ENGINEERS ARE failing to design appropriately for seismically active zones because they do not speak the same 'language' as the seismologists whose data they use, a leading Portuguese engineer claimed last week.
Misinterpretation of seismological information by engineers using non-linear analysis techniques often results in significant under-design, Lisbon's National Laboratory for Civil Engineering principal research officer Professor Ricardo Duarte said last week.
Duarte told members of the Society of Earthquake & Civil Engineering Dynamics (SECED) that the engineering profession must develop a mind-set and language in common with seismologists if it is to deliver structures that are safe but not excessively expensive.
Drawing on the cognitive, linguistic, logical and mathematical theories of philosophers WV Quine and Ludwig Wittgenstein, Duarte argued that data should not be treated as objective fact, but rather as subject to a wide range of subjectivity and variability.
'We should not talk about probability, but about the probability of probability, ' he said.
'Probability is like language: there are degrees of uncertainty.'
The concept is well accepted in seismological circles. But engineers tend to think and express themselves in more absolute terms.
Applying this approach to hazard analysis, risk analysis, structural analysis and design analysis would give a spectrum of performance parameters.
While the mid-range of these fields of probability would normally match the data selected for use by engineers, the upper bands of the probability spectrum would give engineers more extreme criteria to design to.