ABSTRACT
The evaluation of the stochastic character of variables involved in structural analysis has been a subject of interest over recent years (JCSS, 2002). Such as approach allows for a more effective consideration of the design influencing parameters and for the establishment of sensitivity analysis criteria. Consequently, methods of level 3 and level 2 (within the structural safety framework) are then used for the calibration of partial safety factors, thus characterising the limit states semi-probabilistic approach (Leporati, 1979). However, it should be pointed out that the nature of the variables involved cannot be described by means of their statistical characterization only (Möller et al., 2002). A theory that handles other kinds of uncertainty, like the theory of fuzzy sets, also needs to be considered. For this purpose, in the past, convex sets and interval analysis have also been used (Moore, 1966 and Ben-Haim & Elishakoff, 1990). The theory of fuzzy sets enables the mathematical consideration of uncertainties that otherwise could not be taken into account within a conventional structural analysis. The uncertainties are essentially the lack of statistically representative samples for the evaluation of design parameters, the lack of conformity
of conventional tests on materials, the uncertainty in the correlation between the different variables, the difficulties in the evaluation of effective limit state attainment, and the human error influence. Finally, the merging between probabilistic and fuzzy sets theories produces the fuzzy probabilistic analysis.
