Study on the Flow Mechanism of Shale Oil with Different Injection Media

Chinese shale oil has high recoverable resources and great development potential. However, due to the limitation of development technology, the recovery rate of shale oil is not high. In this paper, the effects of different injection media on the development of shale oil reservoirs in Dongying formation, Qikou depression, Huanghua depression, and Bohai bay basin, were studied by means of imbibition and nitrogen flooding. Combining nuclear magnetic resonance (NMR) technology with imbibition and gas displacement experiments, the mechanism of shale injected formation water, active water (surfactant), and nitrogen was reproduced. The displacement process of crude oil under different injection media and injection conditions was truly demonstrated, and the relationship between different development methods and the pore boundaries used was clarified. A theoretical basis for the effective development of shale oil was provided. At the same time, Changqing tight oil cores with similar permeability to Dagang shale oil cores were selected for comparison. The results showed that, as the imbibition time of shale samples increased, the imbibition efficiency increased. Pores with T2 < 10 ms contributed the most to imbibition efficiency, with an average contribution greater than 90%. 10 ms < T2 < 100 ms and more than 100 ms pores contributed less to imbibition efficiency. Active water can change the wettability of shale, increase its hydrophilicity, and improve the efficiency of imbibition. The imbibition recovery ratio of injected active water was 17.56% higher than that of injected formation water. Compared with tight sandstone with similar permeability, the imbibition efficiency of shale was lower. As the nitrogen displacement pressure increased, the oil displacement efficiency also increased. The higher the shale permeability was, the greater the displacement efficiency would be. T2 > 100 ms pore throat of shale contributed to the main oil displacement efficiency, with an average oil displacement efficiency contribution of 63.16%. And the relaxation interval 10 < T2 < 100 ms pore throat displacement efficiency contributed to 28.27%. T2 < 10 ms pore throat contributed the least to the oil displacement efficiency, with an average oil displacement efficiency contribution of 8.58%. Compared with tight sandstone with similar permeability, shale had lower oil displacement efficiency. The findings of this study can help for better understanding of the influence of different injection media on shale oil recovery effect.


Introduction
After tight oil, shale oil has become a research hotspot in the field of exploration and development. Compared with conventional sandstone, shale is dominated by nanoscale pores, with a special pore structure, complex rock composition, and ultra-low permeability [1,2]. In 2016, the proven reserves of shale oil in the United States were 2.222 billion tons, of which Bakken's proven reserves of shale oil were 747 million tons, ranking first. e average daily output of shale oil commercial development in the United States reached 670,000 tons in 2017, and the average daily output in the first half of 2018 exceeded 800,000 tons [3]. In North America, the study of shale oil theory has become increasingly sophisticated [4][5][6]. Loucks et al. classified shale pores by scanning electron microscopy; then, shale pores were divided into three categories: inorganic intergranular pores, inorganic intragranular pores, and organic pores [7,8].
Labani et al. quantitatively characterized the pore structure of shale based on gas adsorption method, high-pressure mercury intrusion method, nuclear magnetic resonance method, and small-angle scattering method [9]. Chalmers et al. carried out a joint characterization of the pore size of several shales in North America by nitrogen adsorption and CO 2 adsorption [10]. In terms of imbibition, the predecessors mainly explored the cause of fracturing fluid retention in shale reservoirs through spontaneous imbibition experiments [11,12], but the spontaneous imbibition process had not been studied in depth. e shale oil revolution in North America has promoted the development of shale oil exploration and development theories in the world, and it also set off a wave of shale oil exploration and development in China [13][14][15][16]. Shale oil resources are mainly distributed in Meso-Cenozoic continental basins, such as Songliao Basin, Ordos Basin, Bohai Bay Basin, and Sichuan Basin.
Due to the difference between the type of continental shale oil resources in China and the type of marine shale oil resources in the United States, there is still a lack of suitable production technology, resulting in relatively low shale oil production in China and great development potential [17]. Although Chinese research pages on shale oil are increasing, the current research is currently focused on the field of microscopic pore structure characteristics. Wang et al. graded and evaluated shale reservoirs by means of average pore size, specific surface area, and chlorite relative content obtained by XRD and cryogenic nitrogen adsorption tests of whole rock [18]. Lu et al. used nuclear magnetic resonance technology to test the dynamic fluid of shale oil [19]. e results indicated that the content of the dynamic fluid in shale oil reservoir was very low and it was mainly stored in micron-level pores, with poor fluid mobility and great difficulty in development. Lu et al. used high-pressure mercury intrusion technology to characterize the micropore throat structure of shale oil reservoirs [20]. On this basis, shale oil reservoir classification evaluation standards and lower limits of reservoir formation were established. e micropore structure of shale oil reservoir is the basis of its reservoir evaluation and effective development [21,22], and the evaluation of development mode and the range of pores is particularly important for the enhanced oil recovery of shale oil [23]. At present, the research on the oil production law under different injection media and different injection conditions is mainly focused on tight oil. Zhang et al. used nuclear magnetic resonance (NMR) technology to conduct spontaneous imbibition experiments on tight oil reservoirs; then, the contribution of pores of different sizes to imbibition recovery was analyzed quantitatively [24]. e conclusion was obtained that three reasonable ratios were conducive to imbibition recovery. Li et al. studied the effects of the pore structure of tight sandstone reservoirs on imbibition characteristics and self-priming oil displacement efficiency by means of mercury injection, scanning electron microscopy, and nuclear magnetic resonance (NMR) [25]. e results showed that the ratio of large pore throat determines the oil displacement efficiency, and the residual oil is mainly retained in the micropore throat. e higher the ratio of large hole throat was, the higher the efficiency of self-absorption oil displacement would be. Mannon and Chilingarian studied the relationship between the rate of water injection and absorption through laboratory experiments using models to simulate fractured matrix reservoirs [26]. e results showed that, under linear countercurrent conditions, the increase in water injection rate led to an increase in the rate of water absorption and an increase in the final recovery rate. Cuiec et al. found that the lower the permeability was, the stronger the permeability would be, and the potential of crude oil in the matrix system relying on imbibition in tight reservoirs was greater than that of medium-high permeability [27]. Li et al. used reservoir wettability contact angle experiments to characterize the lipophilicity of tuffaceous mudstone samples [28]. Zhang et al. found that the surfactants can promote the spontaneous imbibition of matrix shale [29]. Yang et al. applied NMR technology to physical model experiments and obtained a new method to calculate the mixed wettability index of tight oil reservoirs [30]. Zuwa and Omoregle determined the operating pore boundary of the injected medium displacing crude oil by mercury injection method and semipermeable separator method [31]. However, these two methods have difficulties in characterizing the fine pores of the tight reservoir, and they could not truly show the real displacement process of the injected medium and crude oil under different injection media and conditions. rough a large number of literature surveys, it can be found that the previous studies on the flow mechanism of different injected media were mainly focused on tight oil reservoirs or shale gas reservoirs, and there were few studies on shale oil reservoirs [32,33]. In order to solve the above problems, this study took shale samples from Qikou sag of Huanghua depression, eastern Bohai Bay Basin, China, as the research object, and combined nuclear magnetic resonance (NMR) technology with physical simulation experiments to analyze the flow mechanism of shale oil in different injection media. First, physical simulation experiments of spontaneous imbibition and nitrogen displacement were established, and then the degree of crude oil recovery was analyzed by measuring the nuclear magnetic resonance spectra of rock samples under different imbibition times and gas flooding pressures, so as to provide a theoretical basis for the effective development of shale oil.

Principle of Nuclear Magnetic Resonance (NMR)
Experiment. Nuclear magnetic resonance (NMR) is a common test and analysis method for reservoir description and evaluation based on the observation of hydrogen nuclear signals in rock pores, the direct measurement of pore fluid properties, the acquisition of reservoir physical-property related parameters, and the calculation of dynamic fluid parameters. In the field of shale oil reservoirs, the measurement error of many common testing methods has become larger, thus misleading the interpretation of shale oil reservoirs. e nuclear magnetic resonance (NMR) technique can be used as an approximate nondestructive technique to measure the pore structure of shale oil reservoirs [34].
When saturated oil, water, or oil-water rock samples are placed in a magnetic field, hydrogen atoms in the water will be affected by the magnetic field to produce polarized hydrogen atoms. is kind of nucleus has a certain magnetic moment. When this nucleus is in an RF (radio frequency) magnetic field, the hydrogen atoms oscillate in a phenomenon called nuclear magnetic resonance. When the applied RF magnetic field disappears, the oscillating hydrogen atom loses its energy source and relaxes in the pore space, its kinetic energy decreases with time, and the amplitude of oscillation decreases accordingly. e increasingly slow motion signal of the hydrogen atom core can be divided into T 1 and T 2 according to the decay rate, which, respectively, represent the longitudinal and transverse relaxation time.
Since the longitudinal relaxation time is slow and the measurement has a large error, the measurement of NMR is generally the transverse relaxation time T 2 [35].

Nuclear Magnetic Resonance (NMR) Laboratory
Apparatus. Recore-04 core NMR analyzer w developed by the institute of percolation, the Chinese academy of sciences, was used in the NMR experiment in this study. In the main test parameters, the echo interval was three hundred milliseconds. e number of echoes was one thousand twentyfour. e number of scanning was sixty-four. e waiting time was three thousand milliseconds. e gain was fifty. e influence of temperature on NMR results was reduced by measuring the standard sample before core NMR. All the NMR T 2 measurement experiments in this study took the above data as parameters.

Shale Oil Cores in Dagang.
Huanghua depression is a sublevel depression in Bohai Bay Basin in eastern China. It is a rifting and depression basin developed in the mainland. e east and west sides of the depression are sandwiched by Chengning uplift and Cangxian uplift. Secondary fault depression of Qikou sag in Bohai gulf basin tectonic units, of the genus in formed since Cenozoic intracontinental extension basin. e main body is located in the center of Bohai bay basin. Paleogene system includes two sets of Shahejie and Dongying group formation. And Shahejie formation are further divided into a sand, the second member of the sand, and sand three sections. Among them, a sand is further divided into three subsections: sand below, sand in the middle, and sand above. e inclined direction of the slope in the southwest margin of Qikou Depression is the sedimentary environment of "overwater uplift-lake flatunderwater low uplift-lake bay-semideep lake-deep lake." e slope direction is the interphase topography of "upliftdepression," and the overall environment is the closedsemiclosed wide brackish lake bay environment. e Chengning uplift, Gangxi uplift, Kongdian uplift, and Cangxian uplift have a weak influence on this area. Limestone and granulite are developed around the uplift, and lacustrine carbonate and shale deposits are widely developed in Huwan belt. ere is a major development of carbonate shale, felsic shale, mixed shale, and shale. Clayey mineral composition is diversiform. It has higher brittleness mineral content, high-quality hydrocarbon source rocks, and hydrocarbon generation conditions. e shale is 10-40 m thick and 1,900-3,800 m deep and is in the stage of abundant hydrocarbon generation. According to the mass oil content, the shale oil resources amount is 365 million tons. As of early February 2019, the Guandong 1701H and Guandong 1702H horizontal shale oil wells in Dagang Oilfield had been spouted for more than 260 days, and the daily crude oil production had stabilized at 20-30 m 3 . e Guandong area had formed a battlefield for increasing reserves of 100 million tons [36].

Tight Sandstone Oil Cores in Changqing.
e tight oil cores were taken from the Yanchang Formation in the Ordos Basin and mainly developed the semideep lacustrine facies area. e shale in the 7 th member of the Yanchang Formation and the tight sandstone in the 6 th member of the Yanchang Formation in the middle of the lake basin are the most typical ones. e tight oil in Ordos is mainly distributed in the center of a large depression lake basin in the Mesozoic era. e tight sandstone is in direct contact with high-quality hydrocarbon source rock hydrolyzation, and the oil source is sufficient. e generated oil and gas can be accumulated after short distance migration. e tight sandstone reservoirs in the basin are mainly terrigenous clastic rocks with complex lithology, mainly fine sand rocks, fine grain size, poor physical properties, and dense reservoirs. e tight oil in Ordos Basin has two key geological characteristics. First, the source and reservoir are symbiotic. e large area of the banded oil-bearing reservoir is distributed continuously, and there is no obvious trap and oilwater boundary. Second is nonbuoyancy accumulation and continuous filling of oil and gas. It is not affected by the obvious hydrodynamic effect. ere is no unified oil-water interface and pressure system. Ordos Basin is characterized by large shale thickness and wide distribution range, and the interbedded clastic rocks are relatively developed, which has favorable geological conditions for the formation of tight oil. Physical parameters of shale and tight sandstone are shown in Table 1. Shale was taken from the main production horizon of Dagang shale reservoir. Due to the development of reservoir bedding, it was impossible to get a large number of regular cores. erefore, several representative shale blocks were selected to conduct imbibition and gas flooding experiments [37].

Experimental Study on Shale-Formation Water and Active Water Imbibition.
e effects of injected water and injected active water (surfactant) on the development of shale oil were mainly analyzed by the spontaneous imbibition experiment of shale oil cores. e experimental equipment included a beaker (holding water), rubber band (fixing plastic film), plastic film, and mesh bag (holding broken core). Figure 1 shows the experiments setup of Advances in Civil Engineering 3 spontaneous imbibition.
e spontaneous imbibition oil recovery experiment is a physical simulation experiment. e spontaneous imbibition experiment of formation water and active water in saturated oil cores was designed, and the NMR T 2 spectrum was detected according to the preset time interval.
e steps were as follows: ① firstly, the shale cores of no. 1 and no. 2 after oil washing were dried and vacuumed. en, the saturated cores were pressurized with kerosene and finally fully saturated kerosene. After completion, T 2 NMR was performed. ② en, the cores of sample 1 and sample 2 were put into two beakers of 10,000 ppm formation water with heavy water configuration (heavy water has no signal in the nuclear magnetic experiment and its main function is to eliminate the nuclear magnetic signal of water) and 10,000 ppm active water with heavy water configuration to conduct spontaneous imbibition experiment. ③ e no. 3 Changqing tight oil core with a permeability similar to that of no. 1 shale was selected to conduct spontaneous imbibition experiments with formation water injected with 10,000 ppm salinity. e experimental results of the study on the difference of imbibition between tight oil and shale oil are shown in Figures 2-4 and Table 2.
In order to better analyze the microscopic mobilization law of imbibition, kerosene-saturated cores were used. Because the crude oil was a mixture, the magnetic signal in the core of the core was affected by both the pore size and the crude oil composition, so it was impossible to clearly analyze the microscopic law of imbibition. Nitrogen flooding mainly studied the microscopic law of oil production during the displacement process. Live oil mainly had a great influence on the experimental results when studying the effect of dissolved gas flooding. is study did not study dissolved gas flooding, so the use of kerosene did not affect the results. In addition, the Dagang shale had a low gas-oil ratio, and the reservoir had a small gas content, so there was no need to consider the influence of gas.

Experimental Study on Nitrogen Displacement of Shale.
Nitrogen flooding experiments were carried out on saturated oil cores under different displacement pressures successively. When the displacement was over, the cores were taken out for NMR T 2 spectrum detection. e experimental equipment included a nitrogen bottle, core holder, precision pressure gauge, and hand pump. e nitrogen bottle provided the air source, and the hand pump and core holder provided the confining pressure for the cores. Figure 5 shows the gas flooding experiment setup. e steps were as follows: ① Firstly, the shale cores of sample 4 and sample 5 after oil washing were dried, then they get vacuumed, and the cores were pressurized and saturated with kerosene until the cores were fully saturated with kerosene. Finally, the NMR T 2 was performed. ② e cores were placed in the core holder to provide confining pressure. e displacement pressure point of the core holder was controlled by adjusting the nitrogen bottle. e confining pressure was controlled by the manual pump. e gas displacement experiment was carried out under the condition of keeping the net confining pressure unchanged. ③ After being taken out, the cores were weighed. And then the cores with different gas displacement pressures were dislocated for NMR T 2 spectrum detection. ④ e NMR T 2 spectra under different displacement pressures were plotted and analyzed according to the results of different displacement pressures. Sample 4 shale fractured after displacement at a pressure of 1.36 MPa, so it was impossible to obtain the data after displacement at 2.72 MPa. e gas flooding experiment of no. 6 Changqing tight sandstone with similar permeability to no. 4 shale was carried out under the same pressure to facilitate comparison with shale gas flooding results. e test results are shown in Table 3 and Figures 6 and 7.

e Influence of Formation Water and Active Water as
Imbibition Media on the Degree of Recovery. Figures 2 and 3 are NMR spectra of the cores after imbibition. As can be seen from Figure 2, the NMR spectrum gradually changed from a single peak to a double peak with the increase of imbibition time. While the signal peak began to shift to the right, the spectrum span gradually increased. is indicated that the oil in the small pores of shale started to be absorbed into the large pores. e right wing of the right peak began to shift to  the right, especially at the intersection with the relaxation time axis. It indicated that microcracks were produced by the expansion of the shale due to imbibition. It was shown in Figure 6 that microcracks can also be seen by the naked eye on the surface of shale after imbibition. It could be seen from Figure 3 that when the tight sandstone oozed spontaneously, the NMR spectrum decreased steadily, indicating that the pore throat distribution of the tight sandstone was relatively uniform, and the spectrum span did not increase. It showed that the tight sandstone did not cause new microfractures due to water absorption and expansion. It can be seen from Figure 4 that, with the increase of the imbibition time, the shale recovery ratio kept increasing. Table 2 shows the contribution ratio of pore recovery degree at different relaxation times. It can be seen from Table 2 that the ultimate contribution of shale recovery ratio was the most pore with  Advances in Civil Engineering relaxation time T 2 < 10 ms, and the average contribution of recovery ratio was more than 90%. e pore with a relaxation time of 10 ms∼100 ms and the pore with a relaxation time of more than 100 ms contributed little to recovery. e imbibition recovery ratio of shale injected water was 19.54% and that of active water was 37.10%. e imbibition recovery ratio of injected active water was 17.56% higher than that of injected water. Active water imbibition of shale can improve the recovery ratio in the pore range of T 2 < 10 ms, indicating that the oil saturation rate in this pore range decreased.
According to the calculation method of mixed wettability index, the hydrophilicity and wettability of the shale were increased after the active water imbibition, so the final recovery ratio was also improved. Figure 4 shows the trend line of shale formation water and active water imbibition recovery over time. It can be seen that the imbibition rate of    shale in the formation was higher than that in the active water, but the shale in the formation imbibition ratio was balanced first than that in the active water imbibition; that is to say, the amount of imbibition in the active water was higher. Compared with the Changqing tight sandstone with the same permeability, the shale had lower ultimate recovery and less crude was used in the pore range of 10-100 ms and greater than 100 ms.

e Influence of Injected Nitrogen on the Development
Effect. Figures 7 and 8 show the NMR spectra of nitrogen displacement of shale and sandstone under different pressures. As can be seen from Figures 7 and 8, the peak state of shale nuclear magnetic resonance was a single peak state, and the peak state of sandstone nuclear magnetic resonance was a double peak state, which indicated that the shale pore throat distribution was relatively simple, and most of them were small pore throats. Besides small pore throats, sandstone also had the distribution of large pore throats. erefore, compared with sandstone, shale had a large specific surface area, and shale had a stronger adsorption effect on oil, with poor oil availability. Tables 3 and 4 show the cores gas flooding effect and pore recovery contribution of different relaxation times. Tables 3 and  4 show that, as the displacement pressure increased, the oil displacement efficiency also increased. e higher the shale permeability was, the greater the displacement efficiency would be. T 2 > 100 ms pore throat of shale contributed to the main oil displacement efficiency; the average oil displacement efficiency contribution reached 63.16%. And the relaxation interval 10 < T 2 < 100 ms pore throat of shale displacement efficiency contributed to 28.27%. T 2 < 10 ms pore throat of shale contributed the least oil displacement efficiency, with an average oil displacement efficiency contribution of 8.58%. Compared with tight sandstone with similar permeability, shale had lower oil displacement efficiency.

Summary and Conclusions
In this paper, nuclear magnetic resonance was combined with physical simulation experiments such as imbibition and nitrogen displacement to analyze the contribution of different injected media to the ultimate recovery ratio, so as to provide a scientific basis for the effective development of shale oil. e conclusions are as follows: (1) As the imbibition time of shale samples increased, the imbibition efficiency increased. Pores with T 2 < 10 ms   contributed the most to imbibition efficiency, with an average contribution greater than 90%. Active water can change the wettability of shale, increase its hydrophilicity, and improve the efficiency of imbibition. Compared with tight sandstone with similar permeability, shale had a lower imbibition efficiency. (2) As the displacement pressure increased, the oil displacement efficiency also increased. e higher the shale permeability was, the greater the displacement efficiency would be. T 2 > 100 ms pore of shale throat contributed to the main oil displacement efficiency, the average oil displacement efficiency contribution reached 63.16%, and the relaxation interval 10 < T 2 < 100 ms pore throat displacement efficiency contributed to 28.27%. T 2 < 10 ms pore throat contributed the least oil displacement efficiency, with an average oil displacement efficiency contribution of 8.58%. Compared with tight sandstone with similar permeability, shale had lower oil displacement efficiency.
(3) Due to the small number of shale samples taken in this study, the research results cannot represent the entire shale block. It is recommended that the number of samples be increased next time for in-depth analysis.

Data Availability
e data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest
e authors declare that they have no conflicts of interest.

Authors' Contributions
Lanlan Yao conceived the experiment and conducted experimental work, drafted the manuscript, and made a series of modifications. Zhengming Yang helped analyze the experimental data. Haibo Li proposed experimental suggestions for improvement during the experiment. Bo Cai made suggestions for the manuscript, which greatly helped the completion of the thesis. Chunming He participated in the recording and collection of experimental data. All the authors gave approval for publication.