A laboratory study of fluid-driven tensile fracturing in anisotropic rocks

  • Stephan Gehne

    Student thesis: Doctoral Thesis


    Fluid-driven tensile failure is a ubiquitous phenomenon in Earth sciences, as seen in examples ranging from dyke and sill injection in volcanic systems to veining and mineralisation. In the engineered environment, the method has recently been used for the intentional hydraulic fracture of water and hydrocarbon reservoirs. This has allowed the exploitation of previously uneconomic reservoirs by generating tensional fracture networks for enhanced permeability, but with the side effect of generating small earthquakes in the process. This has made the application of the technology controversial, as it generates a clear inherent risk. Although this industrial application has proven itself, it has developed in a largely uncontrolled trial-and-error approach and with little regard to the fundamental science behind the process. This is important, as to understand and predict the fracture process, the various controlling factors must be known, which is challenging in a natural environment. To address some of these gaps in knowledge, this study has developed a novel laboratory-based method to simulate the fluid-mechanical process of hydraulic fracturing. New data are presented that illustrate the combined effects of the inherent rock anisotropy, fabric and initial permeability, and how this is manifested in terms of tensile fracture initiation, propagation and geometry. To achieve this, a new apparatus to generate fluid-driven tensile fractures using a conventional triaxial cell (providing simulated burial depth) is developed. Rock physics data from the experiments (Acoustic Emission, radial strain and fluid pressures) recorded at high speed are combined with post-test micro X-ray CT imaging. For the first time, the generation and propagation of fluid-induced hydraulic fractures is made with respect to the initial rock fabric, and then linked to the generated Acoustic Emission, for direct comparison to field seismicity.

    Fracture orientation is primarily controlled by the principal stresses and their orientation relative to bedding planes. However, inherent rock anisotropy, initial rock permeability and rock fabric are key controlling factors in governing fracture initiation, propagation, and fracture geometry. It has been shown that anisotropy and initial permeabilities affect fracture initiation and can lead to increased or premature failure pressures respectively. Fracture geometry strongly depends on the orientation of the inherent bedding, determining if fractures propagate parallel or normal to the bedding, and the rock fabric, resulting in planar or more tortuous fracture paths. By linking Acoustic Emission and mechanical behaviour with respect to the final fracture network, the hydraulic fracture process is decoded into distinct fracture stages: (i) maximum fluid pressure, (ii) a short period of ‘plastic’ deformation, (iii) fracture initiation, (iv) stable fracture propagation, (v) sample breakdown and finally (vi) unstable fracture propagation. This analysis shows that a combination of seismic activity, fluid injection rates and deformation are reliable indicators for imminent breakdown in anisotropic sedimentary rocks subjected to injection fluid pressures, a critical step towards the development of an updated, engineered approach to hydraulic fracturing in an effort to reduce risks, increase controllability and to optimise gas extraction. Finally, the incremental fracture process is analysed and related to the fracture toughness (KIC) using fracture energy as a proxy to show that fracture extension only occurs when fluid-driven stress increases beyond KIC, whereas fracture initiation is controlled by the tensile strength. Ultimately, relating fracture behaviour in unconventional resource lithologies to induced seismicity and key mechanical and fluid injection parameters may provide for better fracture prediction during field operations, reducing the risk and improve resource exploitation.
    Date of AwardJul 2018
    Original languageEnglish
    Awarding Institution
    • University of Portsmouth
    SupervisorPhilip Benson (Supervisor) & Nick Koor (Supervisor)

    Cite this