Sage Journals: Discover world-class research

Abstract

The Hancheng block in the southeastern Ordos Basin is one of the earliest and the most important areas for coalbed methane exploration and development in China. However, there are significant production variations in different wells or even some adjacent wells in the Hancheng block. To reveal the reasons of production differences in such a small scale, further detailed studies of coalbed methane productivity in the Hancheng pilot test area, a precursor trial area in Hancheng block with mature, well-characterized coalbed methane reservoirs and long-term production database, were conducted. The influence of nine factors (including engineering and geological factors) on gas production was analyzed. By introducing the rough set theory, which is applicable to the vague, imprecise, and incomplete information system, this paper presents a method for quantitative evaluation of the influencing factors on gas production. The results indicate that there are certain distribution characteristics of productivity in Hancheng pilot test area, which can be partitioned into four zones. The degressive order of the influencing degree of these nine factors is (i) the distance between the well and the fault, (ii) the structure curvature of the coal seams, (iii) the gas content, (iv) the critical reservoir ratio, (v) the volume of the fracturing liquids per meter, (vi) the volume of the fracturing sand per meter, (vii) the dynamic liquid level drop rate, (viii) the depth, and (ix) the thickness. Geological factors, especially the influence of fault, structural curvature of the coal seams and gas content, play a major role in controlling long-term gas production. Engineering factors (effect of fracturing and dynamic liquid level drop rate) have always been integral parts of coalbed methane development.

Keywords

Coalbed methane productivity characteristics the rough set Hancheng block southeastern Ordos Basin

Introduction

As a clean energy source and an important strategic complement to conventional hydrocarbon resources, coalbed methane (CBM) has achieved large-scale commercial development in the United States, Australia, and Canada (Al-Jabouri et al., 2009; American Association of Petroleum Geologists, 2015; Moore, 2012; Islam, 2015). China has abundant CBM resources (Yun et al., 2012). Since the beginning of this century, with major investments from the state and related enterprises, the CBM industry in China has developed rapidly. In 2000, the number of CBM wells drilled in China was less than 500, and it increased to 12,574 by the end of 2012 (Ye et al., 2013). However, China still faces some urgent problems in the development of CBM, for example, poor economic benefits due to low average productivity of individual wells (Fu, 2015; Zhu et al., 2015).

Many literatures have reported the CBM production characteristics and its influencing factors in the basin-scale. Kaiser et al. (1994) analyzed and contrasted the CBM gas systems, resources and production between the Sand Wash Basin and the San Juan Basin. Based on statistical analysis, Ayers (2002) revealed the key factors determining CBM resources and output in the Powder River Basin and the San Juan Basin. Su et al. (2005) concluded eight main parameters affecting CBM production in the Qinshui Basin. Qin et al. (2008) conducted a comprehensive analysis of production influencing factors of CBM wells in the Fuxin Basin in China. Gentzis et al. (2008) dissected the geological and engineering factors affecting CBM well performance in Alberta.

With the continuous development of CBM in China, however, it is found that CBM productivity varies significantly in area-scale or even in wells adjacent to (Qin et al., 2012; Ye and Lu, 2016). The research results from the basin-scale have important guiding significance in the exploration period of CBM, but it is becoming more and more difficult to meet the actual demand in the production stage. Therefore, detailed studies about CBM productivity on finer scale (i.e., block-scale or area-scale) are urgently needed. In recent years, minority researchers have conducted fine studies about CBM production and the main influencing factors in block-scale (Lv et al., 2012; Tao et al., 2014; Zhao et al., 2015). However, these studies predominantly focused on early (primarily 2–4 years) production characteristics and its influencing factors, while the producibility performance of CBM wells in different production stages vary greatly. In this study, a further detailed study of CBM producibility was carried out, using the Hancheng pilot test area (HPTA), a precursor trial area in Hancheng block with mature, well-characterized CBM reservoirs and long-term production data (primarily 7–8 years, which can reflect the regional production potential more accurately). The paper begins with a review of the geological setting of the studied area, and continues with a detailed analysis of CBM productivity subarea in HPTA, and then a spectrum of geological and engineering factors was extracted and discussed, followed by a quantitative evaluation of these affecting factors based on the rough set theory. The paper concludes with a comprehensive analysis of the key factors on long-term production performance in HPTA and the implications that are helpful in optimizing the reservoir in proven and emerging CBM plays around the world.

Geological setting

The Hancheng block is located in the Southeastern Ordos Basin, China (Figure 1(b)), with an area of 1120 km² (Chen et al., 2006). The total CBM reserves of the Hancheng block is 1.7 × 10¹² m³ (Ma and Yin, 2002). HPTA is located in the northeast of the Hancheng Block. The basic structural form in HPTA is a monoclinic, inclining to the Southwest, with some small fold growth (Figure 1(c)). There are several small-scale and high-angle faults in the study area, which could cut through the surface (Wang, 2002; Zhang, 2008).

Figure 1.

(a) Location map of the Ordos Basin. (b) Map of the Ordos Basin and the Hancheng block, and (c) Structural map of the coal seams in Hancheng pilot test area.

The Permian strata bearing with coal spread widely in the study area. The Taiyuan and Shanxi Formations in the lower Permian System are the two main coal-bearing strata (Sun et al., 2017), where the No. 3, No. 5, and No. 11 coal seams are the main targets for CBM development (Figure 2). The thickness of the three coal seams is 1–3.8, 1–6, and 2∼10.8 m, respectively. The coals in HPTA are mainly low volatile bituminous (R_o ranging from 1.6% to 1.9%). The coalbed gas content is generally between 2 and 14 m³/t. According to the well test results of parameter wells, the permeability of coal reservoirs in the Hancheng block is between 0.01 and 2.5 md.

Figure 2.

Stratigraphic column in the Hancheng block.

Data and methods

Data

The drilling data, logging data, core samples test data, fracturing data, and production data were collected from 70 CBM wells in HPTA, which belongs to the PetroChina Coalbed Methane Company Limited. These data were used to analyze the production characteristics and its key controlling factors of CBM wells.

A spectrum of structural, sedimentologic, geothermal, reservoir petrologic, hydrogeological, and engineering variables have been understood to affect CBM production (Ayers, 2002; Gentzis et al., 2008; Kaiser et al., 1994; Lv et al., 2012; Qin et al., 2008; Su et al., 2005; Tao et al., 2014; Zhao et al., 2015). In this study, the processes of influencing factors selection are as follows: (1) Listing all factors extensively that have potential impact on gas production. In HPTA, these factors include the structural conditions, the depositional setting, the hydrological conditions, the coal thickness, the gas content, the depth, the coal rank, the initial reservoir pressure, the critical desorption pressure, the porosity and permeability, the well selection, the well completion, the hydraulic fracturing, and dynamic liquid level drop rate. (2) Excluding the unavailable and obviously invalid parameters. In HPTA, the porosity and permeability tests are unavailable for most of 70 CBM wells, so they have to be excluded. Furthermore, the collected data show that the depositional and hydrological conditions, the coal rank change very little, and the well selection as well as the well completion are the same. Subsequently, nine geological and engineering factors (structural curvature of the coal seams (SCCS), effect of faults, burial depth, thickness, gas content, critical reservoir ratio (CRR), volume of the fracturing liquids per meter (VFLPM), volume of the fracturing sand per meter (VFSPM), and dynamic liquid level drop rate (DLLDR)) were analyzed in this study.

The rough set and its application

The rough set theory is a new mathematical method to describe the vague, imprecise, and incomplete information system (Pawlak, 1982, 1992; Zhang et al., 2001). Based on the classical set theory, the rough set theory defines an information system: I ={U, A, V, f}, where U is a set of finite objects, U={u₁, u₂,…, u_n}, A is a set of finite attributes, which include two parts: the condition attributes A_C ={a_C1, a_C2, …, a_Cm} and the decision attributes A_D= {a_D1, a_D1, …, a_Dl}, so A=A_C∪A_D, V is the ranges of the attributes, V={V_a|a∈A}, V_a is the range of an attribute a, f is a function from U × A to V, where n, m, and l represent the number of objects, condition attributes, and decision attributes, respectively. To expose the concepts, 20 CBM wells from all 70 CBM wells in HPTA were taken as an example. Table 1 is called the CBM production information system: I ={U, A, V, f}, where objects set: U={Well1, Well2, Well3, … Well19, Well20}, the condition attributes: A_C={CRR, VFLPM, Gas content, VFSPM, Thickness, Depth, SCCS, DWF, DLLDR}, the decision attribute: A_D= {Gas production rate}, V is the range of attributes, for example, V_Thickness=[2.5,15.9]. Taking Well 1 as an example, function f means that a well with the condition attributes: 0.79 (CRR), 12.94 m³/t (gas content), 6.0 m (thickness), 409 m (depth), −0.000897 m⁻¹ (SCCS), 2.9 km (distance between the well and the fault (DWF)), 130 m³/m (VFLPM), 10.2 m³/m (VFSPM), and 3.39 m/d (DLLDR), respectively, has obtained an average gas production rate of 2279 m³/d in more than 7 years.

Table 1.

The CBM production information system (examples of 20 CBM wells in all 70 CBM wells in HPTA).

Objects set (U)	Condition attributes (A_C)									Decision attribute(A_D)
Well	CRR	VFLPM(m³/m)	Gas content(m³/t)	VFSPM (m³/m)	Thickness(m)	Depth (m)	SCCS(m⁻¹)	DWF(km)	DLLDR (m/d)	Production(m³/d)
Well1	0.79	130	12.94	10.2	6.0	409	−8.97E-04	2.9	3.39	2279
Well2	0.45	159	9.25	12.0	7.5	751	−8.17E-04	3.5	2.79	1318
Well3	0.45	112	8.82	13.7	8.0	619	4.8E-05	2.1	2.08	937
Well4	0.70	132	8.72	10.2	7.0	523	−9.1E-05	3.3	3.98	2654
Well5	0.26	104	6.16	9.9	2.5	445	5.6E-05	2.5	11.72	451
Well6	0.23	125	4.22	7.0	14.5	456	−9E-06	0.5	3.03	215
Well7	0.74	149	12.11	10.5	5.5	400.5	7.65E-05	1.9	2.13	2008
Well8	0.35	77	5.91	11.4	6.9	488.5	0.001102	1.8	2.03	558
Well9	0.67	137	11.37	11.3	5.0	439.5	2.5E-05	1.7	1.32	1654
Well10	0.47	145	8.89	11.7	5.5	518	3.8E-05	2.1	3.94	1095
Well11	0	104	3.32	11.2	10.5	639.5	2.6E-05	0.8	—	0
Well12	0.58	141	3.45	12.4	15.0	484	3E-06	0.4	2.72	255
Well13	0	66	3.27	7.3	15.9	592	1.1E-05	0.5	—	0
Well14	0.44	98	7.18	8.5	8.0	668	−9.9E-05	1.6	1.23	1528
Well15	0.38	126	7.89	10.7	12.6	714.5	3.8E-05	0.9	3.91	725
Well16	0.71	113	11.80	7.1	5.0	422.5	−3.7E-05	1.2	1.63	1130
Well17	0.54	106	8.91	11.5	6.8	427	−0.001136	2.7	2.76	1412
Well18	0.41	125	5.92	9.5	11.4	458	−0.001267	2.3	3.11	1161
Well19	0.27	141	4.91	9.9	8.9	452	1.4E-05	0.6	7.89	245
Well20	0	58	5.70	7.8	4.5	516	0.000231	0.1	—	0

CBM: Coalbed methane; CRR: critical reservoir ratio; DLLDR: dynamic liquid level drop rate; DWF: distance between the well and the fault; HPTA: Hancheng pilot test area; SCCS: structure curvature of the coal seams; VFLPM: volume of the fracturing liquids per meter; VFSPM: volume of the fracturing sand per meter.

Each condition attribute a_C, to some extent, influences the result attribute a_D, defined as the attribute dependency degree: γ(a_C; a_D), where 0 ≤ γ(a_C; a_D) ≤ 1. If and only if γ(a_C; a_D) = 0, can it be said that a_C is independent of a_D. If and only if γ(a_C; a_D)=1, can it be said that a_D entirely depends on a_C. The larger γ(a_C; a_D) is, the greater the influence of a_C on a_C (Yamaguchi, 2009; Zhang et al., 2001). By introducing the rough set theory, this paper presents a method for quantitative evaluation of the influencing factors on gas production. The calculation of attribute dependency degree predominantly includes four steps (Appendix 1). The attributes classification should be determined firstly. In this study, the condition attributes consist of nine elements: CRR, VFLPM, gas content, VFSPM, thickness, depth, SCCS, DWF, and DLLDR, and the decision attribute is gas production rate. Then, the condition attributes needed to be normalized by equation (1). While the decision attribute has only one element, so it doesn’t need to be normalized. The third step was to discretize all attributes. Here, the condition attributes were discretized by equidistance discrete with intervals of 0.1, so the range of each condition attribute: V_C = {1, 2, 3, 4, 5, 6, 7, 8, 9, 10}, where 1 = [0, 0.1], 2 = (0.1, 0.2], 3 = (0.2, 0.3], 4 = (0.3, 0.4], 5 = (0.4, 0.5], 6 = (0.5, 0.6], 7 = (0.6, 0.7], 8 = (0.7, 0.8], 9 = (0.8, 0.9], 10 = (0.9, 1]. The decision attribute was discretized by equidistance discrete with intervals of 500, and the range of the decision attribute: V_D ={0, 1, 2, 3, 4, 5, 6}, where 0 = [0], 1 = (0, 500], 2 = (500, 1000], 3 = (1000, 1500], 4 = (1500, 2000], 5 = (2000, 2500], 6 = (2500, 3000]. Finally, with the help of the Rough Set Toolkit for Analysis of Data (The Linnaeus Centre for Bioinformatics, 2009), the single attribute or multiple attributes dependency degree could be calculated by the equations (2) to (9), and the calculation results of 70 CBM wells are shown in Table 2.

Table 2.

The result from the rough set theory analysis.

Condition attributes	Dependency degree γ
Single attribute
Thickness	0.4878
Depth	0.5095
DLLDR	0.5120
VFSPM	0.5126
VFLPM	0.5213
CRR	0.5348
Gas content	0.5357
SCCS	0.5547
DWF	0.5643
Multiple attributes
Depth, thickness	0.6044
VFLPM, VFSPM	0.6339
SCCS, DWF	0.7873
VFLPM, VFSPM, DLLDR	0.7800
SCCS, DWF, gas content	0.8446
SCCS, DWF, gas content, CRR	0.8685
SCCS, DWF, gas content, CRR, depth, thickness	0.8978
SCCS, DWF, gas content, CRR, depth, thickness, VFLPM, VFSPM, DLLDR	0.9481

CRR: critical reservoir ratio; DLLDR: dynamic liquid level drop rate; DWF: distance between the well and the fault; SCCS: structure curvature of the coal seams; VFLPM: volume of the fracturing liquids per meter; VFSPM: volume of the fracturing sand per meter.

Results and discussions

CBM productivity subarea in HPTA

The CBM production process is thought of as having three stages: (1) dewatering stage, (2) stable stage, and (3) decline stage, and these different stages have different production characteristics (Ayers, 2002; Clarkson, 2013). In more than 7 years of production, CBM wells have experienced a long period of the stable stage, and thus the production data can reflect the production potential more accurately.

The production performance of CBM wells in HPTA is obviously different. Of the 70 wells, 21 wells produce more than 1000 m³/day, 33 wells less than 1000 m³/day, while the other 16 wells cannot produce gas. Additionally, the wells can be divided into three categories: type I (≥1000 m³/d), type II (<1000 and >0 m³/d), type III (no gas production). Furthermore, HPTA can be partitioned into four zones (Figure 3): zone I, zone II, zone III₁, and zone III₂. Zone I is distributed in the central and eastern regions of HPTA, all wells in zone I have gas production rate exceeding 1000 m³/d, while the water production rate is between 1.1 and 5.3 m³/d. Zone II, which surrounds zone I, is an approximate annular region where majority of wells are stripper wells with an average gas production of 541 m³/d, while the water production rate is between 1.4 and 18.2 m³/d. Zone III₁, located around the northern faults, has no gas production. But the water production is relatively large, varying from 9.5 to 29.1 m³/d. Surrounding the southern small-scale fault, the wells in Zone III₂ also have a large water production, ranging from 4.5 to 24.9 m³/d.

Figure 3.

CBM productivity subarea in Hancheng pilot test area.

Geological factors affecting CBM well productivity

Structural curvature of the coal seams

Curvature is a parameter that reflects the curving degree of a curve or surface. In the viewpoint of mathematical mechanics, the distribution of fractures, which is affected by the stress field, can be predicted through the curvature of the structural planes (Bergbauer and Pollard, 2003; Lisle, 1994; Mandujano et al., 2005; Ouahed et al., 2005; Watkins et al., 2015). The larger the absolute value of the structural curvature, the greater the curving degree. Among all regions where the absolute value of structural curvature is large, the positive curvature region above the neutral plane and the negative curvature region under the neutral plane are under the environment of extension, and thus fractures generate and expand more easily in these regions (Figure 4; Shen et al., 2010; Suo et al., 2012). For the coal reservoir, the influence of the fracture (or cleat) system on permeability is significant (Close, 1993; Levine, 1996), so it is reliable to predict coal seam permeability by using the structure curvature method. In HPTA, most of the production wells lack well testing data, and thus SCCS (instead of permeability) was used to analyze the influence on gas production.

Figure 4.

Formation deformation and fracture dominant position (left), sketch of relation between fold and stress strain (right).

According to the calculation method in Appendix 2, SCCS was calculated and the structural curvature map of the coal seams in HPTA was plotted (Figure 5). SCCS in the study area ranges from −0.001635 to 0.0011 m⁻¹. A prominent area where the gas production rate is more than 1000 m³/d in the central of Figure 3 corresponds, in part, with a natural anomaly in Figure 5 where SCCS is a negative number with larger absolute value. The possible reason for this phenomenon is that the natural anomaly is also the region of extensional deformation, so fractures (or cleat) generate and expand easier in coal seams, guaranteeing larger permeability, and thus contributing to a better production performance.

Figure 5.

Structural curvature map of the coal seams in Hancheng pilot test area.

Fault

Faults have an important influence on hydrocarbon accumulation (Markowski, 1998; Xu et al., 2012). During the development of CBM, the impact of fault is clearly worth considering (Gentzis et al., 2008; Sang et al., 2009; Tao et al., 2014). In HPTA, the CBM wells in zone III₁ and zone III₂ always produce a large amount of water and no gas (Figure 3). The following reasons may account for this phenomenon: (1) as a channel, the fault connects the coal seams with the other aquifers. Water analysis data show that the Cl⁻ content of the coalbed water in the study area is generally 800–1300 mg/L, while the Cl⁻ content in the water produced from wells near faults is less than 500 mg/L, the lowest is only approximately 300 mg/L, which indicates that the shallow groundwater flows to the wells through faults (Shao et al., 2014). During the dewatering stage, the depressurization of the coal reservoir in the fault-affecting regions is difficult to realize, thus, the gas could not desorb from the coal matrix, let alone move by Darcy flow to the wellbore. (2) Water is relatively active in the fault area, and thus, gas in the coal seam escapes easily, resulting in a lower gas content, which also has a negative impact on gas production.

Depth

The burial depth of the coal seams (i.e., the middle buried depth) in HPTA ranges from 349 to 771 m. Statistics show that, among all the type I wells, 68% have coal seams buried shallower than 500 m, while only 32% of the wells are deeper than 500 m. Overall, the burial depth has no significant impact on gas production, because the distribution of points in Figure 6 is relatively scattered.

Figure 6.

Scatterplot of daily gas production and burial depth.

Thickness

Coal seam thickness is one of the important parameters for CBM resources evaluation. When other conditions keep the same, thicker coal seams are more productive (Pashin, 1991, 1997). According to the data of 70 wells, total thickness of the coal seams ranges from 2.5 to 15.9 m. Figure 7 show that the thickness of coal seams have no obvious effect on gas production. But it is to be observed that the wells with coal seams thicker than 10 m do not show high yield. This phenomenon, which also exists in the CBM exploitation of Qinshui Basin in China (Lv et al., 2012), is not consistent with the general perception. Compared to a single thin coal seam, the overlay of multiple coal seams with larger total thickness, enhances vertical reservoir heterogeneity (Jin et al., 2004), leading to a serious interlayer interference problem, and thus restricting the production potential of CBM wells (Huang et al., 2014; Qin et al., 2014; Yao et al., 2014).

Figure 7.

Scatterplot of daily gas production and thickness.

Gas content

The gas-bearing volume (gas content) is a material base and essential condition of desired gas production (Moore, 2012). The coalbed gas content in HPTA varies from 2 to 14 m³/t, and are much lower than those in Qinshui Basin (22–25 m³/t; Lv et al., 2012). Figure 8 shows a positive correlation between the gas content and gas production. All of the type I wells have coalbed gas content greater than 5 m³/t. Generally, in a certain productive section, higher gas content makes the CBM resources more enriched, which ensures high gas production.

Figure 8.

Scatterplot of daily gas production and gas content.

Critical reservoir ratio

Gas saturation is one of the key sources of coal reservoir heterogeneity (Bustin, 1997; Cui and Bustin, 2005; Pashin, 2010). To obtain gas saturation, the desorption testing determining gas content and the adsorption isotherm testing determining gas capacity should be taken first. However, most of the production wells in HPTA do not have the above testing data, hence, another related parameter—the critical desorption pressure to original formation pressure ratio, namely the critical reservoir ratio (CRR)—is used to analyze its influence on gas production. Figure 9 shows that CRR ranges from 0 to 0.79 based on the 70 data points, with 0 representing that wells do not produce gas. There is a positive correlation between CRR and gas production rates. Among all the CBM wells with CRR less than 0.4, none of them have a gas production rate greater than 1000 m³/d. However, 62.8% of the CBM wells with CRR larger than 0.4 have a gas production rate greater than 1000 m³/d.

Figure 9.

Scatterplot of daily gas production and critical reservoir ratio.

Engineering factors affecting CBM well productivity

Hydraulic fracturing

The coal reservoir in the Hancheng Block is characterized as low permeability (Yao et al., 2009). To gain commercial gas flow, reservoir stimulation is necessary (Johnson et al., 2002). Presently, hydraulic fracturing stimulation technology is widely used in HPTA. During the process of fracturing, enhancing the discharge capacity and increasing fracturing fluid volume are important measures to enlarge the length as well as the width of the fractures. Similarly, a larger quantity of sand is more conducive to expanding the scale of sand-packed fractures and their conductivity (Zhao et al., 2015). Therefore, VFLPM and VFSPM are used to characterize the fracture scale for lack of fracture monitoring data. Figures 10 and 11 are scatter maps of gas production rates and VFLPM, VFSPM. The maps show that VFLPM and VFSPM of type I wells are >80 and >7 m³/m, respectively. Overall, the effective reservoir stimulation is necessary for CBM wells in HPTA to achieve an adequate production rate.

Figure 10.

Scatterplot of daily gas production and volume of the fracturing liquids per meter.

Figure 11.

Scatterplot of daily gas production and volume of the fracturing sand per meter.

Dynamic liquid level drop rate during the dewatering stage with single-phase water flow

Compared with the conventional sandstone reservoir, the coal reservoir has a significant stress sensitivity (Li et al., 2013; Palmer and Mansoori, 1998; Somerton et al., 1975; Seidle et al., 1992). During the CBM production process (i.e., dewatering stage, stable stage, and decline stage), with the reduction of reservoir pressure, the damage of stress sensitivity to coal reservoir is inevitable, especially in the dewatering stage with single-phase water flow (Clarkson, 2013; Moore, 2012; Tang et al., 2015).

For the CBM wells in HPTA, the bottom-hole flowing pressure, which controls the depressurization of the coal reservoir, is controlled by the dynamic liquid level in the dewatering stage with a single-phase water flow. Hence, the dynamic liquid level drop rate (DLLDR) can reflect the decline rate in the coal seam pressure. The average DLLDR of the 70 wells in the dewatering stage with a single-phase water flow were collected (Figure 12, points (0, 0) representative a well only produce water all the time). Based on current data, DLLDR of type I wells in the dewatering stage are mostly less than 4 m/d, while the wells with unreasonable DLLDR always have low gas production rates. The reason for this phenomenon is that an unreasonable DLLDR in the dewatering stage could cause severe stress sensitivity of coal reservoir, reducing permeability, and has long-term bad effects on production performance.

Figure 12.

Scatterplot of daily gas production and the dynamic liquid level drop rate in the dewatering stage with single-phase water flow.

Comprehensive analysis

The results from the rough set theory analysis show that the degressive order of the influencing degree of these nine factors is DWF, SCCS, gas content, CRR, VFLPM, VFSPM, DLLDR, depth, and thickness. Geological factors play a major role in the control of long-term gas production, especially the effect of fault, structural curvature of the coal seams, and the gas content ( $γ (a_{D W F}, a_{S C C S}, a_{g a s c o n t e n t}; a_{D}) = 0.8446$ ). Engineering factors, such as reservoir fracturing and DLLDR, are indispensable in the development of CBM.

Faults play an obvious passive role for CBM accumulation and exploitation in HPTA. They create passages for gas to escape during the accumulation of CBM causing the coal seams to slide and rupture. Moreover, the coal reservoir adjacent to faults cannot be depressurized effectively, due to the continuous water supplement from other aquifer through faults. Therefore, the wells adjacent to faults, making up to 22.9% of the 70 selected wells, produced excessive water but low or no gas. Accordingly, the attribute dependency degree of DWF (0.5643) is the highest among the nine single attribute. SCCS is most likely to affect the heterogeneity of fracture development in coal seams (i.e., permeability), and thus has vital effects on gas production. Generally, gas content is the fundamental material in gas production, therefore, it is also an important factor.

In the early (primarily 2–4 years) production, CRR and hydraulic fracturing are the most significant factors to CBM production (Lv et al., 2012; Zhao et al., 2015). For long-term production performance in HPTA, however, the dependency degrees of CRR, VFLPM and VFSPM to gas production rate only rank 4–6 in the nine factors. The following reasons can account for this phenomenon: (1) CRR greatly affects the difficulty of the water drainage, because even a small degree of under-saturation can necessitate prolonged dewatering before a large reservoir volume can reach the critical desorption pressure (Lv et al., 2012). In addition, during the early production period, depressurization of coal seams mainly occurs in the fracturing affected areas near the wellbore. Consequently, the CRR and the hydraulic stimulation are the most prominent factors to early production behavior of CBM wells. (2) After a long period of production, the gas in the coal seams near the wellbore is increasingly exhausted, and depressurization extends from the fracturing affected coal seams to the original coal seams. Therefore, original permeability (i.e., the fracture (or cleat) development in original coal seams) and gas content is more important to long-term gas production than the CRR and the hydraulic fracturing. Moreover, the comprehensive dependency degree of the nine factors that influence gas production is 0.9481, which indicates that they cannot completely determine gas productivity due to other unconsidered factors and that the result obtained by the method of the rough set theory analysis are reasonable.

Conclusions

The long-term production data of 70 CBM wells in the HPTA demonstrate that production performance varies significantly in area-scale. There are certain distribution characteristics of productivity, and HPTA can be partitioned into four zones: zone I, zone II, zone III₁, and zone III₂.

The results from the rough set theory analysis show that the degressive order of the influencing degree of these nine factors on CBM production is (i) DWF, (ii) SCCS, (iii) the gas content, (iv) CRR, (v) VFLPM, (vi) VFSPM, (vii) DLLDR, (viii) the depth, and (ix) the thickness. Geological factors, especially faults, SCCS and gas content, play a major role in controlling long-term gas production. While engineering factors, such as effective reservoir fracturing and reasonable DLLDR, are indispensable in the development of CBM.

The CBM wells in HPTA have experienced long-term production, and thus the regional productivity could be reflected more accurately. The high-yield CBM wells (>1000 m³/d) should meet the following conditions: keeping away from the faults; drilling wells in predominant vadose zones (especially in the zones where SCCS are negative numbers with larger absolute values); gas contents greater than 5 m³/t; CRR larger than 0.4; VFLPM of more than 80 m³/m; VFSPM of more than 7 m³/m; DLLDR lower than 4 m/d. While the buried depth and the thickness of coal seams have no obvious effect on gas production.

Footnotes

Acknowledgements

The authors appreciate the constructive comments and suggestions from the reviewers and the editors.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research,authorship,and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research,authorship,and/or publication of this article: This work was ﬁnancially supported by the National Natural Science Foundation Project (grant no. 41530314),the Key Project of the National Science and Technology (grant no. 2016ZX05042-002),the Public Welfare Foundation Project of China Geological Survey (grant no. DD20179362),and the Fundamental Research Funds for the Central Universities (grant no. 2652017320).

References

Al-Jabouri

Johnston

Boyer

et al . (2009) Coalbed methane: Clean energy for the world. Ilxir Publishing 21(2): 4–13.

American Association of Petroleum Geologists (2015) Unconventional energy resources: 2015 review. Natural Resources Research 24(4): 443–508.

Ayers

WBJ

(2002) Coalbed gas systems, resources, and production and a review of contrasting cases from the San Juan and Powder River basins. AAPG Bulletin 86(11): 1853–1890.

Bergbauer

Pollard

(2003) How to calculate normal curvatures of sampled geological surfaces. Journal of Structural Geology 25: 277–289.

Bustin

(1997) Importance of fabric and composition on the stress sensitivity of permeability in some coals, northern Sydney basin, Australia: Relevance to coalbed methane exploitation. AAPG Bulletin 81(11): 1894–1908.

Chen

Wang

Song

(2006) Evaluating the condition of CBM reservoir, Hancheng area. Natural Gas Geoscience 17(6): 868–870 (in Chinese).

Clarkson

(2013) Production data analysis of unconventional gas wells: Review of theory and best practices. International Journal of Coal Geology 109–110(2): 101–146.

(1993) Natural Fractures in Coal: Chapter 5. 180: 119–132.

Cui

Bustin

(2005) Volumetric strain associated with methane desorption and its impact on coalbed gas production from deep coal seams. AAPG Bulletin 89(89): 1181–1202.

10.

(2015) Legal or regulatory barriers to the clean development and utilization of China’s coal-bed methane and policy suggestions. Natural Resource Economics of China. 69–72 (in Chinese).

11.

Gentzis

Goodarzi

Cheung

et al . (2008) Coalbed methane producibility from the Mannville coals in Alberta, Canada: A comparison of two areas. International Journal of Coal Geology 74(3): 237–249.

12.

Huang

Sang

Miao

et al . (2014) Drainage control of single vertical well with multi-hydraulic fracturing layers for coalbed methane development. Journal of China Coal Society 39(S2): 422–431 (in Chinese).

13.

Islam

(2015) Chapter 2 – World gas reserve and the role of unconventional gas. In: Unconventional Gas Reservoirs. Amsterdam: Elsevier Inc, pp.9–69.

14.

Jin

Zhang

Zhao

et al . (2004) Control of Late Carboniferous-Early Permian sea level fluctuation on coal reservoir heterogeneity, Taiyuan, Shanxi. Petroleum Exploration & Development 5(5): 44–49.

15.

Johnson

Thomas

Campagna

(2002) Improving results of coalbed methane development strategies by integrating geomechanics and hydraulic fracturing technologies. SPE Asia Pacific Oil and Gas Conference and Exhibition 77824: 8–10.

16.

Kaiser

Hamilton

Scott

et al . (1994) Geological and hydrological controls on the producibility of coalbed methane. Journal of the Geological Society 151(3): 417–420.

17.

Levine

(1996) Model study of the influence of matrix shrinkage on absolute permeability of coal bed reservoirs. Geological Society London Special Publications 109(1): 197–212.

18.

Tang

Pan

et al . (2013) Characterization of the stress sensitivity of pores for different rank coals by nuclear magnetic resonance. Fuel 111(3): 746–754.

19.

Lisle

(1994) Detection of zones of abnormal strains in structures using Gaussian curvature analysis. AAPG Bulletin 78(12): 1811–1819.

20.

Tang

et al . (2012) Production characteristics and the key factors in high-rank coalbed methane fields: A case study on the Fanzhuang Block, Southern Qinshui Basin, China. International Journal of Coal Geology 96–97(4): 93–108.

21.

Yin

(2002) Analysis of major factor influencing the coalbed methane in Hancheng area. Journal of Xi’an University of Science and Technology 22: 162–165 (in Chinese).

22.

Mandujano

Khachaturov

Tolson

et al . (2005) Curvature analysis applied to the Cantarell structure, southern Gulf of Mexico: Implications for hydrocarbon exploration. Computers & Geosciences 31(5): 641–647.

23.

Markowski

(1998) Coalbed methane resource potential and current prospects in Pennsylvania. International Journal of Coal Geology 38(1–2): 137–159.

24.

Moore

(2012) Coalbed methane: A review. International Journal of Coal Geology 101(6): 36–81.

25.

Ouahed

AKE

Tiab

Mazouzi

(2005) Application of artificial intelligence to characterize naturally fractured zones in Hassi Messaoud Oil Field, Algeria. Journal of Petroleum Science & Engineering 49(3): 122–141.

26.

Palmer

Mansoori

(1998) How permeability depends on stress and pore pressure in coalbeds: A new model. SPE Reservoir Evaluation & Engineering 1(6): 539–544.

27.

Pashin

(1991) Regional analysis of the Black Creek-Cobb coalbed methane target interval, Black Warrior basin, Alabama. Bulletin/Geological Survey of Alabama: Geological Survey of Alabama, Tuscaloosa, Alabama, p.127.

28.

Pashin

(1997) Productivity of coalbed methane wells in Alabama. In: International coalbed methane symposium, The University of Alabama, Tuscaloosa, Alabama. pp.65–74.

29.

Pashin

(2010) Variable gas saturation in coalbed methane reservoirs of the Black Warrior Basin: Implications for exploration and production. International Journal of Coal Geology 82(3): 135–146.

30.

Pawlak

(1982) Rough sets. International Journal of Computer & Information Sciences 11: 341–356.

31.

Pawlak

(1992) Rough Sets: Theoretical Aspects of Reasoning about Data. Boston: Kluwer Academic Publishers.

32.

Qin

Zhou

Chen

(2008) Key influencing factors of productivity of coalbed methane in wells of Liujia area Fuxin Basin. Journal of Liaoning Technical University 27: 28–30 (in Chinese).

33.

Qin

Yuan

et al . (2012) Status and development orientation of coal bed methane exploration and development technology in China. Coal Science and Technology 40: 1–6 (in Chinese).

34.

Qin

Zhang

Bai

et al . (2014) Source apportionment of produced-water and feasibility discrimination of commingling CBM production from wells in Southern Qinshui Basin. Journal of China Coal Society 39(9): 1892–1898 (in Chinese).

35.

Sang

Liu

et al . (2009) Geological controls over coal-bed methane well production in southern Qinshui basin. Procedia Earth & Planetary Science 1(1): 917–922.

36.

Seidle

Jeansonne

Erickson

(1992) Application of Matchstick geometry to stress dependent permeability in coals. SPE Rocky Mountain Regional Meeting 433–444.

37.

Shao

Dong

Tang

et al . (2014) Treatment technology and method of low-to-moderate production coalbed methane wells in Hancheng mining area. Natural Gas Geoscience 25(3): 435–443 (in Chinese).

38.

Shen

Qin

et al . (2010) Application of structural curvature of coalbed floor on CBM development in the Zaoyuan block of the southern Qinshui Basin. International Journal of Mining Science and Technology 20(6): 886–890.

39.

Somerton

Söylemezoḡlu

Dudley

(1975) Effect of stress on permeability of coal. International Journal of Rock Mechanics & Mining Sciences & Geomechanics Abstracts 12(5–6): 129–145.

40.

Lin

Liu

et al . (2005) Geology of coalbed methane reservoirs in the Southeast Qinshui Basin of China. International Journal of Coal Geology 62(4): 197–210.

41.

Sun

Zhao

Püttmann

et al . (2017) Evidence of widespread wildfires in a coal seam from the middle Permian of the North China Basin. Lithosphere 9(4): 595–608.

42.

Suo

Peng

Chang

et al . (2012) A new calculating method of the curvature to predicting the reservoir fractures. Procedia Environmental Sciences 12(Part A): 576–582.

43.

Tang

Deng

Meng

et al . (2015) Characteristics and control mechanisms of coalbed permeability change in various gas production stages. Petroleum Science 12(4): 684–691.

44.

Tao

Tang

et al . (2014) Factors controlling high-yield coalbed methane vertical wells in the Fanzhuang Block, Southern Qinshui Basin. International Journal of Coal Geology 134–135: 38–45.

45.

The Linnaeus Centre for Bioinformatics (2009) Rosetta, a toolkit for analyzing tabular data within the framework of rough set theory. Sweden: Uppsala University. Available at: www.lcb.uu.se/tools/rosetta

46.

Wang

(2002) Geologic structure control of CBM in Hancheng mining area. Coal Geology and Exploration 30(1): 21–25 (in Chinese).

47.

Watkins

Butler

RWH

Bond

et al . (2015) Influence of structural position on fracture networks in the Torridon Group, Achnashellach fold and thrust belt, NW Scotland. Journal of Structural Geology 74: 64–80.

48.

Tang

Liu

et al . (2012) Study on coalbed methane accumulation characteristics and favorable areas in the Binchang area, southwestern Ordos Basin, China. International Journal of Coal Geology 95(2): 1–11.

49.

Yamaguchi

(2009) Attribute dependency functions considering data efficiency. International Journal of Approximate Reasoning 51(1): 89–98.

50.

Yao

(2014) Study on numerical simulation of interferences between stratum interfaces of coal bed methane well in Panzhuang block. Coal Engineering 172(1): 165–172 (in Chinese).

51.

Yao

Liu

Tang

et al . (2009) Preliminary evaluation of the coalbed methane production potential and its geological controls in the Weibei Coalfield, Southeastern Ordos Basin, China. International Journal of Coal Geology 78(1): 1–15.

52.

(2013) Technology and Industrialization of Coalbed Methane Exploration and Development in China. Beijing: Geological Publishing House (in Chinese).

53.

(2016) Development status and technical progress of China coalbed methane industry. Coal Science and Technology 44: 24–28 (in Chinese).

54.

Yun

Liu

et al . (2012) New progress and future prospects of CBM exploration and development in China. International Journal of Mining Science and Technology 22(3): 363–369.

55.

Zhang

(2008) Analysis method of coal bed methane reservoiring conditions: Case study in Hancheng area. China Coalbed Methane 5: 12–16 (in Chinese).

56.

Zhang

Liang

et al . (2001) Rough Set Theory and Method. Beijing: Science Press (in Chinese).

57.

Zhao

Tang

et al . (2015) High production indexes and the key factors in coalbed methane production: A case in the Hancheng block, southeastern Ordos Basin, China. Journal of Petroleum Science & Engineering 130(7): 55–67.

58.

Zhu

Zuo

Yang

(2015) How to solve the technical problems in CBM development: A case study of a CBM gas reservoir in the southern Qinshui Basin. Natural Gas Industry B 2(2–3): 277–281 (in Chinese).

Productivity subarea of CBM field and its key controlling factors: A case study in the Hancheng pilot test area,southeastern Ordos Basin,China

Abstract

Keywords

Introduction

Geological setting

Data and methods

Data

The rough set and its application

Results and discussions

CBM productivity subarea in HPTA

Geological factors affecting CBM well productivity

Structural curvature of the coal seams

Fault

Depth

Thickness

Gas content

Critical reservoir ratio

Engineering factors affecting CBM well productivity

Hydraulic fracturing

Dynamic liquid level drop rate during the dewatering stage with single-phase water flow

Comprehensive analysis

Conclusions

Footnotes

Acknowledgements

Declaration of conflicting interests

Funding

References