Abstract
Hydraulic fracturing in the immediate roof is a suitable method to increase the permeability of soft outburst coal seam. In this work, the propagation law of hydraulic fractures in laminated virgin sandstone and the change of microstructure after hydraulic fracturing were revealed. The results show that the hydraulic fracturing pressure of sandstone with different bedding angles is 15.69~22.17 MPa. In the experiments, the hydraulic fractures extend through the bedding planes when the bedding angle is 45°. When the bedding angle is 0°, 15°, 30°, or 60°, the hydraulic fracture is prone to extend along a bedding plane. The hydraulic fracture in sandstone is jointly determined by the internal fractures and the cementation strengths of the bedding planes. The internal fractures may not necessarily guide the extension of the hydraulic fractures. The hydraulic fracturing pressures of the samples affected by the bedding plane and internal fractures are significantly lower than that of samples with uniform strength. Besides, hydraulic fracturing cannot change the connectivity between pores. The hydraulic fractures are the main channel for gas migration. The results can provide some guidance for the design and application of hydraulic fracturing in the immediate roof of the soft outburst coal seam.
1. Introduction
As a green and clean energy associated with the formation of coal, gas is also one of the important factors affecting the occurrence of coal and gas dynamic disasters. Therefore, gas extraction before coal mining is necessary [1–4]. Due to the low permeability of coal seams in some areas, permeability-increasing measures must be taken to improve gas extraction efficiency [5–8]. At present, hydraulic fracturing [9–11], hydraulic slotting [12], hydraulic flushing [13–15], and mechanical cutting assisted by water jet [16, 17] have been applied in some mining areas. However, in the soft and low-permeability outburst coal seam, field application results were not ideal. Under the existing mining conditions, it is particularly difficult to increase the permeability of soft coal seams [13, 18]. Moreover, with the increase of mining depth, the geo-stress further increases. It is particularly important to identify how to enhance the permeability of the soft coal seam and improve gas extraction efficiency.
In recent years, horizontal well hydraulic fracturing has been used in some mining areas to enhance permeability. However, due to a large amount of engineering and huge capital investment, the application of technology is restricted [19, 20]. Besides, the hydraulic disturbance was carried out in the coal seam or immediate roof through drilling holes across the seam, and hydraulic fracturing was carried out through a long borehole in the roof or floor of the coal seam, which promotes the development of hydraulic fracturing in gas control of the soft coal seam [21–23]. In some mining areas, staged hydraulic fracturing through comb-shaped holes has been tested. However, since the coal seam is relatively soft, it is easy to cause sticking or the loss of drills when drilling the branch borehole. Underground hydraulic fracturing technologies are shown in Figure 1. Nevertheless, field practice shows that it is very difficult to maintain the boreholes in the soft coal seam under the action of geo-stress and water. The immediate roof of the soft coal seam is mostly hard and tight. Therefore, hydraulic fracturing and gas extraction in the immediate roof of the soft coal seam is a better choice. As the fracturing area is located in the roof, the propagation law of hydraulic fractures in the roof is the key to affecting the effect. Some mining areas in China have carried out field tests of hydraulic fracturing in the immediate roof, and some scholars have also carried out relevant research. Sandstone samples were used to carry out hydraulic fracturing experiments. The initiation, propagation, and closure of hydraulic fractures were studied, and it was pointed out that hydraulic fractures were easy to close in fracture evolution [24]. A new loading machine was designed by Chen et al. to improve the traditional hydraulic fracturing experiment in the laboratory and a method to better understand the fracture behavior of solid media was provided [25]. Chitrala et al. monitored the hydraulic fracturing experiment of sandstone under different stress conditions by acoustic emission (AE), and the propagation law of hydraulic fracture was studied [26]. AE and micro-CT were used to study the stress-induced fracturing of sandstone by Pradhan et al. [27]. Through hydraulic fracturing experiments, Stanchits et al. pointed out that the propagation speed of acoustic emission parallel to the bedding plane was faster than that perpendicular to the bedding plane, and shear type and pore collapse type acoustic emission events were the main events when it was close to failure [28]. Fall et al. believed that the formation of fractures was driven by the generation of natural gas and migration to adjacent reservoirs. Meanwhile, the increase in pore fluid pressure led to natural hydraulic fracturing [29]. Tan et al. explored the integrated hydraulic fracturing in tight sandstone coal-interbedded formation through triaxial hydraulic fracturing experiments. The results showed that the hydraulic fracture of sandstone was relatively simple, but the initiation fracturing pressure was high. Besides, the hydraulic fracture was perpendicular to the direction of minimum horizontal stress. Through the research, they suggested hydraulic fracturing in the tight sandstone firstly to improve the extraction effect of coalbed methane [30].
Although some research and tests have been conducted, there are relatively few pieces of research on the hydraulic fracture propagation law in virgin sandstone on the whole. At present, hydraulic fracturing has been widely used in the stimulation of low-permeability oil and gas reservoirs, and some scholars have noticed the influence of shale bedding on the hydraulic fracturing effect [31–34]. However, there are many differences between sandstone and shale in formation and mechanical characteristics. In the field application, the fracturing borehole may cross the bedding from different angles (Figure 2). Since coal seam and the immediate roof are always laminated, their layouts cannot always keep a uniform direction. When the immediate roof is encountered with the hydraulic fracturing borehole, the induced stress is expected to exhibit localized concentrations along the bedding planes which may further lead to the alterations of hydraulic fracture development. However, the influence of the laminated structure on the propagation law of hydraulic fractures in virgin sandstone is rarely studied.
In this paper, hydraulic fracturing experiments were carried out to study the influence of the laminated structure on hydraulic fractures. The propagation law of hydraulic fractures in virgin sandstone with different bedding angles was explored. The research results in this paper can not only provide some references for field design and implementation of hydraulic fracturing in the immediate roof but also have certain guiding significance for increasing the permeability of soft outburst coal seam.
2. Experimental Method
2.1. Samples Preparation
The sandstone used in the experiments was collected from the central coal basin in China. In this mining area, 21 coal seam is the main mining seam. The average thickness, maximum original gas content, and permeability of the coal seam are about 5.53 m, 15.78 m3/t, and , respectively. The coal is soft and the firmness coefficient is 0.1~0.25. The immediate roof of the coal seam is medium or fine grained sandstone.
Lumps of sandstone collected from the working face were processed into standard cylinder samples and disc samples after coring, cutting, and grinding. The mechanical parameters of sandstones with different bedding angles were obtained through the uniaxial compression experiments and tensile strength experiments, as shown in Table 1. In uniaxial compression experiments, the axial stress was loaded in displacement-controlled mode, with a loading rate of 0.005 mm/s. The uniaxial compressive strength and elastic modulus can be obtained from the stress-strain curve. In tensile strength experiments, the axial stress was loaded in force-controlled mode, with a loading rate was 0.5 kN/s.
During the preparation of core samples, the designed bedding angles were planned. The bedding angles were measured from the samples. Even though there have errors in readings, the bedding angles () of sandstone samples are nominal 0°, 15°, 30°, 45°, and 60°, respectively.
In hydraulic fracturing experiments, the size of the samples was . The vertical drilling machine was used to drill a hole with a length of 60 mm and a diameter of 6.2 mm in the center of the end face. The design length of the hydraulic fracturing section and borehole sealing section were both 30 mm, as shown in Figure 3.
2.2. Experimental Apparatus
The coal/rock hydraulic fracturing experimental system was used to carry out the hydraulic fracturing experiments under the condition of hydrostatic stress. Figure 4 is the schematic diagram of the hydraulic fracturing experimental system.
The main role of the KD250 constant speed/pressure pump is to boost and generate hydraulic pressure. Pure water is used as the booster liquid. The pump has two cylinders A and B, and it can achieve constant flow or pressure. The maximum hydraulic flow and alarm pressure of the pump are 80 ml/min and 30 MPa, respectively. The pressure generated by the air compressor is used to push fracturing fluid into the piston container, and the booster liquid in the piston container will be discharged into the discharging container at the same time. The maximum pressure that the piston container can withstand is 40 MPa. The Teledyne ISCO 100DX piston pump is used to provide radial stress for the triaxial stress device. The pump volume, maximum flow, and maximum pressure are 103 ml, 45 ml/min, and 50 MPa, respectively. The JB-80 manual metering pump is used to provide axial stress, with a pump volume of 100 ml. The hydraulic fracturing pipe is made of a 304 stainless steel, with an outer diameter of 3 mm and a wall thickness of 1 mm. The water outlet holes are located on the radial side (Figure 5).
Before and after the hydraulic fracturing, the samples were scanned and tested by the Phoenix v|tome|x s-loaded coal/rock industrial CT scanning system and MesoMR23-060H-I L-NMR experimental system, respectively, as shown in Figure 5.
The Phoenix v|tome|x s-loaded coal/rock industrial CT scanning system can be used to scan the mesostructure of the samples. The system is equipped with two X-ray tubes, which are the high-power micron-scale ray tube and the high-resolution nanoscale ray tube. The former is used for scanning large-size samples. The maximum power and maximum tube voltage are 320 W and 240 kV, respectively. The minimum size of focus and the best scan resolution are 3 μm and 2 μm, respectively [35, 36]. In the experiments, the voltage and current were about 140 kV and 120 μA, respectively. A total of 1000 pictures were collected in each scanning. The exposure time was 1000 ms, and the voxel size was about 56 μm.
The MesoMR23-060H-I L-NMR experimental system is used for T2 spectrum tests. Through T2 spectrum tests, the microstructure characteristics of internal pores and microfractures before and after hydraulic fracturing were analyzed. The system is mainly composed of the host, temperature control unit, RF unit, gradient unit, and magnet. In the experiments, the magnetic field intensity and the H proton resonance frequency were 0.5 T and 21.67 MHz, respectively. And the RF pulse frequency and the magnet temperature were 21.67 MHz and , respectively.
2.3. Experimental Conditions and Procedure
With an increase in mining depth, the vertical stress tends to be equal to the horizontal stress. In hydraulic fracturing experiments, the axial stress and radial stress were designed to be both 1 MPa. Meanwhile, to collect more data, the constant flow mode was selected and the hydraulic flow was set to 2.5 ml/min. Before the experiments, all the samples were put into the DZF6210 vacuum drier to dry at a constant temperature for 12 hours. The specific experimental steps are described below. (1)Industrial CT scanning and L-NMR T2 spectrum test. The detailed experimental processes have been described in the papers and will not be repeated here [35–38]. Then, dry all the samples for 12 hours.(2)Prepare borehole sealing. To prevent the adhesive from blocking the water outlet hole, wrap the white paper tape at a position of 30 mm away from the end of the hydraulic fracturing pipe until the diameter is close to the borehole diameter.(3)Seal the borehole. Insert hydraulic fracturing pipe into the bottom of the borehole and keep the fracturing pipe always in the center of the borehole. After that, use acrylate AB adhesive to seal the borehole. First, squeeze adhesive A into the medical syringe, then squeeze adhesive B into it. Mix them quickly and evenly, and inject the adhesive evenly into the borehole rapidly. The strength of the adhesive reaches the maximum after 24 hours of standing at room temperature.(4)Prepare hydraulic fracturing experiments. Install the sealed sample into the triaxial stress device and connect the pipe. Then, turn on the air compressor, close valves II and IV, and open valves III, V, and VI. Under the pressure produced by the air compressor, the fracturing fluid will be forced into the piston container. When the booster liquid is no longer discharged, close valves III, V, and VI. Open valves I, II, IV, VII, VIII, and IX, and ensure that valve X is closed. Finally, check the status of pipelines and valves again.(5)Conduct hydraulic fracturing experiments. Load axial stress and radial stress to the target. Turn on the constant speed/pressure pump and set the working mode. Start to carry out the hydraulic fracturing experiment and collect data at the same time.(6)Pressure relief and cleaning. After the experiment, turn off the constant speed/pressure pump and then open valve X. Close valve X after removing the residual hydraulic pressure. Then, gradually remove axial stress and radial stress, close valves VIII and IX, and take out the sample. After that, use the dissolving reagent to remove the adhesive, and carefully take out the hydraulic fracturing pipe.(7)Replace the sample and repeat step (2)–step (6).(8)After hydraulic fracturing experiments, dry the samples at a constant temperature for 12 hours.
Then, repeat step (1).
3. Results and Analysis
3.1. Characteristics of Hydraulic Fracturing
To study the propagation law of hydraulic fractures in laminated sandstone, sandstone with different bedding angles was used in the experiments. The experimental results are shown in Figure 6 and Table 2. It can be obtained from Figure 6 that, with an increase in time, the overall trends of hydraulic fracturing curves are the same, which mainly include three stages: slow rise, rapid increase, and instantaneous decrease. Under the condition of hydrostatic stress, the fracturing fluid was firstly injected into the borehole at constant flow, and the hydraulic pressure began to rise slowly. With the injection of fracturing fluid, the hydraulic pressure began to increase rapidly. Finally, under the combined action of stresses and hydraulic pressure, the sample was fractured rapidly, and the hydraulic pressure instantaneously decreased to 0 MPa. The fracturing of the samples shows the characteristics of brittle failure.
(a)
(b)
(c)
Through the experimental results, it can be found that the hydraulic fracturing pressure of sandstone is 15.69~22.17 MPa. With an increase in bedding angle, the average hydraulic fracturing pressure first increases, then decreases, and finally increases. It can be seen from Figure 6(c) and Table 2 that, when the bedding angle is 45°, the average hydraulic fracturing pressure is the lowest. The bedding angles of samples 1-1 and 1-2 are the same, but the hydraulic fracturing pressures are quite different. The reason may be the existence of fractures inside the sample. The same is true for samples 1-3 and 1-4.
The hydraulic fracturing pressures are characterized by short duration and the duration of fracturing pressure is about 2~3 s. It indicates that the hydraulic fracturing pressure is expected to be higher than the pressure required for fracture propagation, which is consistent with the experimental results and field application [23]. Due to the scale of the samples, the propagation of hydraulic fractures has not been obtained. When the hydraulic pressure reached the fracturing pressure, the sample was fractured rapidly and the hydraulic pressure in the sample tended to be stable after failure. At this stage, the injection flow was equal to the filtration flow. To make the extension and development of hydraulic fractures more sufficient, the hydraulic pressure in the sample can be further increased by increasing the hydraulic flow.
The failure modes of sandstone samples after hydraulic fracturing are shown in Figure 7. The failure modes of the samples are relatively simple, and the samples are relatively complete. From Figure 7, it can be found that part of the samples has been fractured along a bedding plane. However, the hydraulic fracture of sample 1-4 shows the characteristics of passing through bedding planes, as are samples 1-7, 1-8, and 1-10. Since the inside of the sample was unknown and the internal fractures may affect the propagation of hydraulic fractures, industrial CT scanning was combined for further analysis.
(a)
(b)
(c)
(d)
(e)
(f)
(g)
(h)
(i)
(j)
3.2. Visualization of 3-D Fracture Network
Before and after hydraulic fracturing, the industrial CT scanning system was used to scan the samples. The 3-D fracture networks of the sandstone samples are shown in Figure 8. In Figure 8, the green area represents the larger volume fractures, the blue-green area represents the small-volume fractures, and the blue area represents the smaller volume fractures. It can be found from Figure 8 that the hydraulic fractures are relatively simple and symmetrical. And the hydraulic fractures mostly have tips, which confirms the fracture tip effect during hydraulic fracturing [39, 40].
(a)
(b)
(c)
(d)
(e)
(f)
(g)
(h)
(i)
(j)
For the bedding angle of 0°, the samples were mainly fractured along a bedding plane. Before hydraulic fracturing, there were no fractures inside sample 1-1, while some fractures were located in sample 1-2. After hydraulic fracturing, the two samples were fractured along a bedding plane. As to sample 1-2, the original internal fractures increased, and some fractures were connected with the hydraulic fracture. The internal fractures have guided the extension of the hydraulic fracture, which may be the reason for the minimum hydraulic fracturing pressure in all samples.
For the bedding angle of 15°, there were hardly any small-volume fractures inside samples 1-3 and 1-4 before hydraulic fracturing. After hydraulic fracturing, the hydraulic fracture inside sample 1-3 has extended along a bedding plane, and then it deflected with the change of the bedding. Sample 1-3 has been mainly fractured along a bedding plane. Besides, the extension of the hydraulic fracture in sample 1-3 does not correlate with the internal fracture. Sample 1-4 was not fractured along a bedding plane. The included angle between the hydraulic fracture and the bedding plane was about 15°, and then it deflected and the included angle increased. From Figure 8(d) and Table 2, it can be found that sample 1-4 has formed larger volume fractures at higher hydraulic fracturing pressure. The reason may be that the cementation strengths of the bedding planes in this sample are large enough. The strength of this sample is relatively uniform, and there is no relative weakness inside the sandstone sample. The hydraulic fracturing pressure of sample 1-4 is the highest among all samples. It indicates that, when the cementation strengths of the bedding planes are large enough, the hydraulic fracturing pressure required is relatively high. Besides, the extension of hydraulic fracture is relatively uncertain.
For the bedding angle of 30°, there were hardly any small-volume fractures inside sample 1-5 before hydraulic fracturing. On the contrary, there were a few fractures inside sample 1-6. In the experiments, the two samples were mainly fractured along a bedding plane. However, the hydraulic fracturing pressures of the two samples are not much different. The reason for it may be that the original fractures inside sample 1-6 were not in the propagation direction of the hydraulic fracture. The original fractures did not guide the extension of the hydraulic fracture.
For the bedding angle of 45°, sample 1-7 had few small-volume fractures before hydraulic fracturing, and sample 1-8 had no fractures inside. The hydraulic fractures of the two samples extended through the bedding planes. However, different from sample 1-4, the hydraulic fracturing pressure was lower for both samples. Through analysis, it is considered that there are some differences. Due to the brittleness of sandstone and the tip effect during hydraulic fracturing, it is considered that the hydraulic fractures of samples 1-7 and 1-8 have been caused by point failures of the bedding planes. The cementation strengths of the bedding planes are all relatively low.
For the bedding angle of 60°, samples 1-9 and 1-10 had a large number of fractures inside before hydraulic fracturing, which was different from other samples. In the experiment, sample 1-9 was fractured along a bedding plane, but the original fractures did not correlate with the extension of the hydraulic fracture. The hydraulic fracture in sample 1-10 has an included angle of about 15° with the borehole, and then the included angle increases. From Figure 8(j), it can be found that there have been fractures near the borehole before hydraulic fracturing, and the hydraulic fracture has connected the fractures after hydraulic fracturing. During the drilling process, a plastic zone formed near the borehole. The fractures in the plastic zone have a guiding effect on the extension of the hydraulic fracture. Besides, it can be found from Figure 7(j) that the hydraulic fracture of sample 1-10 began to extend along a bedding plane when it extended to the bedding plane. The fractures in the plastic zone and the relatively low cementation strength of the bedding plane resulted in the sample being fractured under relatively low hydraulic pressure.
From the above analysis, it is obvious that the laminated structure is one of the important factors affecting the propagation of hydraulic fractures in sandstone. The propagation of hydraulic fractures in sandstone under hydrostatic stress mainly depends on (1) the cementation strengths of the bedding planes and (2) the internal fractures. If the cementation strength of a bedding plane is relatively low, the sandstone sample is likely to form the hydraulic fracture along the bedding plane, and the direction of the bedding plane is the same as the initiation direction of hydraulic fracture. If there are some fractures in the bedding plane, the sandstone sample is more easily to be fractured. If there are fractures in the plastic zone near the borehole, the fractures would guide the hydraulic fractures. However, if the internal fractures are not located in the initiation direction of hydraulic fractures, the internal fractures may not necessarily guide the extension of the hydraulic fractures. On the other hand, if the cementation strengths of the bedding planes are large enough, the overall strength of sandstone is relatively uniform. The hydraulic fracture is relatively uncertain. Moreover, the hydraulic fracturing pressures of the samples mainly affected by the bedding and internal fractures are significantly lower than that of samples with uniform strength.
Overall, the propagation of hydraulic fractures in sandstone is susceptible to the cementation strengths of the bedding planes. Furthermore, without considering the internal fractures, the experimental results show that it is easy to form the hydraulic fractures passing through the bedding planes when the bedding angle is 45°. However, when the bedding angle is less than or equal to 30°, or the bedding angle is 60°, the hydraulic fracture is prone to form along a bedding plane.
According to the 3-D reconstruction model, parameters such as the volume of fracture in the sandstone sample before and after hydraulic fracturing can be obtained through the VG Studio MAX software. Through calculation, the fracture density, fracture rate, and aperture of the fractures before and after hydraulic fracturing can be obtained. The fracture density is an important parameter to characterize the fracture development of coal/rock, and the most commonly used one is utilized to represent that of the sandstone sample [35, 41]:
In Equation (1), is the fracture density inside the sandstone sample, is the surface area of fractures, and is the volume of the sandstone sample.
Before hydraulic fracturing, there may be some fractures inside the sandstone sample. After hydraulic fracturing, the internal fracture network will change. The fracture rate can reflect the change in the fractures. The fracture rate of the sandstone sample can be expressed as follows [35]: where is the fracture rate of the sandstone sample and is the volume of fractures inside the sandstone sample.
The aperture can be used to indicate the size of the fracture or the tightness of the fracture surface. Here, it is used to analyze the opening degree of fractures. The aperture of the fractures can be expressed as
The fracture parameters of sandstone samples before and after hydraulic fracturing are shown in Table 3. It can be obtained from Table 3 that the volume of fractures, surface area of fractures, fracture density, and fracture rate of the sandstone samples have increased significantly after hydraulic fracturing. The fracture volume, fracture surface area, fracture density, and fracture rate increased by 344.25~710.59 mm3, 7189.34~13776.73 mm2, 0.0370~0.0709 mm-1, and 0.18%~0.37%, respectively. It indicates that hydraulic fracturing can significantly improve the fracture network inside sandstone samples. From Table 3, it can be found that the fracture aperture of each sample after hydraulic fracturing is not much different. It indicates that the fracture opening ability caused by hydraulic fracturing with the fracturing fluid is certain in sandstone. After hydraulic fracturing, the fracture aperture of most samples has increased, with a maximum increase of 0.0436 mm. The aperture of samples 1-2 and 1-6 are less than that of the samples before hydraulic fracturing, respectively. The reason may be that the fracture opening ability caused by hydraulic fracturing is certain in sandstone, and the aperture of some fractures becomes smaller under the action of hydraulic pressure.
From the above analysis, it can be found that although hydraulic fracturing can significantly improve the fracture network inside sandstone, the hydraulic fractures are relatively simple, and the apertures of the fractures are relatively small. Under the high geo-stress environment, hydraulic fractures can be easily closed. Therefore, it is necessary to add a certain proportion of abrasives to the fracturing fluid in field application, which can further improve the aperture of hydraulic fractures on the one hand and prevent the hydraulic fractures from closing on the other hand.
3.3. T2 Relaxation Characteristics
The change of microstructure may affect the permeability of coal and rock to a certain extent. After hydraulic fracturing, the macrofracture formed in the sandstone sample. However, the microstructure characteristics of the rest part of the sample are unknown. To study the influence of hydraulic fracturing on the microstructure of sandstone, L-NMR T2 spectrum tests were conducted before and after hydraulic fracturing, respectively. The experimental results are shown in Figure 9. It can be obtained from Figure 9 that the T2 spectra of samples before and after hydraulic fracturing all show the characteristics of three peaks. According to the research of scholars, the left peak corresponds to micropores and transition pores, the middle peak corresponds to mesopores, and the right peak corresponds to macropores and microfractures [42, 43].
(a)
(b)
(c)
(d)
(e)
The amplitude of each peak in the T2 spectrum reflects the development degree of the corresponding range of pores and fractures. It can be found from Figure 9 that the peak amplitudes of all samples after hydraulic fracturing have changed. On the whole, the peak amplitude of the left peak all decreased significantly, and the peak amplitude of the middle peak also decreased, while the peak amplitude of the right peak was significantly higher than that before hydraulic fracturing. It indicates that the micropores, transition pores, and mesopores of the samples are less developed after hydraulic fracturing, and the macropores and microfractures are significantly more developed. From the change degree of peak amplitude, it can be seen that the influence of hydraulic fracturing on the micropores and transition pores is higher than that of the mesopores.
The amplitude at the transition between peaks reflects the connectivity between the two kinds of corresponding range pores. It can be seen from Figure 9 that the connectivities between the micropores, transition pores, and mesopores of samples 1-3, 1-9, and 1-10 are poor before hydraulic fracturing. After hydraulic fracturing, the characteristics have not changed, which indicates that hydraulic fracturing cannot change the connectivity between micropores, transition pores, and other pores.
The integral area proportion of each peak in the L-NMR T2 spectrum before and after hydraulic fracturing is shown in Table 4. The integral area proportion of each peak in the T2 spectrum reflects the quantity of the corresponding range of pores and fractures. From Table 4, it can be found that the peak area proportion of the left peak decreases after hydraulic fracturing, reaching 0.773~0.973 times that before hydraulic fracturing. It indicates that the micropores and transition pores become less after hydraulic fracturing, and part of the micropores and transition pores may have been compacted during hydraulic fracturing. After hydraulic fracturing, the area proportion of the middle peak reaches 0.945~1.826 times that before hydraulic fracturing. The proportions of most samples change little, indicating that the quantity of mesopores in the sample does not change much. Besides, the area proportion of the right peak in the T2 spectrum increases significantly after hydraulic fracturing, reaching 2.866~130.4 times that before hydraulic fracturing. It shows that hydraulic fracturing increases the macropores and microfractures in the samples. Before L-NMR T2 spectrum tests, the samples need to reach the water-saturated state, but the hydraulic fractures could not hold water. This may result in the fluctuation of . However, it does not affect the analysis of the results.
From Table 4, it can be obtained that the tight sandstone samples are dominated by micropores and transition pores before hydraulic fracturing. After hydraulic fracturing, the macroscopic fractures have formed, but the samples are still dominated by micropores and transition pores. It shows that although hydraulic fracturing can result in macroscopic fractures, it cannot change the main microstructure composition of the sample.
4. Discussion
The purpose of this paper is to study the influence of the laminated structure on hydraulic fracturing in the immediate roof. To make the hydraulic fractures extend to the soft coal seam, the extension law of hydraulic fractures in the immediate roof is very important. As the immediate roof may be affected by mining, the internal fractures may affect the extension of hydraulic fractures. Combined with industrial CT scanning, the propagation law of hydraulic fractures was analyzed by comparing the 3-D fracture networks before and after hydraulic fracturing. The distribution law of pores and microfractures before and after hydraulic fracturing was obtained by L-NMR T2 spectrum tests to analyze the influence of hydraulic fracturing on the microstructure of samples.
The extension of hydraulic fractures can be classified into four models (Figure 10): model a: hydraulic fracture extends along a bedding plane (samples 1-1, 1-2, 1-3, 1-5, 1-6, and 1-9); model b: hydraulic fracture extends through the bedding planes (samples 1-7 and 1-8); model c: hydraulic fracture extends along a bedding plane after passing through the bedding planes (sample 1-10); and model d: the hydraulic fracture is relatively uncertain (sample 1-4). Among all the samples, the hydraulic fractures in samples 1-2 and 1-10 have been affected by internal fractures. In model a, the cementation strength of a bedding plane is relatively low, or the direction of the bedding plane is the same as the initiation direction of hydraulic fracture, and there are some internal fractures in the bedding plane. In model b, the cementation strengths of the bedding planes are relatively low, and the point failure at each bedding plane results in the hydraulic fracture passing through the bedding planes. In model c, the fractures in the plastic zone near the borehole guide the hydraulic fracture to extend through the bedding planes and the hydraulic fracture turns to extend along the bedding plane when it extends to a weak bedding plane. In model d, the strength of the sample is relatively uniform, and there is no relative weakness inside.
(a)
(b)
(c)
(d)
Since the sandstone is relatively hard and tight, the pore pressure will not be considered. According to the previous research, the theoretical hydraulic fracturing pressure can be expressed as follows [44, 45]: where is the sandstone tensile strength, is the minimum horizontal principal stress, and is the maximum horizontal principal stress.
According to the study in this work, the samples with the same hydraulic fracture model and affected by the same factor (samples 1-1, 1-3, 1-5, 1-6, and 1-9) were selected for theoretical calculation and regression analysis. The hydraulic fracturing pressures of the samples with the same bedding angle were averaged. The result is shown in Figure 11. The regression analysis shows that the coefficient of determination is 0.9742, and the experimental results have a good correlation with the theoretical calculation results, which verifies the rationality of the previous analysis of the experimental results to a certain extent.
The 3-D fracture networks show that the hydraulic fractures are relatively simple, and the connectivity between hydraulic fractures and other fractures is poor. On the other hand, T2 spectrum tests show that the development degree and quantity of micropores and transition pores decreased after hydraulic fracturing, while that of macropores and microfractures increased. However, the connectivity of pores and the main microstructure composition have not changed. Therefore, hydraulic fractures are expected to be the main channel of gas migration under high geo-stress conditions. As the apertures of hydraulic fractures are relatively small, it is necessary to add a certain proportion of abrasives in the fracturing fluid to further improve the fracturing effect. From the study, it can be concluded that the propagation of hydraulic fractures in sandstone is susceptible to the laminated structure. The hydraulic fractures may not extend toward the coal seam under the condition of hydrostatic pressure. Thus, auxiliary measures should be taken. It is suggested to conduct high-pressure sand-bearing water jets in the direction of the coal seam before hydraulic fracturing. The formed fractures can guide the hydraulic fractures to extend toward the coal seam to increase the permeability.
5. Conclusions
Combined with industrial CT and L-NMR, hydraulic fracturing experiments on laminated sandstone were carried out in this work. The main conclusions of the completed work are as follows: (1)The hydraulic fracturing of sandstone with different bedding angles shows the characteristic of instantaneous failure and the hydraulic fracturing pressure is 15.69 to 22.17 MPa. The hydraulic fractures are relatively simple and characterized by brittle and tensile failure. The propagation of hydraulic fractures can be classified into four models: along a bedding plane, through the bedding planes, along a bedding plane after passing through the bedding planes, and uncertain(2)Under the condition of hydrostatic stress, hydraulic fractures in sandstone are jointly determined by the cementation strengths of the bedding planes and internal fractures. The hydraulic fracturing pressures of the samples mainly affected by the bedding plane and internal fractures are significantly lower than that of samples with uniform strength. When the cementation strength of a bedding plane is relatively low, the hydraulic fracture extends along a bedding plane. Nevertheless, the hydraulic fracture may extend through the bedding planes if the cementation strengths of the bedding planes are all relatively low. The internal fractures may not necessarily lead the extension of the hydraulic fractures, depending on the location of the fractures(3)The development degree of micropores and transition pores decreased after hydraulic fracturing, and the peak area proportion reached 0.773~0.973 times that before hydraulic fracturing. On the contrary, the development degree of macropores and microfractures increased, and the peak area proportion reached 2.866~130.4 times that before hydraulic fracturing. After hydraulic fracturing, the connectivity between micropores, transition pores, and other pores has not changed. The hydraulic fractures are the main channel of gas migration
Data Availability
The underlying data and figures can be found in the manuscript.
Conflicts of Interest
The authors declare no conflict of interest.
Acknowledgments
This work is financially supported by the National Natural Science Foundation of China (52130409 and 51904310) and Zhongyuan Talent Program-Zhongyuan Top Talent (ZYYCYU202012155).