ABSTRACT

ABSTRACT: Some common ideas concerning the correct way to model the constitutive behaviour of unsaturated soils are discussed. Neither theoretical considerations nor experimental evidence may support definite conclusions on the stress state variables that should be adopted in the formulation of constitutive laws. Different approaches may be followed in the constitutive modelling, each of them presenting advantages and shortcomings. Constitutive models for unsaturated soils are usually formulated in terms of two separate stress variable. However, the average soil skeleton stress, defined as the difference between total stress and an equivalent fluid pressure with saturation degree as weighing parameter, may be equally adopted advantageously. The latter choice allows for a natural reproduction of some aspect of the overall mechanical behaviour. Nevertheless, additional constitutive parameters are usually necessary to model the increase in the preconsolidation stress with suction and collapse upon wetting. The main criteria for the development of a model written in terms of a single stress variable are recalled. These criteria are adopted to adapt Modified Cam Clay for unsaturated states. A comparison is then presented between the latter model and the one proposed by Alonso et al. (1990).