ASD vs. LRFD, Who Decides? 5

[vmirat]

I just bought the fifth edition of Salmon/Johnson/Malhas "Steel Structures - Design and Behavior." This text book is focused on the LRFD method. In the preface, they state, "This modern philosophy of design [LRFD], discussed only briefly in one section of the second edition, is moving toward being the predominant approach to design."

There have been many posts to this forum about ASD vs. LRFD. Several people have commented in this forum that they were taught LRFD in school but that their engineering firm uses ASD. It seems apparent that academia has made the decision that LRFD is the best method and therefore the only method to be used. The question is whether academia should be driving this train or the industry? There is a form of social engineering (pardon the pun) going on here if academia is making the decision for us as opposed to the code committees.

Some have commented that, as far as the 13th edition of the steel manual goes, the results are pretty much the same regardless of which method you use. If so, then why the push for one method over the other in the academic world? Since both methods are related to code compliance, perhaps academia shouldn't teach either one and leave the code issues to on-the-job training or offer a separate course that covers all the code aspects.

My own personal desire is for one method for all materials: steel, concrete, wood, masonry, etc. I don't really care which one it is, as long as it's consistent. But, in my opinion, the decision as to what system is used should be made by the industry, not the teachers.

 

[nutte]

AISC, an industry group, began the push for LRFD years ago. It is believed that structures designed with LRFD would be lighter than those designed with ASD, and as such steel would have an advantage over other materials. The practitioners (engineers) have not all followed AISC's lead. I'm glad that AISC recognized this and put out a unified specification, the one found in the 13th edition manual. Now you can use either way without having to rely on 20-year old technology.

I think using solely one method for all materials isn't likely. There are benefits to each method, and some materials are better suited for one than the other.

 

[MiketheEngineer]

Your local building code will dictate which you use.

There are so many of us "old" engineers - I am 56 - that never learned LRFD - that we just stick with ASD - easier to use and understand.

I have "young" engineers that have trouble with LRFD - and they just got out of school!! Parts of it just don't make sense.....

 

[cvg]

AASHTO is also pushing LRFD and I heard ACPA is moving towards LRFD for concrete pipe design also.

 

[hokie66]

vmirat,

Your post sounded very familiar to me. I wrote a similar letter about 20 years ago to the Institution of Engineers Australia about the then proposed AS4100, Steel Structure Code. No answer was received. Academics have won.

 

[vmirat]

hokie66,

So this is a world-wide phenomenon? I wonder if this translates to other fields, like the legal profession.

By the way, I don't suppose you're a VPI grad (hokie reference)?

 

[hokie66]

Yes, it is worldwide. I think the US is resisting a bit, like with the metric system. I am a Hokie, class of 1966. There are a few of us hanging around on this site. You?

 

[vmirat]

VMI class of 80 (hence the vmirat moniker).

My steel design course was just a brief overview of the process of using steel. Although we were required to buy the AISC steel manual (blue edition), we had no clue what we were doing with it. Our instructor told us what to use out of it. Of course, that was all ASD, so it wasn't an issue at the time. But my point is, our instructor really didn't focus on the code aspects of steel design. We could have done without it. Heck, I didn't even know how to read a blueprint until I starting working, let alone understand the intracacies of codes.

 

[JoshPlumSE]

Of course young engineers have trouble with LRFD.... They have trouble with everything! That and they don't have engineers with years of experience in LRFD to explain to them why they are having trouble.

For what it's worth, I have always used both. I learned LRFD in school. Worked under other engineers using mostly ASD, but some LRFD. Then when I became a lead, I ran my projects mostly in LRFD.

Also, I'm not sure the push for LRFD was really related to cost as the two methods are relatively comparable in total project cost... at least for the industrial work that I did. Instead, the basic argument that I heard was related to a more uniform factor of safety based on probabilty theory of load and failure and such. Since you usually have a better idea of what your expected dead load is, then you can use a lower factor of safety for dead load.... et cetera. If brittle failure is dangerous then put a larger factor of safety on that type of failure.

That's the reason why I switched.... Though I have to admit that it is really annoying to have to carry one set of load combinations for deflection and another one for design. That is the best argument I've heard for why ASD is better.

 

[hokie66]

vmirat,

I must be getting slow in my old age not to pick the meaning of your handle. After all, I am only a Hokie. I was around in the days when the VPI-VMI game was the Thanksgiving Day classic. We got in a lot of trouble one year by kidnapping a couple of Keydet rats before the game.

 

[DataAB]

To date I have yet to have anyone explain to me clearly what the advantage is of using Limit States Design (or LRFD) over ASD.

I would suggest that the same can be said for metric versus imperial units.

 

[271828]

I feel your pain because I remember how much gnashing of teeth I did back when I had to lose my SBC and start using the IBC2000. I cussed at it for probably 2 years and my errors in applying it actually did cause problems on at least one job.

As for ASD89 vs LRFD, I really don't see the point, though. It is HIGHLY debatable that the 89 Manual is better organized and easier to use than any of the LRFD or the unified Manual. Earlier manuals were even worse.

For example, depending on web slenderness, one could end up all over the place in the 89 Spec. trying to compute the allowable shear and moment. In the 2005 Spec., these areas are extremely well organized in Ch. F and G. Same for columns. Depending on slenderness, one can end up digging through appendices for Q factors and such. Ch. E is very well organized and has everything one could need for a compression member.

There's also the fact that the newer specifications have provisions for so many more cases than did the 89 Spec. For example, if you have an I-shape with unequal flanges, you end up beating a Ch. F equation to fit the situation even though it wasn't derived for that case. The 2005 Spec. has Section F4 which is directly applicable and leaves little to wonder about.

To me, the biggest enhancement going from ASD89 to LRFD is the way equations are presented. The old ASD equations embed so many different constants into one constant that it is impossible to know what's really going on in the equation. The ASD89 Ch. F Fb's are great examples of this. Check out the Salmon & Johnson derivation for those constants--there are pages of derivations IIRC! The designer has no prayer of looking at those equations and having a clue where they came from. In the LRFD specifications, this is far less of a problem because the variables are still broken out in most cases. THe downside is that equations are longer, but ever designer has a computer nowadays anyway.

As for whether academia or industry drives this stuff, I'd say it's both, but admit that academia plays a big role. (Full disclosure: I'm a professor who teaches the steel classes at a university, but I designed buildings for about 10 years.)

Like it or not, universities force professors to crank out research, and a lot of it, or they're fired in the form of not getting tenure. They also don't get paid for 3 months out of the year if they don't bring in enough to fund summer pay. I'm actually more of a teacher than researcher, so if I had it my way I would just teach and then do research only when there was an actual problem that could be of immediate practical use to industry. Unfortunately, I would not get tenure if I did this.

If any of you guys don't like the way professors have to crank out papers and more research, then I'd suggest voicing your opinions to the universities. The university would point out that their tax dollar budgets are shrinking and that money doesn't grown on trees, though.

 

[vmirat]

hokie66,

I think we both had our share of cadet (or Keydet) antics!

Being a '66 grad, I'm assuming you've had a considerable amount of engineering experience. What was your educational experience like, with respect to steel design? Was it mostly theoretical or did they cover code-related issues such as safety factors (I'm not sure what the steel code was in '66)? How would you compare your educational experience with what their teaching today? Specifically, is it necessary to teach a code system?

 

[vmirat]

271828,
Thanks for your candor. I appreciate hearing from the academic side on this. I would imagine that there are many professors on code committees, so I can see how this can become a circular argument. What latitude do you have in what/how you teach steel design? Would you be run out on a rail if you taught ASD or even no code system at all?

 

[271828]

Yes, I'd say there are a fair number of professors on the Spec. Committee.

Lower level classes have very specific topics that must be taught. My Steel I class is senior level, so I have a lot more leeway. I'm sure the industry advisors would complain if their students came out of school knowing ASD89 instead of LRFD/ASD2005, but I could choose to teach ASD2005. However, I probably wouldn't because I suspect that the ASD part of the Spec. will evaporate over the next 2-3 manuals so there's no sense getting used to omega instead of phi. I wouldn't be doing anybody any favors by teaching them something that's likely to become vapor.

As for teaching "codelessly"(? LOL), I'm sure that the industry advisors would complain about that because their new guys would be even less useful (LOL again) than they are now. The other constraint is the FE exam. Our guys have to be able to answer the steel questions.

In the advanced course, I have complete freedom to do whatever seems most beneficial to the students, simply because this course isn't a prereq for any other class. I try to focus on behavior because those topics are timeless. For example, say I teach a prescriptive part of hte Spec., those provisions might change, making that part of the class worthless. (Other profs teach enough worthless stuff!) If I teach with an emphasis on behavior, then less of that will happen to my guys throughout their careers. In a way, this class is closer to being "codeless" than the first class. I try to teach both as best that I can, though.

This reminds me of anther common industry question: why don't we teach them how to read drawings, develop sections, etc.? I used to ask these questions. The reason is that there is so much variability from office to office that a class like that would likely be considered worthless to many engineers. Also, the way things are done changes all the time. Take BIM for example. Teaching guys to put drawings together might be worthless in 10 years.

Hey, hello from another Hokie! I have 2 pals from VMI. I really respect the toughness of anybody who survived their engineering program along with the rest of the VMI experience. I never would've made it. VT was hard enough without the military activities.

 

[hokie66]

vmirat and 271828,

Nice to have a conversation with fellow Virginians, at least from a uni standpoint. To answer vmirat, my steel design courses at VaTech were woeful in hindsight. I had a steep learning curve on the job. I hope and trust that they do a much better job today. It's just the luck of the draw as to which professor you get. On the other hand, my other structural courses were good, and I had Prof Richard Barker for concrete, and he was tops. It is inevitable that Codes will be included in coursework, but I think that there is too much emphasis on the current code, as it will always change.

By the way, when I was there, the Corps at Tech was bigger than the VMI Corps, and compulsory for 2 years, so this country boy had a big shock the first few weeks, both with the Corps and with the coursework.

Vmirat, what type work do you do? The stereotype of a VMI engineer used to be that they either worked for the Army or the Virginia Department of Highways.

 

[271828]

In 1986, the VT steel design curriculum changed a lot when Dr. Murray came there. Last time I checked, there were 3 main steel design courses, a course in cold-formed steel, and an entire course in connection design. He just retired, so who knows what'll happen with those courses as time goes on.

I also had Dr. Barker years ago. I thought he did a good job. I bet you had Don Garst too, right?

 

[hokie66]

Don was more of a contemporary. Bob Heterick taught some of the general structural courses, then he went to head up the computing center, I think. I understand he recently passed on. Yes, Dr Murray certainly helped the VT reputation.

 

[kslee1000]

Unlike reinforced concrete design, which is largely developed under experiment and controlled testing, the elastic/plastic theories/behaviors of metals were well estabilished and verified for years. I am not familiar with LRFD, but judged from USD for RC, I couldn't resist to post the question: would those factors causing us to lose sight on foundamentals, especially when we need to link the real world phenomenons with the theories behind. If it is the case, we will be much more practical, efficient, with a lesser brain. However, if the change (ASD-LRFD) is similar to the change from slide-rule to hand-held calculator, then what is the difference.

 

[271828]

kslee1000, I don't understand where you're going with your post.

"would those factors causing us to lose sight on foundamentals, especially when we need to link the real world phenomenons with the theories behind."

You're not trying to say that the newer steel specifications make it easier to lose sight of the fundamentals, are you? If so, then I disagree--see the 5th paragraph of my first post. I think the older specifications did a good job at hiding the physical phenomena behind the equations.