Water Imbibition into Sandstones: Influence of Flow Rate on Water Distribution and Acoustic Response

Abstract

Fluid injection into a porous medium in order to displace another fluid is a commonprocedure while dealing with the recovery of subsurface fluids (e.g., secondarywaterflooding, oil recovery). The seismic method has been widely used to imageunderground structures and, consequently, map reservoirs. More specifically, the timelapsemethod consists in acquiring successive 3D seismic data of the same area over aperiod of time using acoustic sensors arrays. The aim of our experimental work is toobserve changes in seismic signals resulting from changes in the acoustic impedance.Since the acoustic impedance is the product of velocity and density, time-lapse signalsare affected by the compressibilities of the reservoir rock and of pore fluids. This acousticmonitoring is also the basis to track saturation front displacements. The interaction ofseismic waves with the fluid-rock system is complex and complicates the interpretationof time-lapse signals. In order to better understand this interaction, experiments atlaboratory scale were performed. Distilled water is injected into a dry sandstone sample(from the Otway Basin, South Australia, Australia), porosity: 19.5%; permeability: 25.0mD. Ultrasonic P-wave velocities (VP) across the sample are calculated using the firstbreakpick of the output signal recorded by an oscilloscope. Simultaneously, the waterdisplacement is visualized using X-ray Computed Tomography (CT). The CT scans aremaps of the fluid distribution from which the water saturation (SW) is estimated. Theevolution of VP and SW with the injection of water at low and high injection rates (LIRand HIR, respectively) compared to the natural imbibition (NI) rate is studied. It is also investigated how flow rate influences the displacement and geometry of the saturationfront. A clear dependence of the evolution of VP and SW with the flow rate isexperimentally observed. The saturation front passing the ultrasonically monitoredposition is associated with a characteristic change in VP. There is a sudden increase in VP,sharper for the HIR but still noticeable for the LIR. The HIR promotes a flat, compactsaturation front and a fast increase in VP and SW while the LIR promotes a diffusivesaturation front and both VP and SW increase slowly as water is injected. NI reveals apattern on the evolution of VP and SW similar to the HIR, sustaining a flat saturationfront. For the NI there is a continuous decrease of the imbibition rate revealing that at acertain point gravity forces overcome capillary forces. A set of experiments is alsoperformed where the injection rate is increased and decreased within a single injection.Once more, a strong influence of the injection rate on the evolution of VP and SW isspotted. Decreasing (increasing) the injection rate directly decreases (increases) VP andSW. Another important result is that both experiments (at constant and variable injectionrates) show that the same SW is associated to very different VP. Since the same SW canrelate to different fluid distributions, this result reinforces that acoustic waves are notonly sensitive to the amount of fluid present in the pore space, but also how it isdistributed within the sample.