Experimental Studies Focused on the Pore-Scale Aspects of Heavy Oil and Bitumen Recovery Using the Steam Assisted Gravity Drainage (SAGD) and Solvent-Aided SAGD (SA-SAGD) Recovery Processes

Increasing energy consumption and continuous depletion of hydrocarbon reservoirs will result in a conventional oil production peak in the near future. Thus, the gap between the global conventional oil supplies and the required amount of fossil fuel energy will grow. Extensive attempts were made during the last three decades to fill this gap, especially using innovative emerging heavy oil and bitumen production technologies. Most of these recovery methods have been developed in Canada, considering the fact that Canada and Venezuela have the largest deposits of heavy oil and bitumen throughout the world. The horizontal well drilling technology opened a new horizon for the recovery of heavy oil and bitumen. Most of the in-situ recovery techniques, including Steam Assisted Gravity Drainage (SAGD) recovery method, take advantage of horizontal injection and production wells. The vacated pores in the reservoir are filled mainly either with steam or with a mixture of steam and solvent vapour in the case of the SAGD and Solvent Aided SAGD (SA-SAGD) recovery methods, respectively. The use of long horizontal wells combined with the reduced viscosity of the produced oil allows economic production with limited amount of bypassed residual oil in the invaded region. 
\nThe macro-scale success of the SAGD recovery technique is greatly affected by its pore-scale performance. It is beneficial to understand the pore-level physics of the SAGD process in order to develop mathematical models for simulating field-scale performance. Available commercial reservoir simulators cannot describe pore-level mechanisms of the SAGD process including mechanisms related to the fluid-flow as well as heat-transfer aspects of the process. A systematic series of flow visualization experiments of the SAGD process using glass-etched micromodels was developed to capture the pore-level physics of the process using qualitative analysis. With the aid of image processing techniques, the pore-scale performance of the SAGD process was qualitatively and quantitatively investigated. The main objective of Chapter 2 of this thesis is to address the relevant pore-scale mechanisms responsible for the in-situ oil mobilization and drainage in a conventional SAGD process. Transport processes, occurred in a conventional SAGD process at the pore-level including fluid flow and heat transfer aspects, were mechanistically investigated and documented. The qualitative analysis of the results revealed that near a well-established oil-steam interface, gravity drainage takes place through a thick layer of pores, composed of about 1-6 pores in thickness, within the mobilized region. The drainage of the mobile oil takes place due to the interplay between gravity and capillarity forces near this mobilized region. In-situ mobilization of bitumen was found to be as a result of both conductive and convective elements of the local heat transfer process. Moreover, the phenomenon of water-in-oil emulsification at the interface was also demonstrated which is due to the local steam condensation and spreading characteristics of water droplets over the oil phase in the presence of a gas phase. Other pore-scale aspects of the process such as drainage displacement as well as film-flow drainage mechanisms of the mobile oil, localized entrapment of steam bubbles as well as condensate droplets within the mobile oil continuum due to capillarity phenomenon, sharp temperature gradient along the mobilized region, co-current and counter-current flow regimes at the chamber walls, condensate spontaneous imbibition followed by mobile oil drainage, and snap-off of liquid films are also illustrated using these pore-level studies. The second objective of Chapter 2 is to quantitatively analyze the production performance of the SAGD process based on the micro-scale measurements. Our pore-scale experiments revealed that the rate of pore-scale SAGD interface advancement depends directly on the pore-scale characteristics of the employed models and the pertaining operating conditions. The average sweep rate data were correlated using an analytical model proposed by Butler (1979, 1981, 1991) and a pore-scale performance parameter was defined for the SAGD process. The measured horizontal sweep rates of the SAGD process at the pore-scale are in good agreement with the theory predictions provided by the performance parameter. In addition, the effect of different system variables on the ultimate recovery factor of the SAGD experiments were investigated and it was found that higher permeability values and lower in-situ oil viscosities lead to higher ultimate recovery factor values for a particular SAGD trial. Moreover, the Cumulative Steam to Oil Ratio (CSOR) data were scaled and a reasonably good fit for the experimental data was achieved by defining a scaling parameter. 
\nAlthough the SAGD process offers several inherent advantages including high ultimate recovery, stable oil production rates, reasonable energy efficiency, and high stable sweep efficiency, there are some drawbacks associated with the SAGD process such as high energy consumption, high levels of CO2 emission, and usage of large quantities of fresh water which make this process uneconomical in reservoirs with thin net pay, low matrix porosity and thermal conductivity, and low initial pressure. The most promising route for improving the SAGD performance appears to be the co-injection of a light hydrocarbon solvent with steam in the context of the Solvent Aided SAGD (SA-SAGD) process. The pore-level aspects of the SA-SAGD process are not yet understood to the extent of incorporating the pore-scale physics into mathematical models. The main objective of Chapter 3 of my thesis is to mechanistically investigate the SA-SAGD process at the pore-level to enlighten the unrecognized pore-scale physics of the process. A methodical set of pore-scale SA-SAGD experiments were designed and carried out with the aid of glass micromodels. The methodology used in this set of the SA-SAGD trials was similar to that of the pore-scale SAGD experiments described in Chapter 2. Normal Pentane and Normal Hexane were used as the steam additives. The pore-level events were recorded on a real-time basis and then analyzed using the image processing techniques. According to the qualitative results, it was obtained that all the condensate and gaseous phases flow simultaneously in the mobilized region composed of about 1-4 pores in thickness. Heat transfer mechanisms at the pore-scale include conduction as well as convection. The mechanisms responsible for the mass transfer at the pore-level include molecular diffusion as well as convection. The mobile oil drains as a result of two active mechanisms of film flow as well as direct capillary drainage displacements at the pore-scale. Due to the near miscible nature of the displacement process, the residual oil left behind in the invaded portion of the micromodels was negligible and asphaltene precipitation and plugging was found to be a temporary phenomenon. The second objective of Chapter 3 is to quantify the pore-scale production performance of the SA-SAGD process using the flow visualization experiments. The horizontal SA-SAGD interface advancement velocity was chosen to be the indicator of the pore-scale performance of the process. It was found that addition of n-C6 as the steam additive was more effective than n-C5 in terms of enhanced pore-scale interface advancement as well as achieving higher ultimate recovery factor when all the other experimental variables are unchanged. The higher the solvent concentration in the injection mainstream is, the higher would be the pore-scale sweep rate as well as the ultimate recovery factor of the process. When oil type with lower in-situ viscosity was used, higher sweep rates as well as higher ultimate recovery factors values were achieved compared to the trials in which the more viscous bitumen was employed as the oil type. In addition, a scaling parameter composed of porous media properties was found by which the pore-scale interface advancement velocity and the ultimate recovery factor of the SA-SAGD trials were scaled when all other experimental variables remain unchanged. 
\nIn Chapter 4 of this thesis, the production performance of the SAGD and SA-SAGD processes were demonstrated and compared at the macro-scale under controlled environmental conditions. A 2D physical model was designed and fabricated and Athabasca bitumen was used as the oil type. According to the experimental results, it was obtained that the average mobile oil as well as dead oil production rates are reasonably constant over the course of the SAGD and SA-SAGD trials. As far as the SAGD experiments are concerned, there is a linear correlation between the mobile oil production rates and the square root of the porous media permeability when all the other experimental variables remain unchanged. In addition, the Steam to Oil Ratio (SOR) values of the SAGD trials correlate reasonably well with the inverse of the square root of permeability when all the other experimental variables are fixed. By introducing the solvent additive to the injection mainstream of the SAGD process, it was found that enhancements of about 18% and 17% were observed in the mobile oil and dead oil production rates of the SAGD process respectively. In addition, the SOR values of the SA-SAGD process was reduced by about 35% compared to that of the SAGD process. Finally, an advanced photomicrography unit with an integrated image processing software was used in order to investigate size of the enclosed water condensate droplets in the continuum of the mobile oil produced during the course of the SAGD and SA-SAGD experiments. The captured microscopic snapshots were analyzed using the image processing techniques and some representative average values of the water condensate droplet sizes were reported for the corresponding SAGD and SA-SAGD

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.