC on this topic after they discussed it during a webinar:

"When checking something to modern standards, you should use the modern design values. The standards and design values are developed in conjunction with each other, and changes over the years have impacted both. The new values do reflect that there is more juvenile wood in the supply chain than in the past, as well as the refinement of various test standards, amongst other things that have impacted design values over the years. There’s really not one particular thing you can point to as the main driver of the changes in design values. I don’t know that there’d be a consistent adjustment factor, and there isn’t one in the NDS. In general it’s best to follow the standard whenever applicable."

For me this was the end of using the old Fb of 1250 psi for the older DF#2 we were running calcs on (most common lumber around here). Current NDS says use 950, I use 950. Unless I get it checked or tested (which is ultra-rare), I'm not going against AWC. They write the code, not me. And if something happens I don't want to be the guy going against the code.