No abstract

Summary Compared with a conventional reservoir, the ultralow permeability in the Bakken Formation makes it very challenging to perform normal waterflooding or gasflooding operations. “Permeability-jail” effects cause low injectivity and prevent injected fluids from sweeping oil out of the matrix efficiently. Two distinguishable flow regimes have been identified in fractured, hydrocarbon-rich shale formations: viscous flow in high-permeability fracture networks and diffusion-dominated flow in the low-permeability matrix with high oil saturation. Improving hydrocarbon transport (and technically recoverable resources) in unconventional reservoirs relies on our ability to enhance diffusion-dominated flow from the oil-saturated matrix to the natural- or induced-fracture network, which is the focus of this study. To unlock the unproduced Bakken and Three Forks oil, high-pressure carbon dioxide (CO2) may be used to enhance the diffusion-dominated flow in the matrix and keep the viscous flow in the fractures under reservoir temperature and pressure conditions (e.g., 230°F and 5,000 psi). Core samples were collected from two Bakken wells, including all oil-bearing intervals: Upper Bakken (UB), Middle Bakken (MB), and Lower Bakken (LB) Members and the Three Forks (TF) Formation. Detailed core analyses were performed to measure petrophysical properties and characterize these units. Ten samples were selected for pore-size-distribution measurement and 21 samples (11-mm-diameter rods) were used for 24-hour CO2 exposures and hydrocarbon-recovery experiments. These experiments were conducted as CO2 “bathing” at reservoir conditions (rather than “flow through” tests) and were aimed at increasing our understanding of the microstructure and diffusion-dominated-flow ability within these tight geologic formations. CO2-exposure and hydrocarbon-extraction experimental results clearly showed the improvement of diffusion-dominated flow in all the Bakken members. The UB and LB samples, characterized by generally high total-organic-carbon (TOC) content (10–15 wt%) and small pore size (approximately 3–7 nm), yielded approximately 60% of the present mature hydrocarbon at the end of the 24-hour exposure. The MB and TF samples, characterized by lower TOC content (<0.5 wt.%) and moderate pore size (approximately 8–80 nm), provided more-favorable flow conditions for CO2 and hydrocarbons, yielding approximately 90% of the mature-hydrocarbon content. Because all experiments were conducted at reservoir conditions, the results demonstrate that diffusion plays a significant role in the mobilization of oil in tight reservoirs. CO2 greatly enhances the diffusion process to improve hydrocarbon transport in the tight matrix. This observation is especially useful for densely fractured shale-oil formations (high surface-area/volume ratio) where CO2 has greater areal contact with the reservoir, enabling CO2 diffusion into the matrix and hydrocarbon diffusion out of the matrix to occur more efficiently (increasing recoverable reserves), and where the fracture networks assist in alleviating potential injectivity challenges.

Reservoir models typically utilized for desorption-controlled reservoirs such as coals and gas shales possess dual-porosity/ single-permeability characteristics. In this case dual-porosity means that two in-situ locations exist that can be used for gas storage, adsorbed within the matrix and in the free form in the cleat system. Single-permeability, which refers to the cleat system, is the only permeability network that gas or water must flow through to reach the wellbore. While this approach to modeling coals and shales has become accepted practice, experience has shown that the models can frequently be in gross error when forecasting well or field performance based on limited reservoir and/or production data; gas production is usually over-predicted and water production under-predicted. The implications for economic decision-making in an exploration mode are obvious, and there are many examples of projects that have suffered from this very problem. Further, reservoir parameters derived from history-matching, when historical gas production does exist, are commonly found to be inconsistent with measured permeability and gas sorption/content data. While there has been considerable effort focused on improved data collection procedures, such as well testing and gas content measurement for example, these problems persist. While performing reservoir studies in the Antrim shale and low-rank coal plays throughout the world, it became clear that the accepted assumption of gas desorbing directly from the coal matrix into the cleat system is not entirely valid. In practice, gas production occurs much later than the models predict, and cannot be adequately explained though the normal parameters of sorption time, permeability, relative permeability, etc. Analysis of core and other data suggests that another porosity and permeability system is required to account for this effect, specifically within the matrix blocks themselves. An advanced, triple-porosity/dual-permeability model has therefore been developed, in which gas desorbs from the internal matrix block surfaces, migrates via conventional Darcy flow through micro-permeability matrix, and into the cleat system where it then flows to the wellbore. Water can also be stored both within the matrix blocks and in the cleat system. In essence, this model requires that desorbed gas must work its way through the matrix before reaching the cleat system, and must establish a relative permeability to gas within the matrix block before it can do so. This geometry is similar to conventional dual-porosity models, with the addition of an adsorbed gas component. Comparisons of this new model versus the historical modeling approach confirm that the new model predicts lower gas and higher water production rates, consistent with field evidence. Further, more accurate production forecasts can be achieved using measured well test information (for the cleat permeability), low cleat porosities (which are known to exist), and lab-derived porosity and permeability data for the matrix block properties. This paper presents the historical accuracy problem with reservoir simulation in desorption-controlled reservoirs, the practical theory behind the new model, comparisons between the new and conventional models, and some example applications.

Oil production grew significantly from 0.2 million barrels per day (bpd) to 1.1 million bpd in the Bakken petroleum system from 2009 to 2014. A large volume of associated gas (1.6 billion cubic feet per day) has also been produced with the oil. A substantial part (>10%) of this produced gas is flared off because of the low natural gas price and limited infrastructure for gathering and transporting the gas from the well sites. Such a large scale of gas flaring not only wastes energy but also emits contaminants such as SOx, NOx, and CO2 to the atmosphere. Reduction of flaring and utilization of produced gas are important steps toward sustained development of the Bakken. The potential for recycled gas enhanced oil recovery (EOR) is being investigated as a method of reducing flaring through utilization. However, large-scale gas flooding might be difficult for the Bakken because of the difference between the low-permeability matrix and the highly conductive hydraulic and natural fracture networks, which may lead to low sweep efficiency. Instead, this research by the Energy & Environmental Research Center (EERC) has aimed to investigate, through a series of laboratory experiments and numerical simulation activities, the potential to extract oil from the tight rocks by taking advantage of diffusion-based processes. Oil and gas produced from Bakken wells were characterized, and the reservoir formation properties were analyzed based upon core samples. A series of oil extraction experiments with varying gas (solvent) compositions were conducted. The minimum miscibility pressure (MMP) of various produced gas components and oil was measured to determine the pressure required for effective extraction. Based on the experimental results, a well-scale model was developed to simulate the performance of recycled gas EOR. Results showed CO2 and produced Bakken gas to be miscible with the oil in reservoir conditions (>5000 psi, 230°F). The measured MMPs for pure CO2 and ethane with typical Bakken oil samples were 2528 and 1344 psi, respectively. The presence of methane in the gas increased MMP, but miscibility was still achievable under reservoir conditions. CO2 and ethane enabled extraction of most oil components from the rocks during a 24-hour experimental period, but methane exhibited strong molecular selectivity for light-end components. Simulation results showed that a single-well CO2 and methane/ethane huff ‘n’ puff operation could increase cumulative oil production as much as 50% for the multistage fractured wells in the Bakken. The results of this study clearly showed that produced Bakken gas could be effectively used for recycled gas EOR. Implementation of EOR may have potential to compensate for the production decline of Bakken wells while reducing the quantity of flared gas.

No abstract

Over 40 rock samples were obtained from six Bakken wells which penetrate through the major oil pay including two shale intervals: Upper and Lower Bakken, and two tight intervals that are the targets for drilling: Middle Bakken and Three Forks. Detailed petrographic and petrophysical analyses were performed on the samples to better correlate the extraction results with the physical and geochemical properties of the rocks. Round rods (11.2-mm diameter X ca. 30–40 mm long) drilled from each of the 40 samples were individually exposed in a "bath" of CO2 for 24 hours at reservoir temperature and pressure of 5000 psi and 230°F (34.5 MPa, 110°C), and the recovered crude oil hydrocarbons were collected periodically and analyzed to determine the rates and efficiencies of oil recovery. For the 26 Middle Bakken and Three Forks rocks, hydrocarbon recovery upon CO2 exposure averaged 86% after 7 hours, and 99% after 24 hours. Recoveries of the crude oil (not including kerogen) from the 15 Upper and Lower shales were surprisingly high with an average of 30% recovered after 7 hours, and 50% recovered after 24 hours. While the Middle Bakken and Three Forks TOC values were ca. 0.3 wt.% (similar to their crude oil content), TOCs for the Upper and Lower Bakken shales were typically 10 to 15 wt.%, with ca. one-tenth of that organic content being crude oil hydrocarbons as opposed to kerogen. The Upper and Lower shales also had significantly smaller pore throat sizes (averaging ca. 3 nm) than the Middle Bakken and Three Forks samples (which averaged ca. 10–26 nm). Additional studies are being performed to determine whether the small pore throat sizes (which approach molecular dimensions) and/or the sorption of crude oil hydrocarbons onto the kerogen in the Upper and Lower shales are responsible for the slower hydrocarbon recovery than that achieved from the Middle Bakken and Three Forks rocks under CO2 exposure. Currently, the main targets for horizontal drilling are Middle Bakken and Three Forks, where thousands of multistage hydraulically fractured wells have been drilled in the past decade. The high oil recovery factor observed in cores from these intervals, especially when compared to the 7% average recovery in the field, indicates the huge potential for oil recovery factor improvement in these units by increasing oil production based upon supercritical CO2 extraction.

No abstract

No abstract