Need a book? Engineering books recommendations...

Return to index: [Subject] [Thread] [Date] [Author]

Re: History of Fy in A36

[Subject Prev][Subject Next][Thread Prev][Thread Next]
Harold Sprague wrote:
> This is a problem that crept up over time.  As the tendency of steel manufacturing evolved from ore to recycled steel, there was a noticeable steady increase in yield strengths.  This then became the marketing impetus to move to the dual standard in the 90's.  I recall in 1988 canvassing various steel suppliers what the difference was between Gr 36 and Gr 50 steel.  It was then that I found that some fabricators had no price differential.
> The mill test reports in the late 70's showed A36 steel Fy as 45 to 50 ksi.  This made sense when you consider developing a standard deviation and assurance of a Fy minimum of 36 ksi steel throughout a rolling.  The exercise is similar in developing the f'cr for concrete where you develop a standard deviation for concrete strength.
> Many steel fabricators maintain records for several years.  I know of older firms that have project files from the '20's.  These records could be researched to establish trends.
> This would be a good University research project funded by AISC.
> Regards,
> Harold Sprague
> KL&A
> -----Original Message-----
> From:   Stan J [SMTP:hawneng(--nospam--at)]
> Sent:   Monday, June 01, 1998 6:41 PM
> To:     seaoc(--nospam--at)
> Subject:        History of Fy in A36
> We are now all well aware of the problem of A36
> steels with yields closer to 50 ksi than to
> 36ksi.  While most of us were oblivious to this
> problem until Northridge; when did this problem
> actually develop?  Had strengths of supplies
> gradually crept upwards over the decade before
> Northridge?  Was there a sudden dramatic change
> during the 70's?  Does anyone have info that could
> be useful in predicting reasonable upper limits on
> Fy based upon time in which the steel was erected.
> Thanks
> Stan Johnson, PE,  This, of course, can be a tough
> one to explain to the layman.  How do you tell a
> building owner that their steel may be too
> strong?!
> BTW:  My problem does not involve 50ksi columns vs
> 36 ksi beams and strong column weak beam criteria;
> but instead, evaluating a section for compactness
> criteria.
The mill test results from the 70's, are you referring to the Galambos
and Ravindra paper "Properties of Steel for use in LRFD", that appeared
on the ASCE Journal of Structural Division, September 1978?

This paper is indeed, a very informative paper.  The mechanical
properties, specifically Fy, was determined using statistics available
at the time, not collected at the 70s.  

On p.1462, 
"...Mill test data on ASTM A7 and A36 steel are listed in Table 2. 
These data are for rolled U.S. structural shapes, with the coupon taken
from the webs.  They represent many shapes, a time span of some 40
yr,.....and do not include thick plates"

Table 2 showed three sources with 3,794, 3,124 and 400(approx.) samples
respectively.  The first two (3794 and 3124) are from A7 steel with Fy
=33 ksi.  In other words, only approximately 400 samples out of the 7300
samples of these data were from A36 steel.  Keep in mind that these
values have become the basis of the LRFD used TODAY.

Furthermore, the paper on p.1465 stated the following:
....there is an effect due to the material thickness, due to the
straining in the inelastic range and due to differences between various
mills, in fact, differences in the same mill with time........The
designer has no control over any of these.......

Keep also in mind that the statistics was prepared for gravity and wind,
elastic design, not for seismic design where the materials are expected
to perform in ductile manner well into inelastic range.  The authors did
state on p.1460 that "...The importance of the material statistics may
be, and often is, overshaodowed by the uncertainties inherent in design,
and this paper should be viewed in this context.."

With regard to the more recent steel, I am troubled by a test done at U.
of Texas by Mike Englehart showing significant variation of strength
within a given cross section.  I believe this variation was published by
SAC.  Too bad I cannot find the source at this moment at this time.

Y. Henry Huang