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?

 

[Tomfh]

Here's a thread which discusses this issue:

http://www.eng-tips.com/viewthread.cfm?qid=52511&page=1

Overstress by definition violates code, hence I can't imagine you'll find a code reference deeming it acceptable practice.

 

[JKStruct]

My vote in this debate is that it's acceptable, and widely practiced. IMHO, it's better to invoke the 5% overstress when checking existing construction than designing new. But I've still allowed 5% overstress on new design at times -- it depends on the condition.

I have never found it codified (someone correct me if I'm wrong). I doubt it ever will be codified, due to the "where do you draw the line" counter-argument.

In your letter, I'd vote for the significant figures argument as to code intent. Often the unity stress equations are given as <= 1.0 1.049 rounds down to 1.0 with 2 significant figures, and 2 sig. fig's is the code language (e.g. not 1) I might also write something about what a "calculated/predicted overstress" ACTUALLY means (to the lay person), with info about load approximations, safety factors and this minor relative encroachment, ductility, "failure" modes, yielding, and NOT "detrimental structural collapse jeopardizing public safety", ... stuff like that. It's true a lawyer might challenge a 3% overstress in court, but your lawyer could use the above as an industry-accepted defensible position.

 

[NS4U]

for new design i practice the 5% under stress approach... simply because i like to sleep soundly at night

 

[abusementpark]

I think it would be easy to justify a slight overstress if you know you are conservatively approximating certain aspects of the design (ie. considering a support pinned when in reality it is does transfer some moment). What is it in particular that you are being questioned about?

 

[msquared48]

I am conservative in my dead loads that I assume, and rarely, if ever, use the code allowable LL reduction factor in my designs. Hence, I have never had a problem using the 3 to 5% overstrength factor in my projects.

Mike McCann
MMC Engineering

 

[jike]

I know of no code implied overstress other than using 24 ksi versus 23.76 ksi and 22 ksi versus 21.6 ksi. This yields a maximum percentage of 1.02. There may be other examples but I don't know what they would be.

When being challenged in court,I heard an expert wittness refer to various textbooks that allowed an overstress of 5% in their examples which was used as evidence of "normal standard of care".

 

[csd72]

I think you will find that if there was a problem and it ever went to court, the opposing lawyers would have a field day. In court it is referred to as a code violation - an automatic loss to you, no contest.

If this is a continuous beam in bending, then you may be able to claim the 5% under load redistribution.

 

[frv]

jike-

I think you're right that a lawyer would have a field day. However, I read an article in the last year relating to liability in design, and it mentioned that the standard is something like "care that another reasonable engineer would take". Obviously, I'm paraphrasing.

There are many reasons for arguing that a 5% over stress is acceptable. Every engineer I know tends to be conservative in their dead load calculations. As abusemntpark mentioned, we frequently assume pinned when almost any connection has the capacity to transfer some moment. A structure is designed for extreme events. The loads at these extreme events (wind and earthquake, specifically) are not really known and are very conservative. You could use exposure C in your wind when in reality you have exposure B. Most materials have well in excess of 5% more yield strength than specified.

At the end of the day, I think allowing a 5% increase in stress would fall under "engineering judgment", which is why we get paid the "big bucks"

Having said all that, I too fall on the side of designing for 5% less stress than allowable.

 

[hokie66]

I doubt that the 5% overstress thing has ever been argued in court. It falls within the purview of engineering judgment, but things don't fail because they are 5% overloaded. They generally fail because of gross error or lack of attention to detail.

 

[JStephen]

On the 24ksi vs 23.76 ksi, this was changed in the AISC manual, as I recalled- older books showed 24 ksi, newer showed the 23.76 ksi or 23.8 or something.

This whole idea goes against the grain with me. I wonder, though, had those same equations been written to show "<1" instead of "<1.0", would everyone be figuring that a 49.999% overstress was okay? That just seems an unreasonable way to interpret a code unless there is something that specifically says to do it this way (which there may be- I think plate thickness tolerances are done this way per ASTM A6). But I note that in the Numerical Values tables in AISC, that they do NOT follow this significant digit usage- 0.60 Fy is always shown to three significant digits, not two. Even in the older book that showed 24 ksi, it is shown as 24.0 ksi, not 24 ksi.

I also find it odd that if there was a requirement that Quantity X be > 1.0, then everyone would agree that 1.049 was definitely greater than 1.0. If the code intent is to round it off, it ought to be done all through, not just where convenient.

 

[RWF7437]

Curious............

No one has mentioned the statistical origin of design loads prescribed by Codes. Or the idea that Code compliance alone is a valid criteria in judging the adequacy of a design.

 

[JAE]

I think there is a big difference here between NEW design and checking existing conditions for new loads.

In most all my new designs I don't ever think of using a 5% overstress because my unity is usually between 0.7 and 0.9.

For existing conditions where I'm asked to check a beam or steel joist or wood joist, etc. I do allow for a 5% overstress simply because I know and understand the reliability methods that went into setting the inherent safety factors and given a structure that is "older" I have more confidence in its design and compentence.

Also, the 5% overstress is usually confined to a discreet area within the span, not across the entire member so the overstress doesn't present a comprehensive violation of the "required" level of safety.

The concept of overstress should also be remembered as a diminishment of the required level of safety factor and not a condition where the loads are 5% beyond a theoretical failure limit.

 

[csd72]

Jstephen,

you make a good point. I must admit that on occassion I have used the rounding of numbers to make things work.

I believe that the code prescribed safety factors allow for this.

I do think, however that stating that you accept 5% overstress is a red flag for lawyers if anything goes wrong.

frv,

You are correct in a way, you are compared to what your peers would do in the same situation, but only when it exceeds the code requirements. Once it is less safe than the code, this does not apply. As I said, a code violation is a free kick for the other side, but if it meets the code and there still is a problem then the other side needs to prove that a reasonable engineer would have done differently in the same situation.

 

[DaveAtkins]

The IBC specifically allows 5% overstress when checking existing construction--I don't have the Code if front of me, but I remember it being in there.

My opinion--it is ridiculous to think a structure is dangerous simply because something is found to be 5 percent overstressed. hokie66 hit it on the head--structures fail due to gross miscalculation or gross misunderstanding. And all the others who mentioned how conservative we are in our assumptions are correct as well. I think a fair amount of overstress--even more than 5%--can be justified by relaxing our conservative assumptions.

DaveAtkins

 

[hokie66]

Those who are concerned what a lawyer would think of an engineer who in his judgment allowed a slight overstress should in my opinion worry about other things. If you are the accused in court, a lot of things can happen to you (most of them bad) whether you have done your job well and complied with all applicable codes or not. The aim should be to stay out of court.

 

[msquared48]

To play devil's advocate to my original post, if the 5% overstress is recognized by all to be a "standard of care" in the engineering industry, then, legally, there should be no concerns with using it. There are some standards of care not specifically published in the codes that fall into the line of engineering ethics and good business.

However, that being said, as we have an obvious disagreement here among competent and experienced practicing engineers in the use of the 5% overstress, it is blatantly obvious to me that it is not a uniformly recognized "standard of care", and, thus, subject to legal scrutiny.

To each his own...

Mike McCann
MMC Engineering

 

[frv]

hokie66 is absolutely right. Failures (at least catastrophic collapse failures) are a result of gross negligence, not something being overstressed 5%.

It is ludicrous to believe we know the true state of stress in any member, mush less to believe that we know that a member will be "dangerous" because it is 5% over stressed based on our assumed loads. Again, these loads are almost always conservative (even live loads). If I have a member 5% "overstressed" in design, I would sleep very well at night knowing that my assumptions have been conservative.

Think about it. Your typical metal roof deck load is usually taken as 20PSF. In reality, it is closer to 15PSF (including MEP). A 5% increase in dead load is 1 PSF!. Same applies to Code LL for roof loads. If any building fails because you failed to take into account 2 PSF, there is something fundamentally wrong with your design.

 

[civeng80]

Whilst I fundamentally agree with frv, mssquare48 and hokie66 on this from an engineering sense, you have to look at it from a lawyers point of view.

Of course I agree that the aim is to try to stay out of court. Sometimes thats not within our control and out of the blue so to speak you get a writ in the mail to appear in a court of law.

A building will not fail by being 5% overstressed, but just imagine being cross examined under oath by a QC (Queens Council) who finds out that technically you the responsible designer have deliberately and knowingly overstressed a member and gone against the Code of practice and therefore against the building act. Believe me its a minefield for a Barrister at law.

They use everything and anything to win the case. I would keep everything strictly within the code of practice.

 

[csd72]

The problem is that when these things go to court the first thing they do is check the calculations. It is easier to prove that the engineer messed up rather than that the builder did.

Take the recent bridge collapse as an example.