Advanced constitutive modelling for soft rocks

The investigation of the soft rocks macro-behaviour through both standard and dedicated laboratory tests has become more and more important and has triggered the interest towards the formulation of constitutive models specifically dedicated to them. However, in some cases, the analysis of the macro-structural features of a soft rock has been combined with the analysis of changes occurring at the micro-scale, highlighting many similarities between soft rocks and natural clays, and therefore, opening for the possibility to explore the applicability of constitutive models formulated for structured clays to the case of soft rocks. In this research, two soft rocks have been taken into consideration, since they have been both studied at the micro- and macro-scale: a soft rock from Abu Dhabi and a soft rock extracted in Malta. Both rocks are characterised by bonding and cementation that progressively degrade with the applied load already at rather small strains. This indicates the necessity of adopting a constitutive model that accounts for at least two surfaces, i.e. a small elastic nucleus and a larger surface representing the locus of gross yield points. These features are common to many multi-surface constitutive models with mixed kinematic and isotropic hardening rules and destructuration, proposed mainly for natural clays rather than soft rocks. Constitutive models for soft rocks are often characterised by a single large surface representing the effect of structure, within which the behaviour is purely elastic. The constitutive model chosen for simulating the behaviour of both soft rocks presented in this research is the RMW model, by Rouainia and Muir Wood (2000), whose features have been extended to account for an increased stiffness of the natural material and different amplification factors for the contribution of plastic volumetric and deviatoric strains to destructuration. It is necessary to distinguish between the model parameters that can be rigorously determined based on experimental results, and the ones that cannot be associated with a specific measurable quantity of the material behaviour, but that have a clear meaning in the model equations. One of the purposes of this research is to provide a clear and comprehensive methodology for calibrating the model specific parameters, with a direct application to the case of soft rocks, but that could be easily extended to the case of other materials and, even more generally, other constitutive models. In addition, this research investigates the capabilities and limitations of a model commonly used to simulate the behaviour of natural clays, to capture the distinctive features of these soft rocks.