Robert Bonomi wrote:

In article ,
J. Clarke wrote:
Michael Daly wrote:

On 28-Apr-2005, Unquestionably Confused wrote:

To overbuild is never a crime

Overbuilt can cause a failure. For example, if something is designed
and built correctly, it will tend to show excessive deflections before
failure,
providing a warning. Overbuilt things can fail spectacularly without
any warning.

Where you run into the kind of problem you describe is when a strong but
brittle material is substituted for a weaker but more ductile material.

*NOT* necessarily true.

Please provide a case in which replacing a weak ductile material with a
strong equally ductile material results in the kind of failure you
describe. Please include the analysis.

The ductile material will bend before it breaks, the brittle material will
simply break.

As for something "designed and built correctly" showing "excessive
deflections before failure", certainly one can design things that way but
that doesn't mean that it's the only correct way. Concrete for example
doesn't flex noticeably before it breaks so by your reasoning concrete
should never be used as a building material.

When dealing with wooden beams, making the beam stronger than called for
is not going to result in sudden failure with no warning unless the
original design would also fail suddenly with no warning at a lighter
load.

*NOT* necessarily true.

Yes, necessarily true. Please provide an example of a case in which a
wooden beam was strengthened and there was subsequently a failure with no
warning while an identical structure subjected to the identical loading
gave warning. Please prove that this was the case, I don't want someone's
opinion.

Engineering for systems under stress, particularly dynamic stresses, is a
_complex_ and _complicated_ subject.

Buildings of the kind where an individual would be installing or removing a
beam are not typically under "dynamic stresses" to any significant extent
unless you want to count wind loading. If you want to talk skyscrapers
it's another story, but they typically have little or no wood in the
structure.

*ALL* the components have to be considered, =both= singly, and in
combination. 'Over-building' _one_ component can result in excessive
transfer of stress
to _other_ components,

How can it "result in excessive transfer of stress" if the loading is the
same? Please demonstrate the mechanics of this. Show me an analysis of a
case where under identical loading increasing the strength of one member
increases the stress in other members.

Leading to failure of _that_ component under
conditions that are _less_ severe -- as measured for the overall system --
than the original design was spec'ed to handle.

Again, show me an analysis that demonstrates that this happens.

There are numerous real-world instances of this *exact* thing happening.
One of the easiest places to find them is in the world of home-built,
plans-
built, aircraft. Firstly, in general, the 'safety margin' on _any_
aircraft
design is extremely small. "1.6" is typical for commercial construction.

I was not aware that the OP was talking about an aircraft. Houses,
workshops, and other buildings typically have much higher margins than
that.

Homebuilts usually are designed with higher margins, because there is more
variability in the quality of construction. However, there are =many=
cases on record, including after-the-fact engineering analyses, where a
home-builder has modified a design -- to =strengthen= some part of it --
where said mods
have led to _premature_failure_ of other areas of the design. Higher
"point stresses" occurred in the modified design, as a result of the
modification, than the original design was designed to handle.

Now why would "higher point stresses" occur under the same loading? That's
a matter of forces and geometry. While I probably could design a structure
in which stiffening one member increased the static stress somewhere else,
I'd have to work at it.

What usually happens in such incidents is that the stiffness of a structural
member was changed, resulting in an altered natural frequency, which put it
into a range to resonate with shed vortices and there by causing a flutter
problem. But putting a heavier beam than required in a house is not going
to cause a problem such as this.

Look, the bottom line on this is that you seem determined to overengineer a
simple problem like spanning a doorway.

Show us how to make a house fall down by making the headers too strong and
maybe someone will listen. In the meantime you're just crying gloom and
doom to no purpose.



--
--John
to email, dial "usenet" and validate
(was jclarke at eye bee em dot net)