A first-principle modelling framework for thermo-hydro-mechanical-chemical processes in fractured rock is developed. The key characteristics of such rocks are incorporated by direct representation: the power law nature of fracture size distribution, the fractal nature of rock surfaces, the scale dependence of discrete contact between these, and the mechanically induced fluid-flow anisotropy. It is shown that hydro-mechanical modelling in fractured rock can be reduced to a single free parameter, the friction coefficient. Properties of fractured rock, such as transmissivity, stiffness and permeability, emerge as direct results of physical processes, and need not be complemented by empirical relationships or distribution functions. Results reproduce the field observation that critically-oriented fractures are likely hydraulic conduits; while the established reasoning of shear activation and associated fracture dilation for this fact is confirmed, it is also shown, for the first time, that chemically-mediated processes have the potential to contribute to this effect. Compaction mechanisms of pressure solution and precipitation act preferably on fractures with stress ratios far from Coulomb failure, which increases the relative contribution of near-critical fractures. New insights also emerge as to the likely orientation of the maximum permeability of a fractured rock mass. Due to shear-induced anisotropy of fracture transmissivity, the preferential flow direction tends to be aligned with the orientation of the intermediate stress. The capability to accurately determine the upscaled permeability tensor is facilitated by a novel algorithm that is independent of geometry and reference system, and accounts for locally anisotropic matrix permeability and fracture transmissivity. The developed framework presents a path forward in fractured rock modelling. It accounts for state of the art knowledge of the internal processes of fractured rock across multiple scales, and perhaps aids to overcome the necessity for using isotropic effective properties that is prevalent in many applications.
Identifer | oai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:712882 |
Date | January 2016 |
Creators | Lang, Philipp |
Contributors | Zimmerman, Robert W. ; Paluszny, Adriana |
Publisher | Imperial College London |
Source Sets | Ethos UK |
Detected Language | English |
Type | Electronic Thesis or Dissertation |
Source | http://hdl.handle.net/10044/1/45644 |
Page generated in 0.0011 seconds