ASD: 1.03 Allowable Overstress 6

[SmithJ]

I have been used to allowing a 3 percent overstress when working with the AISC code. However, I recently realized that I can not adequately identify the basis for this overstress allowance. Does anyone know the origin of this 1.03 or where I can find information that would explain it.

Thanks.

 

[JedClampett]

Never heard of such a thing. I've used a 10% overstress as reasonable by judgement when checking already installed work where there was an error in the original design or a changed condition. But I don't think you'll find that (3% overstress) codified anywhere.

 

[shemp]

Do we really know ANY loads to within 3% accuracy? In design, why would you want to walk so closely to the edge of the minimum design loads? Why not just use the next size up?

 

[scherry]

I was taught to use the three percent rule for checking. in other words if I had a structure that either looked fine and i needed to check it for current code and the current code loads resulted in three percent overstress, i did not necessarily need to reinforce it although I might anyway depending on redundancy. Also if I wqas adding a load to a member and the load was less than three percent of the stength of the member (ASD) I did not worry about doing a full analysis of all of the loads on the member - again this is limited to situations where i had reasonable certainty that the member I was dealing with was not overstressed to start with. My boss who learned structural design in europe and practised in canada and the USA taught me this.

 

[abquad]

check in aisc-asd volume ii - connctions. it's somewhere there.

 

[dbuzz]

For unity checks it's a case of significant figures. If the check states "less than or equal to 1", then 1.01 doesn't comply. However, if the check states "less than or equal to 1.0", then 1.049 rounds down to 1.0 and complies while 1.050 rounds up to 1.1 and does not comply. Therefore you can "accept" up to 5% "overstress".

 

[HgTX]

Not the answer you want to hear, I'm sure, but--

What I was taught in school was not to sweat a couple of percentage points because none of the calculations are that accurate to begin with. However, one should be prepared, in the case that anything goes wrong, to explain in a court of law why one technically did in fact violate the code.

The significant digits argument is much more compelling.

Hg

 

[UcfSE]

I doubt this is actually in a code. The code is worded to give minimums and requirements, not suggestions. I don't think it would also say "but if you want to add 3% go ahead". I believe the 3% is historical in origin and is based on your own judgment and how comfortable you are signing and sealing something that technically violates code. I've heard of the rule for checking retrofits and stuff like that and used it as well. Personally for new construction I say just use the next size up. We all know that our analysis methods and theory aren't accurate to 3% but that isn't necessarily a legal argument.

 

[waytsh]

At one time I thought I read something about this 3% allowable overstress being applied to bring ASD results more in line with LRFD solutions. Has been quite a few years and I can't remember where I saw it. I just thought I would throw it out there as a possible lead for you.

 

[broekie]

Whenever our firm reviews pre-engineered metal building calculations, it seems like the pre-engineered metal building manufacturers always allow a 3% overstress. Not sure if it comes from a code they use or not.

We always reject the calculations for the reason that if, God forbid, something were to ever happen, we would have a hard time defending the overstress in court.

It always turns into a nasty dogfight with the metal building guys.

 

[dbuzz]

It has to do with significant figures. 1.03 (and 1.049!) rounds down to 1.0. None of our calculations are accurate to two decimal places.

 

[FalsePrecision]

Some calculations are more accurate than others. When you have huge steel beam to design, it's your PC - duty to save steel and allow a 3% overstress. When designing a wood beam with a complex, hard to exactly calculate load in a custom home, you would be wise to bump up to the next larger beam size even if your stress or deflection ratio is only 90%. It's all a matter of personal druthers.

 

[ajh1]

Just got back from an AISI spec committee meeting. The concept of 1.03 allowable came up in passing. While it is agreed that most everyone does it at one time or another, the committee wouldn't think to codify it in any way. We've all faced the expert who thinks 1.0001 is the end of the world.

 

[tonyuk]

This is an interesting question for those of us who write software and have to decide whether something is acceptable or not. For example the AISC code refers to 0.66Fy = 23.76 for A36 steel, whilst most textbooks use a value of 24. The difference is just over 1% which, as already stated by several posters, is really irrelevant but when it comes to comparing numbers computers do what they are told. If you work to 3 figures Fb is 23.8. So if fb is 23.9 should this be flagged as failing? I would rather show the exact numbers and allow a designer to use his discretion if it 'fails' but on the stand the lawyer is going to say "if they meant to say 0.67 don't you think they would have done so?".

 

[FalsePrecision]

tonyuk - it says in the AISC manual that 24 ksi is common practice, and acceptable. And regarding the 1.03 "say OK", issue, that is common practice, within the usual standards of care. I wouldn't let a lawyer bully me around with his interpretation of common sense.
What kind of software do you write?

 

[tonyuk]

We produce several programs - mainly for the UK - that handle simple beam and column design: the USA version info can be found at

http://www.superbeam.com

. I think most of us aged 30+ appreciate that any structural calculations will, at best, only be accurate to a few percent if for no other reason than there is some tolerance on member sizing and quibbling over decimal places is ludicrous. But when you write a program, even simple (in engineering terms) ones like ours you have to decide where you draw the line.

 

[FalsePrecision]

ton,
OK I can see the "draw the line" dilemma. Conversely, it seems very unlikely to be concerned about overstressing a steel beam in flexure, say 5%, without first exceeding a deflection allowable.

 

[VirtualEngineer]

I don't know about all you guy's but I don't like my calculations to show any "OVERSTRESS" because any laywer worth his salt can hang you.

If it makes a difference, get more precise on the way you specify your loads. Unless you always design to the gnat's eylash, there is some room to make you results come out under the "ALLOWABLE" load. If you are designing to the gnat's eylash, then you should be particular about "EXCEEDING" the code.

Regards,

JPJ

 

[dbuzz]

JPJ

I think it's important to understand that there is no actual "overstressing" of the material. It's a question of mathematics. As much as we might admire the accuracy of our calculations, they are not accurate to two decimal figures. Consequently 1.03 rounds down to 1.0. So does 1.049 for that matter!

dbuzz

 

[VirtualEngineer]

dbuzz,

From a purely technical point of view I agree with you, however, from a legal point of view, you leave yourself open to a line of questioning in which you will have to admit that according to your calculations, you have exceeded the allowable stress.

That’s what the jury will hear and you will have lost a great deal of credibility with the jury.

I know that when I am designing some structure or another, I typically use conservative but easily applied loads to reduce computation time.

When my results show members exceeding the allowable stress, if I don't want to increase the member size, then I go back in and rework the loads into a more precise configuration.

I am not suggesting you have to increase the member size, I am just saying you should run another iteration with your loads more precisely located so the results do not show any overstressed members.

Of course I don't design to the gnat's eyelash, so I usually have room to get the member stress below the allowable limit.

I just like to keep my calculations clean. During an AISC seminar I atteded, I heard that recommendation given, and so I am just passing it along to all who read this post.

Regards,

JPJ