Frederik Eriksen soutiendra sa thèse le 31 janvier à 14h
Titre : "Flow in evolving fractures and porous media / Ecoulements dans des milieux poreux en évolution"
Thèse en cotutelle, Université de Strasbourg et Université d'Oslo. Projet ITN FlowTrans
Jury members:
- Dr Harold Auradou, DR CNRS, FAST, University Paris-Sud, France, rapporteur
- Pr Anke Lindner, PMMH, University Paris Descartes, France, rapporteuse
- Dr Bjornar Sandnes, University of Swansea, UK, examinateur
- Dr Laurence Jouniaux, DR CNRS, IPGS, University of Strasbourg, France, examinatrice
- Pr Knut Jorgen Maloy, Physics Dept, University of Oslo, Norway, dir.
- Dr Renaud Toussaint, DR CNRS, IPGS, University of Strasbourg, France, dir.
- (co-directeur de thèse, en Norvège: Pr Eirik G Flekkoy, University of Oslo, Norway)
Abstract:
During many subsurface processes in nature and industry, the flow of fluids transforms the surrounding porous medium (reservoir) containing them. Depending on the involved mechanisms, these transformation processes can be fast, lasting only a couple of seconds or less, or slow, lasting from a few hours to several weeks or more. The project of this thesis focuses on both the fast and slow processes of transforming reservoirs due to fluid flow, which are important aspects in for example contaminant transport, improved oil and gas recovery, carbon storage, enhancement of water well and geothermal energy production, or subsurface sediment mobilization. Based on tabletop experiments and numerical simulations, we study the flow of fluids in evolving fractures and porous media; how the flow drives the transformation of the media, and the interplay between the transformation and the flow behavior itself.
Fast transformation processes of reservoirs include sudden mechanical deformation and fracturing due to high overpressure in the pore fluid. In cases where the fluid pressure is high enough to open fractures in the solid, a flow instability (Raleigh-Taylor) is initiated where the fluid flow concentrates into the longest fractures, where there is higher permeability and larger pressure gradients on the tips, such that the more developed fractures propagate on expense of the less developed ones. The stabilizing mechanism of this flow behavior is the resistance of the surrounding material to deform further, which increases in confined reservoirs when the material is compacted during deformation. We study this phenomen by performing lab experiments where we inject air at a constant overpressure into saturated or dry granular media confined inside Hele-Shaw cells. This simplified system is a quasi-2-dimensional rock/soil analog confined between two glass plates. The setup is optically transparent, which facilitates direct observations during experiments, and we record the flow and deformation processes with a high speed camera at a framerate of 1000 images per second. Post-experiment image processing and analysis of the resulting image sequences yield information about the growth and shape of the fracturing/invasion patterns over time. We use the results to characterize typical properties of the emerging structures, such as their fractal dimension, typical thickness and invasion depth, scaling exponents, growth dynamics, as well as flow regimes depending on injection pressure and boundary conditions.
Furthermore, we have tracer particles in the granular packing which provide information of the surrounding displacement and deformation in the porous media over time. This information is quantified by processing the image sequences with frame-to-frame image correlation software, and enables us to characterize the deformation in the material and how it evolves over time. The driving force of the deformation is the pressure gradient across the solid, and we calculate the evolution of the pressure field in the porous medium numerically. In the immiscible case, e.g. when air is injected into a water saturated porous medium, the overpressure is dissipated across the saturated medium, from the overpressured air cluster towards the less pressurized outer boundary. This is a steady-state configuration of the pressure field, which is estimated by solving the 2D Laplace equation with fixed pressures in the air cluster and at the outlet boundary. On the other hand, in a porous medium where the same fluid is injected, the imposed pressure initially diffuses into the pore space with a diffusion constant that is calculated by knowing the porosity of the material and the viscosity of the fluid. This pressure field is calculated over time by solving the 2D diffusion equation with the Crank-Nicholson method. Eventually, the pressure field reaches steady-state, so the diffusing pressure field is compared with the corresponding Laplace solution at each timestep to determine when this happens. By comparing the obtained deformation data and gradients of pressure fields across the porous media, we develop theoretical models of the deformation and fracturing dynamics.
In a closely related project, where the work in this thesis is also involved, acoustic emissions during experiments with air injections into dry granular media are recorded with piezo-electric shock accelerometers. By fourier analysis of the obtained signals, the evolution of the emissions in the frequency domain is characterized during the different stages of deformation and fracturing, and the processes creating the various types of signals are discussed in comparison with the optical data. In addition, microseismic events caused by particle rearrangement in the later stages of the experiments are counted and fitted with real world events. In combination with the optical data of deformations, an energy based localization technique of the sources of these microseismic events is developed.
Slow transformation processes of reservoirs include chemical evolution of existing fracture networks due to reactions between the flowing fluid and host rock, i.e. dissolution and precipitation. Such processes further increase or decrease the permeability of the system, and may change the stresses in the material surrounding the fractures. In the study of such slow processes, we perform reactive flooding experiments through fractured carbonates. Chalk samples, cylinders of 3.8 cm diameter and 1.7 to 2.5 cm in length, are fractured in two pieces through their principal axis, by loading the samples across their diameter in a brazilian strength test apparatus. The fractured samples are put back together, mounted inside a Hassler cell (cylindrical confinement), and flooded with a reactive fluid. Here, we inject distilled water at constant flow rate of 0.1 ml/min through the fractured samples for various amounts of time. A local topographical profile of the fracture surfaces in the samples are measured before and after the reactive flooding by using a white light interferometer, giving a height resolution of a few microns. By analyzing the measured topography and comparing the before/after measurements, we investigate the local evolution of roughness and fracture aperture after different durations of flooding, and investigate the impact of flow direction on the changes. For the study of the surrounding stresses we are constructing a setup to load the samples parallell and normal to the fracture plane during flooding, in order to measure the evolution of stresses in the material depending on these orientations.
The work done in this thesis project is ultimately a part of further understanding the fundamental mechanisms of complex processes involved in transforming reservoirs due to fluid flow, and is also involved in the development of acoustic localization techniques which may have industrial applications.