| Summary: | 2014 Spring. Includes illustrations (some color). Includes bibliographical references (pages 158-172). Natural gas hydrates exist in unconsolidated marine or permafrost sediments and can adopt many morphologies. Hydrates can occur in shapes of nodules, veins, layers, or finely disseminated between sediment grains. In the latter configuration, hydrate may cement sediment grains, act as a load-bearing sediment component, or occur free-floating in the pore fluid. While the sediment lithology dictates whether hydrate is present in massive or disseminated form, the manner of hydrate formation determines whether the disseminated hydrate will act as a cementing or non-cementing element in the system. The physical properties imparted to the hydrate-bearing sediment by hydrate can vary dramatically depending on how the hydrate forms, and in nature, gas hydrates generally form from gas dissolved in water, but can also form from water in the presence of a free gas phase. In this study, the effect of hydrate formation on the wave velocities of unconsolidated sediment was investigated in a series of laboratory studies, with particular focus on the extent to which the initial water saturation controls the manner in which hydrate is distributed, and thus the extent to which hydrate formation increases the wave velocity in sands. Ultrasonic p- and s-wave velocities (Vp, Vs) were measured in conjunction with magnetic resonance imaging (MRI) in hydrate-bearing Ottawa Sand F110 during hydrate formation and dissociation. vp and vs were determined as functions of gas hydrate saturation (Sh). Hydrates were formed out of solution using tetrahydrofuran (THF) and through CH[subscript 4] injection into partially water-saturated samples. For the latter, samples with low and high initial water saturation (Swi) were tested. The recorded velocities exhibited a noticeable dependence on Swi. At low Swi (~20%) the hydrate stiffened the sediment and increased the ultrasonic velocities dramatically. However, the rate at which the velocity increased during hydrate formation decreased with increasing Swi and small changes in the initial water saturation resulted in significant changes in final velocities. At high Swi (~80%), the velocity increased almost linearly with increasing hydrate content even at very low Sh. This behavior differed from that observed for hydrate formed out of solution. Ultrasonic velocities recorded in water-saturated, THF hydrate-bearing sand sediment did not change until a critical hydrate saturation of 35-50 percent was exceeded. Comparing measured velocities to those calculated with existing rock physics models links the initial water saturation, which determines the gas-water distribution in the sediment and hence the location of initial hydrate formation, to the evolution of wave velocity during hydrate formation. We concluded that at low Swi, the water is evenly distributed and located at the grain contacts. The resulting hydrate cements the grains, dramatically increasing the wave velocities even at low hydrate saturations. At high Swi, gas occurs in stiff bubbles within the pore space. The resulting hydrate particles become load-bearing and keep the grains from slipping. At intermediate Swi, gas and water appear to occur as patches, resulting in a mixture of load-bearing and cementing hydrate. The model comparison also reinforces our picture of hydrate-formation out of solution, where the pore-filling hydrate particles do not interact with the sediment until a critical value for Sh is exceeded and the hydrate becomes load-bearing. To test the dependence of the initial water distribution on the initial water saturation, micro X-ray CT images were also acquired of partially saturated glass-bead packs without hydrate but with varying amounts of water. At low water saturations, water occurred as bridges between adjacent glass beads or was located at the glass-bead contacts. At high water saturations, water occurred as a continuous phase and gas was present in the form of bubbles or patches. At intermediate water saturations both types of water distributions, gas bubbles and water bridges, were observed within one sample. With decreasing Swi, the closer the gas/water interface is located to the grain contact, the higher the degree of cementation will be when hydrate forms . Conversely, the higher the Swi the less cementation occurs and the less a given degree of hydrate formation will affect vp and vs in unconsolidated sediments. To extend our observations to hydrate-bearing systems, the distribution of THF and cyclopentane (CP) hydrate packs of glass beads was observed with micro X-ray computed tomography (CT). While THF is completely miscible in water, CP and water coexist as two separate phases. After hydrate formation in the THF hydrate-bearing samples, unconverted water was identified at the grain contacts and was often observed as a thin layer between hydrate and the glass-beads, indicating that the hydrate formation from the dissolved phase had initiated away from the glass beads and the bead-to-bead contacts. In some areas, THF hydrates grew to become amorphous patches that surrounded grains, but in most cases the hydrate was restricted to individual pores and exhibited characteristically sharp edges. In the CP hydrate-bearing samples, the image contrast between water/ice and CP hydrate could not be resolved. However, it appeared as if CP hydrate had formed along the CP-water interface and served as a diffusion barrier between the two phases. As hydrate in known to initially grow at the gas-water interface, the hydrate distribution was most likely affected by the distribution of water prior to hydrate formation. Based on the experimental results and our comparison with rock-physics theory, a conceptual model of hydrate in sediment was developed. The distribution of hydrate can be shown to be primarily a function of the water distribution within the sample prior to hydrate formation. With increasing water content, the distribution of the liquid phase transitions from being located at the grain contacts, to forming patchy networks, and finally becoming the continuous phase. When gas occurs as the continuous phase, hydrate grows into the water phase and the resulting hydrate will cement sediment grains. If water is the continuous phase, hydrate will grow into the gas phase and the resulting hydrate will initially be pore filling, but will become patchy and load-bearing as the hydrate saturation grows beyond 35-50% of the pore space.
|
|---|