The utilization of clean energy is regarded as one of the efficacious measures to alleviate global climate change (Dickey, 2006; Zhu et al., 2021). CBM, characterized by its non-toxic combustion, absence of particulate matter emissions, and lower carbon dioxide release compared to coal, oil, or wood, is acknowledged as a clean fuel (Chen et al., 2017; Peng et al., 2017; Kong et al., 2022; Ni et al., 2023). Over the past 2 decades, CBM has emerged as a significant energy resource and is anticipated to play a pivotal role in meeting future global energy demands (Moore, 2012; Lou et al., 2013; Fu et al., 2017; Gao et al., 2023a; Ifrene et al., 2023; Wang et al., 2024a). China, as the current global leader in coal consumption and production, can benefit from the development of CBM resources. This not only aids in reducing natural greenhouse gas emissions and maximizing energy utilization but also enhances safety measures in coal mining operations by mitigating risks associated with underground disasters such as fires caused by spontaneous combustion of natural gas and coal (Bao et al., 2020; Kong et al., 2021). The CBM reserves in China are distributed among high-rank, middle-rank, and low-rank coal seams, with each accounting for approximately one-third of the total resources. However, the focus of CBM development is primarily on the high-rank and middle-rank seams. A prime example of a high-rank CBM field is located in the southern Qinshui Basin of Shanxi Province (Bustin, 1997; Chen et al., 2016; Gao et al., 2023b; Qiu et al., 2023). In recent years, scholars have conducted comprehensive research on the geological background, physical properties of CBM reservoirs, resource evaluation, reservoir formation mechanisms, and production potential in the Qinshui Basin. Li et al. demonstrated that the tectonic subsidence and weak hydrodynamics in the southern block of the Qinshui Basin created favorable constraints for the occurrence of CBM (Li et al., 2018). Liu et al. discovered that during the formation of CBM reservoirs, the spatial distribution of CBM composition exhibits a gradual change in different locations, and there is relatively weak interference between CBM wells in the southern block of Qinshui Basin (Liu et al., 2018). Lu et al. conducted a study on the production characteristics and key factors of numerous gas Wells in the southern Qinshui Basin using variable correlation analysis and grey correlation theory. They concluded that hydraulic fracturing is an effective method to ensure high production of CBM wells, and identified gas content and permeability as two crucial factors influencing gas production of CBM wells (Lv et al., 2012). Jiang et al. conducted a study on the characteristics and gas production potential of the No. 15 coal seam in the northern block of Qinshui Basin, revealing that the No. 15 coal seam exhibits significant thickness, high adsorption capacity, moderate gas content, low critical desorption pressure, and substantial gas production potential (Jiang et al., 2022). At present, the commercial development of CBM from high-rank coal reservoirs in Qinshui Basin has achieved remarkable results (Nie et al., 2000; Qin et al., 2005; Chen et al., 2015; Wang et al., 2024b).

The P block is situated in the southeastern region of the Qinshui Basin, where CBM exploration activities commenced in the 1970s. This area represents a typical site for CBM development in China, with two primary target coal seams: the No. 3 coal seam of the Shanxi Formation and the No. 15 coal seam of the Taiyuan Formation. The CBM well of No. 3 coal in Block P has a prolonged production history, and the overall gas production of this coal seam has exhibited a noticeable declining trend. Therefore, it is imperative to develop No. 15 coal seam in order to enhance the gas production of the block (Yuanyuan et al., 2012). However, due to the different reservoir characteristics, there are significant differences in the gas production rules of No. 3 and No. 15 coal seam CBM wells, and the development method of No. 15 coal seam cannot be fully used for reference by No. 3 coal seam. Therefore, it is the key to develop efficient development methods to clarify the production characteristics and influencing factors of CBM wells in No. 15 coal seam.

Permeability is a key parameter that determines fluid migration ability, and its influence on the production dynamics of CBM wells cannot be ignored. In the process of CBM production, the permeability of the reservoir will change significantly due to the strong influence of effective stress, matrix contraction and gas slippage (Cai et al., 2014). Scholars have established multiple permeability models to evaluate the dynamic changes of reservoir permeability. These models can be divided into three categories: stress-based models (Palmer and Mansoori, 1998a), strain-based models (Shi and Durucan, 2005a) and production data-based models (Lai et al., 2013). However, these models consider fewer factors and have limited applicability. Fine classification of CBM wells is the prerequisite for accurate analysis of production dynamics. Most previous studies have classified wells based on geostatic parameters, without quantifying the influence of dynamic characteristic parameters (dynamic changes in permeability and pressure) on well classification. Shang et al. classified CBM wells into four types according to gas production and analyzed the forming reasons. The results showed that coal structure was the basic factor affecting the productivity of CBM wells, and gas content and permeability were the direct geological factors controlling the production of CBM wells (Sang et al., 2009). Yi et al. classified CBM wells according to the type of reservoir pressure drop, and divided them into three types: rapid drop type, medium-term stable type and slow drop type. Among them, the rapid drop type is more favorable for high and stable production of CBM (Yi et al., 2019). Hou et al. classified CBM wells into four categories based on the static coal reservoir properties and the dynamic average daily gas production data of a single well, and systematically analyzed the gas production characteristics of each type of well (Hou et al., 2018). At present, most of the studies on CBM well classification do not consider dynamic parameters such as permeability and average reservoir pressure. In addition, the classification results of CBM wells are greatly affected by subjective factors. Therefore, it is the key point of the future CBM development evaluation in Block P to objectively classify the CBM wells according to the dynamic and static parameters, and comprehensively analyze the production characteristics and control factors of different types of CBM wells.

This study focuses on the No. 15 coal seam in the P block as a research subject, investigating the impact of various static factors on CBM production. Based on this, a neural network method is employed to finely classify CBM wells by considering both static and dynamic parameters. The production dynamics and productivity differences among different types of CBM wells are analyzed, providing CBM drainage optimization suggestions that offer technical support for future development of No. 15 coal seam.

The P block is situated in the southeastern part of the Qinshui Basin. The structural configuration within the block is characterized by simplicity, presenting a predominantly west-dipping monoclinic structure overall, with only two faults observed in the northeastern region (refer to Figure 1). Secondary folds are primarily developed within the block, exhibiting a general orientation along the north-south axis on a planar scale. These folds possess distinctive features including broad and gentle characteristics, symmetrical wings, and shallow dip angles. The fold structures in the east-west direction consist of Zhengcun anticline, Panhe syncline, Shigou anticline, Huojiashan syncline, Mashancun anticline, etc. The strata exhibit gentle inclination angles ranging from 5° to 8° on average, with a maximum dip angle of only 15°. Generally speaking, the synclines tend to have relatively high gas content compared to the anticlines.

In the P block, the primary coal-bearing strata comprise the Upper Carboniferous Taiyuan Formation and the Lower Permian Shanxi Formation (refer to Figure 2). The Taiyuan Formation predominantly accumulated in a marine sedimentary environment, consisting of sandstone, mudstone, coal, and limestone. Conversely, the Shanxi Formation primarily deposited in a terrestrial sedimentary setting composed of sandstone, mudstone, and coal. The coal seam within the block comprises the No. 3 coal seam of Shanxi Formation and the No. 15 coal seam of Taiyuan Formation, with a current focus on exploring and developing CBM in the latter. The No. 15 coal seam consists of sediments from barrier coast and delta systems. Furthermore, within the paleo-sedimentary environment, there is well-developed distributary channeling and inter-distributary bay formation, leading to bifurcation or even disappearance of certain coal seams within the block.

The No. 15 coal seam is buried at depths ranging from 350 m to 700 m, with an average burial depth of 510 m. The thickness of the coal seam remains consistent in the horizontal plane, varying between 2.68 m and 7.78 m, with an average thickness of 4.14 m. The coal body structure primarily consists of intact structural coal and fragmented coal, exhibiting a permeability range of 0.08 to 5.707 mD, with an average permeability value of 1.84 mD. Additionally, the raw coal exhibits an ash yield ranging from 9.15% to 15.54%, averaging at around 12.17%, indicating its classification as low and medium ash coal. The desorption test results indicate that No. 15 coal exhibits a high gas content, ranging from 12.57 to 27.71 m³/t, with an average of 20.91 m³/t, and the gas saturation exceeds 90%. Additionally, the reservoir pressure ranges from 1.82 to 2.93 MPa, averaging at 2.21 MPa. The overall development geological conditions of No. 15 coal are more suitable for the development of CBM.

The consideration of permeability as a parameter in dynamic analysis is crucial; however, it is impractical to measure the permeability of CBM wells in real-time during production. Therefore, this section initially derives the calculation model for determining the dynamic permeability of CBM wells. In the actual production process of CBM, reservoir pressure is closely correlated with permeability. Therefore, the dynamic characteristics of coal reservoir pressure are investigated through analysis of the material balance equation. The material balance equation for CBM can be expressed as follows: G p = A h ϕ i 1 − S w i B g i + ρ B A h V L p i p i + p L − A h ϕ i 1 − S w B g − ρ B A h V L p p + p L (1) where, G _p is cumulative gas production, m³; A is well control area, m²; h is coal seam thickness, m; ϕ i is initial porosity, %; S _wi is initial water saturation, %; B _gi is initial gas volume coefficient, m³/m³; ρ B is coal density, kg/m³; V _L is langmuir volume, m³/t; p _i is initial reservoir pressure, MPa; p _L is langmuir pressure, MPa; ϕ is porosity, %; S _w is water saturation, %; B _g is gas volume coefficient, m³/m³; p is reservoir pressure, MPa.

The material balance equation for water can be expressed as follows: A h ϕ S w B w = A h ϕ i S w i 1 + c w p i − p B w + W e − W p (2) where, W _B is gas volume coefficient, m³/m³; c_w is water compression coefficient, MPa⁻¹; W _e is water invasion volume, m³; W _p is cumulative water production, m³.

W_e = 0, Eq. 2 is simplified as: A h ϕ S w = A h ϕ i S w i 1 + c w p i − p − W p B w (3)

The porosity of coal exhibits a self-adjusting effect, which varies in response to the pressure exerted on the coal reservoir. According to the P&M model (Shi et al., 2021), at the current reservoir pressure, the ratio of porosity to initial porosity can be expressed as follows: ϕ ϕ i = 1 + c m ϕ i p − p i + ε l ϕ i K M − 1 p p + p L − p i p i + p L (4) where, c _m is P&M model coefficient, MPa⁻¹; ε _l is langmuir volumetric strain fitting parameter; K is volume modulus, MPa; M is unaxial modulus, MPa. c m = g M − K M + f − 1 γ (5) where, g is a correction factor between 0 and 1; f is a number factor between 0 and 1; γ is coal matrix compression coefficient, MPa⁻¹. M E = 1 − v 1 + v 1 − 2 v (6) K M = 1 3 1 + v 1 − v (7) where, E is young’s modulus, MPa; ν is poisson ratio.

Combine Eqs 4–7 to obtain the equation of porosity under the influence of effective stress and matrix shrinkage: ϕ = ϕ i + g 1 + v 1 − 2 v E 1 − v − 1 + v 3 1 − v + f − 1 γ p − p i + ε l 1 + v 3 1 − v − 1 p p + p L − p i p i + p L (8)

The iterative solution of Eqs 1, 3, 8 enables the calculation of coal reservoir pressure during the production process, thereby providing fundamental parameters for subsequent permeability calculations.

For an unsaturated coal reservoir, when the reservoir pressure exceeds the critical desorption pressure, CBM cannot be released from the coal matrix, resulting in pore saturation with coalbed water. This stage is characterized by single-phase water flow production. During this stage, the permeability of the coal reservoir is solely influenced by effective stress. The mathematical model describing changes in coal permeability during this stage is as follows (Chen et al., 2015): K = K I e − C f 1 + v 1 − v p i − p (9) where, K _i is initial permeability, mD; K is dynamic permeability of single-phase flow stage, mD.

When the reservoir pressure is below the critical desorption pressure, CBM is released from the coal matrix due to the decrease in pressure. The permeability of the reservoir is not only affected by effective stress but also by matrix shrinkage. According to the classical P&M model (Shi et al., 2021), the relationship between porosity and permeability can be expressed as follows: K = K c d ϕ P ϕ c d 3 (10) where, K _cd is the permeability at critical desorption pressure, mD; ϕ c d is the prosity at critical desorption pressure

Combine Eqs 8, 10 to obtain the equation of permeability under the influence of effective stress and matrix shrinkage: K n o w = K c d 1 + g 1 + v 1 − 2 v E 1 − v − 1 + v 3 1 − v + f − 1 γ ϕ i p − p i + ε l ϕ i 1 + v 3 1 − v − 1 p p + p L − p i p i + p L 3 (11) where, K _now is the permeability below critical desorption pressure, mD.

The permeability calculation method is as follows: the change in reservoir pressure over time during CBM well production can be determined using Eqs 1, 3, 8. Equation 9 is employed to calculate permeability when the reservoir pressure exceeds the critical desorption pressure, while Eq. 11 is utilized for calculating permeability when the reservoir pressure falls below the critical pressure.

(1) Utilizing static and dynamic parameters, the SOM neural network is employed for the classification of CBM wells. The CBM wells in P Block’s No. 15 coal seam are categorized into three types: exceptional wells, potential wells, and inefficient wells. The exceptional well exhibits favorable geological conditions and a rational drainage system, thereby fully realizing its gas production potential. Conversely, the potential well possesses advantageous geological conditions but lacks an optimal drainage system, resulting in an underdeveloped gas production potential. Implementing certain measures could potentially transform them into exceptional wells. Lastly, the inefficient well is characterized by unfavorable geological conditions and limited gas production capacity.