Why did they do that? 19

[TehMightyEngineer]

Working under a PE from the red book era of AISC I often see aversion towards "modern" methods of doing things. It made me wonder what people thought of engineering then (before my time) and now. Basically I'm looking to get perspective from other engineers on what they like or dislike about modern methods vs "back in my day" and find out why people think engineering has made the changes they did.

Speifically, LRFD vs Allowable Strength Design vs Allowable Stress Design. Green book AISC vs black book. Hand calcs vs computers (i.e. effective length vs direct analysis). Simple equations vs complex equations.

If you're wondering, this was brought on by my boss and I discussing how the green book has you check base metal at the weld fusion for a fillet weld but the black book does not. We also often have discussions on the merits of ASD vs LRFD load cases. For example, is it unconservative to design a buried concrete structure for LRFD load cases when it's all dead load.

Maine EIT, Civil/Structural.

 

[EngineeringEric]

I am curious what we have to say about this one. I know my office we have some people who still exclusivly use ASD for everything but large Seismic and some special concrete design.
-
I have almost always been told that ASD is more conservative than LRFD and has simpler equations. the simplicity stems from more conservative curve fitting(?). I suppose to me it is a thing of doing what one is comfortable with, and why would 20 years of ASD knowledge be replaced with 0 years experience of LRFD. I can see that as a difficult transition to make a Senior Design Engineer do.

 

[JoshPlumSE]

Black book still has that. Sort of. . . . Page 9-5 (of the manual, not the spec) has a section about base material rupture at welds. Essentially saying that the base material has to have a certain minimum thickness to avoid base material failure.

If not met, in my calcs, I use this to reduce the strength of the weld connection down to a value that will match the base material rupture.

 

[TehMightyEngineer]

Josh: Yes, that's the conclusion we came to as well in our office and elected to use either that or the 9th edition base metal fusion check.

Maine EIT, Civil/Structural.

 

[Archie264]

Speaking only for myself, I don't like:
[ul]
[li]Complications for complication's sake[/li]
[li]Feeling like I have to rely on a computer to get to a solution[/li]
[li]Rube Goldberg load factors so that I can design to within a gnat's whisker yet still have to wonder if I've done it correctly[/li]
[li]Getting lost in the forest for the trees (i.e., feeling so bogged down by the code updates that I wonder if the building will pass muster with the code officials when I know it will stand up to the forces of nature.)[/li]
[li]The implication that the old-timers were a bunch of ignoramuses who couldn't design a building now that we have our nifty new codes[/li]
[li]Feeling like I'm at the mercy of the theoreticians[/li]
[li]Being distracted by minutia causing me to miss the larger principles[/li]
[li]Feeling like we're getting away from the wisdom, elegance, and simplicity of our forebearer's designs for too-clever-by-half "sophisticated" solutions that might prove to be penny-wise and pound-foolish[/li]
[/ul]

I could go on but I'll confine myself to that for the moment. But, hey, you asked.

 

[TehMightyEngineer]

No, no keep going Archie. As a young and impressionable engineer (Hah!) I not only want to know what you think and actually agree with you on many of your points.

Maine EIT, Civil/Structural.

 

[EngineeringEric]

I am glad to hear that I am not the only one who uses the t_min calculation to downgrade the capacity of my welds under the assumption that excess weld will not contribute but is present.

Archie, what do you like?

 

[dhengr]

Archie264:
You speak for me too. Nicely done.

 

[JAE]

Yikes - another ASD vs. LRFD thread. Run. Hide. Stock up your shelter with toilet paper and scotch!

 

[TehMightyEngineer]

Well, I guess, but I probably wasn't clear on what I meant. ASD vs LRFD was merely an example which has been discussed to death. While I'd love to hear more about it I also was hoping people would discuss items similar to what Archie brought up; how we used to design vs how we design now and what they think of it.

In the end I'm basically looking for what people would reply to any engineer who says "I don't like how they do X now...", not "I don't like X because." Though obviously there is a lot of parallels between the two.

Maine EIT, Civil/Structural.

 

[racookpe1978]

Hmmmmn.


Seems like the "Build per Code" answers don't ever answer ..
Does it work?
Did it work?
Will it work? (Under this specific set of conditions, temperatures, maintenance, pressures, loads, movement, soil/building/steel/ground/wind/seismic/dynamic loads conditions?
What have I not allowed for?
What might happen, but not at the same time, have I allowed for that can be eliminated? (Do you really get an earthquake, hurricane and all those people climbing the tower at the same time?)
Is it safe now?
Will it be safe in the reasonable future?
Will it be safe under unreasonable (but foreseeable) conditions?

Can it be built at my client's budget, material, ability, material, labor, and time constraints+

 

[JAE]

OK - sorry - Just seemed like the thread was taking that turn....

I can say that I have designed out of the red AISC manual. That said, I currently use LRFD. I can use both, don't really care that much about either. I understand how LRFD is more accurate in terms of probability of failure and also understand that in many cases that increase in accuracy doesn't amount to a hill of beans....but sometimes it does.

Back when I was a younger EIT, I did see a lot of the older engineers using "old" methods. One guy was still doing concrete design in ASD - so you'r talking perhaps 1950's era designs.

I think what I saw, and understood, was that people don't always feel comfortable with change

Think about yourself and how you react when your favorite word processing program or spreadsheet program gets updated.

People get comfortable with what they know and with their skill set. When new "stuff" comes along this can create a sense of threat - I suddenly don't know what I used to know or I suddenly aren't up to snuff with the latest code requirements - it can be un-nerving for an engineer who has successfully designed things for years only to be told that they aren't doing it right anymore.

 

[Archie264]

Yikes, I don't think I deserves stars for my whining, though I did mean every bit of what I wrote. (But thanks for them.) For me it's not so much an issue of ASD vs. LRFD but, rather, the overall increasing complexity that seems to be occurring across the codes and reference books. In my opinion the benefit from the continual changes and additions are well into the realm of diminishing returns and have become, in some cases, an impediment to good design.

 

[msquared48]

Archie:

If I could give you more stars, I would.

Sometimes I wonder if the code changes are due to the "nitch" effect, i.e., someone carving out a place in history for themselves.

If it ain't broke, don't "fix" it, especially by making it so complicated to decipher that more mistakes are far more likely as that moveable failure envelope is getting smaller and smaller!

KISS!

Sorry, but I guess I am into trite and hackneyed expressions today... Still very applicable though to me.

Mike McCann
MMC Engineering

 

[paddingtongreen]

I learned to propose a structure to fill a need, then to visualize the loads flowing down the load paths, not unlike water down a pipe. I then asked each member (or group of similar members) questions, I asked if they were stiff enough and then if they were strong enough. As we investigated only primary load paths the structures had a level of redundancy. Very occasionally, I would perform a second order analysis but not often, I was discouraged by my bosses from doing so. Using load factors that applied to the whole structure was simple. As different load factors, and different strength reduction factors crept in it was discouraging, I could no longer "see" the loads flowing because of these multipliers and dividers got in the way.

I learned woodworking in high school, all with hand tools. Even though I have used power equipment since they do not feel comfortable. I spend more concentration on safety than on the shaping of the wood. Same thing with structural design, not having to worry about whether the model would run, or even having to tweaking it to make it run, left me free to understand the structure. Sorry if that sounds Zen but it is real.

I said in the first paragraph that the old methods left a level of redundancy, I understand that now we calculate secondary and tertiary load paths to the gnat's ass, we have to deliberately add in some redundancy. Because we took out, what to my mind was a natural redundancy, we have to deliberately and formally introduce an unnatural redundancy.

While we do use less material with LRFD, do we really save money, the material is the least of the costs of structures.

I would like to know if those of you raised on LRFD can "see" the loads flowing down through the structure to the foundations Whether you can see the structure bending a little, swaying a little.

Michael.
"Science adjusts its views based on what's observed. Faith is the denial of observation so that belief can be preserved." ~ Tim Minchin

 

[msquared48]

You knows, many of the large buildings had wind tunnel tests performed on them during the design phase, many 20 or 30 years ago, and were placed in the skylines of the times. However, how many of these structures have been re-analyzed with the additional buildings around them today? How have the forces to the structures changed? Lessened overall? Locally increased? This is just the tip of the iceberg here...

To design these things to a knat's eyebrow just makes no sense to me when the rules are always changing. It's just too dangerous, and lives will be lost. Where is the redundancy, the fudge factor, the extra capacity, or whatever, to stand up to the unknowns of the future?

Mike McCann
MMC Engineering

 

[IDS]

A different perspective (from a 1972 graduate):

I think the hand-calcs vs computer we see so often here is a false dichotomy. Sure you can get over-involved with detail and miss the big picture when using computers, but that can happen with blindly following a text book in a hand calculation too. Computers can be used as a black-box, but they can also be used to get a better understanding of how a structure really behaves than is possible with simple hand calculations, or to do a quick simplified analysis as an order of magnitude check on a more complex one.

As for allowable stress vs LRFD (or limit state design, as we call it in the rest of the World), I really don't understand some of the comments here. How does a procedure that only considers one condition (estimated stresses at nominal maximum loads) lead to a better understanding of structural behaviour than procedures that specifically requires you to consider the strength and serviceability of the structure under expected loads and maximum loads with minimum strength, and also to consider the stability of the structure under extreme loads, far in excess of the design loading?

Certainly it is possible to over-complicate the design process, but it is possible to over-simplify it as well. The question is not computer vs hand-calcs, it is how can we use all the available analysis and design tools most efficiently to produce the best end result?

We should simplify as far as possible, but no further. - Albert Einstein

Doug Jenkins
Interactive Design Services

http://newtonexcelbach.wordpress.com/

 

[SAIL3]

The majority of errors that I have made in my designs were when I became distracted my some detail of the design that was not that important in the overall design and as a result I had less time to devote to the more important areas of the structure. I place the current codes in that catagory with too much time and attention devoted to minutia and squeezing the last ounce out of every detail. Every hour I spend trying to untangle the true intent of the codes is an hour less that I have to analize and evaluate all the more important aspects of the design. This is really a dangerous trend and is imposing on the practicing engineer unreasonable and impractical constraints while trying to ensure that the public is safe when his design is utilized.
Others have mentioned some real serious issues and specifically Paddington about redundancy.

 

[hokie66]

It is certainly not computers vs hand calculations, as computers have been a great tool. But in the computer era, the practice of structural engineering has been overly complicated, partly due to computing power. Because computer programs can be written to solve complex problems, creating complex problems is attractive, particularly to academics. Practicing structural engineers have not been sufficiently vocal or diligent in resisting complication for complication sake. If you just consider reinforced concrete structures, the last 50 years have seen moderate improvement in strength design, but not in serviceability. Problems of cracking and deflection, at least in my experience, are worse than ever. Some of this is due to accelerated construction practice, which the design codes fail to adequately consider.

 

[TLHS]

I don't understand why people focus on the code so much. Design shouldn't really be happening to the code. Code is a way to check at the end that you've met some minimum standard. Even then, it's really easy to use conservative factors in the code where you've sized things for other reasons and prove that everything's okay quickly.

I size 90% of my members to within 90% of their final size on the back of a piece of scratch paper. How I do my final documentation check doesn't really affect how I actually design things or slide loads around.

Whether it's LRFD or ASD or complicated formulas or simplified approximations it's all the same physical behaviour. If you have a reasonable understanding of that you're going to get to the same point.

A lot of the new code stuff just gives you more tools. Sure I've got a complicated interaction equation I can use for steel in biaxial bending with compression loads. I've got it in a spreadsheet and can use it if I have reasons to get that close to theoretical capacity. If I don't need to do that, in most cases I can still just add my three force/capacity percentages and if it's under 100% I know I'm okay and conservative. There are all sorts of fun complicated equations that you have the opportunity to use if you need to, but generally don't have to when you're in the 'figuring out the structure' stage.