Limit State Vs Working Stress Design 5

[civeng80]

Im from Australia and did my course close to 30 years ago when the steel codes were still in wsd. Now we universally use Limit state design. The codes are more sophisticated e.g wind codes and the steel design codes. I did a job not long ago and designed by both the new code and the old steel code. Member sizes came to be the same. Differences were that computer printouts amounted to about 60 pages.

The old code required about 6 pages of hand written computations.

Now it cannot be designed by hand because of the new code complexities.

Why do the code committees make life so hard?

What was wrong with working stess design? Didn't it serve us well over many years?

Id love to hear comments, especially from the older engineers who are proficent with both design philosophies.

 

[patswfc]

i'm a young engineer and have never used wsd. From what i hear from older colleagues they loved the old wsd codes. I think that they are fine to use on traditional type structures, however when you have to design slender and lightweight structures i believe that you need to use the limit state codes and computers to account for the likes of 2nd order effects, etc.

 

[civeng80]

Well Im just wondering if it was all worth changing design philosophy when working stress was working so well.

Any other comments?

 

[youngstructural]

I worked with a very senior (in his late 70s, graduated early fifties) who was very well respected... He maintained that the only reason codes changed was for academics to get their name in the code. I'm pretty sure he was being at least a little tongue in cheek, but he was pretty all fired-up about it.

Personally I'm all for advancement of the state of the art, but just make sure that the simpler methods are still available for the regular every day jobs! And FORCING us to use computers is truly demeaning; Sometimes it takes all the fun right out of analysis and design...

Cheers,

YS

B.Eng (Carleton)
Working in New Zealand, thinking of my snow covered home...

 

[civeng80]

Dont get me wrong I think advancement is always a good thing, but a complete turnaround of design philosophy when it worked so well for so many years really doesn't make sense to me.

I sometimes do the old stress analysis because the maximum permissible values are still in the back of my mind. Our code does not permit this. Why not?

As for computers in my opinion they are fantastic. But we engineers I feel are becoming to dependent on them because we have to. But lets not forget that with computers its rubbish in = rubbish out, and just because we get 60 pages of printout doesn't mean its any better than the 8 pages of manual computations.

 

[JStephen]

I think you're confusing two different things. They have complicated the codes considerably, and they have gone to limit-states design, but those are not the same thing. There has been a lot of refining and improvement in the codes, but had they kept them as allowable stress codes, most of those changes would have still been there, only with some changes in the numerical factors.

A simple example- how about a flat bar turned on edge? The old AISC was didn't include that load case, there wasn't an allowable stress given for it; the new one does. So yes, that made the old code simpler, but didn't make it better.

If you look at the ASME Boiler and Pressure Vessel Code, for example, it was allowable-stres 30 years ago, and still is today, but the size has increased considerably.

 

[civeng80]

There were many textbooks that were with working stress design and these cated for all types of structures. Of course they went into considerable depth and rightly so.

My point was why did the powers that be change design philosophy?

It worked so why change it?

 

[WillisV]

Putting large stone blocks together to form pyramids worked pretty well too.

Advances in the codes are usually to address areas of the code that were not well defined or in some cases could be refined further to gain more capacity etc. Also codes change in response to failures (e.g. Northridge earthquake). Keep in mind that the codes must cover ALL cases - from the typical 1 story strip malls to 80 story buildings, stadiums, and everything in between. The trick is to try to maintain a balance between having the complexity and nuances to properly design a difficult structure while still providing the basics and "simplified" shortcuts to handle the general cases.

 

[dozer]

While I do appreciate the addition of things that weren't previously covered, I do think the steel code (AISC 13th ed. specifically) is becoming too complicated.

Now this suggestion might make it even more complicated but it seems to me there should still be a "simplified" way to do hand calcs and a more complex (I hope more accurate) method for computer generated calcs.

I really like the idea of elimating the K factor in slenderness ratios and running a second order analysis. That makes a lot of sense to me. But what about just doing a single beam with transverse loads and axial compression by hand? I did this the other day and maybe I overthought it but as far as I understood the code I had to do something to account for the second order effect. Notional loads, which seemed like a huge hassle to me or in my case I got lucky and it was like the benchmark test in the commentary, which gives the equation for the magnafied moment. Oh, or I could have modeled it and ran a P-delta (large and small, well really just small in this case).

Anyway my point is this very simple problem became unnecessarlily complicated.

Another problem I have with the new code (and this does tend to be how LRFD approaches design) is the use of forces and moments to indicate the strength rather than stress. Yes, I know the code says you can convert it to stress but they must not think that's such a great idea since they don't write the code that way. So what's my problem? You have no feeling if the answer makes sense or not. If I get 24 ksi allowable bending stress for a A992 beam that's a little longer than Lu I'm going to think, yeah that seems reasonable. If I get 1130 kip-in, guess what I'm going to have to do to see if it makes sense? Yep, divide by the section modulus. Why not just compute the allowable stress in the first place.

Here's my real fear, as these codes get more complicated, engineers are going to loose the "feel" of what makes sense and what doesn't. In my example let's say you slipped a digit and 1130 kip-in results in ,oh say, 62 ksi stress. But an inexperienced engineer is running the calcs and it never occurs to him that this answer may not make sense. He just accepts it and moves on. Or even an experienced engineer who is feeling the constant pressure of going faster and faster, and he just blows and goes.

 

[Dinosaur]

As I understand it, there are several things at work here pushing us closer and closer to the envelope.

First we designed by WSD and the codes looked a lot like a solid mechanics textbook. But engineers understood there were many vague approximations in the code beginning with the problem definition that said for this kind of load we will use 100psf, and for that we will use 50psf, and for this other thing we will use 75 psf, ... These loads were only 'significant' to two digits, and engineers understood that and were OK with it.

But there were folks that wanted to be more precise and 'invade' the 10% space between specific and code allowable loads. Engineers are just like that. They want to make that W36x135 work even though their quick and dirty calcs said the stress was 2% over allowable. What is 2% when the loads themselves are so vague?

But the biggest controversy that started it all was the fact that WSD applied the same factor of safety against dead load as it did against live load. This did not make sense. Eventually, a rational method was developed to address this and it became Ultimate Strength Design, USD. USD allowed a rational method for engineers to design longer spans and taller buildings.

This trend set off a new dynamic. Codes were developed by the different industries. If the concrete industry made it possible for structural engineers to accomplish greated economy in concrete construction, the steel industry had a problem. If the steel industry did something to give them an economic edge, this gave the concrete folks something to think about.

The race to upgrade the code caused us to get very close to new modes af failure which were not 'in play' before such as a number of buckling modes, both local and global buckling. But to analyze the numerous buckling modes, one needed multi DOF models that could only be solved by hand by a few very capable mathmatically inclined engineers. Everyone wanted to take a breath until suitable computers and programs were available.

The PC came into wide spread use in the 90s and software to solve multi DOF problems started to become available along with them. By the time pentium computers were common, we had some wiz-bang computer programs out there with very user friendly interfaces and graphical input-output routines that helped sell the software to even a miserly business major.

The problem is that we have sold our souls to the new era of the box that thinks for us. The pace of work and the desire to please our clients (both internal and external) is so strong that we are unable to review the results of our analysis to the degree we once could. I fear this is dangerous because as the OP alludes, it is the engineer that is doing the thinking, not the box. But we just can't think fast enough to get the subtle problems of the differences between the model and the real geometry of the problem. Now we think about things like 'progressive collapse' and plastic hinges, but we should be thinking about the basic connections. History shows us that structures fail at the connections not in the members. When we designed and detailed by hand, we thought a lot more about the connections. Now we just pass that off to the detailer.

Sorry this sounds like a rant, but I am worried for the future with the engineering business in the state it is in right now. I have caught too many significant problems with engineered designs in the past to be any more positive than I am.

 

[whyun]

Complexities introduced in the newer codes "intends" to allow flimsier structures to work. Complexities also add more opportunities for error.

I too am worried for the future of structural engineering. In the not-too-distant future, number of true engineers will diminish, replaced with a horde of engineering technicians.

I've seen many situations that made me lose confidence in our education system (engineering) and the professional engineer licensing boards.

The leaders in the industry need to stop the p*ssing contest and focus on common sense and K.I.S.S.

 

[Lion06]

whyun-
You're not in PA are you? I know several people who use that acronym - K.I.S.S.

 

[whyun]

I am in CA. K.I.S.S. is an acronym that I picked up from the U.S. Army. I haven't researched the origin.

 

[mudflaps]

Sorry guys, but the code is my fault - and those old dudes like me. We were asked to join code committees and declined. This allowed space for the young guys with fresh PHD thesis papers to get their thoughts into the code. I'd like to see a mandatory 10 years of private work in your chosen field before PHD work could be started. Sounds like sour grapes when I see this in print!

Anyway, there's a lot of ego and three martini decisions in most codes (it's a voting thing. You vote for my ugly dog and I'll vote for yours). The IBC wind design section is a perfect example and is starting to be ignored in my neck of the woods. Even the plan checkers can't tell if it was done correctly.

How do we correct the problem. Easy. Just get involved!

old CA SE

 

[asixth]

I thought K.I.S.S. were a band.

I agree that if something isn't broken, then don't try to fix it. However, we should always be looking to improve it, and I believe that limit state design is a progression from working stress (except for geotechnical applications). For example, we have a better understanding on the behaviour and material strengh of steel as opposed to concrete, so it is appropriate to use a higher strength reduction factor. Likewise, live loads are more variable than dead loads, so a greater load factor should be applied.

To answer civeng80's question directly, from what I understand, AS1250 is written in working stress and I'm not sure whether it is still called up in the BCA, I will have to check that on Monday.

The codes aren't gospel, in our office we have policies about what clauses we should be careful with and how to apply them because their defination is a little "vague".

And to think back in the day, Bill Gates had to argue he's belief that every person needed a personal computer.

 

[JAE]

My three older mentors (started their careers in 1946, 1947, and ...gasp 1927) all thought the red AISC steel manual was too complicated.

It's all relative.

 

[civeng80]

Good old AS1250 is not mentioned in BCA (wish it was!) nor any other recent document that I know of on building design and construction. I bet alot of engineers still do preliminary checks to determine rough in sizes.

Love to know how many structures actually failed because of AS1250 ?