Mill Certificate vs. Design Yield Strength 2

[ekelley]

During a luminaire pole submittal review I noticed a fabricator had specified the material for a square HSS as ASTM A500 Gr. B. Rather than listing the yield strength as 46ksi, they indicated it as 55ksi on the drawing and used 55ksi as the yield strength in the calculations. I commented on the error, noting that the element was not structurally sound when using a yield strength of 46 ksi in the calculations, and returned it.

I received the following in response: "We buy the steel to be certified from the mill to have a minimum 55 ksi for the member. It meets the chemical properties of ASTM A500 Gr. B, but is tested to higher yield."

I'm not very happy with the response. I don't know if the typical yield strength is 55 ksi or not. I know the yield performance of this material is not bracketed like an A992 steel would be, but is the performance affected by the higher yield? The resistance factors (or allowable stress factors) specified in the code are based on the probabilistic properties of the material, right? Is using the precise yield strength indicated on a mill certificate a bad idea for design? It seems they are taking a shortcut rather than just sizing the post appropriately. At a minimum, I would request the mill certificate, but I would rather reject it. How would a typical resident engineer know to check that the mill certificate yield strength exceeded the yield strength spec'd for the material? Seems sneaky to me. Has anyone had any experience with this sort of thing?

Thanks!

 

[BigH]

desertfox - when you have 20,000 tons or more of steel rebar being delivered and much of it hasn't even been manufactured before the design is finalized and construction started - this creates a problem. I would agree with JAE - design as per the specified strength or with the caveat that a specific "non-specified" value be used - one that you can control and the supplier would honour.

 

[desertfox]

Hi BigH

point taken, I don't work with rebar so I won't have that problem

 

[weab]

I didn't intend to include rebar in my discussion. But since the subject is raised, I would say that this is similar to substitution of higher grade rebar. As long as there are no rho max issues and a higher grade rebar has been unknowingly substituted by the contractor, I would have no problem designing at some future date based on the actual installed rebar strength.

JAE- I like the "geeky-technical-literal" issue that you bring up - and it's important. I'm at home and don't have a code with me. But is there a definition of "specified"? If not, then the spec yield is the specified yield. If I have an actual yield, I see no reason not to use actual yield because it is now SPECIFIED.

I absolutely don't agree about the standard of care. I'll leave it at that.

Now with that said: I have never heard of using mill certs in original design, only as design and construction progressed, and extra strength is required.

 

[desertfox]

Hi weab

I've seen it done not in your line of work but for components in mechanical engineering.

desertfox

 

[IFRs]

I have used mill certs for plate products and would now hesitate to use it for shapes for some of the reasons given above. I have found this to be a very interesting and informative thread. Thank you all!

 

[BigH]

IFRs - when you do, do you take in the statistical aspects of using a "single" specimen mill certificate? The specimen tested might very well be on the "high" side of the representative average.

 

[dhengr]

I specifically said that I didn’t make a habit of designing to the mill certs. The usual situation was that I had done a prelim. design, so we could mill order steel, then when doing the final analysis, design and detailing, we might come across a highly stressed area which I hadn’t foreseen, and we were thankful to find that the mill cert. many times gave us the 3 or 4ksi extra Fy that we needed to stay within allowable design stresses which were usually Fy and/or Fu. We were dealing with large orders of steel and large pieces of plate, 2 or 3" thick x 10' wide x 30 or 40' long, or 5 or 6" thick x 4 or 5' wide x 30 or 40' long. And, we did sometimes ask the mill to give us mat’l. that tested toward the upper limit of a spec. for Fy and they would try to do that, and then notify us if they were having trouble meeting their shipping date for that request. If I saw some small, well defined, over stress I might not want to bump the whole plate up a size in thickness. The bigger problem we had with some of this mat’l. was through thickness defects like laminations caused by rolling, since we were ripping these plates and had a lot of through thickness stresses and highly restrained welding details. These types of defects and problems seldom come to light in std. structural work, because most of these defects are parallel to the primary stresses.

And, as I explained our commercial structural fab. shop did not do this, nor did our rebar fab. dept. Obviously, BigH’s example of rebar which is dumped into a bunker by bar size and grade is a good example of where this wouldn’t work very well at all, nuts and bolts from a box or a keg might be another example where individual mill certs. wouldn’t be practical. In fact I do recall examples where we used mat’l. other than that shown on the structural drawings and specs. because we had it in stock and it lead to a connection or detail more to our shop’s liking or the erector’s liking, on a large job with much detail repetition. That was about the only time I really did work with the structural dept. I would get involved doing the joint designs, calcs. and a report which had to be submitted to the EOR to get their approval to change the detail. Furthermore, I knew many of those engineers, otherwise connections came detailed on the plans, or our structural detailing dept. designed their own std. connections.

JAE’s admonition to “follow the code” is well taken, particularly when you have no control over the material to be used other than an ASTM Spec. You could probably be sued these days for using the wrong grade pencil lead for your calcs. too; and the codes and analysis methods are getting so complex that I suspect that there isn’t a design out there that one couldn’t find some fault with if they combed the codes, calcs., plans and actual construction deeply enough. I spend about the same percentage of my time defending engineers as I do arguing against their work, and I don’t often damn an engineer for using well reasoned engineering judgement and experience, that is seldom the cause of a failure or problem. However, the reason “SPECIFIED yield” is used in AISC is because the designer only knows the grade of steel he speced. for his job and his design is done, prior to and, irrespective of the exact Fy that the fab’er. might get in that spec. At the same time the laws of Strength of Materials, etc. do not go out the window or become dormant just because you use an actual Fy which is 8% higher than the spec. minFy, assuming a situation arises where you can make that decision. It could be argued that a prudent engineering working with the care normal to and expected of the profession would not increase one beam, out of dozens of the same sized floor beams, to the next size just because it might be a couple percent over stressed in one small, well understood, area and then stand the chance that it be misplaced during erection. Desertfox and I are often in agreement, and we are in the same camp on this one too.

 

[TXStructural]

An ASTM specification is simply an agreement between the manufacturers, users, and designers who sit on the respective ASTM committee about what they want to produce. I will repeat an earlier statement: There is nothing magical about a number in an ASTM spec. Where a particular spec lists a property, steel can typically be produced and certified to that spec, with a property that is better than the minimum requirement. Generally, ASTM specs cover chemical composition (such as max carbon-equivalence for weldability, limits of materials that increase brittleness), minimum and sometimes maximum yield, yield elongation/ductility, and the like.

Using rebar as an example (all steel production works in a similar way), one mill might produce a ASTM A615 Grade 60 bar which yields at 61 ksi, while another may produce a bar meeting the same spec which yields at 77 ksi. They both meet ASTM A615 Grade 60, and the code would allow design of a structure based on the nominal grade (60 ksi), but would not prohibit design using the actual yield (assuming sample size and QA/QC warranted that increase.) The mill may or may not choose to call it ASTM A615 Grade 75, depending on many factors.

If a designer chooses to design for a known yield strength, and the mill reliably supplies this strength while meeting all other requirements, there is nothing wrong with doing so, as long as that grade is permitted for the use (ACI 318 does limit maximum design strength for certain components.)

The reason we design for ASTM specified values is that we know it can be produced and will most likely be available.

I do not believe it violates a standard of care to use a steel strength which is demonstrated by sufficient testing and QA/QC, since this is how compliance with an ASTM is determined. However, you would not design for 55 ksi yield but specify ASTM Axxx Grade 46 (46 ksi yield) hoping for grade 55. You would specify ASTM Axxx grade 55 (55 ksi minimum yield), even if the ASTM only lists Grade 46 (or Grade B, or whatever applies).

 

[desertfox]

Hi

http://stainlessconstruction.com/ss_info_cntr/pdfs/p291.pdf

Have a look at this link its a code for construction in stainless steel, if you go to section 2.2.2 "Design Values Of Propeties" scroll down and you will find it says you can use the prof stress off a mill spec. and further on it says:-

(iii) Design using mill certificate data
Measured values of the 0.2% proof stress are given on the mill (or release)
certificate. The design value of the 0.2% proof stress can be taken as:
py = sm0.2 / 1.2
where sm0.2 is the average value of the 0.2% proof stress as given on the
mill certificate.
It is suggested that the ultimate tensile strength should still be based on the
minimum specified ultimate tensile strength given in BS EN 10088-2.

 

[TXStructural]

The value shown for yield on US steel may be any of the following:

- halt of the gauge, meaning there is well-defined yield plateau when a sample is tensioned until yield - this is the traditional method of test, but many steels (including stainless and ductile, high-strength steels) do not exhibit well-defined these yield points.

- EUL, or elongation under load, which has been used for rebar here until recently. A value such as 0.35 EUL means that the the yield value presented is the stress at which samples reach 0.35% elongation. Graphically this is determined as the intersection of the test curve for the steel with a vertical line at 0.35% of length as measured from the origin of the stress-strain diagram.

- Percent offset, which is frequently 0.2%. Given a sample with a proportional, elastic curve (essentially all steels), a line parallel to the straight portion of the elastic curve (representing the modulus of elasticity) is drawn 0.2% to the right on a stress-strain diagram, and the point at which this line intersects the test curve for the steel is the 0.2% offset yield strength. (This is the method used by most of the steel industry, including in the cited stainless spec, as well as ASTM A955 for stainless steel rebar.)

The difference between the EUL and 0.2% offset varies from steel to steel, but will be essentially the same for a steel with a well-defined yield plateau. For steel without a well-defined yield, percent offset best represents the actually yield strength of the steel for most applications.
(The current draft of ACI 318-11 attempts to circumvent the ASTM process by specifying EUL as the test method.)

 

[Agricola]

Dear All,

I didn't scrutinize all the responses in detail, but please be aware of the formal meaning of a MTC:
It is a document issued as confirmation that the steel supplier has succeeded in producing the grade mentioned.

The test results are "the evidence" that the production facility still produces within the limits (defined for that specific steel grade) controlled by statisical analysis from the beginning on for that specific production process.

The ultimate consequence is that imho the engineering should be done based upon the minimum guaranteed values as defined for the grade and should not be based upon the reported data on an occasional MTC.

 

[NLENG]

In the AISC ASD design code, Fy, is defined as the specified MINIMUM yield strength for the type of steel being used. Consequently, I would advise against using Mill Test Certificate values as the basis for performing your design.

Nonetheless, I have seen MTC values used for existing structural elements only when there is a great deal of confidence in the design loads and there is full material traceability.