Inconsistency in NDS 1

[JoshPlumSE]

All -

I'm looking at the 2005 version of the NDS. One of the main changes with this version was the tabulation of values of Emin that are to be used for beam and column stabilty calculations (CP and CL).

Now, they give some background that discusses how this factor is calculated. That's in appendix D. In this appendix they give the following equation:

Emin = E*[1-1.645*COV]*1.03 / 1.66

They then go on to say that the 1.03 factor is an adjustment factor to convert E values to a pure bending basis except that the factor should be 1.05 for structural glued laminated timber.

Follow me so far? Well, if you look at the tabulated values for Emin for the glu-lam tables, they appear to be eschewing the 1.05 and using the 1.03 instead.

My belief is that it is an oversight. Perhaps the code committee probably approved the appendix equation, but the tables were put together by a group that didn't understand that sublety at all. Any thoughts on this?

In the end this probably ends up being an academic question (we're talking less than a 2% difference). But, it does raise some follow up questions:

Which value is the more "accurate" value?

Which value is the "legally correct" value to use in design?... meaning which has more legal weight, the NDS tables or the NDS appendix?

Josh

 

[RFreund]

I actually posed a very similar question to AWC a short time back and they directed me to the appendix but I'm not sure if I followed up for the clarification that you are now asking about. I'll check to see if I did get it answered. I'm sure if the question is posed to AWC they would get it answered. They've been very helpful in the past.

 

[JoshPlumSE]

FYI: I got a response from AWC.

The Emin values in the glu-lam tables do use the correct equation (using the 1.05 value). But, they assumed a COV of 0.11 rather than 0.10. This is necessary anytime a Glu-Lam has only 4 or 5 laminations.... Not sure when that would be. But, they acknowledge that it ends up being overly conservative for cases where there are 6 or more laminations.

 

[3doorsdwn]

What I can't figure is the mentality behind the Emin approach (i.e. what warrants such a lower value?). When you look at the values you get when calculating stability factors (i.e. CL) when using Emin vs. using E (just a code or two ago): it's a whole lot less. (In fact, for a beam calculation today I came out with 0.42 for CL; and using E it would have been 0.74.) I don't know if they are adjusting values elsewhere to compensate.......but I just can't figure why they keep changing stuff around when it worked for years.

 

[Cadair]

The difference between E and Emin is a direct result of their uses. E is used to find deflection, which is a serviceability requirement, so E is an average value of the species without a factor of safety. Emin is for stability, so it is a life safety issue. Suddenly the variability of the design values is an issue and a factor of safety is needed. That is why in the 2005 NDS in Section C4.2.4 it states, “Emin represent an approximate 5 percent lower exclusion value on pure bending modulus, plus a 1.66 factor of safety.”

 

[JoshPlumSE]

I'll have to pull up an old NDS to know for sure. But, I thought the Cp and CL equations used to be different back when they were using the regular E value rather than Emin.

Essentially, I believe Emin was in the old buckling equations, but in a more round-about way....

Anyone else what to step up to the plate with a more complete response? I'm just too lazy to look this up this late on a Friday afternoon.

 

[JoshPlumSE]

Okay, I went ahead and looked it up. The 1991 NDS had the concept of KbE and KcE used in the CL and CP calculations.

The old code' appendix H has some discussion about how this value includes a 2.74 reduction from the tabulated modulus of elasticity.


A quick check for sawn lumber shows that the CP value for both of the codes should be essentially the same as the newer codes.

 

[nutte]

JoshPlum, I think you're right, but I'm also not willing to investigate further right now.

The other day I ran a column by hand, using Emin, and got an allowable stress of say 365 psi. I grabbed an old old spreadsheet (89 NDS maybe) and ran the same column to the old code, no Emin, and got say 362 psi. Same answer.

 

[Cadair]

The newer code and older code should give you the same result. I've included an article from the American Wood Council (AWC) that explains the change. It is on page 4.

To skip to the conclusion, the change has to do with combining ASD and LRFD into a single document.

"The value...is algebraically equivalent... Because the design equation for KbE includes a reduction for safety, two different formats of the 2001 NDS equation would be needed to address both ASD and LRFD. Instead, the 2005 NDS utilizes Emin, which is adjusted for safety, so the safety factor is not part of the basic design equation."

In other words, the reduction for life safety versus serviceability has always been there. It just wasn't as apparent.