Fluid pressure diffusion in fractured media: insights from harmonic and non-harmonic periodic pumping tests

Fractures can significantly impact fluid flow and pore pressure distribution in the subsurface. Understanding the mechanisms and conditions influencing their ability to transport fluids and to promote pore pressure diffusion is key for many activities relying on fracture-controlled flow such as, for example, enhanced geothermal systems. In situ characterization of these properties is typically done by performing hydraulic tests in selected intervals of a borehole and their interpretation relies on the solution of a linear pressure diffusion equation. However, it has been shown that the hydraulic behavior of fractures as well as the associated near borehole flow regimes can be largely affected by the coupling between the solid deformation and fluid pressure upon injection/production. In this work, we explore these effects by performing a series of harmonic injection tests (HIT) as well as non-harmonic production tests (NHPT) in a packed-off interval of a borehole containing multiple natural fractures. The borehole is located in the Bedretto Underground Laboratory for Geosciences and Geoenergies in Switzerland and penetrates granitic rock mass. The two kinds of tests consist of a periodic repetition of the same injection or production protocol. Flow rates, interval pressures as well as pressures above and below the double-packer probe are recorded at the surface. An important advantage of periodic testing is that it permits a continuous tracking of hydraulic changes during the test. For our study, we conducted a so-called injectivity analysis, in which the phase-shift (time delay) and amplitude ratio between flow rate and interval pressure are used to infer effective hydraulic properties. We performed over 200 periodic tests including both HIT and NHPT with a large range of periods (7.5 s to 1800 s) as well as varying mean interval pressures (~1300 kPa to 2100 kPa) and flow oscillation amplitudes. As a result, we obtained a robust constraint of the radial flow regime prevailing in the fractures. Overall, we found that results from HIT and NHPT are in very good agreement despite the remarkably different injection protocols. For all cases, a prominent and consistent period dependence of phase shifts and amplitude ratios of flow rates and interval pressure was observed, in which both increase as the oscillatory period decreases. Amplitude ratios showed almost no variation with mean interval pressure regardless of the injection protocol. In contrast, a prominent pressure dependence of the phase shifts is captured by the HIT but not the NHPT data. Using the pressure-independent NHPT results, we reconstruct the general hydraulic response of the tested fractured section, which can be well represented by an analytical solution of the pressure-diffusion equation. This general trend explains the HIT data as well, although evidence of significant variations that are correlated with the amplitude of the pressure oscillations points to the predominant role of hydromechanical coupling effects on the fluid pressure diffusion process.