In western Canada, there have been more than 300 heavy oil waterflooding projects. Most of these projects displayed good economical and efficient variability even though they were operated in marginal pools. Although waterflooding of heavy oil has almost 50 years history, its mechanisms, especially in the situation of high oil water viscosity ratio, are still not well understood. In the situation of high viscosity ratio, fractional flow theory does not work because of severe water fingering and other mechanisms that are different from conventional waterfloods. The operation strategies of heavy oil waterflooding, such as water injection rate, injection pressure and VRR, are still under controversy.
In a water-wet environment, waterflooding (water displacing oil) represents a process of water imbibition. In this paper, the water imbibition mechanisms and their effects on the heavy oil recovery are studied using a water-wet micromodel. The effects of time, viscosity ratio and water injection rate on the imbibition rate are also studied. The imbibition rate of water was found to be proportional to the reciprocal of the square root of time, and inversely related to oil viscosity. The effects of injection rate on imbibition rate are complicated. At low injection rates, waterflooding becomes more efficient, and significant volume of oil is produced discontinuously. Images of the imbibition process were recorded and analyzed from visual micromodel studies. Water broke through quickly because of water fingering, and a considerable portion of recovery comes from post-breakthrough production of oil, under high water cuts. In the cases of low rate water injection, water imbibed into the original oil region perpendicularly to the water channel. In this stage, capillary imbibition was a key factor. Water film thickening and snap-off were the two main mechanisms that made water imbibition work. Emulsification was also another important mechanism observed, with W/O emulsions primarily being formed.
In 2007, Total decided to launch a CO2 injection pilot, its objective being to prove the impact of CO2 injection on a mature depleted carbonate field. This field has been produced since 1982, mainly in natural depletion. The reservoir is Lower Cretaceous, low permeability carbonate rock with 20° API oil. The pilot pattern is an inverted 5-spot, with one central injector, 3 producers and one observer, the well spacing ranges from 300 to 700 m.
The main objective of the pilot was to observe and understand the effect of water and CO2 injection in a depleted carbonate reservoir, far below the Minimum Miscibility Pressure (MMP). This has been done in terms of microscopic oil recovery mechanisms, monitoring, and also field scale simulations. The microscopic work comprises secondary and tertiary CO2 SCAL data, and also a lean HC gas flood. PVT acquisitions include CO2 swelling tests, and slim tube data to estimate the MMP. Considering the risk of high mobility of immiscible CO2 in a heterogeneous reservoir, a WAG CO2 injection scheme was chosen and simulated.
This pilot was started in 2008, and has now 3 years of monitoring data. Breakthrough in HC phase was observed on the two proximal producers, but not on the most remote producer. Interestingly, as low salinity water was injected for a 18-month period after the CO2 injection, CO2 breakthrough was also measured in aqueous phase.
SCAL data were matched by using alpha-factors (also called transport coefficients) in order to limit the stripping effect of the gas. Matching the 5-spot data enabled to quantify a strong multi-layer effect, as well as dissolution of CO2 in the injected water. From the lessons learned, we improved our understanding and capability to simulate and extrapolate WAG CO2 to other fields and conditions.
Objectives of the pilot
Given the context making both greenhouse gas reduction and enhanced oil recovery desirable, the company wanted to operate a CO2 pilot, to ascertain the potential of the technique, to be able to be more precise on the conditions and efficiency of this mechanism. We also wanted the practical experience of operating safely CO2 injection in WAG mode; in order to subsequently being able to avoid the pitfalls of this technique.
Also the request for the reservoir engineers was that the understanding of the recovery mechanism, their monitoring and simulation should be improved and mastered well enough, in order to be able to apply it to other fields with different conditions. Economics were not the driver for this demonstrative pilot, because any potential full field application would not depend only on the results of the pilot, but on other solutions like IOR. As a prominent driller formulated it after some time, the pilot was meant to produce more data than oil. Its cost would thus have to be kept in reasonable boundaries. Therefore a lot of effort was put in monitoring and simulation work, when compared to traditional projects which are usually more focused on short-term production rates.
The seven following success criteria were listed:
- rate increase at one or two producers,
- measure of a reduced residual oil saturation Sor < Sorw,
- increase of oil quality at the producer (swelling effect),
- increase of gas gravity at the producer (stripping effect)
- increase of CO2 content on downhole oil
- capacity to produce at least two of the three producers for at least half of the 18 months
- no safety problems or accidents linked to WAG CO2 operations.
Numerous core-flooding experiments have shown that Low-Salinity Water Flooding (LSWF) could improve oil recovery in sandstone reservoirs. However, LSWF recovery mechanisms remain highly contentious primarily because of the absence of crucial boundary conditions. The objective of this paper is to conduct a
parametric study using statistical analysis and simulation to measure the sensitivities of LSWF recovery mechanisms in sandstone reservoirs. The summary of 411 coreflooding experiments discussed in this paper highlights the extent and consistency in reporting boundary conditions, which has two implications for statistical analysis: (1) Even though statistical correlations of the residual oil saturation to chlorite (0.7891) and kaolinite (0.4399) contents, as well as the wettability index (0.3890), are comparably strong, the majority of dataset entries are missing, and a prediction model cannot be generated; (2) If a prediction model is generated without clay content values and a wettability index, even though LSWF emphasizes wettability modification by virtue of oil aging time and the strong influence of brine cation and divalent ion concentrations on Sor, the prediction model's regression curve and confidence level are poor. Reservoir simulations conducted to examine LSWF recovery sensitivities conclude that LSWF recovery mechanisms are governed based on the initial and final wetting states. In strong water-wet conditions, the increase in oil relative permeability is the underlying recovery mechanism. In weak water-wet conditions, the incremental recovery of LSWF is driven by low capillary pressures. In weak oil-wet conditions, the primary LSWF recovery mechanism is the increase in oil relative permeability, and the secondary mechanism is the change of the non-wetting phase to oil. In strong oil-wet conditions, the underlining LSWF recovery mechanism is the increase in oil relative permeability. In all cases, an appreciable decrease in interfacial tension (IFT) is realized at the breakthrough recovery however that is rapidly overshadowed by the increase in oil relative permeability and decrease in contact angle.
For a given oil and reservoir temperature, the CO2 Minimum Miscibility Pressure (MMP) is largely affected by the CO2 purity. A commonly found contamination gas in CO2 EOR is O2, the existence of which will influence the MMP and thus influence the CO2 Water Alternating Gas (WAG) performance.
Slim tube tests and core flood experiments are conducted in order to study the influence of O2 on the MMP and CO2 WAG performance. Both experiments utilize crude oil from South Slattery Field in Wyoming. The mole concentration of O2 in CO2 gas varies from 0 mol% to 5 and 10 mol% and the experimental temperature is set at 57 °C to mimic the reservoir condition. In the slim tube tests, the oil recoveries at gas breakthrough are used to interpret the MMP. All core flood experiments utilize Berea sandstone cores and synthetic brines. A pressure of 20% higher than the pure CO2 MMP is used to ensure miscible condition and a pressure of 50% of the pure CO2 MMP is used for the immiscible study. WAG is injected in the tertiary recovery mode and its parameters include a WAG ratio of 1:1, a half cycle slug size of 0.1 PV, and a total slug size of 2.0 PV. WAG performance, i.e., percent oil recovery, tertiary recovery factor, and CO2 utilization factor are determined.
Slim tube results show that the existence of O2 increases the CO2 MMP significantly. Five mol percent of O2 in CO2 gas increases its MMP by 20.88%, while 10 mol% of O2 increases it by 61.92%. Core flood results show that the existence of O2 in CO2 gas adversely influence the WAG performance. For miscible CO2 WAG flooding, the higher the O2 content, the lower the incremental recovery and tertiary recovery factor (TRF). The same results are also found in immiscible CO2 WAG with less significant effects.
Similar studies have been mainly focused on N2 and CH4, while the influence of O2 has never been experimentally investigated. This study is essential for better understanding the influence of CO2 contamination on WAG performance, which is a major concern in CO2 EOR utilizing unpurified CO2 gas sources.
Keywords: O2 contamination, CO2 purity, CO2 WAG, Minimum Miscibility Pressure (MMP), Tertiary Recovery Factor (TRF)
Since the first field water alternating gas (WAG) injection reported in 1957 (VanPoollen, 1980), it has been widely used as one of the leading enhanced oil recovery (EOR) techniques all over the world. Christensen, J.R. et al. reviewed the field WAG injections and indicated that the average incremental recovery for miscible CO2 WAG could be around 10% of the Original Oil in Place (OOIP) and somewhat lower for immiscible CO2, nitrogen, and hydrocarbon gases (Christensen, 2001).
A recent global estimate of incremental oil recovery (IOR) also shows that an additional oil production of 450-820 billion barrels could be recovered while storing 130 to 240 Gt of CO2 with full implementations of the CO2 EOR projects (IEA, 2009), but this expectation is based on the assumption that the primary future supply of CO2 is anthropogenic.
It is known that natural sources of CO2 supply contain CH4 or H2S, while other sources contain other contaminants, for example, reproduced gas from reservoir contains hydrocarbons and H2S (sour gas) amine-based post-combustion captured CO2 from conventional coal-fired plants contains almost all of the sulfur, and CO2 from an oxy-combustion plant contains SO2, O2, N2, and Ar (Argon) (John, 2010).