Hydraulic Fracturing Simulation Experiments and Fracture Characterization of Coal Reservoirs in the Dahebian Syncline

Abstract

The coalbed methane resources in the Dahebian Syncline area of the Liupanshui coalfield, Guizhou Province, are abundant, with geological resource reserves anticipated to exceed 20 billion m³. However, this region is characterized by multiple thin to medium-thick coal seams and complex regional geological conditions, which have led to unsatisfactory outcomes from previous hydraulic fracturing and coalbed methane development projects. This study selected underground coal seam cores from this area to conduct large-size physical simulation experiments on hydraulic fracturing of coal specimens. The impacts of fracturing fluid injection rate, in situ stress state, and formation depth on the effectiveness of fracturing were investigated. Additionally, the influences of various factors on fracture pressure, fracture propagation, and fracture morphology were analyzed individually. The research findings indicate the following: (1) The fracturing fluid injection rate exerts the most significant influence on the hydro-fracturing formation of coal specimens. As the injection rate increases, both fracture pressure and complexity of fractures also rise. (2) The coefficient of horizontal stress difference is another critical factor affecting the pump pressure curve and fracture morphology. An increase in this coefficient leads to greater difficulty in fracturing the coal seam, resulting in a fractured area that becomes progressively longer and narrow. (3) With increasing formation depth, the challenges associated with fracturing intensify. The pump pressure curve transitions from a stable rising pattern to one characterized by violent fluctuation, eventually shifting to a fluctuating rising trend. (4) The existence of natural fractures obviously affects pump pressure and fracture characteristics, particularly when these natural fractures are well-developed, as they notably impact the formation of complex local fractures. Consequently, it is essential to enhance the injection rate to facilitate long-distance fracture propagation, while simultaneously creating a complex local fracture network. This study provides valuable insights for designing hydraulic fracturing operations and enhancing productivity within the Dahebian Syncline as well as similar area.

1. Introduction

Coalbed methane (CBM) is a type of natural gas predominantly composed of methane formed through the extensive geological evolution of coal seam reservoirs. It represents a significant unconventional natural gas resource, with particularly abundant reserves located in China, notably in the provinces of Shanxi, Xinjiang, and Guizhou. Due to its unique formation and accumulation mechanisms, CBM presents greater development challenges compared to conventional natural gas. Most coal seams in China exhibit low porosity and extremely low permeability coupled with complex geological conditions. These characteristics necessitate the implementation of reservoir enhancement techniques, particularly hydraulic fracturing, to improve permeability and facilitate economically viable extraction processes. Guizhou Province, a prominent coal-producing region in Southern China, possesses abundant resources of both coal and coalbed methane. As early as 2006, the Ministry of Land and Resources of the People’s Republic of China reported that Guizhou’s geological resources for CBM amounted to 34723.72 × 10⁸ m³, with recoverable resources estimated at 12892.88 × 10⁸ m³. This indicates substantial potential for methane development. However, the province’s current annual production of coalbed methane remains below 400 million m³, highlighting a significant gap between resource advantages and the realization of economic benefits.

The current methods for extracting CBM primarily include surface and underground extraction techniques. The surface extraction techniques encompass hydraulic fracturing, carbon dioxide fracturing, and multibranch horizontal well drainage, among other approaches. In contrast, the underground extraction techniques mainly consist of protective layer mining and gas drainage at the working face. Given that a significant number of coal seams in Guizhou Province and other regions in China are challenging to extract, reservoir reconstruction and permeability enhancement through hydraulic fracturing remain the predominant strategies for contemporary CBM extraction. Furthermore, considering the complex geological conditions prevalent in Guizhou’s coal mines and the associated risks with underground CBM extraction, it is imperative that all newly opened coal mines should prioritize surface coalbed methane extraction. Consequently, for an extended period, both at present and in the foreseeable future, the development of coalbed methane in Guizhou Province will continue to be largely governed by a surface extraction method utilizing hydraulic fracturing.

Scholars, both domestically and internationally, have conducted systematic research on the hydraulic fracturing of coal reservoirs. Representative studies that concentrated on theoretical models include the following: Dos Santos et al. proposed a theoretical model to estimate postfracturing fracture size and transmissivity, taking into account horizontal radial fracture propagation. Li et al. developed a theoretical model that describes the formation and propagation of T-type fractures, which are likely to occur concurrently in the fractured coalbed along bedding planes. Tang et al. established a new criterion under mix-mode conditions to predict the critical initiation pressure of rock, with prediction results showing good agreements with experimental results. Lu et al. introduced a hydraulic fracture initiation model that incorporates a subcritical crack growth law. Because of the inherent time dependence of rock failure, this model can describe the crack propagation at a low stress state due to the inherent time dependence associated with rock failure. Zhao et al. investigated the rheological fracturing behavior and established a rheological fracturing theoretical model for rock fractures based on rheological principles. Zhao et al. formulated an evolution law for stress intensity influenced by hydraulic pressure and far-field stress. Building upon the Griffith criterion, Yao et al. determined the pressure required to cross fractures, open them, or induce slippage. Wang et al. proposed a semianalytical model to quantitatively assess interference between wells. Zhang et al. proposed a semianalytical model with the microscale effects of shale oil flow. The results of this model exhibit good agreement with the numerical solution and field production data.

Representative research on physical simulation of fracturing is summarized as follows: Liu et al. conducted physical simulation tests of hydraulic fracturing on high-order coal specimens obtained from the Zhina coalfield of Guizhou Province and summarized the influence regulations of geo-stress, injection rate, and primary fractures on the fracture network. Due to the heterogeneous properties of the natural rock, artificially created cement specimens or other material specimens are necessary for investigating hydraulic fracturing parameters. Khadraoui et al. utilized PMMA specimens to model rock, enabling direct observation of initiation and propagation of hydraulic fractures. Aghababaei et al. examined how grain size affects fracture roughness and surface toughness. Their findings indicate that bout roughness and toughness values increase with larger grain sizes. Nondestructive testing methods such as CT scanning technology are widely employed for imaging rock fractures. Moghaddam and Golshani investigated the cyclic injection and structural anisotropy. The CT images revealed that cyclic fracturing results in a more complex fracture network. Zenchenko et al. used ultrasonic transmitting monitoring to measure the fracture propagation and opening, so that the effect of natural fractures on hydraulic fractures is revealed. The formation of complex fracture networks is inseparable from the impact of natural fractures. Hu et al. investigated the interaction between the perforating hole and the natural fracture dip angle. Li et al. investigated the influence of stress difference and pre-existing fractures on hydraulic fractures through experiments. Powlay and Zhang et al. used artificial linear notches to control the direction of the fracture direction. Dabbaghi et al. examined the effects of fracturing fluid on the elastic and strength parameters of the rock materials. Butt et al. conducted hydraulic fracture experiments on granitic cubes, employing microseismic monitoring to compare the initiation and propagation regimes induced by oil versus water. They found that water penetrated the injection borehole more easily, facilitating rock fracturing. Li et al. employed acoustic emission monitoring equipment to study crack evolution in Longmaxi Formation shales with a single prefabricated fissure. Novikova et al. , performed large-size hydraulic fracture experiments to assess the minimum horizontal stress and formation breakdown stress. They discovered that these values differ significantly from theoretical estimations. In light of this phenomenon, they proposed a backstress-modified calculation method. ,

The numerical simulation of hydraulic fracturing has experienced rapid advancements in recent years. Commonly employed numerical simulation methods for modeling rock fracture include finite difference methods, extended finite element methods, − phase-field method, − etc. Notably, the attenuation of stress is pronounced at distances away from the crack tip. Zhou and Yang focused solely on elements within an influence region that extends 2–3 times the length of the crack, discretizing the injection rates of units within this region. Their proposed method demonstrated significant advantages in computational efficiency. Sherratt et al. introduced a novel approach to couple a complex fracture geometry model, generated through finite element analysis, with a finite difference solver. The presence of natural fractures significantly affects both the propagation and flow seepage capacity of hydraulic fractures. Ali et al. conducted two numerical simulations to examine the interaction between hydraulic and natural fractures. Their analysis revealed that growing hydraulic fractures exert tensile and shear stresses on adjacent natural fractures. Furthermore, Yi et al. − developed a comprehensive numerical model for multifield coupled hydraulic fracturing that incorporates thermal, elastic, plastic, and flow fields. Their finding indicated more effective hydraulic fracturing in deep geological environments involving rock or coal formations.

The Longtan Formation of the Upper Permian serves as the primary gas-bearing unit of coalbed methane in Guizhou Province. However, the characteristics of these coal reservoirs exhibit obvious differences when compared with those found in northern China or other regions. These differences are mainly characterized by a multitude of coal seams, a distribution of seam thicknesses that range from thin to medium, and complex regional geological conditions. For example, both the thickness and permeability of coal seams in countries such as the United States and Russia are markedly superior to those observed in Guizhou.

However, current research efforts in China predominantly concentrate on hard coals or thick coal seams founded in Northern China. At present, horizontal well technology has achieved breakthroughs in coalbed methane projects in these regions. In fact, it has even been successfully implemented for the development of deep coalbed methane, as well. The thick and relatively uncomplicated structure of the coal seams in Northern China facilitates enhanced productivity through horizontal well extraction. In contrast, the geological conditions for coal seams in Guizhou are notably complex. There is considerable vertical and horizontal heterogeneity among the adjacent coal seams. In addition, many of the coalfields are located within rugged and steep mountainous terrains. Consequently, the implementation of large horizontal wells presents certain challenges. Currently, cluster wells are primarily utilized for the development of coalbed methane, which involves the arrangement of multiple development wells at a single site. Given the characteristics associated with multiple coal seams, a “staged fracturing and laminated extraction” approach is frequently adopted. Nevertheless, there remains limited research concerning fracturing technology and fracture network characteristics specific to complex thin-to-medium-thick coal seams within Guizhou Province. It is evident that the commercial development of coalbed methane in Guizhou Province still lacks a reliable theoretical framework, as well as the technical foundations necessary for further advancements. The Liupanshui coalfield is the block with the highest coalbed methane resources within the Upper Permian Longtan Formation in Guizhou Province. It is recognized as one of China’s most important coalbed methane producing areas, ranking in the forefront of bituminous coal-anthracite coalfields nationwide. The total estimated coalbed methane resources in Liupanshui amount to 1.39 trillion m³, which constitutes 45.47% of the province’s total geological resources, yielding an average resource abundance of 226 million m³/km². The efficient exploration and development of coalbed methane surrounding the Liupanshui coalfield have emerged as a primary focus for coalbed methane development efforts in Guizhou Province. The Dahebian coal reservoir belongs to the typical coal seam of Upper Permian Longtan Formation in the Liupanshui coalfield, exhibiting complex geological conditions and a complex coal matrix structure. It is still a very important research method to carry out hydraulic fracturing simulation tests of a coal reservoir for recognizing the effect of fracture propagation and permeability enhancement of a coal reservoir. Such studies provide direct guidance for subsequent field fracturing design, reconstruction efforts, and drainage control strategies.

In thin-to-medium-thick coal seams, the cluster wells distribute rather densely. The well clusters have a close space distance. Hong et al. propose a new calculation approach for the wellbore separation factor based on the relative position of adjacent wells. Cheng et al. inverted the 3D trajectory of an adjacent well employing the waveform data separated from the scattered P-wave data. The maximum deviation inversion trajectory and actual trajectory are 1.5 m. Hu et al. transformed the skewed raw data into a closure distance probability distribution function based on the Box–Cox transformation method. The distribution function is used to assess the collision risk of well clusters. Wu et al. developed a rotating magnetic ranging method, in order to determine whether the drill is approaching the target well, by analyzing the variation trend of the curve. Wang et al. established the numerical model with the same size as the on-site zones. The research found that multistage fracturing makes adjacent casing deformation accumulate gradually. The machine learning and deep learning approaches have also been extensively employed in processes, such as the design and drilling of cluster wells. Kheirollahi et al. proposed a new automatic well clustering approach with a particle swarm optimization method. The result is conducive to selecting the pilot well. Gu et al. established a forecast model of well drilling limits under different constraints. The comparison results indicated that the optimized cluster well could drill more wells. Diaz et al. used an artificial neural network to predict the rate of penetration based on various drilling-related parameters. The prediction results may guide the drilling process to reduce the cost. Wang et al. optimized the side tracking horizontal well trajectory well in well clusters, based on the DS heuristic algorithm. The research results showed that multiobjective is more advantageous than a single-objective method.

In this study, coal blocks from the 11^# coal seam of the Wangjiazhai coal mine, situated in the Dahebian Syncline of the Liupanshui coalfield, are selected for the preparation of large-sized complete coal cubes. Physical simulation hydraulic fracturing experiments are conducted to investigate the fracturing process under various influencing factors. The effects of injection rates of fracturing fluid, differential coefficients of geo-stress, and varying depth within the coal formation on fracture initiation and propagation during the fracturing process are examined. Furthermore, the influence of natural fractures on the development of the fracture network is discussed and analyzed. Finally, recommendations for on-site fracturing operations are put forward.

2. Fracturing Simulation Research Plan

2.1. Geological Conditions of the Study Area

The Dahebian Syncline is situated within the Liupanshui coalfield in Guizhou Province and serves as a representative example of a typical syncline controlled gas-bearing structure. Other notable syncline-controlled gas-bearing structures include the Santang Syncline and Agong Syncline in Zhina coalfields. Fresh coal cores were collected from the 11^# coal seam of the Wangjiazhai coal mine, which is situated in the Dahebian Syncline Coalbed Methane Block of Zhongshan District, Liupanshui City, Guizhou Province. In accordance with relevant provisions outlined in the ″Regulations for Rock Physical Mechanical Property Tests of the Ministry of Geology and Mineral Resources of the People’s Republic of China (DY-94)″ and the ″Regulations for Rock Tests in Water Conservancy and Hydro-power Engineering (DLJ 204–81)″, various standard specimens required for investigating the mechanical and physical properties were prepared.

2.2. Core Sampling and Preparation

Large-sized coal cores were selected from the 11^# coal seam (as illustrated in Figure ) at the underground working face for the purpose of processing large specimens. Observations of the visible coal reservoir within the underground coal seam work face indicate that jointing fractures are well-developed, which significantly complicates the extraction of large coal blocks.

1.

Work face of the 11^# coal seam.

Therefore, the specifications for coal specimens used in the hydraulic fracturing physical simulation experiments were established at 200 mm × 200 mm × 200 mm. To restore the hydraulic fracturing condition of the actual coal seam as much as possible, the size of the sample should be as large as feasible to accurately reflect the size effect of the cleat fractures. Nevertheless, due to the constraints imposed by the size of the extracted coal sample and the experimental equipment, the sample size cannot be overly large. After comprehensive consideration, the standard of 200 × 200 × 200 mm is chosen as the size of the experimental sample. It can not only encompass abundant primary fractures of the coal body but also facilitate the cutting of a considerable number of samples from the collected coal samples. And there are also devices available to conduct experiments on it.

Due to the challenges associated with preparing the large-sized coal specimens, some corners and edges were found to be missing or chipped during processing, therefore, this issue. The coal cores were initially cut as large as possible using a wire cutter, and any missing corners and edges were subsequently repaired with cement. The raw specimen was then placed in a specialized mold for curing. First, a layer of cement mortar is laid at the bottom of the mold. Subsequently, the coal sample is carefully placed into it. Next, the cement mortar is poured along the edge of the mold, and the mold surface is simultaneously tapped with a hammer to densify the cement mortar. Finally, the upper surface of the specimen is leveled off. After hardening, it is maintained for 28 days. To minimize strength discrepancies between the coal and the cement mortar, several samples of cement mortar were prepared prior to repairing the actual specimen. It was ultimately determined that the cement mortar composition with a quality ratio of ″quartz sand:cement:coal powder = 7:2:1″ exhibited uniaxial compressive strength close to that of the Wangjiazhai coal specimens. Consequently, this specific proportion was employed for preparing the cement mortar used in specimen repairs. Figure illustrates both the pouring and curing processes involved in creating the large coal specimens that will be utilized in the experiments.

2.

Processing and maintenance of large-sized coal.

During on-site hydraulic fracturing, the fracturing fluid enters the coal seams through perforated holes in the wellbore. To accurately simulate this process, a wellbore design was developed following a specific scheme: SUS304 stainless steel was used for the pipe, and its tensile strength is greater than 515 MPa. The steel pipe with an outer diameter of 14 mm and an inner diameter of 10 mm was utilized to represent the casing. Three rows of threaded holes, each with a diameter of 6 mm, were drilled at distances of 72.5 mm, 95 mm, and 117.5 mm from the mouth of the pipe. Each row contains three holes to mimic perforation. The schematic diagram of the fracturing pipe is illustrated in Figure . The bottom end of the steel pipe is welded shut, while its upper end features internal threading for sealing connection with the pump pipeline from the hydraulic fracturing pump pressure servo control system. Additionally, high-strength adhesive is employed to seal the pipe to the prefabricated wellbore. The schematic diagram depicting the drilling process for coal specimens is presented in Figure .

3.

Preparation process of the large-size coal specimens simulation wellbore.

4.

Large-size complete coal specimen and schematic diagram of fracturing simulation (before fracturing).

2.3. Fracturing Scheme Designing

The results of hydraulic fracturing for coal seam 11^# in the Dahebian Syncline coalfield area are presented in Table .

1. Stress Results of Hydraulic Fracturing.

stress type	value of the stress
first failure pressure	p ic1 = 16.389 MPa
second failure pressure	p ic2 = 15.812 MPa
pore pressure	p0 = 8.343 MPa
stable pressure	ps = 14.878 MPa
tensile strength	σt = p ic1 – p ic2 = 0.577 MPa
minimum horizontal stress	σ3 = p s = 14.878 MPa
maximum horizontal stress	σ1 = 3p s – p ic1 – p 0 = 18.616 MPa

Thus, the ground stress value of the coal seam is approximately

$${\sigma }_{\mathrm{v}}=18.56\mathrm{M}\mathrm{P}\mathrm{a};{\sigma }_{\mathrm{H}}=18.6\mathrm{M}\mathrm{P}\mathrm{a};{\sigma }_{\mathrm{h}}=14.88\mathrm{M}\mathrm{P}\mathrm{a}$$ 1

The physical simulation parameters for hydraulic fracturing were determined, and a comprehensive testing plan was developed. Based on the theory of similarity principles, , the actual hydraulic fracturing scenario has been significantly abstracted, to accurately identify the physical simulation parameters related to laboratory-scale hydraulic fracturing. These parameters include well type, wellbore inner diameter, type of fracturing fluid, injection rate, and in situ stress difference coefficient, among others. Following the orthogonal test principle, we establish a systematic testing plan for laboratory hydraulic fracturing simulations. The primary factors under investigation are as follows:

• 1Different fracturing fluid injection rates. The predominant fracturing fluid currently employed for coalbed methane extraction remains active water fracturing fluid, with fresh water as its largest component, supplemented by swelling inhibitors, drainage aids, antibacterial agents, and other additives. In the present experiment, a mixture of fresh water and 1% KCl was utilized as the fracturing fluid. The primary function of KCl lies in its swelling inhibition effect. Considering the presence of clay components in coal seams of Guizhou, China, it is necessary to formulate fracturing fluids with appropriate KCl concentrations to mitigate clay mineral expansion induced by fracturing operations. Surrounding our study area, multiple fractured wells have consistently employed fracturing fluids containing 1% KCl, as evidenced by field practices. Laboratory investigations reveal that the fracture propagation process occurs within an extremely short time frame, during which the majority of fracturing fluid rapidly flows through the fracture network. Notably, KCl concentration exhibits negligible influence on both fracturing fluid viscosity and stimulation effectiveness. To maintain consistency with field conditions, 1% KCl concentration was adopted in laboratory-scale physical simulation experiments. Three different fracturing fluid injection rates of 0.5 mL/s, 1.0 mL/s, and 1.5 mL/s were established for comparison purposes. Given that the primary fractures within the 11^# coal seam are relatively well-developed, analyzing the effects of varying fracturing fluid injection rates can offer valuable insights into on-site dynamic filtration processes, selection of optimal injection rates, pump pressure design considerations, and related applications.
• 2Different in situ stress state. The vertical stress (σ_v) within a specific coal seam is typically considered to be constant. In contrast, the two horizontal principal stresses (σ_H and σ_h) often display significant differences. This variation is quantified by the in situ stress difference coefficient K, as demonstrated in the following equation

$$K=\frac{{\sigma }_{\mathrm{H}}-{\sigma }_{\mathrm{h}}}{{\sigma }_{\mathrm{H}}}$$ 2

where K represents the in situ stress difference coefficient σ_H and σ_h denote the maximum and minimum horizontal principal stresses, respectively. This study establishes three distinct in situ stress difference coefficients of K = 0.25, 0.35, and 0.4 to characterize various stress states of the coal reservoirs.

• 3Different coal formations at different depths. Different in situ stress states are defined to represent formations at varying depths. The properties of the 4^#, 7^#, and 11^# coal seams on-site exhibit notable similarities. Consequently, three corresponding formation depths of 686 m, 763 m, and 817 m have been assigned to these seams. The associated in situ stress is determined based on the formation depth alongside the hydraulic fracturing method employed on-site. Currently, only coal samples from the 11^# coal seam are available for collection. Therefore, coal cores from this seam are utilized for testing purposes.

The fracturing schemes that consider various factors are detailed in Table , which includes a total of 10 experiments.

2. Large-Size Coal Specimen Hydraulic Fracturing Physical Simulation Test Plan.

	specimen no	in situ stress state/MPa			formation depth/m	K	injection rate/mL/s	remark
		σH	σh	σv
different injection rate	7#	18.6	14.88	18.56		0.25	0.5	water + 1% KCl
	8#	18.6	14.88	18.56			1.0
	9#	18.6	14.88	18.56			1.5
different in situ stress	2#	18.6	15.44	18.56		0.2	1	water + 1% KCl
	W-2#	18.6	14.88	18.56		0.25	1
	W-6#	18.6	14.10	18.56		0.319	1
	W-4#	18.6	13.77	18.56		0.35	1
different formations	W-7#	14.49	12.02	16.05	686		1	water + 1% KCl
	W-8#	19.59	11.12	18.17	763		1
	W-3#	20.67	13.77	18.88	817		1

In the process of hydraulic fracturing, cracks begin to initiate at the hole wall when the pressure of the fracturing fluid exceeds the rock’s fracturing pressure at that location. Consequently, the occurrence of fractures is influenced by several factors, including fluid pressure, mechanical properties of rock formation, the in situ stress field, and the drilling direction. Among these factors, the type of in situ stress field serves as a critical determinant for fracture initiation. Based on the relative magnitudes of the three principal stresses σ_h, σ_H, and σ_v, the in situ stress field can be categorized into three types: normal fault type (σ_v > σ_H > σ_h), strike-slip fault type (σ_H > σ_v > σ_h), and reverse fault type (σ_H > σ_h > σ_v).

According to the maximum tensile stress criterion and the fracture mechanics, the initiation pressure for fracturing can be expressed as

$$P_{\mathrm{f}}=\frac{{\sigma }_{\mathrm{t}}-(3{\sigma }_{\mathrm{h}}-{\sigma }_{\mathrm{H}})}{2-\alpha \frac{1-2\nu }{1-\nu }}$$ 3

where P _f is the fracture initiation pressure; σ_H and σ_h are the maximum and minimum horizontal principal stresses of the reservoir; ν is the Poisson ratio of coal; α is the coefficient related to the compression properties of the coal matrix; and σ_t is the tensile strength of coal.

3. Mineral Composition and Microstructure Characteristics

3.1. Mineral Composition

As shown in Figure , The main minable coal seams in this area are all black to grayish black, presenting an asphalt to glass luster and partially a metallic luster. They are mainly in a massive form, with a small amount of fragmented and fine-powdery forms; the main structure is medium-fine banded, with a small amount of wide banded structure. The coal-rock components are mainly bright coal and dull coal, followed by vitrain and fusain. Except for the coal-rock types of some coal seams being dull to semidull, the remaining coal seams are all semidull to semibright. In the middle of the northwest flank and the shallow part of the southeast flank of the syncline, the coal-bearing strata are disrupted by faults and compressed by folds. The coal seams are mostly crushed into powder and fine granular forms, with less block coal, while in the deep part, they are mainly in a massive form.

5.

Rock core diagrams of boreholes in different coal seams.

3.2. SEM Test

The SEM tests were conducted using a Nova 400 scanning electron microscope manufactured by FEI in the United States to examine the development and morphological characteristics of pore structures in coal samples, as illustrated in Figure . The observations reveal that the surface of the coal rock exhibits a scale-like morphology and is relatively smooth. This is primarily attributed to the deformation and even destruction of pore structures induced by tectonic and stress actions, leading to uneven morphological characteristics.

6.

Geometrical features of the coal.

SEM offers the advantage of high magnification and continuous observation for examining microfractures in coal rock. The observational results indicate that the matrix of the top and bottom strata of the Dahebian no. 1 well is generally dense, with fracture width predominantly ranging from hundreds of nanometers to several micrometers. The fracture surfaces are relatively smooth, extend over considerable distances, and are primarily classified as shear fractures (Figure ). This phenomenon can be attributed to the fact that the rock strata have experienced multiple epochs of tectonic movements with varying intensities, leading to a well-developed network of secondary fractures. Furthermore, bedding planes and most fractures are filled with materials such as calcite and clay minerals. These fillings typically manifest in striped patterns or dense mesh formations, resulting in poor connectivity among the fractures.

7.

Fracture feature of the coal.

4. Influence of Different Factors

4.1. Fracturing Fluid Injection Rate

See Table for the fracturing simulation test scheme of coal specimens subjected to varying injection rates of the pump pressure curve. The different settings for the fracturing fluid injection rates are 0.5 mL/s, 1.0 mL/s, and 1.5 mL/s, respectively. The type of fracturing fluid used is fresh water with a concentration of 1% KCl, while the three principal stresses are maintained at identical levels.

4.1.1. Pumping Pressure Curve

The pumping pressure curves corresponding to coal specimen fracturing at different injection rates were obtained and are presented in Figure .

8.

Pump pressure curves of coal specimens fracturing under different fracturing fluid injection rates.

As illustrated in Figure , the pump pressure curve can be broadly divided into five sections.

• 1Pressure accumulation stage. Following the initiation of the servo pumping system, the pumping pressure within the specimen increases linearly as the fracturing fluid injection rate gradually intensifies. For instance, for specimen 7^#, the pump pressure reaches a fracture pressure of 15.15 MPa at 31.94 s. Notably, for specimens subjected to different injection rates, the greater the fracturing fluid injection rate, the shorter the time required to reach the fracturing pressure. This phenomenon occurs because a greater volume of fracturing fluid is injected per unit time due to an increased injection rate. However, there is no significant correlation between fracturing pressure and injection rate itself. For example, specimen 9^# (with an injection rate of 1.5 mL/s) exhibits a fracturing pressure 7.22 MPa and specimen 8^# (with an injection rate of 1.0 mL/s) has a fracturing pressure of 17.31 MPa, while specimen 7^# (with an injection rate of 0.5 mL/s) shows a fracturing pressure approximately equal to 15.15 MPa.
• 2Main fracture formation stage. Following the formation of the main fractures, there is an observable increase in the volume of fractures and fissures within the coal specimen, accompanied by an instantaneous drop in fracturing fluid pressure.
• 3Main fractures propagation and secondary fractures formation stage. With the continuous injection of fracturing fluid, the fracturing fluid pressure gradually rises. Subsequently, the pump pressure curve exhibits sharp fluctuations with significant amplitude and multiple peaks indicating that multiple secondary fractures are being formed either simultaneously or sequentially. The frequent oscillation of pumping pressure serves as a distinct characteristic of network fractures generated during coal specimen fracturing.
• 4Pump pressure maintenance stage. During this stage, the pump pressure remains relatively stable or experiences zigzag fluctuations. However, variations in pump pressure profiles among different specimens are notable and exhibit considerable differences.
• 5Stop pump pressure stage. Upon cessation of pumping operations, there is a gradual reduction in pump pressure from its maintained level.

4.1.2. Result Analysis

Based on the results presented in Figure , the fracturing fluid injection rate significantly influences the fracturing process with the coal specimen. Overall, a higher fracturing fluid injection rate correlates with an increased fracturing pressure of the coal specimen. This phenomenon can be attributed to the nearly incompressible nature of the fracturing fluid. Under identical conditions, a greater injection rate necessitates the overcoming of higher resistance. Furthermore, elevated injection rates lead to more pronounced fluctuations in the pump pressure curve, indicating that the fracturing process becomes increasingly complex. This complexity primarily manifests as an enhanced interaction between the fracturing fluid and natural fractures, resulting in main fractures connecting with a greater number of natural fractures. Considering previous porosity–permeability tests and observations regarding natural fractures in Wangjiazhai coal, it is evident that the 11^# coal seam exhibits relatively high porosity and contains numerous natural fractures. A lower fracturing fluid injection rate may hinder the effective connectivity between main fractures and natural fractures during fluid injection. Therefore, it is advisable to adopt a higher fracturing fluid injection rate for optimal results.

From the shapes of the pump pressure curves, it is observed that at low fluid injection rates, the pump pressure stabilizes following the initiation of fractures. This stabilization facilitates an extension of the fracture length. However, a smooth pressure curve indicates that the fracturing fluid predominantly propagates through the primary fracture with minimal interaction with adjacent natural fractures. Conversely, as the injection rate increases, the pump pressure exhibits sawtooth-pattern fluctuations. These variations suggest significant interactions between the main fractures and the natural fractures during the fracturing process. The observed fluctuations imply that when fluid enters surrounding natural fractures, there is a localized drop in pump pressure. While this phenomenon contributes to forming a complex fracture network, it simultaneously diminishes the driving force for further propagation of the main fractures. With continued increases in fluid injection rates, more favorable conditions arise: the overall trend of the pump pressure curve displays a gradual decline accompanied by minor local fluctuations. This pattern indicates that while extending, the main fractures are effectively connecting with natural fractures. Despite increased fluid loss attributed to numerous natural fractures, a higher fracturing fluid injection rate offsets this loss and prevents stagnation in forward propagation of the main fracture system.

It is crucial to recognize that there exists an optimal fracturing fluid injection rate that aligns with the characteristics of the reservoir’s natural fracture network. This injection rate must be sufficiently high to facilitate the ongoing propagation of primary fractures while enhancing effective communication between these primary fractures and the natural fractures, all while compensating for fluid loss into the natural fracture system.

4.2. In Situ Stress Difference Coefficient

The in situ stress difference coefficients considered are 0.25, 0.35, and 0.4. The coefficient of 0.319 represents the actual stress conditions of formation and is included in this comparative analysis.

4.2.1. Pump Pressure Curve

The pump pressure curves are shown in Figure .

9.

Pump pressure curves of coal specimens fracturing simulations at different in situ stress difference coefficients.

The influence of the in situ stress difference coefficient (K value) on fracture propagation is primarily reflected in its effect on the breaking pressure. As the K value increases from 0.2 to 0.35, there is a general trend of increasing breaking pressure, with values changing from 12.5 MPa, 18.5 MPa, and 16.5 MPa to 23.5 MPa. The smoothness of the pump pressure curves transitions from a gradually decreasing pattern to a slowly decreasing pattern, then to a slowly increasing pattern, and finally to a fluctuating increasing pattern. At a lower K value, three-dimensional stresses are relatively uniform, allowing fractures to initiate from various directions and connect with more natural fractures. This results in the formation of a more complex fracture network. However, as the K value increases, fractures aligned parallel to σ_H gradually become dominant main fractures while propagation in other directions becomes suppressed, thus facilitating easier fracture formation along this direction. When the K value is high, fractures predominantly extend in one direction. If this direction lacks an adequate number of natural fractures conducive for further propagation, then new fracture initiation may be necessary, leading to localized fluctuations during the process. Therefore, it can be concluded that higher K values hinder the development of complex fractures that interconnect with natural fractures but promote long-propagating main fractures. However, when the in situ stress difference coefficient reaches excessively high levels, fracturing becomes constrained solely along the direction parallel to σ_H. This limitation restricts fracture generation in other orientations. Consequently, formed fractures tend toward elongation and exhibit an upward propagation trend. And outcome is also unfavorable for distant fracture formation.

4.2.2. Result Analysis

It can be stated that when the K value is extremely low, although it facilitates the propagation of fractures in all directions, this phenomenon is not entirely advantageous. This limitation is evident in several aspects: (1) the development of natural fractures significantly influences hydraulic fracturing. Specifically, hydraulic fractures may connect with natural fractures in the near-wellbore zone to enhance the reconstruction effectiveness. However, the extension length of these fractures may be insufficient, thereby impacting the overall reconstruction outcome. (2) Due to the omnidirectional spread of fractures, including vertical propagation toward both the roof and floor, there exists a certain degree of ineffective fracturing. Consequently, maintaining a specific K value within the reservoir proves beneficial for increasing fracture network density during reconstruction, while ensuring that formed hydraulic fractures possess considerable length and width. Moreover, the K value plays a crucial role in wellbore location deployment. Based on the elliptical distribution characteristics of hydraulic fracture surfaces, when faced with a low K value, it is advisable to reduce well spacing in alignment with σ_H while simultaneously increasing well spacing along σ_H. Conversely, under conditions where the K value is high, one should increase wellbore spacing along σ_H and decrease it along σ_h accordingly. This strategic adjustment serves to minimize blank zone areas between coalbed methane development wells and enhances overall coalbed methane extraction efficiency.

4.3. Coal Formation Depth

In a given mining area characterized by multiple coal seams, variations in depth can occur even within a single coal seam across different locations. This study focuses on the drilling conditions associated with the 11^# coal seam located within a specific syncline block. Three distinct formation depths of 686 m, 763 m, and 817 m are examined. The fracturing fluid injection rate remains constant at 1 mL/s, utilizing a composition of fresh water combined with 1% KCl.

4.3.1. Pump Pressure Curve

The pump pressure curves obtained from simulation experiments involving the fracturing of coal specimens at varying formation depths are presented in Figure .

10.

Pump pressure curves of coal fracturing simulations at different injection rates.

From Figure , it is evident that the in situ stress increases with the depth of the coal seam, leading to a corresponding rise in breaking pressure. For the simulated depths of 686 m, 763 m, and 817 m, the observed breaking pressures are 6.8 MPa, 12.05 MPa, and 31.3 MPa, respectively.

4.3.2. Result Analysis

Overall, the breaking pressure demonstrates an increasing trend with the formation depth. However, this increase is not linear. The pump pressure curves exhibit transitions through patterns of a slow decrease, slow increase, and stress drop. This phenomenon can be attributed to the fact that as the depth of the coal seam increases, the closure of the coal fractures becomes tighter, necessitating greater resistance to overcome during both fracture initiation and propagation. Consequently, the breaking pressure of the coal increases significantly with increasing depth. At a depth of only 686 m, the pump pressure curve reaches its first breaking point at a pressure of 6.8 MPa within just 8 s. At a greater depth of 763 m, this curve attains its first breaking point at a higher pressure of 11.3 MPa but takes longer, 60 s to do so. In contrast, for a coal seam located at a depth of 817 m, the first breaking point pressure escalates to 31.3 MPa while further extending the time required to reach this threshold to 80 s. These observations indicate that increased depth is a critical factor contributing to both prolonged fracture times and elevated breaking pressure in the coal seam.

It is noteworthy that although the overall pump pressure curve of the coal specimen remains stable, indicating the formation of main fractures, the morphology of these fractures during their extension exhibits relative uniformity. Once the initial fracture is formed (as seen in specimen W-3), it continues to propagate along the main fracture, as evidenced by both specimens W-8 and W-3. This phenomenon can be readily understood: as the depth of coal seam increases, so does the value of (σ_H – σ_h), which facilitates the development of main fractures that are parallel to σ_H. Conversely, at shallower depths, the value of (σ_H – σ_h) diminishes and natural fractures tend to be more open and developed. This results in a more pronounced interaction between propagating fractures and existing natural fractures, leading to significant localized fluctuations in the pump pressure curve.

5. Fracture Cross-Section and Fracturing Mechanism Analysis

5.1. Fracture Cross-Section

For specimens 1, 4, and 7# subjected to a fracturing fluid injection rate of 0.5 mL/s, specimen 1# was chosen for line cutting in order to elucidate the fracture patterns across four distinct faces. Figure presents the cross-section and schematic representation of the fracture network for specimen 1#. Four surfaces were designated as observation planes for cross-sectioning, labeled as A, B, C, and D. The original images of the coal postfracturing are displayed on the upper line, while the schematic representation of the fracture network is presented on the below line.

11.

Cross-section and fracture network schematic of specimen 1# (0.5 mL/s).

On surface A, the fractures are mainly horizontal and align with the bedding direction, accompanied by some fractures that extend along the cleat. However, these fractures exhibit relatively poor connectivity and are generally sparse. In contrast, on surface B, while the fractures remain primarily horizontal along the bedding direction with some extending along the cleat, there is a greater presence of interbedding fractures that connect the bedding planes. This results in a more complex fracture network at the intersections of the cleats. Surface C primarily displays microfractures oriented along the cleat direction, with a localized cross-fracture network developing in the upper portion of the specimen. Nevertheless, the overall connectivity within this fracture network remains limited. On surface D, multiple nearly horizontal main fractures with gentle inclinations formed alongside several secondary fractures. In summary, surfaces B and D demonstrate more intricate fracture networks compared to surfaces A and C. Notably, when observed along the prefabricated cleat direction, a more elaborate fracture network is evident on the left side of surface A, right side of surface B, left side of surface C, and right side of surface D. This observation aligns consistently with the orientation of the prefabricated cleat.

For specimens 2#, 5#, and 8# subjected to a fracturing fluid injection rate of 1.0 mL/s, specimen 8^# was chosen for linear cutting to illustrate the fracture patterns across four distinct faces. As depicted in Figure , it presents the cross-section and schematic representation of the fracture network for specimen 8# (with an injection rate of 1.0 mL/s). Four surfaces were designated as observation planes for cross-sectioning, labeled as surfaces A, B, C, and D. The original coal image of the coal specimen postfracturing is displayed on the upper line, while the schematic presentation of the fracture network is presented on the below line.

12.

Cross-section and fracture network schematic of specimen 8# (1.0 mL/s).

On surface A, multiple main fractures extend along the bedding direction and change orientation at the vertical cleat within the specimen, subsequently connecting with adjacent bedding planes. This face is also characterized by numerous secondary fractures, resulting in a complex fracture network. In contrast, surface B exhibits only a single horizontal main fracture and very few microfractures, which can be attributed to the limited exposure of coal on this side. Surface C features a larger horizontal main fracture alongside multiple main fractures oriented along the bedding direction. Between these main fractures are several vertical or oblique fractures that connect them, indicating excellent connectivity within the fracture network. Surface D resembles surface B in that it forms only a single horizontal main fracture aligned with the bedding direction. In summary, surfaces A and C demonstrate more intricate fracture networks. Observations along the direction of the prefabricated cleat reveal that more complex fracture networks are present on both sides of surface A and on the right side of surface C. This aligns with the orientation of the prefabricated cleat. Conversely, surfaces B and D exhibit less prominent fracturing due to their limited exposure to coal.

For specimens 3#, 6#, and 9# subjected to a fracturing fluid injection rate of 1.5 mL/s, specimen 3# was selected for linear cutting and subsequent analysis of the fracture patterns on its four surfaces. As illustrated in Figure , the schematic diagram depicts the cross-section and fracture network of specimen 3# (with an injection rate of 1.5 mL/s). Four side surfaces, labeled A–D, were chosen for detailed observation. The upper line of this figure presents the original image of the coal after fracturing, while the bottom line displays a schematic of the fracture network.

13.

Cross-section and fracture network schematic of specimen 3# (1.5 mL/s).

On surface A, the fractures are predominantly horizontal, extending along the bedding plane, with some fractures following aligning with the cleat orientation. The resulting fracture patterns mainly consist of nearly horizontal and gently dipping fractures. Surface B features two horizontal fractures that extend along the bedding plane, accompanied by additional steeply inclined fractures forming at significant angles in relation to the cleat direction. These steeply inclined fractures connect the main horizontal fractures, thereby creating a complex fracture network. Surface C exhibits a single inclined main fracture that is interconnected by counteracting fractures, leading to a relatively simple fracture network. In contrast, surface D displays multiple horizontal fractures extending along the bedding direction with numerous secondary fractures developing in alignment with the extension of the main fracture and primarily following natural fractures such as cleats within the coal. Overall, surfaces A, B, and D exhibit more complex fracture networks compared to surface C. Notably, concerning the orientation of prefabricated cleats, complex fracture networks are observed on the left side of surface A, on the right side of surface B, on the left side of surface C, and on the right side of surface D. This distribution aligns closely with the direction of these prefabricated cleats.

In general, specimens subjected to higher injection rates (1.0 and 1.5 mL/s) exhibited a more complex and well-connected fracturing network. In contrast, at a lower injection rate of 0.5 mL/s, the fractures that developed were more segmented, independent, and poorly interconnected. This observation indicates that an increased injection rate (1.0 mL/s or above) facilitates the formation of a more intricate fracture network. However, the distinction between the rates of 1.0 and 1.5 mL/s is not particularly pronounced.

Under the influence of hydraulic pressure, hydraulic fractures initiate and propagate in alignment with the direction of maximum horizontal principal stress, resulting in a main fracture that is perpendicular to the bedding planes. As this main fracture extends, it may deviate at weak bedding planes and continue its propagation along these planes. This process leads to bifurcation and deflection of the main fracture, generating secondary fractures that extend parallel to the bedding, ultimately forming an orthogonal complex fracture network. On the surface of the specimen, multiple hydraulic fractures extend along the bedding direction. These hydraulic fractures interconnect with weak bedding or natural fractures, prompting further bifurcation and turning phenomena. Consequently, an intricate interlaced fracture network is established. In coal reservoirs, hydraulic fractures are oriented vertically relative to the bedding and primarily extend in accordance with the direction of the maximum horizontal principal stress. As these fractures propagate and encounter natural fractures, they exhibit a turning and bifurcating behavior. Concurrently, new hydraulic fractures can be initiated at weak bedding planes during this propagation phase. These connections between bedding planes facilitate a crisscrossing fracture network formation, ultimately achieving enhanced reconstruction of coalbed methane reservoirs.

5.2. Fracture Mechanism Analysis

According to the analysis of hydraulic fractures morphology and propagation behavior in coal, it is evident that the formation of a fracture network is not only influenced by parameters such as the in situ stress state states and fracturing fluid injection rate but also closely associated with fundamental geological conditions, including the presence of natural fractures and the degree of bedding planes development. The latter factors play a crucial role in the formation of complex fracture networks. The correlation model is illustrated in Figure .

14.

Four modes of initiation and propagation of hydraulic fractures in coal.

However, in practice, the mechanical properties of various bedding types exhibit significant differences, and each type of bedding influences the initiation and propagation of hydraulic fractures to varying degrees. Based on experimental results from large-size in situ hydraulic fracturing tests conducted on coal specimens, four distinct modes of fracture initiation and propagation during coal seam hydraulic fracturing have been identified. Additionally, the formation mechanism of network fracture in a coal seam has been identified. Figure a illustrates a diagram depicting coal specimens subjected to three-dimensional in situ stress, while Figure b–e presents the four fundamental modes of hydraulic fracture initiation and propagation observed in coal reservoirs.

Mode 1: as illustrated in Figure b, the hydraulic fracture initiates along the direction of the maximum horizontal principal stress and propagates continuously in that direction, traversing directly through the bedding to form a singular fracture. Specimen WJZ-YL-05 exemplifies this type of failure mode, which occurs when the bonding strength of the bedding is robust, and natural fractures are not present.

Mode 2: as shown in Figure c, the hydraulic fracture initiates and extends along the bedding plane, resulting in a simple fracture that follows the bedding orientation. Specimen WJZ-YL-07 exemplifies this failure mode, which is observed when the bonding strength of the bedding is low or when natural fractures are well-developed.

Mode 3: as shown in Figure d, hydraulic fractures initiate and propagate along the direction of maximum horizontal principal stress. During the propagation process, these fractures may change direction upon encountering bedding planes and subsequently propagate along them, ultimately resulting in the formation of cross fractures. Specimen WJZ-YL-09 exemplifies this failure mode. This fracture mode predominantly occurs when the bedding bond strength is low or when natural fractures are relatively well-developed.

Mode 4: as shown in Figure e, hydraulic fractures initiate and expand in the direction of maximum horizontal principal stress. During their extension, these fractures deflect upon encountering bedding planes and continue to grow along the bedding. While extending along the bedding, local fractures may again change direction due to the influence of in situ stress, subsequently re-expanding in alignment with the maximum horizontal principal stress until a complex fracture network is established. This mode of fracturing predominantly occurs when the bond strength of the bedding is moderate, and fracture pressure remains relatively high.

6. Results and Discussion

6.1. Recommendations on Fracturing Fluid Injection Rate

Luo and Khadraoui et al. hold the view that the injection rate has pronounced effects on breakdown pressure. Zhang et al. carried out detailed investigation on fluid injection rate. Their findings indicate that rapid injection rates result in higher breakdown pressures, and there is a strong correlation between the breakdown pressure and the logarithmic injection rate. However, the leakage effect should be considered due to the permeability. Based on the prior research conducted by the author’s team, it has been established that the 11^# coal seam exhibits relatively high porosity and permeability. The laboratory testing of coal specimens indicates that the average porosity of the 11^# coal seam is comparatively low; thus, formation reconstruction by hydraulic fracturing is necessary. By comparison with the 7^# coal seam, the porosity and permeability of the 11^# coal seam are higher and the natural fractures are more developed. Consequently, the issue of fracturing fluid filtration must be carefully considered throughout the fracturing process. During this process, it is common for the main fractures to connect with natural fractures, allowing fracturing fluid to easily enter these natural fractures and resulting in a drop in fracturing fluid pressure. When the fracturing fluid injection rate is inadequate, a direct consequence is that fracturing only occurs within localized sections of the main fracture while communicating with the natural fractures. This situation makes it challenging to maintain sufficient pressure for main fracture expanding further. Such limitations hinder efforts to expand fracture networks effectively and achieve optimal reconstruction outcomes. Therefore, for coal seams, such as the 11^# coal seam, it is essential to consider a higher injection rate. It is recommended that an injection rate between 10 and 12 m³/min be adopted at operational sites.

For the 4^# and 7^# coal seams, both logging data and laboratory tests indicate that their porosity, permeability, and natural fracture development are significantly lower than those observed in the 11^# coal seam. Therefore, the filtration of the fracturing fluid during the fracturing process is considerably less than that in the 11^# coal seam. The desired fracturing effect can be basically achieved with a medium injection rate. Therefore, an injection rate of 8–10 m³/min is recommended.

6.2. Suggestions on Fracturing Stratification

The coal seams in Guizhou are predominantly characterized as thin to medium-thick and multiseam formations. At present, the predominant methane development approach employed is the cluster well strategy that utilizes “staged fracturing and combined layer drainage″. − Given the significant variations in reservoir properties and the in situ stress states among coal seams 4^#, 7^#, and 11^#, it is advisable to refrain from employing combined layer fracturing. Instead, separate fracturing for each individual seam should be implemented. The recommended fracturing sequence is as follows: 11^#→ 7^#→ 4^#. The bridge plug was utilized to seal off the respective coal seam during this process. Initially, the fracturing of the 11^# coal seam will take place, followed by the sequential fracturing of both the 7^# and 4^# coal seams. It is important to note that within these three seams, there exist certain layers with similar properties located in close proximity. Thus, they may be considered for simultaneous fracturing within an identical interval. During the execution of fracturing operations, a temporary blocking technique can be adopted to enhance efficiency.

6.3. Consideration of the Differential Coefficient of In Situ Stress

The results of three-dimensional geophysical exploration show that this region has a very complex geological structure, which leads to a significant difference in the distribution of in situ stress in space. The maximum horizontal principal stress and minimum principal stress do not conform to the linear distribution in the vertical direction. The in situ stress difference coefficient exhibits heterogeneity across diverse locations and regions, but it increases with the increase of depth as a whole. The overall distribution of the in situ stress state in this area meets the following situation: σ_H ≥ σ_v > σ_h. Higher vertical stress difference coefficient and injection rate resulted in larger fracture vertical extension distance, lower peak pressure, altering the propagation path. The hydraulic fractures extend along the bedding planes at a less in situ stress difference coefficient.

It can be concluded that fracturing the coal seam in the stimulation block will result in the formation of a complex fracture system characterized primarily by vertical fractures. Bedding fractures serve as auxiliary components in the fracture network and will interact with vertical fractures. The implementation of hydraulic fracturing in this area is anticipated to yield an optimal stimulation effect. The larger the K value, the smaller the minimum horizontal principal stress, and the crack surface will preferentially be perpendicular to the minimum horizontal principal stress and expand into an elliptical surface with its long axis aligned parallel to σ_H and its short axis parallel to σ_v. Furthermore, it is important to note that as the difference coefficient of in situ stress increases, the fracture ellipse will elongate and narrow. Consequently, this leads to increased difficulty in the fracturing operations. During construction activities related to fracturing, it is likely that there will be an increase in fracture pressure accompanied by a fluctuating upward trend observed on the pump pressure curve. Therefore, prior to designing a fracturing operation, it is essential to thoroughly assess both the magnitude of in situ stress and the difference coefficient of in situ stress within the relevant area. This assessment provides a foundational basis for making informed decisions regarding the appropriate equipment selection for fracturing and strategic deployment of well locations.

7. Conclusions and Recommendations

By selecting the coal blocks from the 11^# coal seam of the Dahebian Syncline for a physical simulation study of fracturing, this research examines the influences of various factors on fracture reconstruction and propagation. Furthermore, the application is discussed in relation to existing coalbed methane (CBM) development technologies in Guizhou Province. The main conclusions are as follows.

• 1Overall, an increase in fracture pressure within coal is positively correlated with a higher fracturing fluid injection rate. As injection rate rises, the pumping pressure curve exhibits greater fluctuations, suggesting that the fracturing fluid is penetrating more natural fractures or facilitating connections between primary fractures and natural fractures. Coupled with observations from fracture sectioning, it is evident that natural fractures are prevalent in Dahebian coal seams; furthermore, elevated injection rates are more conducive to the development of complex fracture networks.
• 2When the in situ stress difference coefficient is low, it promotes the extension of fractures in all directions to a certain extent. Conversely, as the in situ stress difference coefficient increases, fractures exhibit significant length and a specific width along a particular direction, thereby enhancing the density of the fracture network. Consequently, adjacent wellbore distance should be adjusted according to the direction of minimum horizontal stress based on the value of the in situ stress difference coefficient.
• 3The fracture pressure of the coal seam increases with depth, exhibiting a nonlinear correlation. The patterns of pumping pressure curves can exhibit various patterns, including slow decline, slow rise, and stress drop. At shallower depths, the fracture morphology is more profoundly influenced by natural fractures. In contrast, at greater depths, it becomes increasingly likely for the main fractures to develop parallel to the maximum horizontal principal stress.
• 4Through the analysis of the morphology and propagation characteristics of hydraulic fractures in coal, it is evident that the formation of network fractures is influenced not only by parameters such as the in situ stress magnitude and pumping injection rate but also by the extent of natural fractures development, bedding orientation, and other fundamental geological conditions. The latter factors play a crucial role in the establishment of complex fracture networks. Consequently, recommendations and strategies for field development are proposed.

The authors declare no competing financial interest.