No abstract

Steam Assisted Gravity Drainage (SAGD) is an efficient method for thermal recovery of bitumen from the vast reserves available worldwide and particularly from the oil sands in western Canada. Flow simulators are available for predicting SAGD performance and are used to support reservoir management decisions; however, the high computational time associated with the use of such complex flow simulation makes it impractical for all locations especially when reservoir uncertainty and variable operational parameters are included in the making decision process. The use of a simpler analytical model as a proxy for the reservoir simulator is shown to be a feasible alternative to flow simulation. A proxy model based on the Butler's SAGD theory is developed to predict the oil flow rate, cumulative oil production and cumulative steam injection profiles during both: the rising and spreading steam chamber periods for a confined SAGD well pair. Modifying factors are used to fit the proxy to flow simulation results to account for conformance and reservoir heterogeneity among other factors. A critical aspect of the proxy model is a realistic parameterization of geological heterogeneity. Monte Carlo Simulation (MCS) and the proxy model permit an efficient transfer of the uncertainty in reservoir and operational parameters through to performance variables such as oil production and steam oil ratio. An example application for a single well pair showed the efficiency of the methodology in terms of computation time. The results permit improved reservoir management of complex SAGD projects. Introduction Current high oil prices are boosting the feasibility of bitumen production projects supported on the availability of tested and succeeded exploitation technology as well as on the vast bitumen reserves available worldwide. An important part of these vast resources are in western Canada. According to the Energy Resources Conservation Board 1, at 2007 the established bitumen reserves in Alberta are 27.45x109 m3 and about 82% is considered recoverable by in-situ methods. The successful application of SAGD process as thermal recovery method is one of the pillars in which Canadian oil industry is supporting the exploitation of the in-situ recoverable bitumen leading to a massive expansion around all Alberta oil's sands. SAGD is a thermal recovery process based on steam injection coupled to horizontal well technology. Common implementation consists of two horizontal parallel wells, the first drilled near the bottom of the reservoir with the second located at a short distance, typically 5 to 10 m above it. The upper well provides continuous steam supply into the reservoir and the lower one allows the continuous production of bitumen, gas and condensed water. During the SAGD process, the cold oil is essentially immobile; therefore, an initial preheating stage is necessary to create a uniform thermo-hydraulic communication between the well pair. In this start-up period, steam is injected in both wells to preheat the reservoir between the wells. Once mobility has been established, steam is injected continuously into the upper well and rises within the reservoir, developing a steam chamber. The injected steam will reach the chamber interface, heating the surrounding cold oil sand. The heated oil and condensed water will drain by gravity along the chamber-to-reservoir interface to the lower well in which the fluids are continuously produced. The process follows an initial rising period where the steam chamber rises up to the overburden and then the spreading period which is characterized by the lateral growth of the interface along the well pair.

Corefloods and field investigations confirm that a large amount of incremental tertiary oil can be recovered from dipping water drive reservoirs using gravity assisted tertiary gas injection processes. These processes include the Double Displacement Process (DDP) and the Second Contact Water Displacement Process (SCWD). The DDP consists of injecting gas into waterflooded oil zones. The SCWD process consists of submitting these gas-flooded zones to a new water displacement process. Reservoir simulations performed with an adaptive-implicit simulator were applied to investigate the macroscopic mechanisms of the two processes. The effects of several important parameters on the performance of the DDP were studied to optimize the oil production of the process and to develop a set of screening criteria for selecting candidate reservoirs for the process. Moreover, the SCWD process was simulated to investigate its feasibility. Furthermore, the two processes were simulated physically in a micromodel — transparent cell. The results have shown that both processes are efficient methods for recovering the residual oil to water. A good representation of the laboratory results was obtained through the simulations. It was confirmed that oil film flow plays a very important role in achieving high recovery efficiencies in the DDP. In the SCWD process, trapped gas reduces the possibility of the residual oil being trapped in the center of pores in the secondary water invasion. Consequently, residual oil can be recovered quickly by a second water flood. Therefore, the SCWD process is suitable for application in situations where the source of gas is not sufficient, and where the formation has a high irreducible gas saturation. Introduction Up-dip gas injection into a dipping reservoir is one of the most efficient methods to recover waterflood residual oil. Recoveries of 85% to 95% of the original oil in place have been reported from field tests1,2,3, and higher, even up to 100%, recoveries have been obtained in the laboratory4. The idea of injecting gas to recover the residual oil after a waterflood appeared first in Carson's paper discussing a gas injection project in the Hawkins Field1. He named the process the Double Displacement Process, and defined it as the use of gas to displace a previously water displaced oil column. In the same year, Kantzas et al.4,5,6 showed in the laboratory that gravity drainage played a very important role in this gas injection process, and called it the Gravity Assisted Tertiary Gas Injection Process. They addressed the fact that reservoir wettability and spreading coefficient had a large impact on the gravity assisted tertiary gas injection process. Strongly water-wet porous media and a positive spreading coefficient are preferable in this process, and the process efficiency is dependent on the spreading phenomenon. Oren and Pinczewski7 studied the effect of the spreading coefficient on oil recovery using a network model. Their experimental results showed that oil recovery was significantly higher for positive spreading systems than it was for negative systems. Vizika and Lombard8 and Mani and Mohanty9 confirmed these results by conducting gas gravity drainage experiments in a sand pack and a network model. The incremental oil recovered by the DDP consists of two parts. The first part is the bypassed oil, which exists as a continuous oil phase in the regions of the reservoir unswept by water due to reservoir heterogeneity or well placement. The second part is the residual oil existing at the microscopic scale as isolated oil blobs in the water swept regions of the porous medium due to the capillary and surface forces. The gas injection process improves the sweep efficiency so that the bypassed oil is recovered. The trapped oil can be recovered by oil film flow. If the spreading coefficient of the oil is positive, when gas comes, the isolated oil blobs may form thin oil films. These oil films connect all of the residual oil in the gas swept zone to the oil bank, which is formed in front of the gas front soon after gas injection. The re-established hydraulic continuity of the residual oil provides channels (oil films) for the oil to flow through to the oil bank. When the oil bank reaches the production wells, both the bypassed oil and the isolated oil blobs can be produced.

Gravity assisted tertiary gas injection processes can produce a large amount of incremental tertiary oil from water drive oil reservoirs. These processes include the Double Displacement Process (DDP) and the Second Contact Water Displacement (SCWD) process. A transparent sandpack micromodel was developed to conduct a pore-level observation to investigate the microscopic mechanisms of the DDP and the SCWD processes. Observation of the two processes confirmed that oil films play a very important role in achieving high recovery efficiencies in the DDP. In the SCWD process, trapped gas reduces the possibility of residual oil being trapped in the centre of the pores in the second water flood. Moreover, reservoir simulations at reservoir scale were performed to investigate the macroscopic level mechanisms of the two processes. The results have shown that both processes are efficient methods to recover waterflood residual oil. Introduction A waterflood can only recover 40% - 60% of the IOIP in conventional oil reservoirs. However, it has been shown, in the laboratory, that nearly 100% of the IOIP can be recovered by tertiary gas injection in the presence of connate water(1). Recoveries of 85% to 95% of the OOIP have been reported from field tests(2 –4). This tertiary recovery method involving the up-dip injection of gas into steeply dipping, high permeability, strongly water-wet, light oil reservoirs to recover the residual oil is called the gravity assisted tertiary gas injection process. It is also known as the Double Displacement Process (DDP) because it involves the use of gas to displace the oil remaining after a waterflood(2). The high recovery efficiency made the DDP such an attractive process that numerous laboratory studies(5 –13) of the DDP have been conducted in different media to investigate the mechanisms of the process. Kantzas et al.(5, 6) showed that gravity drainage played a very important role in this process. They suggested that reservoir wettability and spreading coefficient had a great impact on the gravity assisted tertiary gas injection process. A strongly water-wet porous medium and a positive spreading coefficient were preferable in this process, and the process efficiency was dependent on the spreading phenomenon. Oren et al.(8) studied the effect of the spreading coefficient on oil recovery using a network model. Their experimental results showed that oil recovery was significantly higher for positive spreading systems than it was for negative systems. Vizika et al.(14) and Mani and Mohanty(15) confirmed these results by conducting gas gravity drainage experiments in a sandpack and a network model. The incremental oil recovered by the process consists of two parts. The first part is the bypassed oil, which exists as a continuous oil phase in the regions of the reservoir unswept by water due to reservoir heterogeneity or well placement. The second part is the residual oil existing at the microscopic scale as isolated oil blobs in the water swept regions of the porous medium due to the capillary and surface forces. The bypassed oil is recovered because gas injection improves the sweep efficiency. The trapped oil is recovered by oil film flow.