Failure of cohesive geomaterials : bridging the scales – Geobridge

Within the GeoBridge project, we aim to develop several advanced and original experimental techniques that allow for: 1. assessment of the transition from diffuse to localised deformation or fault initiation in rock specimens under realistic 3D loading (three non zero principal stresses), using an innovative loading apparatus and displacement / strain field measurements. 2. quantification of the kinematics at the scale of the grain and characterisation of the micro-mechanisms of deformation at this scale such as inter and intra-granular cracking, through in-situ tests in our X-ray tomography apparatus and correlated 3D image analysis; 3. monitoring of damage activity from acoustic emissions generated by micro-mechanisms, to complement the strain field measurement of experimental objective 1; 4. measurement of local mechanical properties such as elastic stiffness, and their variations due to damage, using ultrasonic tomography; 5. association of permeability and damage, and evaluation of the heterogeneity of fluid-flow properties field in relation to failure, by tracking the passage and shape of fluid fronts through the deformed material. All these new experimental results will nourish the modelling activities of this project. With respect to the modelling activities of this project, several issues are to be considered in relation to the scale effects in geomaterials, as outlined below. 1. For the failure at the macro scale, we will address how to model the two key stages of failure: the initiation of localised deformation/damage and subsequent evolution to a fault. To achieve this requires models that take into account the microstructured aspect of geomaterials, both to model the localisation band (it is well known that the thickness of such a band is controlled by a characteristic length of the microstructure) and to model the fault propagation (which results from a paroxysmal concentration of processes at the small scale). 2. How to make the bridge between the micro- and macro-scales. Models that capture both the micro- and macro-scales need to be constructed using homogenisation schemes within which the pertinent micro-mechanisms of deformation (which will depend on the geomaterial) can be described. Pertinent mechanisms can be micro-crack propagation in an intact continuous matrix (low porosity rocks), or intra- and inter-granular cracks in cohesive granular rocks (high porosity rocks). We aim to develop several homogenisation schemes using asymptotic or numerical homogenisation, depending on the complexity of the mechanisms to take into account. In both cases, behaviours at the micro scale will result from numerical computations (FEM or DEM). 3. How to take into account the hydro-mechanical (HM) coupling and link this with the failure problem. A degree of complexity will need to be added to the previous items to get a macro-scale HM model, this will be achieved either through HM coupling at the micro-scale or by considering the problem of strain localisation or fault initiation in a fluid saturated or partially saturated medium using enriched continua approaches. These theoretical and numerical developments will be performed on two major classes of models, i) microstructured media such as local second gradient models for objectives 1 and 3, and ii) double scale models based on homogenization schemes with several micro structures, for objectives 1, 2 and 3. Experimental results will serve both to define the pertinent micro-mechanisms involved and to calibrate the models by comparison of the numerical and experimental results.