“About twenty-five years ago Battelle was asked to estimate the cost
of fracture to the US economy, we came up with an amount of about a few percent
of gross domestic product. While the
actual figure can easily be disputed, the real cost is probably very
large.
The key to arriving at our result was an appreciation of the role
that material variability plays in engineering specifications. It appears that
typical engineering standards assign properties about two standard deviations
below the mean of alloys when specifying dimensions. For example, if the
coefficient of variation of the yield strength of a particular steel is seven
percent, it is assumed that all pieces of that steel are treated as if their
yield strength is 14 percent below the actual mean. Of course, engineers did not
work this position out statistically -— the various rules-of-thumb just seem to
come out that way. If metallurgists could reduce scatter consistently, we would
have lighter structures with no sacrifice of safety. Extending this idea to
other mechanical properties and plugging our estimates into the economics model
gave rise to the large costs that we reported.
The above argument leads to the non-intuitive conclusion that the
fracture research with the greatest economic benefit would be in the area of
materials production (especially welding). Our report soft-pedaled this point,
since the sponsors had a vested interest in justifying additional funding for
fracture mechanics research.
When I made my point about the importance of production research, at
technical meetings, I was attacked by researchers with the same motivation — they cited all sorts of unsolved analytical
problems and promising experimental approaches. My reply that success in their
approaches would at best lead to a tiny reduction in safety factors only poured
oil on the fire.”
No comments:
Post a Comment