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.

 

[Agent666]

I think it's important to remember that any code is a minimum requirement... This applies whether it's fifty years ago or yesterday....

Our understanding just happens to be more complete in this day and age.

 

[hokie66]

Let me fix that for you...Our understanding just happens to be more INcomplete in this day and age.

 

[fegenbush]

Not to distract from the direction of ASD vs. LRFD, but I have a problem with the complexity of the design codes in reference to seismic design. Is there a statistically relevant difference in the seismic factors just by moving a few miles one direction or the other? The new three or four significant digit lookup tools seem to think so. Is it not possible that the seismic zone maps were of acceptable accuracy? Perhaps going back to that system and updating the parameters would be a good compromise.

/end rant

 

[Agent666]

Some individual engineers understanding is incomplete because they see the codes of the day as the only thing they need to now satisfy. We all know there is more to structural engineering than satisfying equations, to understand how the structure works and behaves are some of the best qualities an engineer can possess.

Otherwise we know more collectively about engineering, the behaviour of materials etc than we ever have. This allows us to push the limit so to speak when we understand what we are doing.

Unfortunately in my experience the vast proportion of Structural engineers just don't get the basics anymore. Design is dumbed down by black box solutions that take everything to the limit where the feel for the design is lost. I am sure everyone has seen calculation sets that are just composed of spreadsheets with little demonstration of load path or even satisfying fundamentals like equilibrium.

 

[IRstuff]

"Practicing structural engineers have not been sufficiently vocal or diligent in resisting complication for complication sake."

I don't think that necessarily the case. Computers has allowed architects to design structures that could previously only be dreamt about, since there was no way to do the math to even prove that the structure could meet its requirements. Asymmetrical structures were not easily analyzed, but computers can not make short work of them. Many of the taller buildings in seismic zones would have been impossible to design 30 yrs ago, because hand calculations would not have been able to model how active or passive dampers could protect the bulding.

TTFN
faq731-376

Need help writing a question or understanding a reply? forum1529

 

[hokie66]

Agent666,
Now that you put it like that, I agree with you.

IRstuff,
I didn't argue about computers, but rather about unnecessary complication. The type structures you referred to probably involve 1% of our work, so should not be the tail which wags the dog.

 

[XR250]

I am sure the pre-engineered metal building folks welcome ANY change in the code that will let them knock a few more pounds out of their spaghetti.

 

[Ron]

Well said Archie and hokie66....similar era, similar thoughts.

As hokie66 noted, strength design is what it is, but serviceability has been woefully neglected. I understand that serviceability changes are on the cusp in the next couple of years (code-wise). Getting things to stand up is easy. Getting things to last is more difficult. Apply that as you wish!

 

[SteelPE]

Ron,

Yeah, more changes!

 

[TehMightyEngineer]

Well, since not all changes are bad, who else is eagerly awaiting the new ACI code? About time they fixed that disorganized pile of lawyer dung. I can't believe it took them this long to realize that designing a column per ACI involes like 10 different sections in the spec whereas in steel or wood it's one freaking section (two or three if you count lateral stability and connections) and AISC and NDS include a manual for your troubles at almost the same cost!

Maine EIT, Civil/Structural.

 

[TopKnot]

Does anyone have an example of LRFD giving less conservative results, and then problems from using the smaller size in the field? I see undesigned and drastically undersized components inexplicably work just fine. It is funny to tell someone that their 100-year-old beam is overstressed by 100% and needs to be reinforced.

LRFD is trying to respect the fact that things don't fail because they are 30% overstressed in ASD. Things fail when they are 400% stressed. Personally, deflection and serviceability mean more to me than a unity check. I see less experienced engineers super-concerned with strength and ignore deflection and vibration.

_________________________
TKE

 

[TehMightyEngineer]

Buried box culverts under AASHTO. City was using 3 rows of precast box culverts with end caps as underground storage tanks for combine sewage and stormwater overflow. Contractor placed one row and then excavated for the next placing the fill over the previous row (almost 200% the original soil load). No vehicle loads (too deep) so the engineer (not me) correctly designed them to a 1.35 soil load factor (vs 2.1 live load AASHTO factor) using the expected soil load. Note that AASHTO does not have a "dead load only" load case except for one that increases the self-weight of structures only. The box culverts were also cast with the wrong bar spacing by the precaster. Anyway, long story short they broke. Had they been done with a uniform safety factor they wouldn't have broken. We got called in to do an after the fact analysis and calculated that the bars were about at the expected yield stress if I recall correctly and that failure was ultimately caused by the soil overload and not the misplaced rebar.

Maine EIT, Civil/Structural.

 

[VoyageofDiscovery]

@TME
That's called experience!

 

[TehMightyEngineer]

Indeed. One thing my boss has taught me, and I've taken to heart, is very similar to what people have said here. Just because it meets the code doesn't mean it's good engineering. For example, we do a lot of work in industrial facilities and my boss never has me bring a column beyond 50% actual/allowable ratio if we can help it. Reason being that 50 years from now they'll probably hang 2 pipe racks off the column, want to add another floor, and bust up the flanges with forklifts. The key is to know what people are most likely to abuse and what they aren't. In an office building, sure, take the floor beams to 99%. However, don't put a custom lifting beam out there with the minimum weld required, it'll cost them a few dollars more to make sure the connection doesn't control and might save someones life.

Maine EIT, Civil/Structural.

 

[NOLAENG]

As a younger engineer, I would like to give me perspective even though this post was originally for more seasoned engineers.

Most of the senior engineers I work with would prefer to never use LRFD or the more complicated codes created only for revenue.
As someone who learned ASD, LRFD, finite element, etc. during college, I thought the senior engineers were just stubborn to change when I first started working.
Now, I often times spend too much time focusing on the potential load combination sets and factors (ASD - Steel, LRFD - Concrete, Service - Foundation) required for code prescriptions rather than the intent of a safe design.

After working for a couple years, I've come to realize that most of the basics and fundamentals of structural engineering have been lost in some ways with the advent of sophisticated computer analysis programs, spreadsheets, etc with my generation.
Some of my coworkers in the same experience bracket will often overlook things from a computer analysis or spreadsheet due to garbage in = garbage out.
I'm not saying we are doomed or I am an exception to the rule.
I still utilize the software tools available, but I've become much more diligent in checking them with hand calcs and common sense as I've gained increased experience.
Hopefully one day in my career we will reach common ground between fundamentals, computers, and methodologies.

As a black book engineer, please don't hit me too hard with your green/red one.

 

[TehMightyEngineer]

I've had the exact same experience uncblue. When I first starting out of college many of my designs were technically correct but looking back I focused way too much on doing every last detail and check in a design trying to "prove" it was "safe" where a simple, conservative back of the envelope calculation could confirm that and all the little details could be checked after I verified it with my envelope calcs. Forests and trees and all that. I've also found that I now do a lot more diligent in checking my spreadsheets and software with rough calcs. I've found a fair amount of my own garbage in = garbage out issues that should have been found much easier with back of the envelop calcs.

Definitely would love to see more younger engineers chime in here.

Maine EIT, Civil/Structural.

 

[dhengr]

Uncblue & TME:
For younger guys, I think you’ve got it right, and do see the trees and the forest for what they are..., This from an older guy. My hat’s off to you for this knowledge and understanding. None of us old guys are saying computers and software aren’t important. For all my ranting about the use of computers, computers and various software have allowed us to do things that we could only approximate or imagine before. In that, I agree with Doug and Hokie. They are a wonderful addition to our box of tools, in the right hands. But, they have also allowed way to many people who have little education or experience to pretend to be engineers, on things they have no business tackling, without some very serious and experienced guidance. Way too many of the users of this stuff have no idea how it works, or where the potential glitch points are in the programing. Most of this software tries to do way too much in one full swoop/or program, and consequently is pretty difficult to check or follow. The users have a very limited understanding of how the structure or detail works, which is actually caused by the very way they have designed and detailed it, so it’s pretty hard to imagine that they know how model it to get reasonable output. If it converges or it doesn’t blink in big red letters, ‘this’ll never work, you dummy’ they think they’ve got it right and they’ve done a good job. Never mind that the details can’t be fabricated or constructed.

None of us old guys are saying bldg. codes and standards aren’t important. But, the rate of change of all the codes and stds., with very little real improvement in the final infrastructure produced from these code changes; and the lack of coordination btwn. referenced codes and stds., and that pertains to content and date of issuance, certainly is no help to our entire industry. It is a hindrance, continuous headache, and very expensive, except for those who are enriched by the continuous publication and testing and research on so much minutia. I have never resisted an addendum or change when something really proved to be deficient in its current form, but change for change’s sake in dribs and drabs, scattered all over the place, mostly just confuses the process. There is no doubt about the comfort issue with what you know and use regularly. I taught WSD and USD in conc. back in the mid 60's, USD showed up and was used seriously for the first time in the ACI code. I have used many different steel codes over the years and followed the development of AISC’s LRFD with considerable interest. But then, by the nature of what I was doing, I missed regular practice with a couple cycles of the AISC codes, and have ended up behind the eight-ball ever since. I think I have a pretty good handle on the intent of all of the codes we use, but I’m becoming increasingly nervous about the rate of change and the minutia involved in that change, for the sake of change, which one could easily miss until he had to explain that code variation to a judge, in our litigious society.

It would seem to me that our educational systems are not teaching enough of the fundamentals any longer. That can all be done by computer, so who needs that fundamental understanding of how structures work any longer, a real feeling for how they work, that’s Paddington’s point. And, who cares about good clean design and simplicity of detail, the computer can handle the complexity even if we don’t know how to model it or check the results.

Mentoring seems to have gone out the window, we have the internet now, google it. Most of the OP’s here should be directed to the boss, so he knows what you know and what you don’t know, so he can make some effort to keep you and the company out of trouble. I’m all for giving a young engineer all he/she can handle, to force them to think a bit beyond their comfort level; but then watching and guiding carefully so they don’t go to far astray.

Good for you, I would like to work with you.

 

[TehMightyEngineer]

"Most of this software tries to do way too much in one full swoop/or program, and consequently is pretty difficult to check or follow."

My goodness, THIS! You made a bunch of good point dhengr (star for you!) and once this thread runs its course I hope to write up my impressions and summarize everything that was said here. That said, I definitely want to expand my personal experience on the above.

Back when I first started out of college I expanded a few spreadsheets I had created during college into design templates. The two most advanced were my wood beam/column spreadsheet and my foundation spreadsheet. The wood spreadsheet took input loads and sizes and ran through the bending and axial checks for a wood beam. While I never really had much I've changed in it over the years (still use both) if I were to make it today I would break it up into sections so that it's much easier to follow and much easier to utilize when I do something that's not very standardized (like a T-shaped wood column or something).

For the foundation spreadsheet though I've editing that thing about 10 times (most recently last week) and it's become this behemoth that can do rectangular footings, square footings, strip footings with unbalanced soil loads, wind and seismic moments at the base, etc. And I've still barely scratched the surface of what I could try to put into it. However, with my last revision just this week (added rectangular footing rebar calcs) I realized that it was very unwieldy. For example, I used it for a rectangular footing, great. But then when I used it for one that had a slightly different loading configuration it wasn't right and I had to re-write a bunch of load cases to make it correct. Tons of checking later and I was satisfied but I pity whoever in the office will be checking this because it's so convoluted now.

In the future I plan to be much more systematic with my calcs and my own spreadsheets. No "all in one" calculations (unless I'm doing something 10,000 times), rather one spreadsheet for one calculation. Then, if I'm doing something unusual, a hand calculation can replace one of my spreadsheets easily without throwing them all out.

I see the same thing you mention in software. They're so "all in one" that it's way too easy to screw up your inputs, way too easy to use something without having the experience to back it up, and way WAY too easy to misunderstand what the black box is doing.

"Good for you, I would like to work with you."

Thanks! That's very flattering.

Maine EIT, Civil/Structural.

 

[Agent666]

I think in parallel with 'improving' the codes, you also need to consider that the actual typical quality you get during the physical activity of building the structure from the typical contractor is usually follows a trend in the opposite direction...

 

[TehMightyEngineer]

Really what they should do is publish code revisions every 10 years or so that contain improvements to address any deficiencies (emergency code addendums could be added if needed). The remaining "new" material could go into non-mandatory addendums which get added to the code every 10 year cycle or so. MUCH simpler, much easier, much more practical. However, then we buy less codes... they gotta make their money somehow.

Honestly though, I'd take a $1000 code every 10 years than a $300 code every 3 years. Only once in the past 8 years have I ever wanted a new code because I wanted to utilize something "new" (GFRP bars in this case).

Maine EIT, Civil/Structural.