Fuel 139 (2015) 257–267

Contents lists available at ScienceDirect

Fuel

journal homepage: www.elsevier .com/locate / fuel

Evaluation of coalbed methane potential of different reservoirsin western Guizhou and eastern Yunnan, China

http://dx.doi.org/10.1016/j.fuel.2014.08.0540016-2361/� 2014 Elsevier Ltd. All rights reserved.

⇑ Corresponding author at: Coal Reservoir Laboratory of National EngineeringResearch Center of Coalbed Methane Development & Utilization, China Universityof Geosciences, Beijing 100083, PR China. Tel.: +86 10 8232 2011; fax: +86 10 82326850.

E-mail address: lisong8585@gmail.com (S. Li).

Song Li a,b,⇑, Dazhen Tang a, Zhejun Pan b, Hao Xu a, Lele Guo a

a Coal Reservoir Laboratory of National Engineering Research Center of Coalbed Methane Development & Utilization, China University of Geosciences, Beijing 100083, PR Chinab CSIRO Earth Science and Resource Engineering, Clayton, Victoria 3168, Australia

h i g h l i g h t s

� Coal reservoir properties are characterized for coals in the study area.� CBM potential was evaluated using fuzzy optimization model.� Coal rank in the study area is a key control for CBM development potential.

a r t i c l e i n f o

Article history:Received 15 May 2014Received in revised form 31 July 2014Accepted 20 August 2014Available online 2 September 2014

Keywords:Coal rankAdsorption capacityPermeabilityPore structure

a b s t r a c t

The coal and coalbed methane (CBM) resources are abundant in western Guizhou and eastern Yunnan,South China. However, commercial CBM production in this region has not been achieved. Reservoir prop-erties are the prerequisites in determining the possibility of CBM exploration and its development poten-tial. Thus, to help to select the most favorable block and to prioritize CBM development in the study area,a comprehensive program of experimental work has been carried out to study the physical properties ofcoal reservoirs in different blocks. Experimental results show that the properties of coal reservoir changesignificantly with respect to coal rank, and the coal rank in the study area is very uneven, with the vitr-inite reflectance (Ro) of coal samples ranging from 0.68% to 3.31%. In detail, low rank coals have welldeveloped seepage pores but undeveloped adsorption pores, resulting in the low adsorption capacityand high porosity and permeability. With burial depth increase, the metamorphic degree and compactiondegree of coals grow accordingly, as a result pores and fractures are gradually closed under stress, leadingto a sharp reduction of porosity and permeability in medium rank coals. However, most of the high rankcoals in the study area have experienced a large number of tectonic thermal events, which not onlyincreased the metamorphism degree and adsorption capacity, but also improved the porosity and perme-ability of the coal reservoirs. Based on experimental results, the CBM potentials of coal reservoirs in dif-ferent blocks in the study area were evaluated using multi-objective and multi-level fuzzy optimizationmodel, and the most prospective zones for CBM production were suggested.

� 2014 Elsevier Ltd. All rights reserved.

1. Introduction Basin where the commercial breakthrough of CBM production

Now China attaches an unprecedented importance to the explo-ration and development of coalbed methane (CBM) due to coalmining safety, greenhouse gas emissions and demand for naturalgas [1,2]. However, the CBM development activities in China havean extremely uneven geographical distribution. Most of them areconcentrated in North China, especially in the Southern Qinshui

has been achieved [3–5]. Western Guizhou and eastern Yunnan,the largest coal-producing region in South China, is rich in coaland CBM resources, and it has been considered to have high poten-tial for CBM recovery [6,7]. Nevertheless, the CBM exploration anddevelopment in this region has not reached any commercial break-throughs. One of the main reasons is the lack of understanding ofthe coal reservoir heterogeneity that resulted from the complexgeological conditions in the study area. Previous studies mainlyfocused on sedimentary characteristics, tectonic evolution andgeological conditions [8–10], however, the maturity evolution ofcoals and the variety in coal reservoir properties in the study areahave not been well reported.

Nomenclature

A, d ash (dry basis)CI consistency indexCR random consistency ratioE exiniteFC, ad fixed carbon (air-dried basis)I inertiniteL length of the core sampleM mineralsM, ad moisture (air-dried basis)n order of the matrixr radius of the core sampleRI random consistency indexRo vitrinite reflectance under oil immersionSBET BET pore surface area

ST area of T2 distributionT2 transverse relaxation timeTE echo spacingTW waiting timeV vitriniteVA actual valueVB bulk volumeVBJH BJH total pore volumeVN normalized valueV, daf volatile (dry, ash free basis)W eigenvectorkmax maximum eigenvalues

258 S. Li et al. / Fuel 139 (2015) 257–267

With different coal ranks, the CBM generation and accumula-tion processes, as well as coal reservoir physical properties varysignificantly [11–13]. It is worth mentioning that the differencesin coal reservoirs’ physical properties will affect the whole processfrom CBM accumulation to production. Consequently, evaluationof coal reservoir physical properties is the prerequisite for thedetermination of the CBM exploration and development potential[14–16]. Based on geological setting analyses of western Guizhouand eastern Yunnan, this research is focused on studying theeffects of metamorphism degree on coal reservoir physicalproperties, and proposing an optimized method using coal reser-voir physical properties to evaluate the CBM potential by usingthe multi-objective and multi-level fuzzy optimization model.

2. Geological setting

Western Guizhou and eastern Yunnan, located in the west of theYangtze Block, covers an area of approximately 2.58 � 104 km2. TheCBM resource is estimated at 2.2–2.75 � 1012 m3 [9]. The maincoal-bearing strata in the study area are Longtan Formation andChangxing Formation of Upper Permian, which were developed innonmarine, marine–terrestrial transitional and shallow marineenvironments from west to east [17]. After coal seams depositedin the Late Permian epoch, the area studied in this work underwentthe Indosinian, Yanshannian and Himalayan orogeny movements.Not only did the tectonic movements generated a large number offolds and faults but also controlled the distribution patterns ofthe CBM reservoirs [6,10]. Now the main coal-bearing formationsare developed in almost 30 sedimentary basins that with complexanticline and syncline structures, in which, Zhina, Gemudi, Pan-guan, Enhong and Laochang areas are the target CBM blocks [9].

The evolution degree of coals in western Guizhou and easternYunnan fluctuates from bituminous coal to anthracite coal. In addi-tion, the coal rank is high in northeast and southwest, but low inthe centre of the study area (Fig. 1). Coal metamorphism in thestudy area is dominated by the burial history besides later tectonicthermal events. The coalification in the study area can be dividedinto two stages: the pre-Yanshan period and the Yanshan period.In the pre-Yanshan period, the coalification was mainly evokedby burial metamorphism; while in the Yanshan period, the coalifi-cation was primarily caused by tectonic thermal metamorphism.Resulted from the various burial history or tectonic thermal events,different evolution process and physical–chemical structure ofcoals generated, which in return determined the variation in phys-ical properties of coal reservoirs in western Guizhou and easternYunnan [7].

3. Experiments and methods

3.1. Samples collection

A total number of 32 coal samples were collected from under-ground coal mines in western Guizhou and eastern Yunnan. Thesesamples have basically covered 5 key target CBM production areas,including 8 in Zhina block, 6 in Gemudi block, 7 in Panguan block, 6in Enhong block and 5 in Laochang block. Moreover, the collectedsamples also fully account for the main coal seams in the study area.

3.2. Experimental work

To characterize the physical properties of coal reservoirs in wes-tern Guizhou and eastern Yunnan, three sets of experiments wereperformed.

Set-I experiments were the tests including the vitrinite reflec-tance (Ro) measurement, coal maceral composition test and proxi-mate analysis. Ro measurement and coal maceral analyses werecarried out on the same polished specimen using the Leitz MPV-3 photometer microscope with the MPS 60 photo system manufac-tured by Leitz Company of Germany (following China NationalStandards GB/T 6948-1998 and GB/T 8899-1998, respectively).Proximate analysis (following China National Standards GB/T212-2001) was performed for all samples to measure the ash con-tent, moisture content and volatile material content of the coals.

To investigate the difference among the adsorption pore struc-ture and CBM storage capacities of coal reservoirs with differentcoal ranks, Set-II experiments (including low-temperature N2 iso-therm adsorption/desorption analysis and methane isothermaladsorption measurement) were performed. The low-temperatureN2 isotherm adsorption/desorption analysis was carried out witha modified Micromeritics ASAP-2000 automated surface area ana-lyzer following the Chinese Oil and Gas Industry Standard SY/T6154-1995. To carry out the low-temperature N2 isotherm adsorp-tion/desorption experiment, all coal samples were crushed andsieved to the size ranging between 0.23 and 0.45 mm, and thendried at 105 �C for 24 h in a vacuum oven. Each sample was mea-sured under the relative pressure from 0.01 to 1 to obtain the BETpore surface area (SBET), the BJH total pore volume (VBJH), and thepore size distribution. With the intention of analyzing the adsorp-tion capacities of coals, methane adsorption isotherm experimentswere performed following the China National Standards GB/T19,560-2004. Coal samples were crushed and sieved to obtainthe particle size of 0.18–0.25 mm, weighing up 100–125 gfor moisture-equilibrium treatment. The moisture-equilibrium

Fig. 1. The evolution degree of coal in western Guizhou and eastern Yunnan.

S. Li et al. / Fuel 139 (2015) 257–267 259

treatment was lasted for at least four days for each sample till thesample reached the moisture equilibrium condition. Then coalsamples were placed into the sample cell of the Terratek IS-100adsorption rig for isotherm measurement. The temperature andmaximum equilibrium pressure were 30 �C and 10 MPa, respectively.

Set-III experiments, including nuclear magnetic resonance(NMR) tests, porosity and permeability analyses, was used todescribe the seepage capacities of coal reservoirs. Before theseexperiments, each coal sample was cut parallel to the bedding sur-face for a horizontal cylindrical core about 25 mm in diameter and30 mm in length. The NMR measurements in this research used theRec Core 2500 instrument with a resonance frequency of 2.38 MHz.And the NMR measurement is in the condition of a very low exter-nal magnetic field. There are two reasons: first, coal is a weak mag-netic substance, thus, in the very low magnetic field condition, theparamagnetic minerals within the coals cannot influence the mea-sured results; second, with this low resonance frequency, the mag-netic information of solid state proton (13C and 1H) can be shielded[18]. Thus, the change of relaxation time that resulted from thecoal samples itself can be ignored. In the NMR experiments, 100%water-saturated coal samples were used to obtain the T2 spectrumdistribution that corresponding to the connected pores and frac-tures in coals. The main NMR measurement parameters includethe echo spacing (TE) of 0.6 ms, the waiting time (TW) of 5 s, theecho numbers of 2048, the scanning numbers of 64, and the envi-ronment temperature of 25 �C. The porosity was measured using agravimetric method. The core samples were initially dried at 80 �Cfor at least 24 h in a vacuum oven, then the weight were measuredand the bulk volume were calculated using:

VB ¼ pr2L ð1Þ

In which VB is the bulk volume, r is the radius of the core sam-ple, and L is the length of the core sample.

Hereafter, dry densities of the samples were calculated. After-wards, these coal samples were saturated with 100% distilledwater for at least another 24 h, and bulk volume and weight wereremeasured again, and wet densities were calculated. Finally, theporosity was calculated using dry densities and wet densities.

The permeability was measured using the ULTRA-PERMTM 200permeameter following the Chinese Oil and Gas Industry Standard(SY/T) 5336–1996. In detail, coal permeability to air was measuredunder simulated in-situ stress conditions, using a tri-axial cell (iso-tropic ambient pressure of 2.5 MPa), and the permeability wasdetermined by flowing air through the core sample until the vari-ation of permeability becomes relatively stable.

3.3. Optimization methods

Coal reservoir is a typical complex system because its propertiesare controlled by many factors. The multi-objective and multi-levelfuzzy optimization model is a useful tool for multiple-factors sort-ing of complex systems, and it has been successfully applied inmany areas [19,20]. This model combines the analytic hierarchyprocess and the fuzzy comprehensive evaluation; the former is uti-lized to calculate the weight of various parameters and the latter isused to obtain a comprehensive evaluation of all the elements.

The multi-objective and multi-level fuzzy optimization model isaimed at achieving the ultimate goal by means of integrating var-ious factors, and the procedure can be divided into the followingthree steps:

(1) The ultimate goal was divided into several levels more easilycomprehended sub-goals which can be analyzed indepen-dently. Using Um to represent the first level evaluation index,and thus the set of the first level evaluation index isU = {U1, U2, � � �, Um}; the second level evaluation index is rep-resented as Uij(i = 1, 2, � � �, m; j = 1, 2, � � �, n), and the set of thesecond level evaluation index is Ui = {Ui1, Ui2, � � �, Uim}. Con-structed matrix Ui based on the normalized second levelevaluation index:

2 3

Ui ¼

Ui11 Ui12 � � � Ui1n

Ui21 Ui22 � � � Ui2n

..

. ... . .

. ...

Uim1 Uim2 � � � Uimn

66664

77775; 0 6 Uimn 6 1 ð2Þ

260 S. Li et al. / Fuel 139 (2015) 257–267

(2) Once the hierarchy is built, the analytic hierarchy process isused to evaluate its elements by comparing all the elementsin a given level to one another with respect to their impacton the elements of the level above [21,22]. Then, weightsfor each of the decision alternatives are calculated. Theweight of the first level evaluation index is W = {W1, W2,� � �Wm}, and the weight of the second level evaluation indexis Wi = {Wi1, Wi2, � � �Wim}.

(3) Fuzzy evaluation result of the second level evaluation indexis:

2 3

Table 1Basic te

Targ

Zhin

Gem

Pang

Enho

Laoc

V: vitrinbasis).Ro data

Bi ¼Wi�Ui¼ðWi1;Wi2; � � �WimÞ�

Ui11 Ui12 � � � Ui1n

Ui21 Ui22 � � � Ui2n

..

. ... . .

. ...

Uim1 Uim2 � � � Uimn

66664

77775

¼ðBi1;Bi2; � � �BimÞð3Þ

All above is the basic principle of single-level fuzzy evaluation.Following the above idea, multi-level fuzzy evaluation can be real-ized by transforming level by level. The calculation of the fuzzyevaluation starts from the lowest level, and the evaluation matrixof a certain level index is constituted by the level below it untilthe recursive computation arrive the target layer. The final result

sting data of coals in western Guizhou and eastern Yunnan.

et CBM blocks Sample no. Ro (%) Coal maceral compositio

V I

a area HJG* 1.64 76.7 21.5HF* 2.03 35.6 62.4NF* 2.51 92.4 6.0FX* 2.54 74.2 25.4QS 2.64 88.3 9.3DSY 3.03 73.4 26.2LS 3.14 85.1 14.8ZH 3.31 87.4 11.0

udi area DWZ* 1.16 85.1 12.5CJG 1.41 73.1 17.5SYK* 1.52 87.6 11.8WTP* 1.54 88.2 11.6HC* 1.55 73.9 24.2YS 1.89 51.8 16.6

uan area HP 0.68 48.8 20.7LZD 0.71 54.0 20.6SJS* 0.81 53.4 14.4XLQ 0.87 68.8 19.5LWJ* 0.89 74.4 16.3HPW* 1.23 93.1 6.4DL* 1.27 89.6 10.2

ng area ZZS* 0.88 74.4 6.8HD* 1.22 81.2 16.3HLW* 1.26 88.8 8.6DAZ 1.37 76.2 8.1JA* 1.38 90.2 4.3XDW* 1.39 84.2 15.5

hang area LL* 1.48 54.5 16.1AG* 1.65 92.4 7.4JL 1.75 54.4 32.1LX* 2.04 66.2 30.7AGL 2.32 87.6 4.5

ite; I: inertinite; E: exinite; M: minerals; M, ad: moisture (air-dried basis); A, d:

of coal samples with symbol "*" cited from [38].

of the fuzzy evaluation can be expressed as: B = W � U = (B1, B2,� � �, Bm). Then, the goal corresponds to B

0= max(B1, B2, � � �, Bm) is

the optimal goal.

4. Experimental results

4.1. Composition of coals with different coal ranks

The coal samples in western Guizhou and eastern Yunnan rangefrom high volatile bituminous to anthracites, and their Ro areamong 0.68–3.31% (Table 1). Coal is formed from the remains ofvegetation, and variant vegetation or different parts of the vegeta-tion will form different macerals during coalification. Moreover,coal metamorphism is also one of the major factors which leadto the differences in macerals [23]. Results of coal maceral compo-sition test show that: contents of vitrinite, inertinite and mineralshave no trend with the increase of coal rank; however, the exinitecontent reduces drastically until it finally disappears after the Ro

higher than 1%. In the low metamorphic stage, due to shallow bur-ial depth, the temperature is not high enough to decompose theexinite in higher plants such as the reproductive organs, barks,and secretions, leaving the enriched exinite in low rank coals. Withthe increase of coal metamorphism, temperature becomes higher,and thus the exinite is decomposed and produces large amountsof hydrocarbon gases, resulting in the disappearance of exinite inhigh-rank coals.

n (%) Proximate analysis (%)

E M M, ad A, d V, daf FC, ad

– 1.8 1.22 15.43 6.89 76.73– 2.0 1.09 19.21 7.22 72.77– 1.6 1.06 12.98 5.84 80.32– 0.4 1.18 11.90 6.15 80.98– 2.4 – – – –– 0.4 0.18 6.71 9.63 84.15– 0.1 0.34 11.13 7.42 82.00– 1.6 1.24 10.37 5.63 83.54

– 2.4 0.38 16.58 19.67 63.50– 9.4 0.14 8.99 24.82 68.32– 0.6 0.54 11.33 16.00 72.28– 0.2 0.42 31.07 16.77 51.94– 1.9 0.34 23.44 16.94 59.42– 31.6 0.20 16.97 20.38 65.97

17.9 12.6 0.28 16.12 36.28 53.3014.6 10.8 0.28 9.27 30.75 62.6631.6 0.6 0.56 12.83 31.78 55.08

5.8 5.9 – – – –6.9 2.4 0.62 19.25 26.28 54.13– 0.5 0.45 10.08 20.31 69.30– 0.2 0.50 7.46 16.92 75.24

18.4 0.4 0.43 13.92 25.53 60.291.6 0.9 0.48 11.06 18.81 69.791.2 1.4 0.43 12.99 19.80 66.93– 15.8 0.16 7.03 19.24 74.96– 5.5 0.36 18.37 16.74 64.66– 0.3 0.31 10.36 16.33 73.08

– 29.4 1.22 33.90 20.58 44.96– 0.2 0.48 6.44 15.43 77.75– 13.4 0.16 17.38 30.09 54.27– 3.1 0.80 16.13 19.11 64.20– 7.9 0.87 9.00 11.01 80.28

ash (dry basis); V, daf: volatile (dry, ash free basis); FC, ad: fixed carbon (air-dried

0

10

20

30

40

0.5 1 1.5 2 2.5 3 3.5

V, d

af (%

)

Ro (%) Ro (%)

a

40

50

60

70

80

90

0.5 1 1.5 2 2.5 3 3.5

FC, a

d (%

)

b

Fig. 2. The coal quality of coals with different coal ranks.

S. Li et al. / Fuel 139 (2015) 257–267 261

Coals with different ranks also show distinct coal qualities.With the increase of coal rank, the volatile content reduces andthe fixed carbon increases (Fig. 2), however, moisture and ash con-tents vary with no trend. Volatile refers to the volatile organic mat-ter yield by coals under vacuum heating, which represents thenon-aromatic fractions (aliphatic functional groups and alicyclicfunctional groups) in coals; and fixed carbon refers to the residuesafter excluding water, ash and volatile from the coal, and it canrepresent the carbon content in coal. With the increase of coalmetamorphism, the aliphatic functional groups and alicyclic func-tional groups gradually reduce, resulting in the decrease of the vol-atile content in coal; on the other hand, the aromatic ringarrangement becomes more orderly and the fixed carbon contentincreases gradually.

4.2. CBM storage capacities of coal reservoirs with different coal ranks

Coal reservoir is a kind of dual-porosity system, which includespores and fractures [24]. Pores in coal reservoirs can be classifiedinto micropores (less than 10 nm in diameter), transition pores(10–100 nm in diameter), mesopores (100–1000 nm in diameter),and macropores (greater than 1000 nm in diameter) [25]. Frac-tures, macropores and mesopores (are so-called seepage pores)are main parts of the flow pathes during CBM production. How-ever, transition pores and micropores are together called adsorp-tion pores, and they would be the main pore systems for gasadsorption and diffusion [26]. Since CBM are mainly adsorbed inpores of the coal, thus the adsorption pore structure and adsorp-tion ability of coal reservoir have a decisive impact on the CBMaccumulation and storage [20].

0

0.3

0.6

0.9

1.2

1.5

0.5 1.5 2.5 3.5

S BE

T(m

2 /g)

Ro (%)

a b

Fig. 3. Characteristics of SBET and VBJH of coals with diffe

4.2.1. Adsorption pore structure of coal reservoirs with different coalranks

The low-temperature N2 isotherm adsorption/desorptionmethod can quantitatively detective the pore whose diameter is1.7–130 nm. This method has been commonly used to studyadsorption pores [27,28]. Results show that SBET of the 32 analyzedcoal samples ranges between 0.168 and 1.266 m2/g, and VBJH variesfrom 0.6 � 10�3 to 12.5 � 10�3 mL/g. The SBET as well as the VBJH ofadsorption pores decrease firstly and then increase with coal rankincrease (Fig. 3), which indicates that adsorption pores of coal res-ervoirs decrease firstly and then increase with the increase of coalrank. This ‘U’ shape behavior is consistent with previous studies[29–31].

It has been suggested that the adsorption pore structure of thecoal can be described by the shape of N2 adsorption/desorption iso-therms [20,32]. The adsorption and desorption curves of coal sam-ples in western Guizhou and eastern Yunnan fall into three types(A–C) in this research (Fig. 4a). Type A is commonly dominatedby adsorption pores with a large maximum adsorption volume,as well as high SBET and VBJH values. There is a hysteresis loop intype A, reflecting that the adsorption pores are mostly open pores.Type B has medium SBET and VBJH values, and the maximumadsorption volume is smaller than that of type A. Moreover, notice-able hysteresis loop in type B illustrates that adsorption pore struc-ture of type B is characterized by open pores. For type C, theadsorption and desorption curves are relatively flat, with the low-est maximum adsorption volume, SBET and VBJH values of the threetypes. There is no hysteresis loop in the curves of type C, indicatingthe adsorption pores in type C are mostly closed. The adsorptionpore structure of type A has the best adsorption capacity and con-nectivity, and is the most favorable for CBM adsorption, desorption

0

0.003

0.006

0.009

0.012

0.015

0.5 1.5 2.5 3.5

VB

JH(m

L/g

)

Ro (%)

rent coal ranks (data of symbol ‘‘*’’ cited from [38]).

0

0.5

1

1.5

2

2.5

0 0.2 0.4 0.6 0.8 1

Vol

ume

adso

rped

(10-3

mL

/g)

P/P0

Type A

Type B

Type C

a

0.5 1.5 2.5 3.5

Typ

e of

cur

ves

Ro (%)

Type AType BType C

b

Fig. 4. Classification and distribution of nitrogen adsorption/desorption curves.

0

5

10

15

20

25

30

0 2 4 6 8 10

Gas

Con

tent

(m3 /t

)

Pressure (MPa)

HP(Ro=0.68)DAZ(Ro=1.37)JL(Ro=1.75)YS(Ro=1.89)AGL(Ro=2.32)QS(Ro=2.64)DSY(Ro=3.03)LS(Ro=3.14)

Fig. 5. CH4 adsorption isotherms for moisture-equilibrated coals with different coalranks.

262 S. Li et al. / Fuel 139 (2015) 257–267

and diffusion. Adsorption and flow properties of type B and type Cbecome worse. Type A distributes in coal reservoirs with Ro higherthan 1.5%, while Type B and Type C distribute in coal reservoirswith Ro less than 1.9% (Fig. 4b), Therefore, the adsorption porestructure of high rank coal is better than that of low rank coal.

4.2.2. Adsorption capacities of coal reservoirs with different coal ranksThe Langmuir parameters of the adsorption isotherms are usu-

ally used to evaluate the adsorption capacities of the coal. Lang-muir volume is the maximum adsorption volume of the coal, andLangmuir pressure is the pressure when the actual adsorptionvolume reaches 50% of the maximum adsorption volume. Inwestern Guizhou and eastern Yunnan, Langmuir volume is 9.59–39.62 m3/t, with an average value of 20.42 m3/t (Fig. 5). From

0

9

18

2 7

36

45

0.5 1.5 2.5 3.5

Lan

gmui

r vo

lum

e (m

3 /t)

a

Ro (%)

Fig. 6. Adsorption capacities of coal

low rank coal to high rank coal, the adsorption pores show aincrease tend, and thus Langmuir volume increases with theincrease of coal rank (Fig. 6a). Owing largely to the difference ofadsorption capacities (Langmuir volume), the average in situ gascontent of each zone varies largely, ranging from 5.85 to24.50 m3/t, and the gas saturation ranges from 30.9% to 85.00%.Langmuir pressure in the study area is 1.19–4.14 MPa, with anaverage value of 2.07 MPa, and the Langmuir pressure of high rankcoal is larger than that of low rank coal. Generally, Langmuir pres-sure of coal reservoir is larger than 2 Mpa when Ro is higher than2.5%; and Langmuir pressure is less than 1.6 Mpa when Ro is lowerthan 2.5% (Fig. 6b). Compared with low rank coal reservoirs, highrank coal reservoirs have relatively higher potential for CBMdevelopment.

4.3. Seepage capacities of coal reservoirs with different coal ranks

Seepage pores (including fractures, macropores and mesopores)in coals are the main gas flow paths during CBM recovery. Thus, theseepage pore structure is the main controlling factor of the seepagecapacity of coals, and has a significant impact on the permeabilityof the coal reservoir and the production of CBM [33,34].

4.3.1. Seepage pore structure of coal reservoirs with different coalranks

NMR experiments can detective connected pores and fractures,making it an effective method in characterizing the seepage porestructure. Water in different pores and fractures show diverserelaxation times, because smaller pores have shorter relaxationtime and larger pores have longer relaxation time. Thus, poresand fractures of various sizes can be distinguished using NMR T2

distributions [7,35]. It was also found that a much higher certain

1

2

3

4

0.5 1.5 2.5 3.5

Lan

gmui

r pr

essu

re (M

pa)

Ro (%)

b

rocks with different coal ranks.

0

40

80

120

160

200

0.1 1 10 100 1000 10000

Am

plitu

de (-

)

T2 (ms)

XLQ: Ro = 0.87%HPW: Ro = 1.23%JA: Ro = 1.38%QDG: Ro = 1.83%QS: Ro = 2.64%LS: Ro = 3.14%

Fig. 7. T2 distributions of coals with different coal ranks.

Table 2Information of the coal samples and results of NMR analysis.

Sample no. Ro (%) ST (�103 ms) Porosity (%)

T2 < 10 ms T2 > 10 ms Sum

XLQ 0.87 0.47 2.93 3.40 7.7HPW 1.23 1.39 0.63 2.02 6.9JA 1.38 0.82 0.22 1.04 5.2QDG 1.83 0.41 0.50 0.90 4.8QS 2.64 1.56 0.03 1.60 5.3LS 3.14 1.74 0.61 2.35 6.6

S. Li et al. / Fuel 139 (2015) 257–267 263

T2 peak reflects much better developed pores (or fractures) thatcorresponding to that T2 peak, and vice versa.

As shown in Fig. 7, T2 distribution of the low rank coal (XLQ)shows a continuous trimodal characteristic, indicating that allthe pores and fractures are developed in this sample. The mainpeak that locates at the high T2 value section represents thewell-developed seepage pores in low rank coal samples. VariousT2 distributions display for medium rank coals (HPW, JA andQDG), however, their peaks are all located at the medium T2 valuesection, meaning that the seepage pores and adsorption pores areequally developed in medium rank coals. Compared with othercoals, the T2 distribution amplitude of the medium rank coal isthe lowest, revealing that the development degree of the poresand fractures in medium rank coals is lower than other coals. Forhigh rank coals (QS and LS), the T2 distributions show independentbimodal or trimodal characteristics. The main peaks are located atthe low T2 value section, and this implies that the adsorption poresare well developed in high rank coals.

With coal rank increases, the main peak of T2 spectra turn leftand the area of T2 distribution (ST) decreases firstly and thenincreases, indicating the pore and fracture structure change pat-tern during the metamorphic process. Low rank coals have low

0

2

4

6

8

10

0.5 1.5 2.5 3.5

Poro

sity

(%)

Ro (%)

a b

Fig. 8. Porosity and permeability of coals with differen

compaction degree and loose structures, as a result, the volumeof pore and fracture in low rank coals is the largest. Take sampleXLQ for example, its porosity is 7.7%, and its ST is up to3.40 � 103 ms. In addition, seepage pores are more developed thanadsorption pores in low rank coals: ST in the high T2 value section(>10 ms, corresponding to seepage pores) is 2.93 � 103 ms, and ST

in the low T2 value section (<10 ms, corresponding to adsorptionpores) is only 0.47 � 103 ms (Table 2). With the burial depthincrease, the metamorphic degree as well as the compactiondegree of coals continuously grow. As a result, seepage pores aregradually closed under pressure, and the molecular structurechanges continuously accompanies with coalification, resulting ina slight increase of adsorption pores in medium rank coals[36,37]. And in high rank coals, the adsorption pores are extremelydeveloped: the ST less than 10 ms is up to 1.56–1.74 � 103 ms,however, the ST more than 10 ms is only 0.03–0.61 � 103 ms(Table 2).

4.3.2. Characteristics of porosity and permeabilityThe porosity obtained by the gravimetric method represents the

ratio of the connected pore and fracture volume to the bulk volumeof the sample, consequently, the porosity mentioned in this paperdescribes the seepage capacities of coal reservoirs. In westernGuizhou and eastern Yunnan, the porosity of coals ranges from1.8% to 9.5% with an average value of 5.1%, and the permeabilityvaries from 0.01 to 42.6 mD. The porosity and permeabilitydecreased firstly and then increased with the climb of Ro (Fig. 8),which is consistent with the NMR testing results.

The evolution of coals in the study area is dominated by burialmetamorphism and tectonic thermal metamorphism. Before Yan-shan period, the burial metamorphism had turned the coal intomedium rank coal. With the increase of burial depth, pores andfractures of the coal were closed in respect to the increasing pres-sure of overlying strata. As a result the porosity and permeability ofthe medium rank coal reduced. In Yanshan period, the intense tec-tonic events uplifted the Moho surface, resulted in massive magmaintrusions, and formed a high temperature field in this region. Thisnot only speeded up the coal evolution and turned the coal intohigh rank coal rapidly, but also produced some epigenetic poresand fractures in the coal. Moreover, when the coal metamorphismdegree increased under the rapidly growing temperature, coal res-ervoirs began to regenerate methane, and a great number of ‘‘hotair holes’’ with larger diameter formed, which greatly improvedthe porosity and permeability of the high rank coals (Fig. 9).

5. Discussions

CBM resources is abundant in western Guizhou and easternYunnan, while the uneven degree of coal metamorphism resultsin the heterogeneity in coal reservoir properties, in addition,

Ro (%)

0.01

0.1

1

10

100

0.5 1.5 2.5 3.5

Perm

eabi

lity

(mD

)

t coal ranks (data of symbol ‘‘*’’ cited from [38]).

Fig. 9. Microscopic characteristics of pore and fracture structures in western Guizhou and eastern Yunnan (a–c: epigenetic pores, d–f: exogenetic fractures, g–i: air holes).

264 S. Li et al. / Fuel 139 (2015) 257–267

increases the difficulty of CBM exploration and development in thisregion. With the intention of selecting the most favorable block forpriority development, this research optimized the CBM potential ofdifferent coal reservoirs in the target area based on multi-objectiveand multi-level fuzzy optimization model.

The evaluation can be divided into three hierarchies: the ulti-mate goal is the coal reservoir physical properties, and nine indices

Table 3Parameter system in five CBM blocks.

Assessment indices Laochang

VA VN

Coal composition Vitrinite content (%) 71.0 0.186Ash content (%) 16.6 0.195

CBM storage capacity SBET (m2/g) 0.9487 0.284VBJH (mL/g) 0.0065 0.254Langmuir volume (m3/t) 15.14 0.157Langmuir pressure (MPa) 1.22 0.127In situ gas content (m3/t) 6.55 0.106

Seepage capacity of coal reservoir Porosity (%) 5.05 0.195Permeability (mD) 0.509 0.037NMR ST (ms) 0.90 0.116

VA: actual value; VN: normalized value.

under three categories were identified. Vitrinite content (U11) andash yield (U12) are the indices of coal composition factor (U1),which is the foundation of coal rocks. Adsorption factor (U2) of coalreservoir includes the indices of SBET (U21), VBJH (U22), Langmuir vol-ume (U23), Langmuir pressure (U24), and in situ gas content (U25),exhibiting the storage capacities of coal reservoirs. In addition,porosity (U31), permeability (U32), and NMR ST (U33) are the indices

Enhong Panxian Gemudi Zhina

VA VN VA VN VA VN VA VN

82.5 0.216 68.9 0.181 76.6 0.201 82.5 0.21612.3 0.205 12.5 0.204 18.1 0.191 12.5 0.204

0.4455 0.133 0.4000 0.120 0.5128 0.154 1.0316 0.3090.0023 0.090 0.0028 0.109 0.0041 0.160 0.0099 0.38718.89 0.195 13.83 0.143 16.16 0.167 32.72 0.3382.40 0.249 1.97 0.205 1.36 0.141 2.68 0.2785.85 0.095 11.76 0.191 12.93 0.210 24.50 0.398

5.74 0.221 6.80 0.262 4.05 0.156 4.30 0.1660.789 0.057 4.216 0.303 0.946 0.068 7.434 0.5351.04 0.135 2.71 0.351 1.11 0.144 1.97 0.255

Table 4Judgement matrix and weights of the parameter system.

U U1 U2 U3 WU

U1 1 0.67 0.5 0.22U2 1.5 1 0.75 0.33U3 2 1.33 1 0.45

Consistency check kmax = 3.0008 CR = 0.0008 < 0.1

U1 U11 U12 WU1

U11 1 0.8 0.44U12 1.25 1 0.56

Consistency check kmax = 2 CR = 0 < 0.1

U2 U21 U22 U23 U24 U25 WU2

U21 1 1 0.67 0.67 0.5 0.14U22 1 1 0.67 0.67 0.5 0.14U23 1.5 1.5 1 1 0.75 0.22U24 1.5 1.5 1 1 0.75 0.22U25 2 2 1.33 1.33 1 0.28Consistency check kmax = 5.003 CR = 0.0007 < 0.1

U3 U31 U32 U33 WU3

U31 1 0.5 1.25 0.27U32 2 1 2.5 0.52U33 0.8 0.4 1 0.21

Consistency check kmax = 3.0002 CR = 0.0002 < 0.1

S. Li et al. / Fuel 139 (2015) 257–267 265

of seepage factor (U3), reflecting the seepage capacities of coal res-ervoirs. These parameters in the five CBM blocks in westernGuizhou and eastern Yunnan are listed in Table 3, and the normal-ized parameters are also listed in Table 3.

The judgement matrix was achieved through comparisonbetween two parameters in the same level with one another. Then,Matlab was used to calculate the largest eigenvalue and the corre-sponding eigenvectors of the judgment matrix (Table 4). To ensurethe reliability and accuracy of the results, this study used the

0.1910.210

0.194 0.196 0.210

0

0.07

0.14

0.21

0.28

Com

posi

tion

of c

oal

CBM producing areas

a

0.096 0.118

0.302

0.108

0.377

0

0.07

0.14

0.21

0.28

0.35

0.42

Seep

age

capa

city

CBM producing areas

c

Fig. 10. Assessment of favorable zones for CBM exploration and development in the studyOptimized by seepage factor; d. Comprehensively optimized).

eigenvalue method proposed by T.L. Saaty to determine the consis-tency of the matrix [21,22]. In detail, the determination of the con-sistency of the judgment matrix utilized the random consistencyratio (CR), which is the ratio between the consistency index (CI)and the random consistency index (RI). In which, CI = (kmax � n)/(n � 1); CR = CI/RI; and n is the order of the matrix. WhenCR < 0.1, inconsistency of the judgment matrix is acceptable; andwhen CR P 0.1, the judgment matrix needs to be re-assigned andcalculated until the consistency is acceptable.

0.167 0.156 0.162 0.171

0.344

0

0.07

0.14

0.21

0.28

0.35

0.42

Ads

orpt

ion

capa

city

CBM producing areas

b

0.140 0.150

0.232

0.148

0.329

0

0.07

0.14

0.21

0.28

0.35

Com

preh

ensi

vely

fact

or

CBM producing areas

d

area (a. Optimized by coal composition factor; b. Optimized by adsorption factor; c.

266 S. Li et al. / Fuel 139 (2015) 257–267

Based on multi-objective and multi-level fuzzy optimizationmodel, favorable zones for CBM exploration and development inwestern Guizhou and eastern Yunnan have been assessed (Fig. 8).Take the fuzzy evaluation of adsorption factor (U2) as an example,evaluation indices include SBET (U21), VBJH (U22), Langmuir volume(U23), Langmuir pressure (U24), and in situ gas content (U25), andthe corresponding weights are W2 ¼ 0:14 0:1ð 40:220:220:28Þ,the judgment matrix of the fuzzy evaluation is:

U2 ¼

0:284 0:133 0:120 0:154 0:3090:254 0:090 0:109 0:160 0:3870:157 0:195 0:143 0:167 0:3380:127 0:249 0:205 0:141 0:2780:106 0:095 0:191 0:210 0:398

26666664

37777775

ð4Þ

And using Eq. (3) for the fuzzy evaluation of the second-levelindex:

B2 ¼Wi � Ui ¼ 0:14 0:14 0:22 0:22 0:28ð Þ

�

0:284 0:133 0:120 0:154 0:309

0:254 0:090 0:109 0:160 0:387

0:157 0:195 0:143 0:167 0:338

0:127 0:249 0:205 0:141 0:278

0:106 0:095 0:191 0:210 0:398

26666664

37777775

¼ 0:167 0:156 0:144 0:162 0:344ð Þ ð5Þ

B ¼max 0:167 0:156 0:144 0:162 0:344ð Þ ¼ 0:344, corre-sponds to the Zhina area (Fig. 9b), indicating that the adsorptionfactors in Zhina area are superior to other CBM blocks. The fuzzyevaluation process of other two second level indices are the same,and the results are shown in Fig. 10. The most favorable zone forCBM development is Zhina area (with a comprehensive assessmentindex of 0.329), because it has an optimum CBM production poten-tial, relatively high metamorphic degree, high gas content, andhigh CBM resource concentration. The comprehensive assessmentindex of Panguan area is 0.232, which is lower than that of Zhinaarea. Compared to Zhina area, Panguan area has an equal seepagecapacity but less gas content, and less CBM resources concentra-tion. Enhong, Gemudi and Laochang areas are the most unfavorableareas in the five target CBM blocks. In these areas, coal seams havemoderate metamorphic degree, medium gas content, poor porosityand permeability. And their comprehensive assessment indices areall less than 0.15, which can be considered as not favorable for CBMproduction.

6. Conclusions

The coal metamorphism in western Guizhou and eastern Yun-nan is very uneven. Composition, pore and fracture structures,adsorption capacity and seepage capacity of the coal rock signifi-cantly change with the increase of coal rank. Moreover, resultedfrom the various burial history or tectonic thermal events, differentevolution process and physical–chemical structure of the coal gen-erated, which in return determined the variety in physical proper-ties of coal reservoirs.

Based on the multi-objective and multi-level fuzzy optimizationmodel, this research optimized the CBM potential of different coalreservoirs in the main five CBM blocks in the western Guizhou andeastern Yunnan. The most favorable zone for CBM production isZhina area, and the potential of CBM development in Panguan,Enhong, Gemudi and Laochang areas is unfavorable in sequence.

Acknowledgements

This work was financially supported by the National NaturalScience Foundation of China (41272175), the National BasicResearch Program of China (973) (902009CB219600), the Key Pro-ject of the National Science & Technology (2011ZX05034), and theFundamental Research Funds for the Central Universities(2652013056). This work was also partly supported by the ChineseScholarship Council.

References

[1] Luo D, Dai Y. Economic evaluation of coalbed methane production in China.Energy Policy 2009;37(10):3883–9.

[2] Tao S, Wang YB, Tang DZ, Xu H, Lv YM, He W, et al. Dynamic variation effects ofcoal permeability during the coalbed methane development process in theQinshui Basin, China. Int J Coal Geol 2012;93(1):16–22.

[3] Tang SH, Sun SL, Hao DH, Tang DZ. Coalbed methane-bearing characteristicsand reservoir physical properties of principal target areas in North China. ActaGeol Sinica (English edition) 2004;78(3):724–8.

[4] Xu H, Zhang SH, Leng X, Tang DZ, Wang MS. Analysis of pore system model andphysical property of coal reservoir in the Qinshui Basin. Chinese Sci Bull2005;50(z1):45–50.

[5] Li J, Liu D, Yao Y, Cai Y, Guo X. Physical characterization of the pore-fracturesystem in coals, Northeastern China. Energy Explor Exploitation2013;31(2):267–86.

[6] Gu CL. Geological characteristics and prospect evaluation on coal-seam gas inEast Yunnan and West Guizhou areas. Xinjiang Pet Geol 2002;23(2):106–11 (inChinese with English abstract).

[7] Li S, Tang DZ, Xu H, Yang Z. The differences of physical properties of coalreservoirs and their origin mechanism between Zhijin and Panxian areas,Western Guizhou, China. Energy Explor Exploitation 2012;30(4):661–76.

[8] Shao LY, Zhang PF. Braided delta depositional system and coal accumulationduring early Late Permian period in eastern Yunnan and western Guizhou,southwest China. Acta Sedimentologica Sinica 1994;12(4):132–9 (in Chinesewith English abstract).

[9] Gui BL, Wang CD. Structural characteristics of coalbed methane in East Yunnanand West Guizhou area. Yunnan Geol. 2000;19(4):321–51 (in Chinese withEnglish abstract).

[10] Gao D, Qin Y, Yi TS. CBM geology and exploring developing stratagem inGuizhou Province, China. Proc Earth Planet Sci 2009;1(1):882–7.

[11] Cuadrat A, Abad A, García-Labiano F, Gayán P, de Diego LF, Adánez J. Relevanceof the coal rank on the performance of the in situ gasification chemical-loopingcombustion. Chem Eng J 2012;195:91–102.

[12] Muhammad AB, Abbott GD. The thermal evolution of asphaltene-boundbiomarkers from coals of different rank: a potential information resourceduring coal biodegradation. Int J Coal Geol 2012;107:90–5.

[13] Zheng GQ, Pan ZJ, Chen ZW, Tang SH, Connell LD, Zhang SH, et al. Laboratorystudy of gas permeability and cleat compressibility for CBM/ECBM in Chinesecoals. Energy Explor Exploitation 2012;30:451–76.

[14] Tang SH, Cai C, Zhu BC, Duan LJ, Zhang JZ. Control effect of coal metamorphicdegree on physical properties of coal reservoirs. Nat Gas Ind 2008;28(12):30–3(in Chinese with English abstract).

[15] Chen Y, Mastalerz M, Schimmelmann A. Characterization of chemicalfunctional groups in macerals across different coal ranks via micro-FTIRspectroscopy. Int J Coal Geol 2012;104:22–33.

[16] Pan ZJ, Connell LD. Modelling permeability for coal reservoirs: a review ofanalytical models and testing data. Int J Coal Geol 2012;92:1–44.

[17] Gui BL. The selected area and evaluation of exploration target of seam gas inEast Yunnan and West Guizhou. Yunnan Geol 2004;23(4):410–20 (in Chinesewith English abstract).

[18] Li S, Tang DZ, Xu H, Yang Z. Porosity and permeability models for coals usinglow-field nuclear magnetic resonance. Energy Fuel 2012;26(8):5005–14.

[19] Lin DY, Ye JP, Qin Y, Tang SH. Characteristics of coalbed methane resources ofChina. Acta Geologicu Sinica (English ed.) 2000;74(3):706–10.

[20] Liu DM, Yao YB, Tang DZ, Tang SH, Che Y, Huang WH. Coal reservoircharacteristics and coalbed methane resource assessment in Huainan andHuaibei coalfields, Southern North China. Int J Coal Geol 2009;79:97–112.

[21] Saaty TL. Axiomatic foundation of the analytic hierarchy process. Manage Sci1985;32(7):841–55.

[22] Saaty TL. Decision-making with the AHP: why is the principal eigenvectornecessary. Eur J Operational Res 2003;145(1):85–91.

[23] Crosdale PJ, Beamish BB, Valix M. Coalbed methane sorption related to coalcomposition. Int J Coal Geol 1998;35(1–4):147–58.

[24] Gamson PD, Beamish BB, Johnson DP. Coal microstructure and micropermeability and their effects on natural gas recovery. Fuel 1993;72:87–99.

[25] Xolon BB. Coal and gas outburst (Chinese Translation). Beijing: China industryPress; 1996.

[26] Yao YB, Liu DM, Tang DZ, Tang SH, Huang WH, Liu ZH, et al. Fractalcharacterization of seepage-pores of coals from China: an investigation onpermeability of coals. Comput Geosci 2009;35:1159–66.

S. Li et al. / Fuel 139 (2015) 257–267 267

[27] Sing KSW. Reporting physisorption data for gas solid systems with specialreference to the determination of surface-area and porosity. Pure Appl Chem1982;54(11):2201–18.

[28] Groen JC, Peffer LAA, Pérez-Ramírez J. Pore size determination in modifiedmicro- and mesoporous materials. Pitfalls and limitations in gas adsorptiondata analysis. Microporous Mesoporous Mater 2003;60:1–17.

[29] Gensterblum Y, Busch A, Krooss BM. Molecular concept and experimentalevidence of competitive adsorption of H2O, CO2 and CH4 on organic material.Fuel 2014;115:581–8.

[30] Prinz D, Littke R. Development of the micro- and ultramicroporous structure ofcoals with rank as deduced from the accessibility to water. Fuel 2005;84(12–13):1645–52.

[31] Prinz D, Pyckhout-Hintzen W, Littke R. Development of the meso- andmacroporous structure of coals with rank as analyzed with small angleneutron scattering and adsorption experiments. Fuel 2004;83:547–56.

[32] Burgess CGV, Everett DH, Nuttall S. Adsorption hysteresis in porous materials.Pure Appl Chem 1989;61:1845–52.

[33] Pan ZJ, Connell LD, Camilleri M. Laboratory characterisation of coal reservoirpermeability for primary and enhanced coalbed methane recovery. Int J CoalGeol 2010;82:252–61.

[34] Shi JQ, Durucan S. A model for changes in coalbed permeability during primaryand enhanced methane recovery. SPE Reservoir Eval Eng 2005;8(4):291–9.

[35] Yao YB, Liu DM, Che Y, Tang DZ, Tang SH, Huang WH. Petrophysicalcharacterization of coals by low-field nuclear magnetic resonance (NMR).Fuel 2010;89:1371–80.

[36] Busch A, Gensterblum Y. CBM and CO2-ECBM related sorption processes incoal: a review. Int J Coal Geol 2011;87:49–71.

[37] Van Krevelen DW. Coal: typology–physics–chemistry-constitution. Elsevier;1993.

[38] Li S, Tang DZ, Pan ZJ, Xu H. Influence and control of coal facies on physicalproperties of the coal reservoirs in western Guizhou and eastern Yunnan,China. Int J Oil, Gas Coal Technol 2014;8(2):221–34.