Need a book? Engineering books recommendations...

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

RE: Structure Magazine - Computers and SE Judgement article

[Subject Prev][Subject Next][Thread Prev][Thread Next]
I'm not sure Chris, but I think we are in agreement here. My argument is not
how the numbers are processed, but what the results yield. This result
should be compared to what we understand of the original equations and this
is what leads to professional judgement.
For example, I recently designed a concrete lintel for a masonry wall. The
lintel turned out to be 13" thick (URM wall) and 21" deep. The steel was
more that I would have guessed - (4) #7 on the bottom, if I recall
correctly. The opening was 16'-0". At first glance I thought this was an
unreasonable solution worked out by a popular program I used - much more
conservative than I expected.
When I reviewed the numbers I understood why. The load originating at the
roof and transferred to the beam by the URM wall (triangular loading) was
not the most critical load applied. The load that created the greatest
effect upon the member originated from a 12'-0" tributary mezzanine which
was connected to the lintel by a ledger. The mezzanine was designed for
light storage - 125 psf.
My point is that my intuition told me that a 21" deep beam was too much -
however, I forgot about the storage load and mezzanine application. I did
what anyone should do - satisfy myself that the results are justified by
checking the input.
I don't believe it was actually necessary to rerun the analysis manually
considering that thousands of others have used the software for years.
Considering this a rather simplistic load pattern it would be highly suspect
for the software to be at fault. The truth was that my intuition had been
correct, but that I failed to consider the additional applied load.

What does this mean? It means that intuition need not be formed by manual
analysis if the tools are trustworthy. If the assumption that the tools are
correct - which requires a leap of faith for many - then our intuitive
judgement can be developed from comparing the input to the output. Overtime
this will provide the repetition we need to develop our judgement.

Nothing can replace the need of education to understand the principles
applied to structures. I just don't believe that the actual arithmetic is
all that important if the tools are accurate. We made this same leap of
faith when we accepted the results of a scientific calculator over a
sliderule or hand analysis. Did any of us really question the output of a
calculator when we started to use one?

Now, to be fair, lets assume that the software was created within the
engineers office and contained a flaw in the programming which yielded
improper logic and thus inaccurate results. If the engineer believes his
software to be faultless, his professional judgement will be built upon
flaws and his judgment will question those result found to be accurate. On
the same token, assume that the constant in an engineers manual analysis is
incorrectly evaluated and used over and over until the engineer is informed
by a peer of his error. Intuitive judgement does not simply occur - it must
be developed over repetitive tasks performed to the same end. Therefore, our
judgement can be flawed if our thinking is flawed.

I wonder if there is more mistakes made by computer programs than made by
arithmetic?

Personally, I believe these same arguments have been debated any time a new
method develops. Was it difficult to accept an Abacus, Sliderule, Calculator
or computer? I would think so although the complexity of the tool did not
become questionable until the electronic era.

It is important to understand that I do not endorse any person who does not
posses the skills taught in an engineering ciriculum to interpret
engineering results and formulate a non-professional judgement. The
education is the key and the work experience under a qualified professional
is what is needed to develop proper judgment. This can't be replaced by any
mechanical calculating device.

Dennis Wish PE