Microvisual investigations to assess and understand enhanced oil recovery processes using etched silicon micromodels

Advisors

Anthony R. Kovscek, primary advisor
Sally M. Benson, advisor
Louis M. Castanier, advisor

Abstract

Conventional oil production is in decline and demand for enhanced oil recovery (EOR) is increasing. Highly calibrated simulation models are built as decision tools for investments and further field developments. Because EOR reservoir mechanisms are more complicated than primary and secondary recovery mechanisms, a more detailed physical understanding is required to design accurate simulation models. EOR flow processes need to be investigated, represented accurately, and calibrated at multiple scales before testing on a field-wide project. Experimental results at different scales deliver the basic construct of each simulation model. Examples are micro-scale pore size phenomena observations for pore network simulations and core-scale for material balance calculations.This study focuses on the microscale investigation of multiphase fluid flow using etched-silicon micromodels to assess the flow behavior on a pore scale. Micromodels have the pore network patterns of a porous medium etched to a silicon wafer and hence are representative of the two dimensional structure of the porous medium. The patterns used in the construction of the porous medium may be prepared from thin sections of any given rock or reservoir type. They represent the medium or, in several cases, are geometrically constructed as a series of repeatable simple or complex geometric figure aggregates. Geometrical and topological properties and pore roughness are close to the original core sample. The various micromodel pore networks (sandstone, unconsolidated sandstone, carbonate and fracture models) are tested with different fluid and fluid pairs and pore scale behavior like sweep efficiency, snap off, micro scale saturations and so on are qualitatively described and characterized. Measured parameters and descriptions aid simulation development to create a fully functional physical model. Experiments reported in this thesis are relevant to a variety of EOR topics: a.)gas trapping and dissolution of CO₂ water systems during carbon sequestration or a WAG EOR process, b.)gas exsolution behavior of supersaturated CO₂ water when traveling from a high to a a low pressure region, c.) front stability and micro displacement efficiency of unstable displacement process during gas injection, d.) foam injection in fractured reservoirs to control mobility after premature gas breakthrough, e.) rheological behavior of polymer solution at near critical conditions in porous medium and in fractures, and f.) multiphase flow behavior in an intermediate wet dual porosity system similar to an ARAB-D carbonate rock. Understanding the immobilization and trapping of carbon dioxide is not only crucial in estimating storage capacity and security during CCS but also an important factor to operate a CO₂ EOR flood in the most efficient manner. Residual and dissolution trapping are time dependent and need to be better understood for better predictions. A set of CO₂-water imbibition experiments were conducted in micromodels whose homogeneous pore space is geometrically and topologically similar to Berea sandstone to investigate the pore-scale events of residual and capillary trapping. Microvisual data, photographs and video footage, describes the trapping mechanism and, especially, the disconnection and shrinkage of the CO₂ phase in various phase conditions. Results show that, depending on the flow rate of the imbibing water, different trapping mechanisms are observed. Lower flow rates, comparable to the trailing edge of a CO₂ plume, lead to more snap-off events. During snap off, the wetting fluid swells at the pore walls until the critical capillary pressure is reached, where the interface collapses. The non wetting fluid is then forced into the pore and the wetting fluid fills the pore throat, resulting in greater trapped residual saturation. Rates comparable to the near wellbore area during enhanced sequestration showed sweep out displacement of gas bubbles. Sweep out is characterized when the interface does not collapse and instead the whole non wetting phase is displaced by the wetting phase leaving no trapped saturation behind, and greater dissolution that ultimately leads to very low or zero gas saturations. Furthermore, complete dissolution events showed that homogeneous as well as heterogeneous dissolution occurs. Whereas the latter is subdivided into microbubble formation and dissolution on crevices or pore roughness, the former occurs without the influence of pore walls. After sequestration, CO₂ concentrations of 50 g/l or more may to be found in saline aquifers. Although dissolved CO2 does not bear an obvious risk there are plausible mechanisms by which the CO₂ laden brine could be transported to a shallower depth, where the CO₂ would come out of solution/exsolve, and form a mobile CO₂ gas phase. This significant mechanism for drinking water contamination has received little attention, and there are basic science and reservoir engineering questions that need to be addressed in order to reduce risks to underground drinking water supplies. This study investigates the conditions under which dissolved CO₂ brines can impact drinking water aquifers. It develops a fundamental understanding of the fate of dissolved and exsolving CO₂ at pore scale, called nucleation using micromodel experiments. Exsolution experiments showed similar pore scale events as in the dissolution study. Bubble nucleation was observed for three different types homogenous and heterogeneous type I and type II. The injection of CO₂ into saline aquifers exhibits a strong unstable displacement due to the viscosity difference of the water and the CO₂ phase that leads to unfavorable mobility ratios (M> 1). Although the subsurface flow of different fluids has been investigated in a large scale in the oil and gas industry, the characteristics of the water-CO₂ fluid pair that lead to highly unstable fluid fronts is still not fully understood. So far, most modeling of carbon capture and sequestration (CCS) relies on the linear displacement theory from Buckley and Leverett. Based on In this work, laboratory experiments using a wide range of mobility and capillary numbers to show displacement fronts of stable and unstable drainage process are reported. Experiments were conducted in etched silicon micromodels with Berea sandstone-like pore structures and geometry. Experimental data in the form of macroscale front displacement videos and micro scale saturation pictures were collected and analyzed. Drainage results showed that there was an increase in finger number and finger size with an increase in capillary number. Capillary number did not influence areal sweep efficiency but showed significant effects on micro saturation where low capillary numbers led to snap off and small pores left undrained whereas large capillary numbers swept out small and large pore structures leaving less wetting saturation behind. Fractal analyses were used to evaluate unstable displacement fronts. Results showed that the average saturation does not scale with wave speed. Moreover the displacement pattern follows a fractal pattern. Foam as a gas-mobility control agent is successful in enhanced oil recovery processes. In fractured reservoirs, foam acts as a blocking agent slowing and redirecting the transport of the aqueous phase in high transmissibility fractures. Foam allows more time for the liquid/foamer agent to imbibe into the matrix blocks and drain remaining oil. In this work, the behavior of foam flow in fractures at various foam qualities and liquid and gas velocities is investigated. Laboratory experiments with different fracture replicates etched in silicon micromodels were used. Different micromodel fractures (smooth surfaces, rough surfaces and different apertures) were used to observe pre-generated foam in terms of texture, pressure drop and flow behavior. Mobility reduction factors for a wide range of foam qualities and flow rates were analyzed. Measured pressure drops increase linearly with an increase in foam quality up to 90%. At qualities greater than 90%, mobility reduction is only slightly reduced further. In general mobility reduction factor (MRF) of 10-400 times were measured for low to high quality foams respectively. Additionally, video footage of foam at micro and macro scale is used to tie rheology to bubble shape and size. Polymer flooding has the potential to recover bypassed oil faster and therefore boosts the economics significantly in an EOR project. The success, however, depends on the injectivity of polymer solution volumes. Injection into porous media at conditions above a critical rate may lead to mechanical degradation of the polymer in solution resulting in a loss of viscosity. The resulting increase in mobility ratio may result in an uneconomical project. Therefore, the investigation of the rheological behavior of polymer solutions at different rate conditions is critical in designing a polymer flood project. Micromodel experiments were used to assess degradation of polymer solutions in fractures as well as in porous media. Only minor, mechanical degradation was found. Polymer solutions exhibit, depending on fracture roughness, shear thinning behavior. In contrast, polymers exhibit shear thickening behavior when flowing through porous media up to a factor 10 when comparing with an equivalent reservoir shear rate in the rheometer. In addition results showed that plugging, that leads to loss of injectivity, can be a critical issue in polymer injection. Currently around 6.25% of the world oil production are delivered from the Ghawar field in Saudi Arabia. The majority of the estimated 100 billion barrels of oil in place are trapped in an Arab-D carbonate formation. The creation and testing of an etched-silicon micromodel that has the features and characteristics of a dual porosity pore system such as might be found in a Arab-D carbonate rock was investig...