Combination of complex sedimentation environment and carbonates’ diagenetic alterations render carbonate reservoirs to be very heterogeneous and often naturally fractured. Heavy oil trapped in carbonate reservoirs worldwide is estimated to be over 255 billion cubic meters. The drastic differences in flow and storage capacity of the fracture and matrix networks trigger the need for larger volumes of steam to be injected to produce one cubic meter of heavy oil production, making the application of thermal recovery processes in these reservoirs to be more economically challenging and highly energy intensive.
Recovery of heavy oil from naturally fractured reservoirs is weakly controlled by viscous forces and mostly controlled by capillary and gravity forces, in addition to convection and conduction heat transfer. This makes proper representation of fracture and matrix characteristics to be of paramount importance to the reliability of modeling thermal recovery processes.
This paper discusses the effect of variable fracture characteristics including fracture typing, spacing, and apertures on the performance of dual permeability models under different thermal recovery processes. Further, the paper demonstrates the necessity to characterize partially mineralized fractures and shows its effect on thermal conductivity from fractures to matrix blocks.
The effect of wettability on the recovery process is also discussed. Oil recovery in fractured reservoirs is highly influenced by imbibition process which is most effective under strong water wetness. The paper addresses the performance under water, oil and mixed wettability conditions. Furthermore, the paper shows that proper representation of capillary pressure hysteresis could cause considerable change in oil recovery performance. The modeling work in this paper helps to curtail any extra optimism that may originate from wrongly characterized dual permeability models. Finally, and in case reservoir characteristics are not well determined then forecasting based on uncertainties of reservoir characteristics should give operators a way to estimate the risk involved in the exploitation process of the carbonate resources.
Lu, Chuan (China University of Petroleum-Beijing) | Liu, Huiqing (China University of Petroleum-Beijing) | Liu, Qian (Exploration and Development Research Institute of Liaohe Oilfield Company) | Lu, Keqin (No.1 Oil Plant of Huabei Oil Field) | Wang, Limei (No.1 Oil Production Plant of Huabei Oil Field)
Steam Assisted Gravity Drainage (SAGD) is widely used as an in-situ technology to recover heavy oil and bitumen. But due to serious heat losses, large energy requirements and enormous CO2 emissions, pure SAGD may not be sufficiently efficient to allow economical production.
The main objective of addition of viscosity reducer is to couple viscosity reduction capability of heat and surfactant. It strengthens the formation of emulsions of O/W type that have much lower viscosity than the oil from which they are formed, and enhances oil production under the same amount of steam to make the thermal process more profitable. With regard to the non-condensable gas, its distribution and movement is essential for the expansion of the steam chamber.
In this paper, the performances of a set of selected viscosity reducers under high temperatures are firstly evaluated. Then based on a high temperature and high pressure two-dimension physical model, different injection processes are conducted to study the separate and combined effect of nitrogen and viscosity reducer on the vertical and horizontal expansion rates of steam chamber. The results show that the injection of nitrogen helps the steam chamber expand more quickly in the horizontal direction. And the combination of viscosity reducer and nitrogen makes the steam chamber expand more quickly than the pure SAGD and Nitrogen-SAGD processes. The following sand pack physical models are conducted to further investigate the mechanisms of viscosity reducer and nitrogen in porous media, such as the steam profile correction relying on the blocking capacity of emulsions, heat loss reduction and the combination effect of gas expansion and viscosity reducer on the displacement of residual oil. In addition, numerical simulations are conducted to demonstrate the feasibility of SAGD assisted by viscosity reducer and nitrogen in Du 84 ultra-heavy oil reservoir. Besides optimizing operation parameters, such as nitrogen steam ratio, effective nitrogen injection volume, interval time between viscosity reducer injection and nitrogen injection, the study results show the effect of this technology on the improvement of thermal oil recovery.
In summary, this study is beneficial for the application of viscosity reducer and nitrogen in the process of SAGD.
Solvent-based enhanced heavy oil recovery techniques are promising alternatives to thermal-based methods. A better understanding of the solvent-diluted heavy oil gravity drainage is important to practical field-scale applications of the solvent-based heavy oil recovery processes. In particular, the transmissibility, which is defined as the product of the absolute permeability k and pay-zone thickness h of an oil reservoir, is an important parameter to affect the solvent-diluted heavy oil gravity drainage. In this paper, a series of experimental tests were conducted in a visual high-pressure physical model to study the permeability effect. The physical model was packed with the Ottawa sands of different sizes to obtain different absolute permeabilities in the range of k = 7–54 Darcy. The original heavy oil sample collected from the Lloydminster area in Canada was used to saturate the sand-packed physical model and pure propane was used to extract the heavy oil. The drainage height effect was studied by performing the solvent-diluted heavy oil gravity drainage tests in another visual physical model with a different drainage height. The produced heavy oil and solvent were collected and recorded in each test. The produced oil samples obtained at different times were flashed and the flashed-off heavy oil viscosities were measured. It was found that heavy oil was produced much faster in a higher permeability physical model with a larger drainage height. A new correlation was developed to take account of the reservoir permeability and drainage height effects (i.e., the transmissibility effect) on the heavy oil production. In addition, the average produced solvent–oil ratio (SOR) was determined to be in the range of 0.3–0.6 g solvent/g heavy oil. The average SOR was increased as the reservoir permeability and drainage height were decreased. It was also found that the produced heavy oil viscosity was reduced significantly. Furthermore, numerical simulations were undertaken by using CMG STARS module to match the experimental results. This study provides much-needed physical and practical understanding of the solvent-based heavy oil recovery.
Gu, Hao (China University of Petroleum,Beijing) | Cheng, Linsong (China University of Petroleum,Beijing) | Huang, Shijun (China University of Petroleum,Beijing) | Zhang, Huideng (China University of Petroleum, Beijing) | Lin, Menglu (University of Calgary) | Hu, Changhao (Liaohe Oilfield Company, Petro China)
An excellent design of steam injection projects requires accurate prediction of bottomhole steam pressure, temperature and quality. However, it is not always easy to meet the requirement when we design concentric dual-tubing steam injection schemes due to the complexity of downward steam/water flow in annuli. Also, previous methods for estimating pressure gradient in annuli, such as mechanistic models and empirical correlations, are either time-consuming or inaccurate.
In this study, we present a new semi-analytical model to predict steam pressure and temperature in annuli. It is based on Coulter-Bardon equation and on mass and energy balances in the wellbore. A more rigorous thermodynamic behavior of steam/water mixture is taken into account. More importantly, one-to-one correspondence between pressure gradient and temperature gradient of saturated steam is reasonably developed and applied in our further derivation and simplification. It is because of the simplification that we do not have to use mechanistic models or empirical correlations to separately calculate the pressure drop in annuli, which is significantly different from previous work, including Sagar et al. (1991), Alves et al. (1992) and Hasan et al. (1994) models. Our solution procedure is straightforward, the equations of steam pressure, temperature, quality, steady-state heat transfer in the wellbore and transient heat transfer in the formation just need to be coupled and solved iteratively for each segment.
Our model is validated by comparison with measured field data from Liaohe Oilfield, Petro China. The results indicate that the direction of heat transfer between inner and outer tubing depends on wellhead conditions and temperature drop in each tubing. We also show that the equivalent hydraulic diameter is not always a suitable characteristic dimension for steam/water flow in annuli. Moreover, the paper shows that our method can also be applied to single-tubing steam injection design. The predicted results from our modified model are also compared with those from CMG simulator and previous work in our study.
The depositional environment at Statoil’s Leismer SAGD demonstration project is dominated by a fluvial-deltaic system with complex geological features which present challenges for SAGD operations. An efficient monitoring tool that could combine geoscience data with production knowledge is vital for optimizing SAGD operational strategies. Time lapse (4D) seismic survey is one such monitoring tool. It can be integrated with wells and production data to prepare and optimize operational strategies. The pattern of growth of steam chamber is critical for well performance optimization. Steam chamber development dictates which reservoir areas are using heat efficiently and contributing to drainage of bitumen and increased production.
At Leismer, two 4D monitor seismic surveys have been conducted since the start of operations. The 1st monitor was acquired in 2012 to observe one year of SAGD performance and the 2nd monitor in 2013 to observe the performance after 2 years of operation. These surveys have been used to identify challenges to maximizing well pair productivity. Of primary concern is the effective distribution of steam along the entire horizontal section of the SAGD well pair. Other 4D monitoring benefits include, monitoring subcools with respect to reservoir heterogeneities, bottom water communication, steam chamber coalescence, baffle or barrier to steam chamber growth identification.
In this study, examples are presented to show how combining production and 4D seismic data has contributed to well pair optimization. Communication with bottom water in the early stages of SAGD, while rare, occurred in specific Leismer well pairs – analyses of production data combined with 4D seismic data helped identify steam loss into bottom water which drove steam injection strategy and improved well productivity. Another Leismer example shows how 4D seismic has been used to confirm a low reservoir roof at the toe of a particular well pair. This knowledge drove a change in injector well completion to optimize production. In the final example, it is shown how 4D seismic has helped analyze and confirm well conformance (heat distribution along SAGD well pairs) which is critical for uniform steam chamber development.
Based on Statoil’s Leismer experience of applying results of 4D seismic analyses for well optimization, it can be concluded that 4D seismic has proven to be a very useful and value adding monitoring tool.
Electrical downhole heating has been used for many years for flow assurance and now is being adapted for reservoir stimulation, viscosity reduction and “in situ” conversion of heavy oil. This paper starts with a short review of flow assurance applications in Alaska and Canada as described in SPE-165323-MS. It then reviews the current and developing technology and some of the heat transfer parameters for use of high voltage high power electrical heaters in a number of types of applications. In the past heater voltages have been limited to below 600 volts for mineral insulated cable heaters. Significant material and processing advances have now permitted operation at 4160 volts. This has a number of operational advantages in providing longer length heater capabilities and less parasitic heating loss in the overburden. MI cable production technology is now available to fabricate MI cable heaters that can produce 1600 meters lengths without external splices. The thermal heat transfer from the well casing to the reservoir is usually the limiting factor on the amount of energy that can be transferred from the electrical heater to the formation. Both constant power and constant temperature heaters are explained with the emphasis on in operation reliability of each type of heater. The paper concludes with an economic analysis of the opportunity provided by a high voltage MI cable heater system in a horizontal well.
Although heavy oil reserves are abundant, recovering them efficiently and economically remains a crucial technical challenge as a result of their high viscosity. Solvent based non-thermal recovery processes are designed to reduce the heavy oil viscosity through mixing and dilution with solvent. Solvent and heavy oil mixing occurs over a narrow zone, so localized viscous fingering can have a significant impact on the effectiveness of the solvent. In this study, direct pore scale modeling was used to simulate viscous fingering phenomenon during unfavorable mobility ratio miscible displacement of heavy oil in a three dimensional heterogeneous porous medium pattern.
In direct pore-level modeling, Navier-Stokes, Diffusion-Convection and Continuity equations, as the governing equations of dispersion, are directly applied and solved on the 3-D porous medium without any simplification in medium geometry.
To study the impact of unfavorable mobility ratio on the miscible displacement at the sub pore scale level, simulations have been run to model miscible displacement at five different unfavorable mobility ratios on the same porous medium pattern. Additional simulations were run to investigate the effect of pore pattern and different injection rates on the patterns, which were generated based on the process/object based reconstruction method. Base line simulations also have been done to model miscible displacement on the same medium when the mobility ratio is equal to one.
Heterogeneity of the pattern and lower viscosity of the solvent leads to appearance of some fingers just after starting solvent injection. The results show that growth rate of the fingers become smaller by decreasing mobility ratio. Finger transitions are the same for different mobility ratios but the fingers size and growth rate of the fingers are different for different mobility ratios. Generated fingers accelerate concentration spreading, so the solvent is mixed faster than that predicted by Convection-Dispersion equation. As the mobility ratio decrease toward one, growth of mixing zone length tends to 0.5, which is the growth rate caused by dispersion alone. By increasing the mobility ratio, fingers causes the mixing zone length growth tends to 1, so, for large mobility ratio, mixing zone grows because of two mechanisms: Dispersion and Fingering.
Reservoir recovery processes are complex and typically entail several physical or chemical mechanisms. Polymer flooding has often been depicted to be dominated by one mechanism: water viscosification reduces the mobility ratio, and stabilizes the displacement front to increase oil recovery. Increasingly though the contribution of mechanisms other than water viscosification is becoming understood. This paper points out two novel insights into polymer flooding: (1) the intrinsic value of the polymer is likely being overestimated and (2) the practice of operating the polymer flood with incomplete voidage replacement may indeed be optimal. The conversion of a conventional waterflood to a polymer flood entails significant injectivity reduction, up to 50% or more. The maintenance of complete voidage replacement (VRR = 1) would thus require an increase in the number of injectors, or a reduction of total production rate or both. As both interventions reduce the economic returns, most projects operate with incomplete voidage replacement (VRR < 1). We have previously reported that a VRR < 1 improves the waterflood response of heavy oil reservoirs. Thus using the VRR = 1 waterflood as the comparison benchmark to the polymer response may overestimate the value of viscosification – the intrinsic value of the polymer. To quantify this, we have performed numerical simulations of polymer flooding for VRRs ranging from 0.4 to 1.4, deconvolving the relative contributions of the viscosification and VRR < 1 mechanisms. We observe that a polymer flood operated with VRR > 1 (above the oil bubble point) underperforms a polymer flood with a VRR < 1 by as much as one third. We conclude that the intrinsic value of the polymer is overestimated.
Mirzabozorg, Arash (Computer Modelling Group Ltd.\University of Calgary) | Nghiem, Long (Computer Modelling Group Ltd.) | Chen, Zhangxin (University of Calgary) | Yang, Chaodong (Computer Modelling Group Ltd.) | Li, Heng (Computer Modelling Group Ltd.)
Population-based optimization algorithms are shown to be excellent candidates for improving the speed and solution diversity of history matching and optimization workflows, based on their successful track records for solving real-world problems.
The incorporation of reservoir engineering knowledge within these workflows, however, has been somewhat neglected. In particular, there is a lack of capability for guiding the optimization algorithms to specific regions of the search space. In a previous study, we introduced a framework for helping reservoir engineers incorporate their knowledge into history matching and optimization frameworks, by coupling a rule-based fuzzy system with a population-based sampling method.
The question is how the use of this type of information in history matching affects the performance of the reservoir study during the prediction stage. This paper investigates the effect that the incorporation of reservoir engineering knowledge during the history matching of the Teal South model production data has on reservoir performance in the prediction stage.
Two scenarios are considered. In Case I, we augment the history matching with reservoir engineering knowledge and then produce a forecast. In Case II, production data is history matched using differential evolution (DE), without fuzzy-logic-based engineering knowledge, then a forecast is produced
The results show that incorporating engineering knowledge of the reservoir under study during the history matching process can significantly reduce the uncertainty in the forecast, compared with the case where unrealistic parameter value ranges are used.
Cyclic steam stimulation (CSS) is a commercial in situ recovery method that involves injecting steam into formations at high pressure to reduce the viscosity of heavy oil and bitumen. The injection pressure is typically above the reservoir fracture pressure to induce fractures, which can improve reservoir permeability and fluid mobility. Geomechanics and heterogeneities near a well significantly affect reservoir fracturing, particularly under thermal conditions. In addition, oversimplified descriptions of fracture geometry may fail to represent actual reservoir performance. Thus modeling near-well flow effects coupled with geomechanics is essential for detailed large-scale thermal simulations with fractures. Conventional methods have exhibited limitations in capturing interactions among wellbore/near-well flow, geomechanics, and fluid flow during CSS development. Therefore, flow/stress/near-well flow coupling under different fracture geometries is investigated in this study.
A facies-controlled geostatistical model is first constructed for a Cold Lake reservoir to obtain more reliable results. A fully coupled model is further generated from this geological model. After local grids near wells are refined, we construct different fracture geometries by changing the fracture length and direction. In this study, uncertainty analyses on fracture geometry near the vertical wells are performed. The simulation results show that omitting geomechanics and wellbore modeling increases oil recovery. Moreover, oversimplified fracture geometry with simple planes overestimates an oil rate. Furthermore, fractures with complex geometries and geomechanics exhibit high conductivity and provides effective channels for steam and bitumen but at the same time, it may cause steam channeling and decrease the efficiency of steam injection.