ProSwell

Despite their importance in geotechnical applications and in geological processes, the swelling behavior of rocks is still insufficiently understood, while its effects at the scale of engineering projects are significant. This joint research will bring together the experimental and modeling skills of French and German partners to characterize, understand and model the macroscopically observable effects of rock swelling from microscale processes. The planned research will focus on clay-sulfate rocks, where swelling of the clay phase and chemical swelling can occur together and influence each other. If the volume increase is hindered, high swelling pressures can occur. The thermal (T), hydraulic (H), mechanical (M) and chemical (C) processes that take place in this process, and in particular their mutual influence (THMC coupled processes) are not yet sufficiently understood. Consequently, construction projects in clay-sulfate rock today are still prone to imponderable risks. The overarching goal of the proposed research is to generate experimental results that quantitatively characterize the THMC coupled swelling behavior of geological material, using clay-sulfate rocks as experimental material, and translate those results into constitutive equations usable in numerical simulations at the geological structure scale. To reach the project goals, we have structured the planned research into four tasks. The first three tasks are experimental and will be used to inform the engineering models developed in Task 4. Task 1 includes the preparation of the samples, i.e., natural samples of clay-sulfate (anhydrite) rocks. We will characterize in Task 2 how those samples swell and their permeability evolves when water flows through them. These experiments will be performed with flow-through cells. Task 1 will also include characterization of the samples before and after the swelling experiments in Task 2, to detect mineralogical changes. Both quantitative mineralogical analysis (by X-ray diffraction, Rietveld method, and thin section analysis) and chemical composition analysis (by X-ray fluorescence spectrometry) will be performed. Fluid samples taken during the swelling experiments of Task 2 will also be analyzed (by microwave plasma atomic emission spectrometry and titration). To understand and interpret the processes behind the macroscopically observable swelling behavior, advanced imaging of the samples at a smaller scale will be performed in Task 3. We will use X-ray computed microtomography (XRCT) and magnetic resonance imaging (MRI) to make direct observations of pore structure and fluid flow and their changes during swelling. Moreover, chemical reactions can be tracked in space and time contemporaneous with swelling by observing changes in pore space, density and water content. These experiments will be complemented by scanning electron microscopy (SEM), mercury intrusion porosimetry (MIP) and, potentially, NMR cryoporometry. The results will enable us to tell where and when both fluid flow (e.g., rock matrix versus discontinuities) and chemical swelling (location and time of conversion) occur, and how they are related to mechanical (e.g., density/volume) and hydraulic (e.g., porosity, flow rate) parameters. Data gathered in Tasks 1, 2, and 3 will serve as a basis for the model development, calibration, and validation in Task 4. Elastoplastic THMC constitutive laws will be either calibrated on the set of data obtained in Tasks 2 and 3 for the specific rock tested, or predicted by Fast Fourier Transform from the microstructural observations performed in Task 3. These constitutive laws will be implemented in two open-source finite elements codes that make it possible to perform THMC calculations (code Bil and OpenGeoSys coupled with the geochemical calculator PhreeqC) and validated on the flow-through swelling experiments of Task 2. The two numerical approaches will be compared with each other.