yield stress and steel design

huangyhg

yield stress and steel design
hello all,
i am having an issue at work that i need some input in. the manufacturing company i work for gets it's steel from a number of different sources. apparently, our 36ksi steel comes in with mill certs from 45 thru 49 ksi. the practice here is to design at the mill cert yield stress, but i remember reading that the design yield stress for steel is the statedfff"> design yield strentgh, not the mill cert test results.
is this accurate? am i unduly penalizing the design if i insist on using 36ksi design when the mill cert comes in at 45ksi? also, if i design using 50ksi steel, but we find a 36ksi beam that comes in at 48ksi, is that an okay substitute? (strength wise it would be okay).
thanks in advance for your toughts.
smithj,
it depends on what type of steel members you are getting. for example most of the wide flange beams come in astm 992, grade 50. if you request a36, you will get grade 50 with mill cert that is dual certification.
you can design to yield of ultimate stress as long as you use safety factor that is acceptable. i am not sure what steel   
per aisc, the value fy to be used in design is the specifiedfff"> yield strength - not the actual.
jae, where did you find that?
aisc defines fy in the nomenclature table.  as jae says, it's the specified yield strength.
indirectly in asd, and directly in lrfd, usable design strengths are determined statistically. if you pick off the top 50% of the sample, then you have drastically lowered the mean and the 95% confidence level. as a consequence, the formuli and design factors that have been developed to use these design strengths are no longer valid.
for a comparison, consider wood design. wood is a hugely variable natural product, and even the older standard test samples (1"x1"x12"of clear wood) can have failure stresses at the 5% confidence levelbe several times the value at the 95% confidence level. sawn wood is divided into grades such as select structural, #1, #2 and so on, based on a certain level of defects, and given its own set of design stresses.
there is also a grade of "#1 and better", a random mix of select and #1 materials, with its own set of statistics, and with design stresses between those of #1 and select - even though some of the pieces in that group only qualify as #1. (note that this is not the same as sorting through the pile at the lumber yard to pull out the sticks marked "select".) hand-picking the select   
haynewp - it is simply the definition used in the aisc specification - if you look under chapter e for instance, i believe under the formulae for columns there is something like:  fy - specified steel yield strength.
i know i've read something about this before - relating to specified yield vs. actual - but i don't recall where - perhaps its somewhere on the aisc website.
so from what i am reading, it seems that we can only use the specified minimum yield stress as defined by code. if i design using 50 ksi flat bar, then plate that has coupon strength of 48.5 ksi is not acceptable because  it is still a36 steel.
now my issue is convincing the company i just joined to change it's standard practice. this is one of the times i wish i was still in the consulting world. is there anyone else out there in manufacturing that has a similar difficulty?
thanks,
js
smithj - you may find the information in this report, "variability in tensile testing of structural shapes", useful for your situation. here is a link
so would it apply to this situation as well?
i am reviewing drawings during the time when a992 was being phased in, and the drawings specified wf beams to be 36 ksi and the dual certified a992 was actually supplied. i can't check the beams using 50 ksi?
haynewp
i would think that if you originally specified the steel to be a36, but as the project progressed, the contractor supplied a dual certification steel, then it would be up to you as eor to accept this alternative product as a substitution to the a36.
now once that happens (the eor saying ok to dual) then the new specified steel yield could be considered as 50 ksi.  but the original poster was implying (i think) that they were looking at the mill certs and simply using an actual yield...like 48.3 or something.  
steel must meet the full requirements of its associated astm criteria (which is more than just yield).  and in doing so - is classified as falling under a particular type of material that is either acceptable to the eor or not.  and each astm has its related and specified yield value that must be used in design.