Probabilistic and possibilistic methods for the elastoplastic analysis of soils

The finite element method is extensively used for solving various problems in engineering practice. The input requires properties, which are generally imprecise. In this paper an elastoplastic analysis (Drucker-Prager yield criterion) is performed, and the properties are considered as uncertain parameters (elastic modulus, Poisson's ratio, cohesion and angle of internal friction). Two different methodologies are studied and compared: classical probabilistic approach in which the properties are treated as random variables: stochastic finite element methods using Monte Carlo simulation. possibilistic approach, by a model based on the fuzzy sets theory. Some results are presented to point out the main characteristics of the two methodologies. [Lima BSLP. DSc thesis, (1996).] (C) 2001 Elsevier Science Ltd. All rights reserved.