Shale formations in North America such as Bakken, Niobrara, and Eagle Ford have huge oil in place, 100–900 billion barrels of oil in Bakken only. However, the predicted primary recovery is still below 10%. Therefore, seeking for techniques to enhance oil recovery in these complex plays is inevitable. Although most of the previous studies in this area recommended that CO2 would be the best EOR technique to improve oil recovery in these formations, pilot tests showed that natural gases performance clearly exceeds CO2 performance in the field scale. In this paper, two different approaches have been integrated to investigate the feasibility of three different miscible gases which are CO2, lean gases, and rich gases.
Dheiaa Alfarge1, Mingzhen Wei2, Baojun Bai2
1Iraqi Ministry of Oil, Baghdad, Iraq. 2Missouri University of Science and Technology, Rolla, MO, USA
Received: 31 May 2017 / Accepted: 13 August 2017
©The Author(s) 2017.
Firstly, numerical simulation methods of compositional models have been incorporated with local grid refinement of hydraulic fractures to mimic the performance of these miscible gases in shale reservoirs conditions. Implementation of a molecular diffusion model in the LS-LR-DK (logarithmically spaced, locally refined, and dual permeability) model has been also conducted. Secondly, different molar-diffusivity rates for miscible gases have been simulated to find the diffusivity level in the field scale by matching the performance for some EOR pilot tests which were conducted in Bakken formation of North Dakota, Montana, and South Saskatchewan.
The simulated shale reservoirs scenarios confirmed that diffusion is the dominated flow among all flow regimes in these unconventional formations. Furthermore, the incremental oil recovery due to lean gases, rich gases, and CO2 gas injection confirms the predicted flow regime. The effect of diffusion implementation has been verified with both of single porosity and dual-permeability model cases. However, some of CO2 pilot tests showed a good match with the simulated cases which have low molar-diffusivity between the injected CO2 and the formation oil. Accordingly, the rich and lean gases have shown a better performance to enhance oil recovery in these tight formations. However, rich gases need long soaking periods, and lean gases need large volumes to be injected for more successful results.
Furthermore, the number of huff-n-puff cycles has a little effect on the all injected gases performance; however, the soaking period has a significant effect. This research project demonstrated how to select the best type of miscible gases to enhance oil recovery in unconventional reservoirs according to the field-candidate conditions and operating parameters. Finally, the reasons beyond the success of natural gases and failure of CO2 in the pilot tests have been physically and numerically discussed.
The Energy Information Administration (EIA) reported that US tight oil production including shale formations will grow to more than 6 million bbl/day in the coming decade, making up most of the total US oil production as shown in Fig. 1. Oil production from tight formations including shale plays has just shared for more than 50% of total oil production in US (Alfarge et al. 2017). Hoffman and Evans (2016) reported that 4 million barrels per day as an increment in US oil daily production comes from these unconventional oil reservoirs. From 2011 to 2014, Unconventional Liquid Rich (ULR) reservoirs contributed to all natural gas growth and nearly 92% of oil production growth in the US (Alfarge et al. 2017).
Fig. 1 Shale and tight oil production in North America from U.S. EIA (Feb-2017).
Specifically, Bakken and Eagle Ford contributed for more than 80% of total US oil production from these tight formations (Yu et al. 2016). This revolution in oil and gas production happened mainly because shale oil reservoirs have been just increasingly developed due to the advancements in horizontal wells and hydraulic fracturing in last decade. Several studies have been conducted to estimate the recoverable oil in place in these complex formations indicating large quantities of oil in place. The available information refers to 100–900 billion barrels in Bakken only. However, the predicted recovery from primary depletion could lead to 7% only of original oil in place (Clark 2009).
Furthermore, some investigators argued that the primary recovery factor is still in a range of 1–2% in some of these plays in North America (Wang et al. 2016). For example, the North Dakota Council reported that “With today’s best technology, it is predicted that 1–2% of the reserves can be recovered” (Sheng 2015). The main problem during the development of unconventional reservoirs is how to sustain the hydrocarbon production rate, which also leads to low oil recovery factor. The producing wells usually start with high production rate initially; however, they show steep decline rate in the first 3–5 years until they get leveled off at very low rate.
According to Yu et al. (2014), the main reason beyond the quick decline in production rate is due to the fast depletion of natural fractures networks combined with slow recharging from matrix system, which is the major source of hydrocarbon. Therefore, oil recovery factor from primary depletion has been predicted typically to be less than 10% (LeFever and Helms 2008; Clark 2009; Alharthy et al. 2015; Kathel and Mohanty 2013; Wan and Sheng 2015; Alvarez and Schechter 2016).
Since these reservoirs have huge original oil in place, any improvement in oil recovery factor would result in enormous produced oil volumes. Therefore, IOR methods have huge potential to be the major stirrer in these huge reserves. Although IOR methods are well understood in conventional reservoirs, they are a new concept in unconventional ones. All the basic logic steps for investigating the applicability of different IOR methods such as experimental works, simulation studies, and pilot tests have just started over the last decade (Alfarge et al. 2017).
Fig. 2 Performance of natural gas-EOR in Canadian-Bakken conditions (Schmidt and Sekar 2014).
However, most of the studies in this area focused on CO2 due to different reasons. CO2 can dissolve in shale oil easily, swells the oil and lowers its viscosity. Also, CO2 has a lower miscibility pressure with shale oil rather than other gases such as N2 and CH4 (Zhang 2016). Furthermore, experimental studies reported an excellent oil recovery factor could be obtained by injection of CO2 in small chips of tight-natural cores (Hawthorne et al. 2017). Unfortunately, the results of pilot tests for CO2-EOR, huff-n-puff protocol, which have been conducted in unconventional reservoirs of North America, were disappointing as shown in Fig. 3.
Fig. 3 CO2 pilot tests performance in Bakken (Modified from Hoffman and Evans 2016). a Pilot test#1. b Pilot test#2.
This gap in CO2 performance between laboratory conditions versus to what happened in field scale suggests that there is something missing between microscopic level and macroscopic level in these plays. Most of the experimental studies reported that the molecular diffusion mechanism for CO2 is beyond the increment in oil recovery obtained in laboratory scale (Alfarge et al. 2017).
Furthermore, most of the previous simulation studies relied on the laboratory diffusivity level for these miscible gases to predict the expected oil increment on field scale (Alfarge et al. 2017). One of the main reasons for the poor performance of CO2 in the pilot tests might be due to the wrong prediction for CO2 diffusion mechanism in these types of reservoirs. A detailed study for determining the level of CO2 diffusivity in the real field conditions have been conducted in this work. Also, comparing CO2 performance with lean gas and rich gas according to different levels of diffusivity has been investigated to clarify the flow and recovery mechanisms for different gases in shale reservoirs.
Gravity drainage, physical diffusion, viscous flow, and capillary forces are the common forces which control the fluids flow in porous media. However, one force might eliminate the contributions of other forces depending on the reservoir properties and operating conditions. Molecular diffusion is defined as the movement of molecules caused by Brownian motion or composition gradient in a mixture of fluids (Mohebbinia et al. 2017).
This type of flow would be the most dominated flow in fractured reservoirs with a low-permeability matrix when gravitational drainage is inefficient (Moortgat and Firoozabadi 2013; Mohebbinia et al. 2017). The role of molecular diffusion flow increases as far as the formation permeability decreases. It has been noticed and approved that gas injection is the most common EOR process affected by calculations of molecular diffusion considerations. Ignoring or specifying incorrect diffusion rate during simulation process can lead to overestimate or underestimate the oil recovery caused by the injected gas.
This happens not only due to the variance in miscibility process between the injected gas and formation oil but also due to the path change for the injected gas species from fractures to the formation matrix. Da Silva and Belery (1989) reported that molecular diffusion process is happened by three mechanisms:
- Bulk diffusion where fluid–fluid interactions dominate.
- Knudsen diffusion where fluid molecule collides with pore wall (happens when molecular mean free pathway closer to pore size).
- Surface diffusion where fluid molecules transported along adsorbed film (minor unless thick adsorbed layer).
The Péclet number (Pe) is a class of dimensionless numbers which have been used to measure the relative importance of molecular diffusion flow to the convection flow. This number can be calculated as shown in Eq. 1. If Pe number is less than 1, diffusion is the dominant flow. However, if Pe is greater than 50, convection is the dominant flow. The dispersion flow is dominant when Pe is in the range of 1–50 (Hoteit and Firoozabadi 2009).
Figure 4 explains the flow regimes according to Péclet number cutoffs. The movement of fluid components in field conditions is equal to the integration of diffusion, dispersion, and viscous forces. Therefore, the total average velocity of any component is equal to the sum of dispersive velocity and bulk phase velocity (Da Silva and Belery,1989).
where v is the bulk velocity, L is a characteristic length, and D is the diffusion coefficient.
Fig. 4 Flow regimes according to Péclet number cut offs.
CO2 molecular diffusion mechanism
Different mechanisms have been proposed for the injected CO2 to improve oil recovery in unconventional reservoirs as shown in Table 1. However, since the matrix permeability in these unconventional reservoirs is in range 0.1–0.00001 md, CO2 would not be transported by convection flux from fracture to matrix (Yu et al. 2014). The main transportation method for CO2 is happening by the difference in concentration gradient between CO2 concentration in injected gases and the target oil. This process of transportation is subjected to Fick’s law. Hawthorne et al. (2013) extensively investigated the CO2 diffusion mechanism in Bakken cores and proposed five conceptual steps to explain it.
Table 1 Proposed CO2 EOR mechanisms for improving oil recovery in unconventional reservoirs.
These conceptual steps include: (1) CO2 flows into and through the fractures (2) unfractured rock matrix is exposed to CO2 at fracture surfaces, (3) CO2 permeates the rock driven by pressure, carrying some hydrocarbon inward; however, the oil is also swelling and extruding some oil out of the pores, (4) oil migrates to the bulk CO2 in the fractures via swelling and reduced viscosity, and (5) as the CO2 pressure gradient gets smaller, oil production is slowly driven by concentration gradient diffusion from pores into the bulk CO2 in the fractures.
Most of the previous experimental studies reported that CO2 diffusion mechanism is beyond the increment in oil recovery obtained in laboratory conditions. Then, the observed increment in oil recovery and/or the CO2 diffusion rate obtained in laboratory conditions were upscaled directly to field scale by using numerical simulation methods. Although modeling of the diffusion effect on ultimate oil recovery in shale reservoirs is very important to develop these marginal shale oil projects, evaluation of the recovery contribution from diffusion will help in understanding the recovery mechanisms (Wan and Sheng 2015).