Gas Injection in the Bakken Formation: An Overview
Peng Luo and Sheng Li, Saskatchewan Research Council
For decades, the low-oil-mobility of the Bakken formation has been the main challenge for development. Significant progress has been made since the application of horizontal drilling and hydraulic fracturing. The estimated recovery factors of the Bakken formation, however, still remain quite low. The application of enhanced-oil-recovery (EOR) technologies is now a necessity for further reducing residual oil. One of the possible approaches is gasflooding. Parallel to some pilot tests being conducted in the Saskatchewan Bakken field, which mainly use produced gas, the Saskatchewan Research Council (SRC) has evaluated the technologies of gasflooding for the Bakken formation in their laboratories. Their approach involved extended pressure/volume/temperature (PVT) tests, geochemistry studies, coreflooding tests, and numerical simulations. They concluded that gasflooding is a promising EOR technology for the Bakken formation. This article presents an accurate summary of their activities and progress. Your feedback regarding the technical content of this article would be highly appreciated. We encourage you to send your comments to: JCPT@spe.org.
Jian-Yang Yuan, Osum Oil Sands Corporation
JCPT Executive Editor
Light/tight oil, such as that in the Bakken formation, is playing an increasingly important role in Saskatchewan’s oil production boom in recent years. The SRC is one of a few technology institutions that work alongside industry to address the technical challenges of its continued production. The most immediate challenge for tight oil development is the fast rate at which primary production declines and its low recovery factor.
For more than 30 years, SRC has been leading and participating in a large number of EOR initiatives. One of the EOR processes that has been proved repeatedly in the field is miscible or immiscible gasflooding for a wide range of oil reservoirs. For tight oil reservoirs, gasflooding provides an inherent advantage over waterflooding in terms of injectivity. The laboratory evaluation process, critical for screening, understanding, and predicting gasflooding scenarios for tight oil, is described here.
This Technical Report is not peer-reviewed, but highlights a topic that may be of interest to JCPT readers.
The Bakken formation, with at least 16 billion m3 (100 billion bbl) of hydrocarbon resources, is one of the largest oil-bearing formations discovered in North America since the 1950s. Characterized as a “tight” formation by its extremely low permeability and porosity, the Bakken traps oil in its tightly compacted grains, making it difficult to flow. Oil production from the Bakken had been marginal until the early 2000s, when technology advancement in horizontal well drilling and multistage hydraulic-fracturing completion unlocked enormous quantities of high-quality light oil. In southeast Saskatchewan, Bakken oil production has been from the middle dolomitic sandstone/siltstone member, with thickness varying from zero to 25 m and permeabilities predominantly less than 1.0 millidarcy (Kreis and Costa 2005). The Saskatchewan Bakken oil boom is evidenced by the surge in oil production from an average 760 B/D in 2004 to 61,420 B/D in 2014 (Yurkowski 2014).
Because of its tight nature, the Bakken experiences fast and significant decline during primary production, which relies on the internal energy of reservoir fluids and compressibility of the formation rock. The substantial amounts of residual oil (>90% original oil in place) left behind in the Bakken demand technically viable and cost-effective EOR technologies. For conventional oil reservoirs, waterflooding (i.e., secondary recovery) is the most widely used technology after primary production because of its low facility and operating cost and effectiveness. However, in most of the Bakken fields, because of their low permeability, the injectivity for waterflooding can be impractically low. Gasflooding, as another major branch of EOR technologies, offers inherent advantages over waterflooding because of the much lower gas viscosity compared with water (Jarrell et al. 2002). At present, several pilot tests using produced gas have been implemented in southeast Saskatchewan Bakken fields.
The SRC started research and development on immiscible/miscible gasflooding technologies for various types of reservoirs nearly 3 decades ago. Its extensive experience has been applied to tight oil reservoirs in recent years. This short review highlights guidelines for laboratory evaluation of gasflooding for the Bakken formation and other tight oil fields.
Besides providing significant injectivity improvement, gases dissolve into the reservoir oil phase to reduce the oil viscosity and swell the oil. Bakken reservoir conditions and oil properties facilitate high gas dissolution for condensable gases (e.g., carbon dioxide and ethane). Many Bakken reservoirs are able to reach a miscible flooding condition with a suitable injection gas stream. On the other hand, the drawbacks of gasflooding are also closely related to the characteristics of injection gas. The rather low viscosity of gas creates an unfavourable mobility ratio that may cause gas breakthrough at an early stage of production. High-pressure gas injection in the field requires considerable upfront facility and operating costs (e.g., pipeline, compressors, and injection pumps). CO2, as an effective EOR injectant, forms carbonic acid and may react with reservoir rock and formation brine, changing the reservoir connectivity and fluid pathways in positive or negative way. Finally, gas source availability is always a practical concern for a field gasflooding project. The best gas candidate in terms of recovery factor in the laboratory may not be the first option in the field from an economic point of view.
A successful gasflooding field project always requires well-designed laboratory experimental and simulation studies. It is vital to conduct all experimental work with representative reservoir fluids and rocks. A laboratory evaluation methodology is described as follows.
1. Live Oil Recombination and PVT Tests. In most cases, bottomhole fluid is unavailable because of its high sampling cost. The live oil is often reconstituted in the laboratory using separator oil and gas. The reconstitution can be based on either a prespecified saturation pressure (Psat) or the gas/oil ratio (GOR) of the reservoir oil. Several important properties (i.e., Psat, formation volume factor (FVF), GOR, density, and viscosity) are then measured for the live oil reconstituted in this method.
A differential liberation test simulates the solution-gas drive mechanism in the reservoir or near the production wellbore, when the reservoir pressure or bottomhole pressure is lower than the live oil’s bubblepoint pressure. In the laboratory, this is accomplished by depressurizing a live oil at several stages. The gas cap at each stage is removed, then the oil phase is depressurized to the next stage. Oil and gas compositions and other PVT properties are measured at each stage.
2. PVT Tests of Live Oil/Injection Gas Mixture. The phase behaviour of the reconstituted live oil and injection gas is studied through PVT tests in a pressure range covering the targeted injection pressure in the field. During these tests, the live oil is saturated with the injection gas at several incremental concentrations. The properties of the oil/gas mixture (i.e., Psat, density, viscosity of mixture, GOR, FVF, and swelling factor) are measured after equilibrium is reached at each saturation pressure. Important gasflooding recovery mechanisms (e.g., viscosity reduction and oil swelling) can be quantified in these tests. Finally, all the PVT test results for the live oil and live oil/gas mixtures are used to create an equation-of-state (EOS) model. A well-tuned EOS model gives reservoir engineers strong confidence in conducting follow-up compositional simulation for gasflooding.
3. Minimum Miscibility Pressure (MMP).
Gas injection for light oil reservoirs like the Bakken is usually a multiple-contact miscible flooding process. Depending on the compositions of the oil and injected gas and the reservoir pressure and temperature, the injection gas dissolves into the oil and simultaneously extracts light ends from the oil until miscibility is achieved (Green and Willhite 1998), which causes the oil and gas flow as a single phase. To obtain maximum recovery performance, it is essential to operate the gasflood as a miscible process by maintaining the reservoir pressure above the MMP.
The MMP for the live oil/injected gas system can be determined using the rising bubble apparatus (RBA) method and a slimtube test. In the RBA method, MMP is determined by observing the behaviour of an injection gas bubble as it rises through a column of live oil under increasing pressure. This method is quick and easy to operate, but provides only the thermodynamic MMP in the absence of porous media. In a slimtube test, gas is injected into a coiled slimtube (usually 18 m or longer) to displace live oil saturated in a sandpack. MMP is determined by the intersection of two straight lines drawn through recovery points in the lower, steeply climbing region and the higher leveled-off region. Compared with the RBA test, the slimtube test is more representative of a field scenario but requires significantly more time and effort. A good practice would be to conduct RBA tests to find the approximate range of MMP, then to conduct slimtube tests at several pressure points covering this range to pinpoint the accurate MMP for the live oil/gas system.
4. Geochemistry During CO2 Flooding. The middle member oil-bearing zone of the Bakken formation has extreme lithological variability, both vertically and along strikes; and it is commonly host to highly saline formation brines. Complex geochemical and geophysical processes from CO2/rock/brine interactions under reservoir conditions have to be accounted for in any design of a CO2 flooding process in Bakken reservoirs. Experimentally, scanning electron microscopy-energy dispersive spectroscopy, X-ray diffraction, and petrographic analysis are powerful tools to analyze morphological and mineralogical changes to samples following CO2/rock/brine interaction for both long-term static saturation tests and dynamic-injection tests. Detailed results of porosity, permeability, relative permeability, and wettability can help oil producers determine the viability of, and select optimal locations for, CO2 flooding in Bakken reservoirs.
5. Core Displacement Tests. Core displacement tests are the most direct laboratory approach to evaluate the recovery performance of a given flooding process, when they are conducted at the reservoir temperature and pressure with actual reservoir plugs and fluids. In a typical coreflood, formation brine and recombined live oil are sequentially injected into a core stack and aged for more than 4 weeks before the flood. This aging process helps partially restore original reservoir rock wettability and re-establish the fluid distribution. During a coreflood, different field-production scenarios (e.g., primary depletion, secondary waterflooding, and/or gas injection) can be implemented at the desired injection pressure or rate comparable to the field case. Produced fluids are collected and measured, and the pressure difference along the core stack is monitored. It is worthwhile to note that a linear coreflood in a stack of core plugs has an inherent drawback for representing tight oil reservoirs, where fluids flow across the matrix and fractures and travel in the fracture network. As well, the capillary end effect causes sudden changes in saturations at the end of each plug, creating experimental uncertainty and in turn influencing the oil recovery and relative permeability endpoint. A 3D, dual-permeability physical model in which fluids can flow from matrix to fracture or vice versa is under development at SRC.
6. Numerical Simulation: From EOS Modelling to Field-Scale Reservoir Simulation. The goal of numerical simulation is to build a reliable reservoir model that can be used for production prediction and sensitivity analysis, using all the information obtained from the previously conducted laboratory experimental studies. The first step is to generate and tune the EOS model for reservoir fluids. The tuned EOS model is imported into a coreflood model that is built on the basis of the actual core stack size, petrophysical properties, and injection sequences. The production data from the coreflood test are history matched by tuning oil/water and gas/liquid relative permeability curves to further generate three-phase relative permeability. Finally, a dual-permeability model can be developed to simulate a naturally fractured reservoir. The relative permeability curves are then incorporated into the field model. Well trajectory and hydraulic fracture properties can be set up based on the real situation. Parameters in the model that are hard to measure (e.g., well skin factor, fracture network, natural fracture spacing, and permeability) are often set up as tuning parameters. After the model is tuned with available well production data, the model can be further used to predict the performance of gas injection and search for the most economic injection strategy.
Gasflooding is a promising EOR technology for recovering oil from the tight Bakken formation. A systematic program of laboratory evaluation studies is required before a gasflood can be implemented successfully in the field. The advantages and disadvantages of gasflooding should be analyzed carefully with consideration of specific field conditions. The laboratory approach described in this review can also be adopted to develop other tight oil reservoirs (e.g., Bakken’s underlying Torquay formation) if specific field geological factors and fluid properties are considered.
Green, D. W. and Willhite, G. P. 1998. Enhanced Oil Recovery, first edition, 6. Richardson, Texas, USA: Spe Textbook Series, Society of Petroleum Engineers.
Jarrell, P., Fox, C., Stein, M. et al. 2002. Practical Aspects of Co2 Flooding, edition. Richardson, Texas: Monograph Series, SPE.
Kreis, L.K. and Costa, A. 2005. Hydrocarbon Potential of the Bakken and Torquay Formations, Southeastern Saskatchewan. In Summary of Investigations, Saskatchewan Geological Survey, Saskatchewan Industry and Resources, Misc. Rep. 2005-4.1.
Yurkowski, M. 2015. Petroleum Geology of Saskatchewan. http://wbpc.ca/pub/documents/archived-talks/2015/Presentations/Yurkowski%20-%20Petroleum%20Geology%20of%20Sask.pdf, accessed June.