Localized buoyant convection in porous media: plumes and dispersion

点击翻译按钮获取中文摘要

Examples of buoyancy-driven flow abound in nature, industry and everyday life from atmospheric convection to the steam that rises from a hot kettle on the stovetop. Less obvious, but equally important, is buoyancy-driven flow within a porous medium whereby hot/cold fluid navigates through the interstitial spaces between solid grains. Indeed, buoyant convection within porous media is relevant to energy systems e.g. thermal energy storage, methods of enhanced oil recovery such as steam-assisted gravity drainage or cyclic steam stimulation, and the geological sequestration of either carbon dioxide (CO2) or acid-gas (a mixture of CO2 and hydrogen sulfide), which result from the production of electricity and oil. The research proposed here, though fundamental in focus, is motivated most especially by this latter example. To this end, one must define the precise conditions under which the CO2 or acid-gas injected into formation will remain permanently sequestered. Although there are a variety of mechanisms that allow the injectate to become immobilized (capillary trapping, etc.), CO2 and acid-gas are more buoyant than, say, aquifer brine and tend therefore to rise from one geological layer to the next. The competition between immobilization and vertical migration, far from being of purely academic interest, determines whether the large outlays of public and private capital necessary for geological sequestration projects are, in fact, justified. In my research, I will use theory and laboratory experiment to advance our current understanding of buoyant convection in porous media. Three interconnected projects will be pursued. Firstly, I will consider the impingement of a buoyant plume on a sloping interface where the fluid has an easy and difficult time, respectively, flowing through the layers below and above. This mimics a scenario in which a plume of buoyant CO2 or acid-gas strikes a (possibly fractured) cap rock layer having a comparatively small permeability. To this base flow, and consistent with real geophysical scenarios, I will then allow some fraction of the injectate to dissolve into the ambient reservoir fluid. Although the volume of dissolving injectate is small, the impact on the flow dynamics is expected to be large. Finally, and again consistent with real flows, I will then specifically account for the limited miscibility of CO2 or acid-gas in aquifer brine. Doing so requires accounting for surface tension among other effects. Model predictions and measured data may be adapted into basin-scale geological algorithms that lack sufficient resolution to fully describe flow details at fine spatial scales. Instead, parameterizations of the tradeoffs between injectate drainage, advection and dissolution are required. To this end, a broader objective of the research is to collect, interpret and synthesize data in a way that is meaningful to other researchers, industry practitioners, regulators and policy makers.