New AISC Combined Spec 1

[JAE]

I just went to a recent seminar where the new AISC Specificaton was introduced - it is a combined spec where ASD and LRFD are mixed in together so you have one spec and two different design methods within.

I thought it was a good solution overall for AISC as they have had all sorts of issues with many engineers just refusing to convert to LRFD.

A couple of observations on it all:

1. One point brought out was that if all the updates over the last few years had occurred in ASD (i.e. if there never was an LRFD) then the ASD spec would be a lot more complex than the current one. So therefore many of the arguments about ASD being so much simpler aren't quite as accurate as we initially thought.

2. In the ASD - ALL equations have changed.

3. Do you think that some engineers will get the two systems mixed up during design? Will this cause problems, errors, collapses?

4. The speaker asked our room of about 150 to raise their hands if they currently use ASD and then if they currently use LRFD - the result was about 80% ASD and 20% LRFD

 

[bjb]

JAE,
Forgetting seismic for a moment, I wonder how important all the updates to the steel specification really are. Did they mention if there were deficiencies in the ASD spec that could cause unsafe designs even if you followed the spec in every way? Or are they just getting rid of some of the approximations, such as the calculation in ASD for allowable bending stress in unbraced beams. It seems to me that ASD in its currrent form has a sucessful history. I know that in some of the derivations for the formulas in the current ASD spec, approximations were made to make hand calculations easier. I am concerned that todays academics are taking those approximations out, so we can have a "theoretically correct" solution, and that with computers, rigorous equations are not a problem. If so, I strenuously disagree. I read an article where Robert Disque said that while he was still with AISC, he was in favor of the new LRFD spec, but that when he left AISC and started using it in the real world, he didn't like it.
I have found LRFD cumbersome to use for design. For example, designing a crane girder with a cap channel by ASD is relatively easy, but it is a lot harder with LRFD. I am going to order myself another 9th edition ASD manual so I can always have it around, because of the shoddy build quality they used for that book.

 

[JAE]

I personnally have always resisted converting to SI (or metric) due to my familiarity with "english" units and my fear that I would lose the feel of my designs, the gut instinct that tells me there's a goof in my calculations.

However, I bit the bullet some time ago and converted to LRFD in steel. Yes, its a bit more difficult, but I felt that after using ASD for so many years, and the fact that all my young engineers were coming into our office with LRFD on the brain, I figured that I'd better "keep up" and learn LRFD. After all, I had been using ultimate design with ACI for years and liked it.

So many engineers I talk to sound just like me when discussing metric, but it really isn't that hard, it isn't that difficult to use once you get it under your belt. And it is more accurate in regards to level of safety.

But I don't condemn any engineer for wanting to stay with ASD and I don't disagree with your points about ASD providing safe designs for years. The combined spec seems to be a nice compromise.

 

[bjb]

I agree with what you are saying, and I do think we need to resist a knee-jerk reaction against change. I also think that the solution AISC is trying is a good idea. I have used the new spec for cold formed steel which is a combination of load factors and allowable stress, and like the approach so far. In terms of user friendliness, I would be much happier with LRFD if it didn't carry material constants as variables, except maybe in an appendix in the back. For the majority of cases, E=29000ksi and G=11,200ksi, and I wish they would use these values and simplify the equations where possible. It may not sound like a lot, but with increasingly tighter schedules and budgets, the fewer keystrokes on my HP and the less I have to write the better.

 

[RUSSRASCAL]

I USE LFRD WHEN DESIGNING STRUCTURES THAT ARE PROTO TYPICAL FOR MANY LOCATIONS. LFRD APPEARS TO PROVIDE SLIGHTLY LESS COSTLY STRUCTURES. HOWEVER I DO NOT PREFER IT AS A MATTER OF COURSE.

 

[earlofbedford]

When I started to work in the local steel industry (Perú,2000) I realized that ASD method was used by almost all P.E. and fabricators, and the 9th edition of Manual of AISC was the most consulted book. At the university, I learned LRFD method and during the early months as P.E. I had to learn the ASD method. It seemed to me that ASD was easier and, why not, very simple respecting of LRFD. Sometimes I prefered to use LRFD method like "checking" purposes and found that LRFD contains many issues and criterions that ASD doesn't consider. Why?
For industrial buildings and industrial facilities (example:mining and gas facilities, pipe racks,etc), I think that ASD would be a good method for design ( I use it) due to the load and resistance factors for this type of buildings are diferents (and sometimes not well defined) from those used in residential, malls or offices buildings.
What do you think about that?
Regards.
EoB.

 

[JAE]

"and found that LRFD contains many issues and criterions that ASD doesn't consider. Why?"

As I mentioned above in my first post, the AISC has indicated that LRFD has numerous other checks and more complexity because it has been the subject of all the research over the years while ASD has just been left alone. AISC thought, incorrectly, that ASD would be abandoned so no new research and development was done to it.

But with the stubborn engineers out there who want to keep ASD, AISC finally saw the writing on the wall and has now updated ASD to correspond to LRFD in the same spec.

That means that the "simple" ASD that everyone likes will be a bit more complex in the new spec.

 

[RUSSRASCAL]

MORE COMPLEX METHODS MAY MAKE BETTER STRUCTURES, BUT LEAVE MORE ROOM FOR ERROR, HENCE MORE POINTS TO ARGUE IN COURT. ADDITIONALLY THE MORE COMPLEX THE METHOD MAKE THE BETTER THE TEACHING OPPORTUNITIES.

THE REALLY BASIC QUESTION IS DO ANY OF THESE COMPLEXITIES MAKE SAFER STRUCTURES. CAN ANY ONE ANSWER THIS QUESTION.

 

[Taro]

In general, complex methods do not necessarily produce SAFER designs. They can, however, produce more ACCURATE and ECONOMICAL designs. Simple methods require larger factors of safety to cover the simplifying assumptions.

If you are designing a small number of elements or a noncomplex structure, the simple method may be quite appropriate. But if you are mass producing elements or designing a very unusual structure, the complex method could be very useful.

Most codes and specs now provide dual procedures to allow the designer the option. A simple, conservative procedure is allowed with some restrictions on applicability. Alternately, a more complex procedure is allowed without restrictions.

 

[haynewp]

I have always used ASD until recently, I started LRFD for the major components and used ASD for small stuff due to speed. I am not sure it is correct to do this, though I see everybody doing it. Doesn't the IBC for example state either ASD or LRFD to be used on a project? Or does it say you can use either or both?

Did AISC mention mixing of the 2 methods in design, particularly in lateral components design?

I think everyone will be finding the easiest parts of each method and applying them from this combined book.

 

[JAE]

Whoooo, haynewp - I'm not sure about mixing ASD and LRFD in the same project.

I guess as long as you are very VERY consistent in your calcs and careful not to mix things up you could probably use, say, ASD for gravity stuff and LRFD for lateral - or something like that.

But it seems that you would be adding a layer of complexity that would just expose you to a higher probability of errors.... is it really worth it?

 

[TFL]

OUR MONGROL DESIGN PROCESS

1. DESIGN THE WHOLE STRUCTURE IN LRFD USING RAM STEEL INCLUDING WIND AND QUAKE.

2 OUTPUT GRAVITY REACTIONS TO PLAN w/ AN INCREASE OF 1.7 +/- AND ALLOW FABRICATORS TO DESIGN CONNECTINOS IN ASD- BECAUSE THAT IS WHAT THEY ALL WANT.

3. OUTPUT LATERAL BRACE CONNECTION LOADS IN LRFD AND REQUIIRE FABRICATOR TO DESIGN CONNECTION AND SUBMIT CALCS.

4. DESIGN ALL SIMPLE ADDITIONAL ELEMENTS IN ASD, i.e. MECANICAL FRAMES, SINGLE BEAMS, WIND GIRTS ETC.

NO PROFESOOR WOULD EVER AGREE TO IT BUT THAT IS THE WAY IT WORKS IN MY AREA

 

[CSEllc]

The two major difference between tfl's "mongrol design process" and mine is:
1. Depending on the size of the structure I use ASD in RAM
2. The lateral brace forces are shown in ASD if design is ASD or LRFD if design is LRFD. But I allow the connections to be designed using ASD (applying a "phi" factor to the LRFD loads) if fabricator wishes. Signed and sealed calc's a must regardless of what method is used. (Most engineers in my area still prefer ASD....me included... .....that is why I allow the connection design as ASD)

 

[haynewp]

I have seen several other engineers that are mixing ASD and LRFD. If the project is over the estimate in weight and you used ASD for the floor layout, then the first thing my boss would ask me is why didn't you use LRFD for the floor design?

I think the reason why I along with the other people I know are using ASD for the smaller things, is that we feel more comfortable in the results obtained since we more likely performed the calcs correctly using ASD. This is since it is easier and we have experience with it. And there is also the speed of the CURRENT ASD.

 

[broekie]

I haven't had time to go through the whole thing, but in the new steel code, most (if not all) of the formulas are from the LRFD code. If you use the LRFD option in the code, you have load factors and resistance factors. If you use the ASD option, you divide by a safety factor "omega".

In other words if you want to use the ASD option in the new steel code, you're still going to have to learn the LRFD formulas.

 

[JAE]

I agree - the ASD is simpler NOW because it hasn't been updated for a while. With the new combined spec, much of the original simplicity of ASD will be gone.

 

[CSEllc]

I guess wheter or not we want to use LRFD we will be forced into using LRFD.

The "young" engineers out of school are learning only LRFD. If you want them to use ASD you need to show them how.....

In my opinion the days of ASD will be numbered once the combined specis out.

We might as well get on the LRFD bandwagon now instead of being forced onto it later.

 

[14159]

I've designed structures both ways and I can tell you that I can't tell a lick of difference in my speed when I'm using either method. It's just whatever you get used to.

I remember how painful it was to adapt to the IBC, so I sincerely feel sorry for hardcore ASD folks. I think that it would be a lot easier to adapt to LRFD than the IBC.

The LRFD equations are sometimes more complicated but everybody I know owns reputable software that's used for all but the simplest calcs. These simple calcs can be yanked out of AISC tables anyway and it's no harder to compare Z to Zrequired than S to Srequired.

As a counterexample of LRFD difficulty, I don't ever hear ASD folks bragging about the ease of ASD composite beam design versus LRFD!! Not to mention the better results.

In a few short years, 89 ASD will be but a distant memory just like my beloved SBC.

DBD