Speaker
Description
The growing complexity of coupled flow-deformation processes in geosystems calls for experimental methods that resolve hydro-mechanical responses with spatial and temporal detail beyond the reach of traditional instrumentation. Conventional element-level laboratory tests rely on point-based sensors and therefore cannot resolve how deformation is distributed along a specimen. As a result, tests are often interpreted under representative elementary volume (REV) assumptions and struggle to capture localisation, anisotropy, and heterogeneity. This study demonstrates how distributed fibre optic (DFO) sensing can overcome these limitations by enabling element tests to be interpreted and exploited as boundary-value hydro-mechanical experiments, providing a new level of observability for porous geomaterials.
A high-pressure triaxial device was developed at EPFL (Fig. 1a) to conduct long-duration multiphase flow experiments on Opalinus Clay, the host rock selected for the Swiss radioactive waste repository. DFO sensors installed along the specimen surface delivered more than 1500 spatially distributed strain measurements with sub-millimetric spacing under high-pressure conditions (Fig. 1b). A dedicated data processing workflow, based on machine learning outlier detection, converted raw data into reliable strain profiles.
During water resaturation, injected from both ends, DFO measurements captured the advance of the saturation front through symmetric swelling strains propagating toward the specimen centre. During subsequent gas injection, deformation localised near the gas entry region and evolved with time, directly capturing the coupling between porewater displacement, gas migration, and the mechanical response of the clay (Fig. 1c). Importantly, spatially continuous strain data enabled direct observation of bedding-induced anisotropy, revealing deformation modes and evolving gradients that would otherwise remain unseen with standard point measurements.
By extending spatial resolution far beyond conventional sensing, DFO transforms geomechanical element testing interpreted under REV assumptions into direct boundary-value observation. This shift provides additional constraints for constitutive, numerical and data-driven model development and offers a step change in experimental geomechanics. The approach improves the interpretation of coupled processes and opens new pathways for the analysis and modelling of complex geomaterials across a wide range of subsurface engineering applications.
| Country | Switzerland |
|---|---|
| Acceptance of the Terms & Conditions | Click here to agree |
