Direct characterisation of transport and mixing in unsaturated porous media

点击翻译按钮获取中文摘要

Multiphase flow and hydrodynamic dispersion in porous media are the key processes in many subsurface and engineering applications, such as vadose zone contamination, soil remediation, and hydrocarbon recovery. To secure sustainable access to clean water and energy, which are identified as the global challenge areas, we urge to enhance our understanding of fundamental processes and improve our predictive capability. Hydrodynamic dispersion in porous media is controlled by the spatially and temporally variable velocity field, described by the advection-dispersion equation for saturated porous media. However, in the case of an unsaturated porous medium (where two or more immiscible fluids are present), the Fickian advection-dispersion equation is not anymore valid. In unsaturated porous media, not only the velocity field varies with the change of saturation topology, but also the spatial variation of velocity can change by orders of magnitude, such that even in homogeneous media some regions become hydrodynamically stagnant. Transport time scales in the stagnant and flowing regions can be different by orders of magnitude, which makes the concentration profiles very skewed. This feature is referred to as the non-Fickian behaviour that varies dramatically with the stagnant saturation and flow dynamics. The existing non-Fickian theories (e.g. the MIM theory) have been used to inversely model the externally-measured concentration profiles. The satisfactory "inverse modelling" results are generally regarded as an indication of the validity of these theories, although this perception has been challenged in the recent 2D micromodel experiments.In this project we articulate the major misconceptions and gaps in our understanding and we hypothesize that the concept of stagnant saturation has been incorrectly employed in the existing theories. The stagnant saturation is a two-phase flow variable, which depends on fluids topology. It should be regarded as the key "variable" to couple two-phase flow and the non-Fickian transport modelling in a hydrodynamically-consistent way, which is absent in the literature. To address the gaps and misconceptions in understanding, we have envisaged a novel experiments and modelling in two work packages (WP). In WP1, we will deliver the first direct visualisation of dispersion and mixing of a tracer in oil-wet and water-wet unsaturated porous media using the fast elapsed-time (4D) X-ray micro-tomography. All experiments will be computationally designed using the in-house developed pore-scale models and the experimental setup will be tested in Henry-Moseley X-ray Imaging Facilities. The 4D experiments are envisaged to take place in the Diamond Light Source (DLS), facilitated by the Diamond Manchester Collaboration. In WP2, the 4D X-ray images will be analysed to extract the concentration field, saturation field at different flow conditions. Transport properties including stagnant saturation, dispersion coefficient, and mass transfer rate will be quantified from the obtained images. Finally, the validity of the existing established models will be tested and a new theoretical framework will be developed. The project will benefit from the collaborations with world-renowned scientists, Prof. M. Celia (MC) from Princeton University, Prof. H. Steeb (HS) from Stuttgart University, and Prof. Brian Berkowitz (BB) from Weizmann Institute of Science and the partial support of two PhD students of the PI's team. MC will bring his key knowledge in multiphase flow and transport in porous media and their applications in large-scale problems of environmental engineering. BB has a renowned track record in mathematical analysis of non-Fickian transport in heterogeneous porous media. HS is an expert in microCT imaging of complex porous media processes and supports the WP1 by offering experimental equipment. Also, the DLS beamline scientist, Dr Nghia Vo, will support the project in data acquisition and reconstruction.