Fy = Specified Minimum Yield Stress 6

[enginerding]

I know how Fy is defined in AISC and I know how we are "supposed to" use that for all our design equations.

I have a condition where an existing piece of steel pipe was tested for strength. The yield strength measured in the test is a touch over 40 ksi. If I assume this is 35 ksi steel (ASTM A53 Gr B or ASTM A252 Gr 2), the section does not check out for strength and requires reinforcements to carry proposed loads. But if I use the actual 40 ksi, the section is OK. Avoiding the reinforcement on the job will save around $20,000.

It seems wasteful to throw $20,000 at a "problem" that is only a problem because of the way a term is defined. Unless I am missing something.

What are some things that I could be missing that would give me reasons not to take advantage of the higher strength measured for this steel?

Or do you think I should just go ahead and use the higher strength?

 

[dcarr82775]

I would be alright using the tested yield with a qualifier. If it is a single column, and that column was tested I would be good with the 40ksi, but if there are a number of columns and only (1) test I might want more testing to verify the other columns have similar properties.

 

[dhengr]

Buy all means do what you are proposing. That’s called good, well reasoned, engineering judgement. We are called on to design safe structures so the member shouldn’t yield, buckle, deflect too much or fail. But, we are also called on to be economical and practical in our designs. You are just using the material to its full capacity or strength. Fy min. is set so suppliers and designers have some basis to group materials into a class that designers can use without testing every plate or piece individually before design. Obviously, all other calcs. and checks on failure and functionality must be done and satisfied, to show a complete consideration of the problem on your part. Otherwise, go for it, that’s called good design.

Save the mill cert. or test report and make it part of your calcs. along with a well reasoned explanation of what you are doing and why. i.e. use material to its full potential and save $20k, for a load which only happens on occasion, and even then does not cause a failure mechanism. Not many good engineers would damn you for that, most would defend you.

I never start my designs thinking I am going to allow a 5 or 10% overstress, any place. But if the top .25" of a 1.5" flange, on a built-up member is slightly above yield under max. load, I might let that pass, under the right conditions. And, I would certainly go look at the mill certs. for that plate and take advantage of the fact that it tested at 10% of the spec. min. Fy. Obviously, there would be cases where you would think long and hard before you did this too.

 

[ToadJones]

well said dhengr

 

[BAretired]

You should not base your calculations on a single test. If you took a series of tests, you would find a certain amount of scatter in the results, so your test of 40,000+ could be an anomaly.

The guaranteed yield is only 35,000 psi and that is what you must use in your calculations. To do otherwise is to accept a risk for a design which you could not defend in court in the event of failure.

To permit a 5 or 6% overstress could possibly be defended on the basis of normal engineering practice.

BA

 

[TXStructural]

There will be variation throughout a single member and sample, which is why certain tests specify things like percent elongation over a sample of a certain length. Wall thickness will also vary. These are the items considered in the "resistance factor".

As said by others, I would not base anything on a single test, on a single sample, from a single member.

The nominal design yield strength assumes the strength is valid for design purposes, using established methods. You should use the design assumptions in the Code, and with the design characteristics of materials.

Are you comparing to existing loads or design loads, factored or unfactored?

 

[enginerding]

Thanks for your thoughts.

To answer TX:
This is a structure that has been in service for a number of years. The client would like to add some eccentric dead load to this structure. This magnitude of load is not enough to cause any problems, but the new wind loads required push this thing beyond acceptable overstress. To be clear, the wind area is not increasing at all (so in reality we are not increasing any wind loads), but this version of the code requires significantly higher wind pressures than the previous version. It is the change in code that makes this overstressed at 35 ksi.

 

[ToadJones]

cue the age old debate on wind loads !!!!

 

[enginerding]

I was not going to argue about the wind loads, but it is only one member, that has been tested, and the test shows 40 ksi - which provides an allowable code check. I was just looking for the thoughts of the community.

 

[ToadJones]

enginerding-
I was just interjecting some sarcasm....or, at least, I was trying to :-D

 

[Ron]

enginerding...the stated yield strengths in ASTM standards are minimums. If you read the entire standard, you will also get a sense of the statistical variation you can expect. If your tested value is above the standard deviation, then you can logically do as you propose, have a rational basis for doing so, and save your client some money. As BAretired noted, don't base anything on a single test. Again, look at the standard. It is based on multiple tests.

My rule of thumb in many years of testing, is that I require a minimum of three tests on ANYTHING! One test can be an anomaly. Two tests can confuse you. Three tests will almost always provide you with a "referee".

 

[mbullism]

Put me down for being leary of a single test... we went along for sometime designing new at Fy=36, knowing full well that if tested it would be close to the 50 ksi that we'd have to pay a premium for at the time... any way to increase your data set?

On a side note regarding your "increased" wind loads, we have a building that was designed "by the book" a few years ago. The code has since changed and the snow load requirements have gone up. The structure is not old enough to qualify as an existing building, so by the letter of the law, it is and will be deficient for the next two years. When it turns five years old, it will qualify as an existing building and be allowed a reduction, at which point it will meet code again...

 

[paddingtongreen]

40 is 14% over 35ksi, a little more than I would allow on engineering judgment alone, and a single test doesn't cut it. I'm always leery of testing unless I am ready for an answer I don't like.

Michael.
Timing has a lot to do with the outcome of a rain dance.