Study on Fracture Propagation Rules of Shale Refracturing Based on CT Technology

Abstract

Reactivating oil and gas wells, increasing oil and gas production, and improving recovery provide more opportunities for energy supply especially in the extraction of unconventional oil and gas reservoirs. Due to changes caused by well completion and production in pore pressure around oil and gas wells, subsequently leading to changes in ground stress, and the presence of natural and induced fractures in the reservoir, the process of refracturing is highly complex. This complexity is particularly pronounced in shale oil reservoirs with developed weak layer structures. Through true triaxial hydraulic fracturing experiments on Jimsar shale and utilizing micro-CT to characterize fractures, this study investigates the mechanisms and patterns of refracturing. The research indicates: (1) natural fractures and the stress states in the rock are the primary influencing factors in the fracture propagation. Because natural fractures are widely developed in Jimsar shale, natural fractures are the main influencing factors of hydraulic fracturing, especially in refracturing, the existing fractures have a greater impact on the propagation of secondary fracturing fractures. (2) Successful sealing of existing fractures using temporary blocking agents is crucial for initiating new fractures in refracturing. Traditional methods of plugging the seam at the root of existing fractures are ineffective, whereas extensive injection of blocking agents, forming large “sheet-like” blocking bodies in old fractures, yields better sealing effects, promoting the initiation of new fractures. (3) Moderately increasing the pumping rate and viscosity of fracturing fluid is advantageous in forming “sheet-like” temporary blocking bodies, enhancing the complexity of the network of new fractures in refracturing. (4) When there is a high horizontal stress difference, after sealing old fractures, the secondary hydraulic fractures initiate parallel to and extend from the old fractures. In cases of low horizontal stress difference, the complexity of secondary hydraulic fractures increases. When the horizontal stress changes direction, the secondary hydraulic fractures also change direction. It is recommended to use high-viscosity fracturing fluid and moderately increase the pumping rate, injecting blocking agents to seal old fractures, thereby enhancing the complexity of the network of refracturing. These findings provide important technical guidance for improving the efficiency of shale oil reservoir development.

1. Introduction

With the continuous development of unconventional oil and gas wells, the decline in pore pressure, closure of fractures, and reduction in oil saturation contribute to a continuous decline in production from oil and gas wells. Refracturing is identified as a key technology for reactivating unconventional oil and gas wells and enhancing the sustainability of shale oil and gas development to a certain extent [1,2]. Despite the board application prospects of this technology confirmed by theoretical analysis and fields cases, it remains associated with high investment risks. One of the significant factors impeding the application of this technology is the complex pattern of fracture propagation [3]. The expansion of hydraulic fracturing fractures in rock is jointly constrained by various factors, primarily including three aspects: (1) Structural geological factors: including stratigraphic morphology, dip, dip direction, fault, etc.; (2) Engineering geological conditions: mainly including in-situ stress state, natural fractures, proportion of thin interbeds, formation temperature and heterogeneity; (3) Rock and fluid physical factors: mainly including elastic modulus, Poisson’s ratio, rock bedding, fracturing fluid viscosity and pumping rate, etc. [4,5,6,7]. Stress state, bedding, and natural fractures are the most significant geological factors on the artificial fracture expansion [8,9,10]. Typically, hydraulic fractures initiate perpendicular to the minimum principal stress and propagate along the direction of the maximum principal stress. When there is a significant stress difference, a single simple fracture is formed; with a smaller stress difference, a complex fracture network is more likely to form with a strong element of randomness. When the minimum principal stress is constant, the fracture pressure increases with increasing stress difference [11,12,13,14]. When hydraulic fracturing encounters bedding or natural fractures, it may activate, penetrate, activate before penetrating, or penetrate before activating natural fractures [15,16,17,18]. However, it is also influenced by reservoir heterogeneity, especially when the reservoir has widespread development of micro-fractures. In such cases, the primary factor influencing fracture expansion is the reservoir’s heterogeneity rather than the stress state, and strong heterogeneity significantly consumes the energy of hydraulic fracturing construction [19,20,21]. Another impact of heterogeneity arises from its structural characteristics. For example, shale reservoirs often have thin interbeds with sandstone and the pore characteristics in carbonate reservoirs. The greater the proportion of sandstone layers, the easier hydraulic fractures can expand and penetrate in these layers. An increase in cementation strength proportionally increases the peak pressure in hydraulic fracturing experiments, and at low cementation degrees, the pressure curve does not always exhibit peak pressure, with the rock fracture surface being relatively rough due to the presence of numerous micro-fractures [22]. Large cavities have an obstructive effect on hydraulic fractures, and under high stress differences, fractures easily traverse cavities to continue extending, while under low stress differences, the fracture morphology becomes complex, with high randomness [23]. Moreover, the transfer of fluid between pores and fractures also has a significant effect on fracture propagation. The permeability of rock, the type and proportion of clay minerals, and the pore microstructure are the main factors affecting fluid flow. The pressure transfer efficiency increases with the increase of permeability and capillary radius. With the increase of clay mineral content, especially the increase of expansive clay mineral content, the pressure transfer efficiency decreases gradually [24,25]. Additionally, when relatively cold fracturing fluid is injected into a high-temperature reservoir, thermal stresses induce thermal fractures and reduce the fracturing pressure, leading to a more complex distribution of fracture networks. In hot dry rocks, elevated temperatures are advantageous for reducing fracture pressure, cumulative acoustic emission energy, and the total number of acoustic emission events. When the temperature increases from 30 °C to 150 °C, the three characteristic parameters decrease by 18.49%, 33.9%, and 39.3%, respectively [26,27]. However, the influence on the spatial distribution of acoustic emission events and fracture complexity is not significant. The fracture pressure generally exhibits a linear decrease with increasing temperature. Reservoir characteristics are often challenging to alter, making the selection of appropriate fracturing fluids, temporary plugging agents, and optimization of fracturing parameters the main means of enhancing fracture complexity [28]. Commonly used fracturing fluids include slickwater, variable viscosity slickwater, liquid or supercritical CO₂ (LCO₂, SCO₂), and liquid nitrogen (LN₂). Both CO₂ and N₂ can significantly reduce rock fracture pressure. It has been observed that the fracture pressure with liquid nitrogen is reduced by 21.3% to 67.2% compared to slickwater, and it tends to form a more complex fracture network. The total length, average curvature, total surface area, total volume, and average fracture conductivity of fractures exhibit significant changes, increasing by 210.8%, 19.6%, 143.6%, 142.6%, and 773.6%, respectively [29]. Pre-fracturing with liquid nitrogen further enhances fracture network complexity. Compared to LN₂ fracturing, pre-fracturing with LN₂ increases the average fracture aperture, the number of fractures, and the total fracture volume by 28.45%, 10.48%, and 34.47%, respectively. When compared to water fracturing, LN₂ pre-fracturing induces an increase of 55.02% in the number of fractures and 42.16% in the total fracture volume [30]. Upon selecting the fracturing fluid, increasing its viscosity appropriately can raise the rock fracture pressure, facilitating hydraulic fractures to penetrate upper and lower sandstone interbeds and achieve larger fracture widths. However, when hydraulic fractures have connected multiple shale layers, to maintain the extension range of hydraulic fractures, it is advisable to increase the proportion of low-viscosity fracturing fluid to enhance the ability to transmit hydraulic energy along the length of the fracture. Excessive viscosity is not conducive to increasing fracture network complexity [31,32,33]. With pre-flushing during hydraulic fracturing, the proppant is introduced into the fracture along with the flush fluid. The transport and settling of proppant are crucial for the conductivity of hydraulic fracturing. Proppant mainly accumulates in the main fractures near the wellbore, accounting for 5.8% to 19% of the total fracture volume and 14.1% to 65% of the total fracture area. Small-sized proppants (200/400 mesh) are more likely to enter branched fractures, and the conductivity of fractures is directly proportional to the concentration of proppant [34]. In thin interbeds, differences in interlayer mechanical properties do not significantly affect fracture extension but do affect fracture width, making it difficult for proppants to migrate through layers. When rock strength is high, fracture pressure is high, and the main fractures initiate sufficiently, resulting in a larger fracture width. The migration and settling effects of proppants are good, making it less prone to sand plugging. Proppants mainly concentrate in the main hydraulic fractures with larger widths near the perforated layers, and only a small amount (or none) of proppant is found in the activated layers, branched fractures, and adjacent layer fractures. The limit width for proppants to enter fractures is approximately 2.7 times the proppant particle size [35,36,37]. Among numerous controllable engineering factors, completion type, pumping method, pumping volume, and fracturing sequence are the primary considerations. Compared to open-hole hydraulic fracturing, perforation fracturing can reduce the initiation pressure for shale oil reservoirs. The bedding surfaces in shale oil reservoirs are more easily activated during perforation fracturing, forming a “barrier-type” hydraulic fracture network. As the perforation direction deviates from the local maximum horizontal stress direction, the fracture pressure tends to increase, resulting in a rougher surface and simpler and more distorted hydraulic fracture morphology [38]. Pumping methods mainly include cycling pumping and pulse pumping. In hot dry rocks, as the number of cycles increases from 3 to 5, the average aperture, total length, average curvature, total surface area, and total volume of fractures in LN2 fracturing increase by 15.0%, 72.72%, 48.1%, 266.6%, and 308.7%, respectively. In pulse pumping, as the pulse frequency increases, the fracture pressure initially decreases and then increases, indicating the presence of an optimal pulse frequency [39]. The pumping volume of fracturing fluid is positively correlated with fracture pressure, and as the volume increases, the complexity of the fracture network initially increases and then decreases. In fracturing construction, to overcome the influence of stress shadows, simultaneous fracturing, step-by-step fracturing, and temporary plugging turning fracturing are increasingly applied, although controlling these techniques presents significant challenges. In shale temporary plugging fracturing, four types of plugging occur: first fracture plugging, natural fracture plugging, fracture tip plugging, and bedding plugging. By optimizing the formulation of temporary plugging agents and the timing of plugging, different types of plugging can be achieved. According to the degree of plugging by temporary plugging agents, it can be categorized into tip plugging, incomplete plugging, and complete plugging [40,41,42]. Especially in refracturing, the pattern of fracture expansion becomes more complex. Previous fractures have an attracting and interfering effect on new fractures. Under the influence of the induced stress from first fracturing, the expansion path of refracturing deviates. However, further research is needed in this area [43,44,45].

To investigate the patterns of refracturing in shale, laboratory fracturing experiments were conducted on outcrop samples of Jimsar shale using a combination of triaxial compression tests, micro-CT scanning, and a low-temperature 3D laser scanning system [46,47]. Initially, triaxial compression experiments were performed on outcrop cores of Jimsar shale to determine the mechanical parameters of the rock. Subsequently, two hydraulic fracturing experiments were carried out on shale samples, and CT scans were performed before and after each fracturing, totaling three scans. This process allowed for the characterization of the fracture morphology of the rock before and after fracturing. Through analysis of the fracture and pump pressure curves, the expansion patterns of refracturing fractures in Jimsar shale were comprehensively understood. The findings from this study provide valuable technical guidance for enhancing the development of oil and gas resources.

2. Experiments

2.1. Preparation of Experimental Samples

In this study, outcrop shale from Jimsar was used as the experimental sample. This sample was obtained from the Lucaogou Formation in the Changji area, Xinjiang, China (Figure 1a), through a series of processes including extensive geological surveys, measurements, rock sample analyses, as well as field excavation and sample preparation. The enrichment area of shale oil layer in Jimsar sag is located in the eastern part of Junggar Basin in Xinjiang, 150 km away from Urumqi City. The surface is grassland, farmland, public welfare forest and residential area, etc. The terrain is relatively flat, and the ground elevation is 580–660 m. The Lucaogou Formation is divided into two sections from bottom to top. In the longitudinal direction, there are two concentrated development sections of desserts, in which the thickness of the lower dessert body is 17.5–67.5 m, with an average of 41.5 m; the thickness of the upper dessert body was 13.4–43.0 m, with an average of 37.8 m. According to the seismic data, the structural morphology is generally characterized by a west-dipping monocline with a high east and low west, and the dip angle is 3–5°. To understand the mechanical characteristics of the outcrop shale, the rock was cut into cylindrical specimens with a diameter of 38 mm and a height of 76 mm, as shown in Figure 1b. Triaxial compression tests on the rock specimens were conducted, accompanied by acoustic emission monitoring, to obtain the elastic modulus, Poisson’s ratio, compressive strength, fracture stress ratio, and volumetric expansion stress ratio of the shale. As shown in

Table 1. Rock mechanical properties of Lucaogou shale outcrops in Jimsar, China.

Sample Number	Confining Pressure	Elastic Modulus	Poisson’s Ratio	Compressive Strength	Cci	Cvi
	MPa	GPa	-	MPa
J3	20	28.1	0.395	245.104	0.322	0.589
J4	60	34	0.376	378.881	0.307	0.696
J2	80	34	0.377	443.03	0.227	0.538
Average	-	32.03	0.383	-	0.285	0.608

, three rock cores were experimented with confining pressures of 20/60/80 MPa. The average elastic modulus was determined to be 32.03 GPa, and the Poisson’s ratio was 0.383. The compressive strength of the specimens increased significantly with the increase in confining pressure, measuring 245.104/378.881/443.03 MPa, respectively. The fracture stress ratio, defined as the ratio of deviatoric stress at the beginning of rock failure to the peak stress, was calculated. The volumetric expansion stress ratio (C_vi) was determined based on the critical point when the specimen volume transitions from compression to expansion [48]. Therefore, the volumetric expansion stress ratio represents the ratio of deviatoric stress sustained at the minimum volume to the peak deviatoric stress, as expressed in Equation (1). With increasing confining pressure, the volumetric expansion stress ratio first increased and then decreased, with an average value of 0.608, indicating that the rock specimen began to expand at axial stresses much lower than the peak stress. The fracture stress ratio based on the rock fracture volume strain, denoted as C_ci, represents the ratio of deviatoric stress at the minimum volume strain of the fracture to the peak deviatoric stress, as expressed in Equation (2) [49]. With increasing confining pressure, the fracture stress ratio based on the fracture volume strain gradually decreased, with an average value of 0.285 for the three specimens.

$$C_{vi}=\frac{{\sigma }_{vi}-{\sigma }_c}{{\sigma }_{CCS}-{\sigma }_c}$$

(1)

$$C_{ci}=\frac{{\sigma }_{ci}-{\sigma }_c}{{\sigma }_{CCS}-{\sigma }_c}$$

(2)

where: ${\sigma }_{vi}$ represents the axial stress at the maximum volume strain of the rock, measured, MPa; ${\sigma }_c$ is the confining pressure, MPa; ${\sigma }_{CCS}$ denotes the compressive strength, MPa; ${\sigma }_{ci}$ stands for the axial stress at the maximum volume strain of the fracture, MPa.

The stress-strain curves and acoustic emission energy curves of the three specimens are shown in Figure 2. At a low confining pressure of 20 MPa, the J3 specimen first undergoes an elastic deformation stage with a relatively uniform distribution of energy and a small number of acoustic emission events. As the axial stress approaches the peak stress, the rock shows slight plastic deformation, with a slight increase in acoustic emission energy and an increase in the number of events. When the axial stress reaches the peak stress, the rock undergoes shear failure, and the axial stress, rock volume strain, and fracture volume strain curves show a vertical decay trend. A large energy acoustic emission event occurs, and the normalized cumulative acoustic emission energy curve rises sharply, with the acoustic emission energy at the time of failure accounting for 72% of the total energy. The rock fracture surface is relatively flat. At a medium confining pressure of 40 MPa, the J4 specimen undergoes almost no acoustic emission events during the elastic deformation stage. As the axial stress approaches the peak stress, the specimen shows significant plastic deformation, with stress remaining essentially constant while radial strain rapidly expands. The number of acoustic emission events increases rapidly, and the event energy gradually increases, accounting for approximately 65% of the total energy in the normalized cumulative acoustic emission energy curve. Meanwhile, the rock volume strain and fracture volume strain curves show a rapid decreasing trend. After radial strain expands over a certain distance, the rock undergoes shear failure, accompanied by a high-energy acoustic emission event, accounting for about 30% of the total energy. At a high confining pressure of 80 MPa, the J2 specimen primarily undergoes elastic deformation from the initial stage to failure, with no apparent plastic characteristics. The number of acoustic emission events shows a staged distribution, and the normalized cumulative acoustic emission energy curve exhibits a stepped pattern. The rock fracture surface is rough, indicating multiple times of fracture along the failure surface, ultimately accumulating into shear failure. The rocks from the Jimsar outcrop exhibit overall elastic behavior.

2.2. Experimental System

The schematic diagram of the true triaxial hydraulic fracturing system is shown in Figure 3, which mainly includes the true triaxial confining pressure loading frame, confining pressure pump, intermediate container, data monitor, temperature control device, ISCO pump, and connecting pipelines. The shale specimens used in this experiment have dimensions of 100 × 100 × 100 mm, as shown in Figure 1c, and a schematic diagram of the core wellbore structure is presented in Figure 1d. The rock core is placed in the true triaxial core loading chamber, and predetermined confining pressures are applied to the three directions of the core using the confining pressure control system, keeping the confining pressure constant. Fracturing fluid is injected into the intermediate container, connected with 6 mm diameter pipelines to ensure the smooth circulation of the blocking agent without blockages in the pipelines. The fracturing fluid is injected into the bare-eye wellbore using the ISCO pump. A pressure sensor is installed at the outlet of the ISCO pump and connected to the data collector. The temperature control device is mainly used to maintain the temperature of the rock core loading frame, avoiding the influence of temperature changes on the fracturing fluid during the experiment. Before and after each fracturing, the rock core is scanned using a CT scanning system (Brivo CT385 produced by General Electric) to describe natural fractures, first fracturing fractures, and refracturing fractures. Commercial software Avizo2019 is used for fracture characterization.

2.3. Fracturing Fluid and Diverters

The indoor physical modeling experiment of refracturing is conducted in two stages: first fracturing and refracturing. In the first fracturing stage, a slickwater solution containing 0.1 wt% DR-800 (a conventional friction reducer) is used, and the apparent viscosity of the fracturing fluid is 10 mPa·s. The low-viscosity solution is advantageous for the initiation and extension of fractures. In the refracturing stage, a viscoelastic slickwater solution with a viscosity of 50 mPa·s and a crosslinked guar gum solution with 0.3 wt% guar gum and 0.3 wt% crosslinking agent, yielding a viscosity of approximately 500 mPa·s, are used to investigate the impact of viscosity. A blocking agent consisting of fibers with a diameter of 10 µm and a length of 5 mm is employed at a concentration of 0.1 wt%, as shown in Figure 4.

2.4. Experimental Design and Procedure

Based on extensive research and considering the results of similar experiments, the parameters for this experiment were designed [50]. The injection rate was set at 20 mL/min. Taking into account the stress state of the formation and the size of the experimental core, the confining pressure was designed as 25/21/18 MPa. The experimental plan is shown in

Table 2. Experimental scheme.

Test	Specimen	First Fracturing			Refracturing			Research Content
		Stress (X/Y/Z)	Flow Rate	Fracturing Fluid	Stress (X/Y/Z)	Flow Rate	Fracturing Fluid
		MPa	mL/min		MPa	mL/min
1	J2	25/21/18	20	Slickwater	25/21/18	20	Slickwater	The stress difference of twice fracturings
2	J8	25/21/16	20	Slickwater	25/21/16	20	Slickwater
3	J4	25/21/14	20	Slickwater	25/21/14	20	Slickwater
4	J6	25/21/18	20	Slickwater	25/17/15	20	Slickwater	The stress decreases before refracturing
5	J3	25/21/18	20	Slickwater	25/14/14	20	Slickwater
6	J10	25/21/18	20	Slickwater	25/11/13	20	Slickwater
7	J5	25/21/18	40	Slickwater	25/21/18	40	Slickwater	pumping rate
8	J7	25/18/18	20	Slickwater	25/18/18	20	Slickwater	bedding orientation
9	J1	25/21/18	20	Vicious Slickwater	25/21/18	20	Vicious Slickwater	fracturing fluids
10	J9	25/21/18	20	HPG fracturing fluid	25/21/18	20	HPG fracturing fluid

, investigating the influence of first fracturing stress state, refracturing stress state, injection rate, viscosity, and bedding orientation on the fractures. Experiments 1–3 have consistent stress states for both first and refracturing, studying the effect of stress difference on fractures. Experiments 4–6 have the same first fracturing stress state, with different stress states for refracturing, examining the impact of refracturing stress states. Experiment 7 investigates the influence of pumping rate, Experiment 8 studies the effect of bedding orientation (vertical well), and Experiments 9–10 explore the impact of different fracturing fluids (high viscosity). In all experiments involving refracturing, fiber diverter was added to the fracturing fluid.

The main steps of the experimental procedure include:

(1)

Prepare the cores according to the design dimensions and configure the fracturing fluids: slickwater fracturing fluid, variable viscosity slickwater fracturing fluid, and crosslinked guar gum fracturing fluid (HPG) with added fiber diverter.

(2)

Infuse the prepared fracturing fluids into different intermediate containers.

(3)

Place the cores that have undergone CT scanning into the true triaxial core loading frame, tighten the baffle screws, and connect the pipelines.

(4)

Apply confining pressure to the true triaxial core loading frame in three directions through the confining pressure control system and maintain it at a constant level.

(5)

According to the experimental plan, pump the fracturing fluid using the ISCO pump, initiate the fracturing experiment, and record the pressure values of the first fracturing until the pressure curve sharply drops to nearly zero, then stop the experiment and record the data.

(6)

Unload the confining pressure, dismantle the pipelines, loosen the screws of the core loading frame, and extract the core after the first fracturing. Mark the fractures on the core surface, take photographs, and conduct CT scanning.

(7)

Repeat steps 3–6 to complete the refracturing experiments.

(8)

Repeat steps 3–7 for all cores to complete the first and refracturing.

(9)

Import the CT scanning data into Avizo2019 software for describing the fractures.

3. Results and Discussion

The data collected during hydraulic fracturing experiments mainly include core photographs, CT slices, pump pressure curves, and three-dimensional fracture morphology obtained from processing CT data. Core photographs provide a visual observation of the orientation of fractures on the core surface. CT slices and three-dimensional fracture surfaces allow for the observation of fracture morphology, and pump pressure curves reflect the initiation and expansion of fractures during the fracturing process.

3.1. Impact of Stress State on Refracturing

In the process of hydraulic fracturing, overcoming the minimum principal stress and the tensile strength of the rock is essential for the initiation of hydraulic fractures. Therefore, the stress state of the rock is one of the most crucial mechanical parameters. Experiments 1–3 involved open-hole completions, with a vertical stress of 25 MPa, a maximum horizontal principal stress of 21 MPa, and minimum horizontal principal stresses of 18 MPa, 16 MPa, and 14 MPa, respectively, resulting in an increasing stress differential. The hydraulic fracturing fracture morphologies for Experiments 1–3 are depicted in Figure 5, and the injection pressure curves are illustrated in Figure 6. In Experiment 1, a noticeable natural fracture (in blue) existed in the rock, and after the first fracturing, the hydraulic fracture first intersected this natural fracture and then expanded obliquely in a direction perpendicular to the minimum principal stress, eventually forming a slanted fracture (in green). The fluid pressure required to intersect the natural fracture was 2.9 MPa. Subsequently, as the pressure continued to rise, the new hydraulic fracture ruptured at a pressure of 20.22 MPa. This indicates that the natural fracture in this rock has high strength and is not easily fully opened. During the refracturing, due to the effective sealing of the first hydraulic fracture by the temporary blocking agent and the influence of the natural fracture, a longitudinal fracture (in red) formed at the intersection of the first and natural fractures, with a significantly increased rupture pressure of 43.15 MPa. In Experiment 2, with an increased horizontal stress differential, multiple natural fractures developed around the wellbore. The first hydraulic fracturing resulted in two artificial fractures: one passed through a natural fracture, and the other intersected a natural fracture before extending along it and deviating due to stress, ultimately forming a longitudinal and an oblique fracture. The injection pressure curve exhibited two small peaks in the early stage, indicating early activation of natural fractures and micro-fracturing. Subsequently, the rupture occurred at a peak pressure of 14.8 MPa, which is less than the minimum principal stress.

The first hydraulic fracture was mainly controlled by natural fractures. During refracturing, the newly formed fractures were perpendicular to the first hydraulic fracture and parallel to the direction of the minimum principal stress. This may be influenced by the orientation of natural fractures around the wellbore. The rupture pressure was 39.54 MPa, and the extension pressure was 26.1 MPa. The injection pressure curve was relatively smooth, with no “micro-fracturing”, indicating effective sealing of the old fracture by the temporary blocking agent. As the horizontal stress differential increased, the fracture morphology tended to become simpler. In Experiment 3, with an increased horizontal stress differential of 7 MPa, the rock developed multiple parallel natural fractures and micro-fractures. The first hydraulic fracturing resulted in a simple fracture perpendicular to the vertical stress, primarily due to the presence of numerous natural micro-fractures acting as “perforations” around the open-hole wellbore. The injection pressure curve exhibited multiple pressure peaks, indicating that the formation of a single hydraulic fracture underwent multiple rupture processes. The rupture pressure for the first hydraulic fracturing was 18.1 MPa. During refracturing, the hydraulic fracturing fluid quickly flowed out along the first hydraulic fracture, and the injection pressure curve did not show a significant increase, indicating unsuccessful sealing. No new fractures were formed during refracturing. Under high stress differentials, the complexity of hydraulic fractures decreased, making it more challenging to seal old fractures, and hydraulic fracturing fluid tended to flow along pre-existing fractures during refracturing.

In actual field production of shale oil reservoirs, as production proceeds after hydraulic fracturing, both pore fluid pressure and in-situ stress gradually decrease. However, the pressure and in-situ stress parallel to the fracture direction (maximum horizontal principal stress direction) often decrease more rapidly, and the in-situ stress differential gradually decreases, sometimes even leading to stress reorientation [51]. Experiments 4–6 maintained consistent first hydraulic fracturing stress states, and during refracturing, the horizontal stress was decreased to varying degrees to simulate different horizontal stress states at different production times. Experiment 4 simulated the stress state in the early stages of production, Experiment 5 simulated a stress state after a longer production period, and Experiment 6 represented a stress state with further prolonged production time. The hydraulic fracturing fracture morphologies for Experiments 4–6 are depicted in Figure 7, and the injection pressure curves are illustrated in Figure 8. In Experiment 4, where the rock had only a few natural fractures, an first hydraulic fracturing formed a transverse fracture and a longitudinal fracture perpendicular to the minimum horizontal principal stress. The transverse fracture cut across a natural fracture without significant deflection, indicating that the first hydraulic fracturing fracture morphology was primarily controlled by the stress state. The injection pressure curve exhibited two peaks, indicating that the fracture extension occurred in two stages. The fracture surface was relatively rough, and it eventually extended to the edge of the rock at a rupture pressure of 18.4 MPa. During refracturing, the horizontal principal stress was adjusted from 21/18 MPa to 17/15 MPa to simulate the stress state after some production time. The fracture initiated from the toe end of the open-hole section, forming a distinct transverse fracture accompanied by multiple parallel fragmentation fractures, collectively forming a fragmented fracture zone. In the upper part of the core, away from the wellbore, another fracture initiated, possibly influenced by the communication with natural fractures. The injection pressure curve showed multiple peaks during the refracturing process, reflecting the multi-stage initiation and extension of the fracture zone. A major fracture formed at 21.86 MPa. Compared to the first hydraulic fracturing fracture, the refracturing fracture was significantly more complex, and the network of fractures formed during the fracturing process was highly effective for shale oil production. In Experiment 5, where the rock developed only a few natural fractures, and they were away from the wellbore, the first hydraulic fracturing formed two nearly parallel and mutually repulsive transverse fractures. Due to the stress shadow effect, the two transverse fractures were formed approximately simultaneously. The injection pressure curve exhibited two peaks at 13.1 MPa and 16.59 MPa, both below the minimum principal stress, indicating the presence of weak planes in the rock. During refracturing, the horizontal principal stress was adjusted to 14/14 MPa to simulate the stress state after a longer production period, reducing the horizontal stress differential to an equal state. Two orthogonal fractures initiated during refracturing, both parallel to the wellbore, leading to increased complexity in the fracture network. The injection pressure curve exhibited only one peak, and it was very smooth, indicating that the first hydraulic fracturing fracture was well sealed, and the two orthogonal fractures initiated simultaneously. In Experiment 6, where the rock had a clear natural fracture parallel to the wellbore, the first hydraulic fracturing formed two transverse fractures approximately parallel to each other and perpendicular to the wellbore. The formation of these fractures was primarily controlled by the stress, and the natural fracture around the wellbore did not significantly influence the direction of hydraulic fracture extension. The injection pressure curve during first hydraulic fracturing exhibited two peaks at 20.35 MPa and 25.51 MPa, with a clear pressure drop after the first peak. After a longer cumulative time, the pressure reached the second peak, indicating that the two fractures initiated successively. During refracturing, the horizontal principal stress was adjusted to 11/13 MPa (stress reorientation), simulating the stress state during stress reorientation after a prolonged production period. A longitudinal fracture initiated during refracturing, approximately parallel to the wellbore. The fracture ran parallel to the wellbore at the toe end and deviated obliquely through the first hydraulic fracturing fracture.

The refracturing fracture was primarily influenced by the stress and pre-existing rock fractures. The injection pressure curve exhibited a micro-fracture at 23.34 MPa, followed by a sharp drop, indicating the re-initiation of an old fracture. Subsequently, the pressure continued to rise, indicating the effective sealing of the old fracture by the temporary blocking agent. As production progresses, pore pressure and horizontal principal stress decrease. During the process of decreasing stress differential to equality and then reorientation, the complexity of refracturing fractures initially increases and then decreases. From the perspective of enhancing fracture network complexity, smaller horizontal stress differentials are more favorable.

3.2. Impact of Pumping Rate on Refracturing

The pump displacement of fracturing fluid is one of the crucial parameters in hydraulic fracturing operations. Generally, the scale of hydraulic fracturing is directly proportional to the displacement, and, within the limits of maximum pressure, it is advisable to increase the displacement as much as possible. However, as the displacement increases, the induced stress at the fracture tip of existing fractures changes, causing a decrease in stress at the tip of the fracture during refracturing. This reduction weakens the squeezing effect on refracturing fractures, thereby reducing the complexity of refracturing fractures. In other words, as the displacement increases, the scale of hydraulic fracturing increases, but the complexity of refracturing fractures decreases. In shale refracturing, the presence of a temporary blocking agent complicates the mechanism by which displacement affects the expansion of shale refracturing fractures. The hydraulic fracturing fracture morphology for Experiment 7 is depicted in Figure 9, and the injection pressure curve is illustrated in Figure 10. In Experiment 7, the rock had well-developed natural fractures, evident in CT slices and on the rock surface, displaying a clear fracture. During first hydraulic fracturing, the slickwater quickly flowed from the edge of the rock, and a natural fracture was connected to the wellbore, with the wellhead pressure never significantly increasing. During refracturing, the fracturing fluid with the temporary blocking agent was initially swapped without increasing the displacement. The injection pressure sharply rose to the rated pressure of the ISCO pump (50 MPa), with no visible fracturing fluid flowing out on the rock surface. Subsequently, after dismantling the pipeline and clearing the accumulated temporary blocking agent at the wellhead, the displacement was increased to double the original value for continued fracturing. Through multiple micro-fracturing events, the injection pressure curve peaked at 25.59 MPa, indicating the appearance of a clear longitudinal fracture parallel to the wellbore on the rock core. Upon extracting the rock core, it was discovered that it had split into two halves, and the temporary blocking agent effectively sealed the natural fracture. During refracturing, when the displacement was low, the temporary blocking agent tended to accumulate at the wellhead, making it ineffective in injecting and sealing fractures. Increasing the displacement allowed the temporary blocking agent to be effectively injected into the fractures, and the pressure micro-fracture in the pressure curve indicated the process of fracture extension and temporary blocking agent formation. Ultimately, in the refracturing fractures, a “sheet-like” temporary blocking agent was formed. The area of the fracture surface is a crucial indicator of fracture complexity. An optical scanning system composed of an OKIO-B non-contact optical 3D scanner and 3D scanning software 3D Scan was used to project specific grating fringes onto the fracture surface using visible light. Two high-resolution CCD digital cameras captured images of the grating interference fringes, generating complete point cloud data for the fracture surface of the rock core, as shown in Figure 11. The white substance on the fracture surface represents the fiber temporary blocking agent. The calculated point cloud data yielded a fracture area of 11,250 mm² (projection area of the fracture: 10,000 mm²).

3.3. Influence of Shale Bedding on Refracturing

Shale commonly exhibits bedding planes, and due to the effects of sedimentation, erosion, and tectonic movements, the inclination angles of shale bedding planes are heterogeneous both laterally and vertically. Additionally, differing cementation strengths on bedding planes often result in a weak plane effect, influencing the expansion of hydraulic fractures. Therefore, the angle between bedding planes and the wellbore significantly affects the morphology of hydraulic fracturing fractures. The hydraulic fracturing fracture morphology for Experiment 8 is shown in Figure 12, and the injection pressure curve is illustrated in Figure 13.

In Experiment 8, the rock core had some natural fractures around the wellbore, and the wellbore was perpendicular to the bedding planes. The horizontal stresses in both directions during the two fracturing stages were set at 18 MPa to control the horizontal stress difference and investigate the influence of bedding planes. During the first hydraulic fracturing, a single fracture parallel to the wellbore was formed by intersecting natural fractures. The injection pressure curve indicated the presence of micro-fractures at the beginning, which then smoothly increased to 19.81 MPa before fracturing. In the re fracturing, two transverse fractures initiated along the parallel bedding direction (perpendicular to the wellbore), and between the two transverse fractures, there was a fractured zone with intersecting fractures. The weak plane effect of the bedding planes was evident, as multiple pressure peaks and sharp drops were observed in the injection pressure curve. After multiple stages of fracturing, the final fracture surface was formed with a rupture pressure of 19.06 MPa.

3.4. Impact of Fracturing Fluid Viscosity on Refracturing

Fracturing fluid viscosity is a crucial parameter reflecting the performance of the fracturing fluid and significantly influences the expansion of hydraulic fractures. Generally, higher viscosity results in greater frictional forces, slower flow velocity, and reduced fluid loss within the fracture. This enhances the carrying capacity of proppants and temporary blocking agents, favoring the formation of complex fracture networks. Conversely, lower viscosity fluids are generally advantageous for initiation and extension of fractures but often result in lower fracture complexity. In contrast to the previous eight experiments, Experiment 9 employed a variable viscosity slickwater (80 mPa·s), while Experiment 10 used a crosslinked guar gum solution (500 mPa·s). The hydraulic fracturing fracture morphologies for Experiments 9 and 10 are illustrated in Figure 14, and the injection pressure curves are depicted in Figure 15. In Experiment 9, natural fractures were present on the left side of the rock core away from the wellbore and parallel to the wellbore around the wellbore. During the first hydraulic fracturing, the hydraulic fracture connected the natural fractures around the wellbore. Multiple low-pressure peaks and clear indications of micro-fractures were observed in the early stages of fracturing, ultimately resulting in an first fracture at a rupture pressure of 7.6 MPa. During refracturing, a fracture parallel to the wellbore but perpendicular to the first fracturing fracture was formed. The influence of viscosity was evident in the injection pressure curve, showing multiple pressure peaks and sharp drops, with a final rupture pressure of 19.86 MPa. In Experiment 10, where a higher viscosity crosslinked guar gum fracturing fluid was used, the rock core had fewer natural fractures. During the first fracturing, a transverse fracture initiated perpendicular to the wellbore in the toe section of the bare-eye section, then turned parallel to the wellbore by intersecting with natural fractures. The injection pressure curve was smoother, with a rupture pressure of 20.84 MPa. During refracturing with the high-viscosity fracturing fluid, two nearly parallel transverse fractures initiated perpendicular to the wellbore, with a fractured zone connecting them. The injection pressure curve was more complex, exhibiting six pressure peaks and sharp drops. The pressures were lower than those observed with lower viscosity fracturing fluid. These pressure peaks and drops represented both the expansion of fractures and the formation and dissipation of temporary blocking agents. The higher viscosity fracturing fluid resulted in greater fracture network complexity, with a final rupture pressure of 30.54 MPa.

4. Conclusions

Refracturing is an important means to re-activate oil and gas wells and improve the sustainability of oil and gas wells. Refracturing experiments were carried out for the Jimsar shale outcrop, and the following conclusions were obtained:

(1)

Refracturing is an important way to effectively develop shale oil and gas. The propagation law of re-fracturing fractures is complex, and the fracture morphology and fracture network are affected by stress state, natural fractures and first fracturing fractures.

(2)

Temporary plugging is an important method for refracturing to create new fractures. The ‘point’ temporary plugging body cannot effectively inhibit the expansion of fractures along the old fractures, while the ‘sheet-like’ temporary plugging body can effectively block the old fractures.

(3)

With the decrease of stress (stress difference), the complexity of re-fracturing fractures increases gradually. With the progress of production, pore pressure and horizontal principal stress decrease. In the process of horizontal stress difference decreasing to equal and then to steering, the complexity of refracturing fractures increases first and then decreases. Only in terms of increasing the complexity of fracture network, the smaller the horizontal stress difference is, the more favorable it is.

(4)

During refracturing, the pumping rate of fracturing fluid has an important influence on the effective pumping of old fractures and the formation of temporary plugging. When the pumping rate is low, the temporary plugging agent is easy to form ‘point’ temporary plugging body at the root end of the old fracture, but can not effectively block the old fracture. When the pumping rate is increased (twice as much as that of the initial fracturing in this experiment), the fiber temporary plugging agent is effectively pumped into the old fracture to form a ‘sheet-like’ temporary plugging body.

(5)

Shale bedding has a weak plane effect. When natural fractures are blocked, the shale bedding plane is easy to be opened, and a broken fracture zone is easily formed between the bedding planes.

(6)

Pumping the temporary plugging agent into the deep formation to form a ‘sheet-like’ temporary plugging body requires a certain degree of fracturing strength. Appropriately increasing the displacement and viscosity is conducive to the formation of a ‘sheet-like’ temporary plugging body.