Eng. practice of allowing 5% overstress 33

[radair]

It seems to be standard structural engineering practice to allow up to 5 percent overstress in structural design. It's been this way since I graduated college in 1980 and I've seen the practice commonly used in the tower analysis field for the last 15 years.

I've been asked my opinion by a government agency as to why this is a safe and acceptable engineering practice, including citing any relevant structural codes. They are not questioning my work but are asking me for a signed & sealed letter of opinion. It seems to me that this would be a better question for their state engineering board of licensure.

Can any of you help?

 

[frv]

OK. Perhaps "exquisitely" is a bit of an overreach.

It's very well written, but I'll reserve that adjective for other forms of literature.

 

[Dinosaur]

One other thing I would like to throw out there ...

I have spent most of my career designing bridges. Before you decide 5% overstress is OK, go out and look at the condition of some of the structures in service near the end of their design life. The corrosion loss, damaged members, frozen expansion bearings, ... might shock you. On the other hand, I believe I heard early in my career that any structure that doesn't fail in the first five years of service is considered an adequate design. I wish I had asked for a citation on that assertion back then.

I still fear the consequences of the slippery slope. One guy decides 105% is OK. Another fellow later decides 110% is OK. Another fellow thinks he will let 112% go on a member. Where does it end?

I have enjoyed reading this discussion. Thanks.

 

[civeng80]

Quite an interesting article from Wooten. Applies to alot of things today and not just in engineering. I think the amount of uncommon knowledge around today is enormous with the internet especially, maybe at the expense of not using the old common sense approach in solving problems.

I think Ive learnt alot from this post.

Cheers!

 

[JStephen]

A very interesting discussion.

It looks to me like there are three separate questions that get combined, although they don't have the same answers.
1) Will it fail?
2) If it somehow failed, would I be considered negligent in a court of law?
3) Does it meet the code?

Now, it seems that everyone agrees pretty much on #1. That a 3% or 5% increase, by itself, with current codes, isn't going to cause a failure. And this seems to be the point that the Wooten article is getting at.

Question #2 seems murkier, with conflicting opinions given above.

Question #3 seems to me to be the sticker point, though. When you have a set of drawings that purports to meet a particular code, and seal those drawings, are you certifying ONLY that the item will not fail?

Let me give you an example from my field. In pressure vessel codes, you might have an equation given for wall thickness along the lines of t=PR/S, with P, R, and S being pressure, radius, and allowable stress. But the exact equation to be used will be specified. And somewhere in that same code, you'll find tolerances for allowable plate underrun. So when you calculate the thickness as 0.52", can you use 1/2" plate? Generally, not.

Now, if you look at the failure of that 0.52" plate, you'll find that they recently changed the code safety factor from 4 to 3.5, and that other codes use different safety factors where perhaps that same plate could be 3/8" plate. So the issue of it failing for lack of that 0.02" is pretty much a non-issue. But the question of whether you've met the code still remains.

Switch from this to say, a building, where you have very involved loads and analysis, and could approach it quite a few different ways. Suppose you calculate the beam stress as 25,000 psi and the code calls out 23,760 psi maximum. Can you say you met the code requirement simply because the problem got more complicated? Or that you meet the code because the loads are more approximate? Or that you meet the code because the code failed to specify a tolerance, or specified it in an ambiguous manner? Maybe you suppose that you could go back and re-work your loads and analysis to make the numbers work, but you haven't actually done so, either. What then?

 

[BRGENG]

What would you rather do? Bump up a member size to avoid being overstressed or explain to a jury why you designed a building to fail?

Why risk your license in order to save a few pounds of materials. I really can not believe this is being debated by so many professionals. Design to fail should not be an option. If you want a reason why you should not design to a stress level of 1 just look at how many Pre Engineered Metal Buildings fail every time it snows or the wind blows hard.

 

[apsix]

"Design to fail should not be an option." No one is suggesting that. Alarmist statements like this don't help the debate of course.

The term 'over-stress' is also misleading. Using a limit of 1.05 reduces the factor of safety written into the codes, as noted by JStephen. It does not lead to over-stress of the member.

However, I would very rarely design at or beyond member capacity. In the industrial field no one will thank me for saving just a few dollars at the risk of serviceability issues. Having the robustness to cope of some corrosion, an unforeseen vehicle impact or additional loading is part of good design in my book.

 

[ThePenguin]

I would just like to point out that the AASHTO LRFD Bridge Design Specifications 4th Edition has 5% adjustments built explicitly into the code. These "load modifiers" effectively allow an additional 5% capacity for structures with exceptional ductility, highly redundant load paths, or structures of low importance. The code also likewise imposes a penalty for structures with low ductility, lack of redundancy, and high importance. So it would seem that the AASHTO folks are comfortable with a 5% adjustment to take advantage of well behaved structures. I would further argue that this would establish a 5% adjustment based on engineering judgement as a standard practice, atleast with regards to bridge design. Similar reasoning, although not explicit in the code, could be applied to buildings. A 5% bonus for ductile members that are part of a highly redunant load path does not seem unreasonable. However to apply the same 5% to a brittle member that is required for the stability of a staticaly determinant structure would be ill advised. Do not get me wrong, I am not advocating that we should strive for 105% for every member as a means of cost cutting, as this does not involve any engineering judgement when applied in such a blind manner. Just my two cents for what its worth...

 

[frv]

BRGENG...

"Designed to fail" has got to be one of the most idiotic statements on this thread.

There is certainly room for disagreement, but your statement is disingenuous at best, or negligently uninformed at worst.

 

[youngstructural]

If 1.05 design limit does not make sense to you then do not apply it. If you feel that this is genuinely a poor design practice, I fear for the state of engineering education...

We must, above all other things, be practical. In my view you are little better than an "Applied Scientist" with arguments like that...

To purposefully paraphrase a famous quote, I think an apt saying for this situation might be "Engineering without art is science". Personally my degree and job titles have all contain the word "Engineering"; Not one contains "Science", nor would I ever want them to...

Regards,

YS

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

 

[hokie66]

Hold on there, YS. My engineering degrees do have science in the wording, and I have gotten over that, except for the physics part, of course.

 

[civeng80]

frv

I think BRGENG was just imitating the Lawyer cross examining the engineer in a legal suite. Believe me they could very well ask the question in that way. Remember the Jury members are not engineers nor is the Judge.

 

[frv]

If that's the case BRGENG, I apologize for the harsh wording.

 

[youngstructural]

Sorry Hokie; I really didn't mean to insult anyone, and it really isn't fair to generalize... I just tend to think of theory based eduactions (ie: Schools that concentrate very heavily on theory) as Applied Science degress, which schools that emphasise practical abilities as being Engineering degrees...

It does seem to be that way in Canada, despite our much vaunted standardized engineering education system.

If I have insulted you at all, I appologize profusely. I really hold you in great estime; Please understand this is my prejudice, and not my honest view.

Regards,

YS

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

 

[hokie66]

YS,

I know no offence was intended, and none was taken. It is just that in the US, a lot of the major universities call their engineering degrees Bachelor of Science in Engineering, etc. I've always thought it a bit strange myself.

 

[Lion06]

I've heard that saying as "Engineering without art is calculating, Art without engineering is dreaming"

 

[HgTX]

I do also agree that shooting for 103% doesn't make sense. You shoot for 100% and if you get something in your model that cracks 103%, squint your eyes, think about what that particular member is doing and how it fits in to the big picture and make a call - yeah, it's OK, or no, I'm gonna have to whittle my loads or move the member or upsize the member.

(I just thought it bore repeating.)

Hg

Eng-Tips policies: faq731-376

 

[JoeTank]

I've been a PE for over thirty years and can count on one hand the number of times that I've observed others using an "overstress" of any percentage as a basis of design. I'm somewhat shocked at the huge response to this academic question couched in practical terms. Very interesting reading. As the overwhelming response has been supportive of this practice, I must ask if it is normally employed in your daily practice?

Joe Tank

 

[Lion06]

We never use it for new design. We typically keep stress ratios < 0.95, though after design smoothing I'm many are < 0.85.
We do however use 1.05 for existing structures when required.

 

[MechEng2005]

JaredS -
I completely understand that the code is very conservative. However, it is written that way for a reason. I question who/when gets to decide when it is acceptable to reduce the "factor of safety" (in quotes since I think it has been agreed that a structural member at 1.05 is not necessarily less safe than a member at 0.95, depending on the case). I question why a code written by engineers should be able to be violated based on engineering judgement. Isn't it then a case of the judgement of the engineers writing the code versus the engineer designing a specific structure? I believe that if something is designed "to code," it should meet ALL the code requirements. The amount of engineering judgement used should be specified by the code. If the code recommends a range of values (i.e. safety factor of 3-4) and the design engineer accepts the lower end of the range, fine. However, if the code says greater than 4, then I don't think "3" is an acceptable option.

I'm not a structural engineer and am not familiar with the code. I do not doubt that they are over-conservative and a hassle. I just disagree with the "theory" of exceeding code limits. I do not have the experience and knowledge to determine the implications (and thus if acceptable) in practice. I understand most of those posting here are practicing this type of work, and therefore have more credability on this topic than I do. However, I hope that my opinion is considered and can provide some insight into an opposing view point. I'm just saying take it for what it's worth (which may be nothing to some reading this, and that's fine by me).

JStephen-
Couldn't agree more! Star for you!

-- MechEng2005

 

[frv]

JoeTank-

Perhaps this discussion gives the impression that 1.05 is the basis for design. I don't think that is what I and others are arguing.

Quite frankly, most of my designs are well below .95. What we are arguing is that for any design, if a few members come out at 1.05, there is really nothing to be concerned about and it is perfectly acceptable to allow it if you know (and why shouldn't you) the full extent of your design- where you were conservative, how realistic your loads are, etc.

If, however, your entire design is right at 1.0, then common sense would tell you that perhaps it's a better idea to beef up some of your more critical members.