NDS vs. EuroCode

NDS is easier for engineers without much understanding of equations or theories. This is good to certain extend, as they can look up values in tables, but can also be dangerous sometimes if they don't what they are doing or they think they know all. 

Right, NDS is more prescriptive. Eurocode is more flexible if you know what you are doing...

I personally found NDS a better "entry drug" for undergrad students... 😆 

@ling - It would be interesting to do a lit review to see why the NDS code is the way it is.  Back before my time in the 1990s I understand there was a concerted effort to develop LRFD for wood.  Doug Rammer could give you a good history.  ASCE 16-95 "Standard for Load and Resistance Factor Design for Engineered Wood Construction" is published.   But it never was adopted as Code.  Why?   I don't know for sure.   

SPDWS has been updated as a "nominal strength" basis from which ASD and LRFD designs are based on.    NDS does not do this -- as you know.  Still ASD reference Design Value.

One problem is that all of the ICC Acceptance Criteria for fasteners and beam hangers are ASD reference design value based.   No point in trying to fix NDS if the ICC Acceptance Criterias are not also updated b/c the large majority of connection designs in Timber use proprietary hardware, and not designs out of NDS.  This would probably cost millions of dollars to all of the current wood connectors companies with ICC ES reports.  I suspect this is apart of the reason that current wood industry is not motivated.

Are other international codes for wood real a true "LRFD" with robust reliability basis to them?  Or are they only partially so where the reference nominal strengths of members and connections till have a good bit of judgement by the Code Committees that write them.  I suspect this is so.  So maybe all codes are sort of an LRFD "format conversion" approach in reality, just not presented as such.

That is my 2 cents.  

HEB

@heblomgren Great! another research/paper idea that I could not find anyone to fund 😆  

I know exactly who I should pull in to make a comment...  I will try to find the AWC folks to provide some input.

In the meantime, My take on LRFD vs ASD

I kind of agree with you the ASD stays main stream for U.S. code due to historical built-up that shaped our industry (connector companies you talked about @spryor , as well as design offices). There is just not enough incentive to force the change. The current LRFD with conversion factor is not really a full LRFD.

However, there must be some benefit. I remember talking to Phil Line about teaching just ASD in my classes, but he encouraged me to do LRFD. So I think that is where AWC wants to go....

Now to Eurocode, I do believe they are not just a conversion but really want to do reliability based LRFD (which they call Partial Factor method). Since there is an overarching EN1990 that dictates design values (characteristic values) as percentiles (either 5% or 95% depending on if it is favorable) from manufacturers. Then there are testing standards which I am not familiar with, but should in principle adhere to EN1990's reliability framework. 

In the end the fundamental question is what is the benefit of a full LRFD on wood? is it worth the cost?

One benefit is that if we are doing more and more concrete-wood hybrid, you have to do LRFD on both systems to be consistent. Maybe that is something AWC and FPL should think about

Let me see if I can pull more folks into this to share their thoughts.

@ling, yes, I was around when then AF&PA rolled out the new LRFD for wood.  It's been awhile, but if I recall correctly, to answer @heblomgren, there was a very concerted effort by Dave Gromala and others to make it a "true" LRFD design.  In fact, doing a quick internet search turned up this article (I have not read it all) that the reader may be interested in: https://awc.org/wp-content/uploads/2021/12/WDF-2006-RecalibrateLRFD-0604.pdf.   I also seem to recall that early implementation of the LRFD code led to vibration concerns in floors, and when that was addressed the design then resembled an ASD design anyway.  I think there was considerable "if it isn't broke, don't fix it" thinking amongst practicing engineers (included me at the time).  While it didn't catch on for the mainstream, whenever I've designed connections for large seismic research projects my go-to approach is use MCE level demand resisted by LRFD capacity, and it's never let me down (yet!).

thanks @spryor , I need to pull Dave into this:) 😆 I never know this part of his work but considering he learnt from Bruce Ellingwood, it is not surprising. The paper is relatively short but I can see the basic road map for calibration there with the duration factor... The mean to LRFD strength ratio is also quite informative. Interesting 🤔 

@ling and @spryor.  Thanks for adding to this thread.  I'd value getting a better understanding of the past and help clear the air on what is still out there as unknown with regard to the NDS becoming a "true" LFRD provision vs. ASD reference design value -> LRFD Format Conversion approach.  I'm currently still just speculating.  Here are 2 papers I've found that point to inconsistencies in the connection design approach and reliability factors.   I don't fully understand these papers or the background for what motivated them being authored.  But it does appear to point to a gap in knowledge.   Also - I still believe b/c ICC AC 233 is not a reliability based test standard and is firmly rooted in test methods that only establish an ASD reference design values that Code Alt fasteners would all have to be tested to using a revised version of AC 233 that is reliability based that confirms thing like Beta factors.  I don't think industry is on board with doing this leg work.  (i.e. time and money for what real benefit).  Does this make some sense to you?  Thanks again for the discussion.

@heblomgren Interesting papers. I know Bill who authored the first one and he's already retired, but I will try to see if he wants to comment. I guess for connections it is even more complicated. My comments are only my opinion, which could be wrong, but here you go:

I dont think current NDS connector design capacities are intentionally calibrated to a certain reliability index. They might fall within a reasonable range of indices, but it is not a back-calibration.

The use of Yielding Model equations meant that the connection capacity is determined by several parameters such as wood bearing strength, dowel yielding strength, geometry, etc. Some reference strengths were included for typical species and those might be a mean value or a certain percentile value, when you combine those in the equations, the reliability indices will be determined, but it is hard to control that reliability index as most of those are nonlinear functions. 

I think this is likely the most logical approach given that there is no apparent incentive to try to unify reliability index for all these connections. In the future, this might change.

I guess a simple question is: what is the reliability index for all the connection strength value listed in NDS now? is there an answer for this question? or do we really need this answer practically?

I have attached a paper that addresses many of the posts.

Some further comments. 

Back in the 80s, the decision to go to LRFD was a middle of the road choice.  ASD was too weak to be reliability-based because all variability in the design parameters (e.g., loads, dimensions, and material properties) was mushed (technical term) into one variable:  the factor of safety.  The other possibility was a partial-factor code that had factors on all design parameters.  LRFD has factors on loads and capacities with such things as dimension variability implicit in in the load or capacity factors.  This seemed like a choice that might prove acceptable.

The present NDS has evolved a lot since 1995 when the first and only attempt at a wood LRFD code was released:  ASCE 16-95. 16-95 was a first iteration of a reliability-based engineered wood design code.  It was certainly different than the NDS of that time.  I don't know all the reasons for its rejection, but apparent ease of use was a big part.  It always seemed to me that its rejection was very shortsighted.  16-95 was a much better place to begin the evolution of wood design post 1995 than was the NDS.  With wood now being used in a wide range of applications, including for fairly tall buildings, we should have a modern code that is reliability-based and allows more flexibility in both design and basis for future evolution.  The NDS and it's weak form of LRFD is not it.

My two or three cents.  The paper has more.

Thanks @wbulleit . It is the first time I heard of ASCE 16-95, I never knew this was attempted! Now I am intrigued about why it did not go through... for pure curiosity:) As @heblomgren indicated, the industry is not currently on board to change the format due to the lack direct benefit. Maybe it will happen when more hybrid structures and systems become main-stream and a true LRFD will be proven to save costs.

I like your term "weak form"... It is a good analogy, if you cannot solve a differential equation exactly, you can use Weak Form to approximate the solution. In fact most of the practical applications are based on "Weak form", which is not the real thing but the best next option.... 

@wbulleit Thank you for the post and the paper. (and thank you @ling for the introduction) I read it with much delight.   Engineers can always be wrong, as long as they are conservative.   😃

I've been on the AWC Wood Design Standards Committee for about 6 years now.  There are many needs for mass timber and Code improvement we could discuss.  Thank you for helping fill in the history of NDS and the LRFD incorporation.  That history is very hard to find as 30yrs have now passed.  -Hans-Erik