Journal of Natural Gas Science and Engineering 92 (2021) 103969

Contents lists available at ScienceDirect

Journal of Natural Gas Science and Engineering

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

Evaluating production behaviour of CBM wells from Raniganj Coalfield

through reservoir characterization under constrained field data conditions
Sujoy Chattaraj a, c, Rajeev Upadhyay b, Debadutta Mohanty a, *, Gopinath Halder c,
Tarkeshwar Kumar b
a
Non-Conventional Gases Department, CSIR – Central Institute of Mining and Fuel Research, Dhanbad, 826 001, India
b
Department of Petroleum Engineering, Indian Institute of Technology (ISM) Dhanbad, Dhanbad, 826 004, India
c
Department of Chemical Engineering, National Institute of Technology Durgapur, Durgapur, 713 209, India

A R T I C L E I N F O A B S T R A C T

Keywords: There remains a good level of uncertainty in peak gas rate and the time to reach the peak for coalbed methane
Coalbed methane (CBM) wells owing to reservoir heterogeneity. The novelty of the present study lies in predicting the production
Reservoir behavior of wells in the emerging CBM block of Raniganj through extensive reservoir characterization coupled
Gas content
with reservoir simulation and validation. Ten regional (R–I to R-X) and two local seams intercepted in the six
Permeability
exploratory boreholes drilled in Raniganj Coalfield were investigated. The reservoir characterization approach
Isotherm
Recovery was used for uncertainty analysis of the key parameters viz. isotherms, gas content, permeability and porosity,
and their impact on the production behavior of the CBM wells. Low, mid and high cases of the parameters were
determined through regression analysis.
Isotherm and gas content data were experimentally determined while in situ permeability data were obtained
from the pressure transient analysis of injection/fall-off test. Reiss model was used for porosity-permeability
transformation and Corey’s model was used to define the range of relative permeability. The experimental
values of Langmuir volume varies between 17.4 and 29.8 cc/gm (daf); Langmuir pressure 2451–7827 kPa; gas
content 2.07–13.03 cc/gm (daf); and fracture permeability 0.79–7.12 mD. The model was set up in CMG-GEM
simulator to represent a single well producing under commingled completion strategy for all the 12 seams
being perforated. Production profiles of both gas and water show a wide range of variation up to 56,201 m3/day
and up to 129.84 m3/day, respectively. Data analysis reveals that Raniganj coals are having a wide range of
saturation distribution (between 27% and 100%) with a huge upside potential going towards 100% saturation
case. Gas content and permeability are found to be the two key factors controlling the CBM production in
Raniganj Coalfield. The model results were validated with actual well performance.

1. Introduction from coal matrices. The water-filled cleat system inhibits the scope for
evaluating the potential of CBM wells in terms of gas production rates at
Methane is stored within the porous network of the coal. The CBM the time of the exploration. Unlike conventional gas wells, the CBM
reservoir model is considered as a dual-porosity model based on the exploration wells cannot test the gas deliverability of the coal reservoir.
Warren and Root principle (Warren and Root, 1963). Gas flows to the This makes the CBM prospect assessment process markedly different
wellbore generally under two-phase conditions in three stages viz., than conventional gas reservoirs. The CBM reservoir deliverability is
desorption from the coal surface due to pressure gradient, diffusion accomplished by the next phase pilot testing in which the pilot’s mul­
through coal matrix due to the concentration gradient, and laminar flow tiple wells can interfere with each other to locally diminish the pressure
described by Darcy’s Law through the fracture/cleat network (Kumar in a significant area of the reservoir to the critical sorption pressure
et al., 2006; Shi and Durucan, 2008; Mohanty, 2011; Bao et al., 2020). allowing desorption and production of consequential gas. Therefore,
CBM reservoirs initially contain a water-filled cleat system leading to the CBM operators are always faced with the dilemma of correct decision
requirement of dewatering of coal seams for effective gas desorption making at the exploration stage about the next phase course of action.

* Corresponding author.
E-mail addresses: drdmohanty@cimfr.nic.in, drdmohanty@ymail.com (D. Mohanty).

https://doi.org/10.1016/j.jngse.2021.103969
Received 23 July 2020; Received in revised form 12 April 2021; Accepted 13 April 2021
Available online 27 April 2021
1875-5100/© 2021 Elsevier B.V. All rights reserved.
S. Chattaraj et al. Journal of Natural Gas Science and Engineering 92 (2021) 103969

Production behavior of CBM wells depends upon several parameters characterization and methane adsorption capacity of Raniganj coals
– both operational control and reservoir characteristics. Major reservoir were carried out by different workers (Laxminarayana and Crosdale,
parameters that dictate the well productivity include seam thickness, 2002; Singh and Mohanty, 2015; Mohanty et al., 2018). However, a
sorption time, gas content, Langmuir isotherm data, porosity, perme­ detailed evaluation of the production behavior of Raniganj coals based
ability, and relative permeability (Fu et al., 2009; Mohanty et al., 2017; on the reservoir characterization has not yet been carried out. The paper
Wei et al., 2019). Pan and Wood (2015) have presented a detailed re­ presents a methodology to predict the production behaviour of a CBM
view of CBM exploration, coal characterization, CBM well simulation field based on detailed reservoir characterization that can be imple­
and production issues, reservoir performance and modeling. A para­ mented in other CBM fields that are in the nascent state of their pro­
metric study to understand the influence of the individual parameters on duction life. The data on in situ gas content, sorption time, proximate and
the production behaviour of a CBM well is of immense importance for ultimate analyses, adsorption isotherms, and in situ permeability of the
the CBM production planning and field development. Zuber and Ols­ coal seams intercepted in six boreholes viz. BH#A, BH#B, BH#C, BH#D,
zewski (1992, 1993) predicted that among all the parameters adsorbed BH#E and BH#F in the Raniganj Coalfield have been analyzed to
gas content, sorption isotherm, water saturation, coalbed thickness, characterize the CBM reservoir. Experimental analyses performed on
permeability, and porosity have the greatest impact on CBM production. twelve (12) number of fresh borehole coal cores in addition to the data
Similar observations are also made by Karacan (2008) for methane flow for thirty-nine (39) number of samples from previous work by the au­
in the underground coal mine workings. Permeability, which is closely thors (Mohanty et al., 2018) for better analyses of the rock, fluid, and
related to the coal fabric (i.e., cleat spacing and aperture width), varies fluid-flow characteristics.
significantly due to reservoir geomechanical effects and fluid pressure A better understanding of regional variation enables the distribution
changes during coal seam gas production (Cui and Bustin, 2006; Pan and of reservoir properties in the reservoir model. In order to investigate the
Connell, 2012; Chen et al., 2015; Liu et al., 2020; Kou and Wang, 2020; techno-economic feasibility of the commingled gas production from the
Wang et al., 2021) and strongly influence the gas production profile and multiple coal seams through modelling and simulation, it is very much
well performance. important to understand the effect of each important reservoir property
The variation of all the parameters is significant across a reservoir on the gas production (Wu et al., 2018). Authors have undertaken a
and monitoring or analysis of every single parameter at each location is statistical approach to characterize the CBM reservoir and to establish
not practicable. Further, due to the complex reservoir geometry and a the regional depth trends of key reservoir properties such as the varia­
wide variation of the reservoir parameters on a regional scale, the un­ tion in cleat porosity, permeability, gas content, and adsorption iso­
certainty in the estimation of the CBM reservoir parameters prompts for therms etc. BH#B has been chosen for modeling purposes due to its
the reservoir simulation studies (Jalali et al., 2010; Karacan et al., 2014). central location among all the boreholes. The interlayer partings are
Though the simulation of hydrocarbon production from the conven­ isolated and the perforation jobs carried out in all the coal seams
tional reservoir is very common since the middle of the last century, the intercepted in BH#B were considered to model the commingled gas/­
application of reservoir simulation in the coalbed methane (CBM) res­ water production from the well. The available subsurface data –
ervoirs is relatively more recent. CBM reservoir simulation combines the geological, petrophysical, and reservoir data - from core wells in Rani­
reservoir geology with the physical phenomenon and the transport ganj Coalfield has been integrated to represent the reservoir in miniature
phenomenon within the porous network to predict the reservoir per­ form of a 3D subsurface model. The producibility of a CBM well and
formance. Few researchers have assessed the uncertainty of the pa­ related uncertainties in production behaviour has been evaluated in a
rameters associated with CBM production (Karimi, 2005; Zhou and four-step process – 1) establishing the most likely depth trends of various
Guan, 2016; Keles et al., 2019). Karimi (2005) had performed sensitivity coal properties; 2) defining a reasonable range of low and high case
analysis considering base, low and high values of each parameter ob­ trends using statistical regression analysis; 3) construction of low, mid,
tained from analysis, and concluded that permeability is the most and high case numerical simulation model in CMG with grid cells
influencing parameter that affects the gas recovery. In a parametric populated with properties established by low, mid, and high case trends;
study, Lv et al. (2012) on Southern Qinshui Basin, China found that the and 4) generation of production profiles with all possible combinations
gas content and permeability are the key parameters that control the gas of low, mid, and high cases on individual parameters.
production as these are the sources of gas volume and the conductivity, One of the biggest problems CBM operators face in the development
respectively. Xu et al. (2015) also inferred that permeability is the most of CBM fields is the limitation posed by the lack of knowledge on gas
influencing parameter for CBM production in Hancheng area, south­ deliverability of a CBM well during the early stage of well production,
eastern Ordos Basin, China. The Spearman rank analysis of the reservoir which compels the operators to delay their business decisions and wait
parameters of Junggar Basin studied by Kang et al. (2018) had shown for the well production to realize its peak gas. The application of the
that critical desorption pressure and permeability play a major role in present investigation will help CBM operators to know the expected
controlling the CBM production. Zou et al. (2018) proposed a model for range of gas deliverability of a CBM well in the area in advance to make
the prediction of CBM production considering critical desorption radius wise business decisions early. The present study has brought uniqueness
and critical desorption pressure. Considering the dynamic changes in that adds considerable value to the existing knowledge through a
permeability during the CBM production process, Meng et al. (2018) structured investigation as outlined below:
proposed an improved model for the prediction of CBM production
through effective stress, coal matrix shrinkage effect and gas slippage • Extensive reservoir characterization using state-of-the-art analytical
effect. More recently, Keles et al. (2019) concluded that changes in techniques
water saturation, Langmuir constants, coal porosity, reservoir pressure • The data integration and rigorous regression analysis of the reservoir
and cleat permeability have a pronounced effect on methane parameters is conducted to define a reasonable uncertainty range to
production. depict the regional as well as depth variation to achieve a robust
Though Raniganj Coalfield has witnessed coal-mining activities since reservoir characterization from CBM perspective of Raniganj
the 18th century, activities related to CBM are very recent and a sig­ Coalfield.
nificant part of field development is yet to come. CBM reservoir

2
S. Chattaraj et al. Journal of Natural Gas Science and Engineering 92 (2021) 103969

• The learning from the reservoir characterization has been imple­ patchy Karharbari Formation. However, the western and southern
mented in the reservoir simulation study to predict the production boundaries are faulted ones. The throw of the southern boundary fault is
behaviour of CBM wells in Raniganj Coalfield. All possible scenarios nearly 2750 m. The Barakar, Barren Measures, Raniganj and Panchet
with a combination of low and high cases for each reservoir Formations are exposed successively from north to south. Out of these
parameter have been considered in the CBM reservoir simulation Barakar and Raniganj Formations are represented mainly by sandstone,
study to quantify the range of production behaviour that CBM wells shales and coal seams. The coal seams of Raniganj Formation are usually
are expected to exhibit that has wide applicability in the future consistent in thickness and quality over a large area while the coal seams
development of Raniganj CBM field. belonging to the Barakar Formation vary both in the thickness and in
• Assessment of key parameters affecting CBM production and vali­ quality within a short distance in Raniganj Coalfield. The iron stone
dation of model data using the actual production data to examine the shale bearing Barren Measure in the Raniganj basin separates the two
efficacy of the simulation. coal measures viz. Barakar and Raniganj (Ghosh, 2002). The Supra
Panchet, exposed in two patches in the south-west and south-central
The workflow adopted for the present work is presented in Fig. 1. parts of the basin, rest unconformably over the Panchet.
The Raniganj Coalfields have commercial coal deposits in both the
2. Geology of the study area Raniganj and Barakar Formations. Barakar Formations comprise of
fluviatile sediments deposited with irregularly thick eight regionally
Peninsular India hosts the bulk of coal deposits of Permian age in the correlated coal seams, B–I to B-VIII. However, the Barakar Formation
lower Gondwana sequences along the major river valley lineaments. coal seams are exposed over a very small extent area in this coalfield.
Among all, the prospective CBM blocks are situated in Damodar Valley. Raniganj Formation is best developed in the Raniganj having a thickness
Commercial CBM projects exist in Parbatpur block in Jharia Coalfield; of more than 1000 m (Gee, 1932). Ten regional coal seams are recog­
Raniganj South, Raniganj East and Raniganj North blocks in Raniganj nized in Raniganj Formation, R–I to R-X, in the ascending order having
Coalfield; and in the East and West Bokaro Coalfields (Chattaraj et al., an average thickness of more than 1.2 m. The lower seams are relatively
2019). The Raniganj Coalfield has a prognosticated CBM resource of 7.7 thicker in Raniganj Formation. The outcrops of the Raniganj measures
TCF (DGH report, 2019). are found in the eastern part of the coalfield and consist of laterite and
Raniganj basin, the easternmost coalfield in the Damodar Valley, alluvium. East of the Barakar River, the succession is well exposed, and
opens toward the south along with the younger sedimentary rocks to­ most of the softer shales and coal seams are hidden. The general strike of
ward the south-southeast. Two prominent elevations namely Panchet the coal-bearing formation is nearly east to west in the western and
and Biharinath are conspicuous in the southern part of the coalfield. eastern parts of the coalfield. However, the strike varies to NW-SE or
Three perennial rivers Ajay, Barakar and Damodar along with their even NNW-SSE in the central part. The general dip is towards the south
tributaries drain the northern, eastern and southern parts of the coalfield with the amount ranging from 3◦ to 20◦ with an average of less than 10◦ .
as shown in Fig. 2. The northern margin represents the normal deposi­
tional boundary of the basal Gondwana strata over the basement of 3. Reservoir characterization
Archaeans. The basal Talchir Formation is the oldest in the Gondwana
Supergroup, is exposed mainly in the northwestern part of the coalfield, Coal core samples were collected from the six exploratory boreholes
and passes upward to the Barakar Formation through a transitional, drilled in the Raniganj Coalfield. The gas contents were determined by

Fig. 1. Workflow adopted in the present work.

3
S. Chattaraj et al. Journal of Natural Gas Science and Engineering 92 (2021) 103969

Fig. 2. Map showing geology of Raniganj Coalfield along with the location of the boreholes.

collecting the coal cores and measuring the volume of gas released parameter i.e., 90% confidence band for depth vs moisture and depth vs
following USBM Direct Method developed by Bertard et al. (1970) and gas content plots; 80% confidence band for depth vs permeability plot;
modified by Kissell et al. (1973). After that the proximate and ultimate and 70% confidence band for depth vs Langmuir volume and tempera­
analyses were carried out following ASTM D3172 (ASTM D3172, 2013) ture vs Langmuir pressure plots. It is seen from Table 2 that, moderate to
and ASTM D3176 (ASTM D3176, 1979), respectively on the crushed high correlation coefficient was found between the studied parameters.
samples. Adsorption isotherm construction was carried out at the The coefficient was highest for temperature vs Langmuir pressure
reservoir temperature. Detail analytical methodology has been dis­ (0.838) followed by depth vs moisture (0.78053), depth vs Langmuir
cussed in detail at Mohanty et al. (2018). Experimental data is presented volume (0.7558), depth vs permeability (0.66000), and the lowest for
in Table 1. The reservoir characterization was carried out based on the the depth vs gas content (0.65841). Similar trends were also observed
experimental data. for the regression coefficient and all the plots were found to be statis­
A workflow of reservoir characterization has been used to define the tically significant as the P values are all less than 0.05. In addition to
depth trends of various coal properties based on the regression analysis that, 95% confidence band (upper and lower) has been defined for all
performed on the observed/experimental data from the set of corehole the parameters (Table 2). For example, depth vs moisture plot has slope
samples. The depth trend, thus established, is used to get a reasonable values as − 0.00320 & − 0.002 and the intercepts as 4.01094 & 4.90522
value of the reservoir parameters at a particular depth where no actual for 95% confidence interval. Whereas, at a 90% confidence interval
data is available. While constructing the simulation model, it was kept in these values are − 0.00310 & 0.00200 for slope, and 4.08500 & 4.83110
mind that the model may have grids where no actual existing well is for intercept. The upper limit of the confidence interval is considered as
present and therefore no observed data is available for those grids. In the high case and lower limit as the low case while the trend line is
such a scenario, the established depth trends have been used to populate considered as the mid-case for modeling purpose. Analyzed data has
the grid cells with values for the respective reservoir parameters. been used as input of the GEM reservoir simulation software, developed
The field as well as laboratory generated data were analyzed to by the Computer Modeling Group (CMG, 2003).
define the range of variation of individual parameters to build a reser­
voir model. First, regression analysis was performed to observe the
statistical significance of the plots and to calculate the high (upper limit 3.1. Moisture content
of confidence interval) and low value (lower limit of confidence inter­
val) of the variables from the plots. Details of the regression analyses The relation between coal maturity and depth is well established.
performed for each relationship is tabulated below (Table 2). Three The moisture content of the coal decreases with depth as the maturity of
different confidence intervals have been considered to calculate high coal increases (Suggate, 1974; Van Krevelen, 1993; Singh et al., 2016).
and low values depending upon the number of data points for each Mohanty et al. (2018) have found a negative relationship between depth
and moisture content for the set of Raniganj coals. The variation in

4
S. Chattaraj et al. Journal of Natural Gas Science and Engineering 92 (2021) 103969

Table 1
Reservoir characteristics of coal seams from Raniganj Coalfield.
Sample Depth Ash, A Moisture, M Temperature, T Gas Content Langmuir Volume, VL Langmuir Pressure, PL Saturation

No. Seam m % % ( C)
◦
(daf, cc/g) (daf, cc/g) (kPa) %

BH#B/01 R-X 237.47 26.6 3.8 37.1 2.49 – – –

BH#B/02 R-X 254.12 17.6 4.2 37.6 2.07 – – –
BH#B/03 R-IX 313.79 31.9 5.2 39.4 4.30 – – –
BH#C/05 R-IX 360.50 26.7 4.0 40.8 5.65 18.2 7119 93.63
BH#D/06 R–VIII 321.45 22.5 3.8 39.6 3.08 23.7 6010 37.80
BH#C/08 R–VIII 425.49 26.1 3.4 42.8 5.03 22.4 6531 57.64
BH#B/07 R–VIII 428.26 29.6 3.2 42.8 3.82 – – –
BH#B/08 R–VIII 428.75 18.2 3.6 42.9 3.77 – – –
BH#B/09 R–VIII 430.26 22.4 3.3 42.9 3.88 – – –
BH#E/04 R–VII 329.25 29.2 2.7 39.9 4.84 21.2 6540 69.13
BH#F/07 R–VII 485.31 30.8 2.7 44.6 7.10 22.9 6057 70.51
BH#B/10 R–VII 504.49 19.2 3.1 45.1 2.33 – – –
BH#B/11 R–VII 505.69 25.0 2.6 45.2 5.77 23 4878 49.79
BH#E/07 L-2A 420.20 30.8 3.2 42.6 5.50 19.1 7121 78.61
BH#B/L2A1 L-2A 537.50 – – 46.1 – – – –
BH#A/12 L-2A 547.28 25.4 3.2 46.4 13.03 22.8 5900 120.05
BH#E/09 L-2 447.26 37.3 2.3 43.4 5.67 17.4 7827 90.80
BH#B/12 L-2 566.12 23.2 2.9 47.0 6.67 21.8 6701 67.57
BH#B/13 L-2 567.41 25.0 2.8 47.0 4.78 – – –
BH#B/14 L-2 568.95 28.9 3.0 47.1 3.33 – – –
BH#D/09 R–VI 470.95 21.6 3.4 44.1 4.60 23.4 5502 43.10
BH#A/17 R–VI 632.29 27.2 3.0 49.0 7.86 22.2 5828 68.72
BH#B/15 R–VI 668.11 28.8 3.1 50.0 7.55 – – –
BH#F/10 R–VI 687.55 19.4 3.6 50.6 7.40 23.9 5019 54.04
BH#D/11 R–V 529.63 26.1 3.8 45.9 4.31 22.3 5843 41.09
BH#C/17 R–V 727.31 25.9 2.7 51.8 8.91 23.7 4842 63.15
BH#B/18 R–V 756.69 22.4 2.7 52.7 10.07 23.1 5276 74.62
BH#F/11 R–V 758.43 15.4 2.2 52.8 8.78 25.6 3516 50.53
BH#E/12 R–IV 573.73 17.0 1.3 47.2 4.49 23.6 4280 33.51
BH#D/15 R–IV 612.30 22.0 3.1 48.4 4.42 24.8 4128 30.09
BH#B/20 R–IV 916.17 13.7 1.6 57.5 7.34 23.2 4406 47.17
BH#C/20 R–IV 824.22 19.5 2.4 54.7 7.32 25.3 3420 41.19
BH#F/13 R–IV 828.71 12.8 1.9 54.9 5.56 23.3 4182 36.16
BH#A/26 R–IV 892.81 12.9 2.2 56.8 10.69 25.2 2810 56.05
BH#B/19 R–IV 914.64 30.4 1.8 57.4 7.92 – – –
BH#B/21 R–IV 917.41 21.0 2.0 57.5 7.03 – – –
BH#E/14 R–III 746.12 23.2 3.4 52.4 5.41 24.5 3297 32.04
BH#D/16 R–III 749.20 21.7 1.2 52.5 6.74 23.8 4899 47.22
BH#A/29 R–III 983.86 24.7 1.3 59.5 7.69 26.5 3315 39.00
BH#B/22 R–III 1019.88 17.5 1.8 60.6 5.31 23.9 4238 31.64
BH#E/15 R–II 759.68 20.8 2.6 52.8 4.27 23.9 3841 27.09
BH#D/17 R–II 771.63 16.1 3.3 53.1 6.68 25.7 3637 38.50
BH#C/25 R–II 1050.88 24.6 2.2 61.5 12.41 26.8 4194 65.17
BH#A/32 R–II 1098.01 26.8 1.6 62.9 8.86 25.8 3495 45.50
BH#B/24 R–II 1118.10 25.3 1.9 63.5 9.43 25.9 3470 47.94
BH#E/16 R–I 866.07 20.3 1.6 56.0 6.49 23.9 3882 39.58
BH#F/14 R–I 965.23 17.2 1.8 59.0 6.40 24.1 3758 37.11
BH#C/26 R–I 1076.27 22.5 1.1 62.3 8.24 29.8 2451 34.08
BH#A/34 R–I 1122.95 24.5 2.2 63.7 11.40 28.4 2653 49.82
BH#B/26 R–I 1162.08 21.3 1.5 64.9 6.41 – – –
BH#B/27 R–I 1162.68 24.5 1.5 64.9 9.89 25.3 3352 50.60
BH#B/28 R–I 1163.89 36.2 1.9 64.9 8.63 – – –

moisture content (M) with depth for Raniganj Coalfield supports the %Moisture (high) = − 0.0021 × Depth + 4.8311 (2)
generally accepted trend (Fig. 3). The relationship is moderate for the
studied coals (R2 = 0.61, linear) and is statistically significant as P < %Moisture (low) = 0.0031 × Depth + 4.0850 (3)
0.05. The uncertainty related to moisture content was captured by
establishing low, mid, and high case trends with depth. The linear fit These low, mid and high set of moisture data were used for the
trend line equation reasonably defines the moisture variation with conversion of dry-ash free estimates of Langmuir volume (VL) and gas
depth. Whereas, low and high values of moisture are generated from the content to in situ estimates.
trendline equation of 90% lower and 90% upper confidence band
respectively (Table 2). Hence, the mid, high value of the moisture for 3.2. Gas content
each sample has been calculated with the below equations respectively
(Eqs. (1)–(3)): In situ gas content of coal is defined as the volume of gas that remains
% Moisture (mid) = − 0.0026 × Depth + 4.4581 (1) in a unit mass of coal. Based on the mechanism of gas storage it is the
sum of the adsorbed and free gas content. The gas content (GC) generally

5
S. Chattaraj et al. Journal of Natural Gas Science and Engineering 92 (2021) 103969

Table 2
Regression analyses of the input variables of the model.
Parameters Plots/Co-relations

Depth vs Moisture Depth vs Depth vs Temperature vs Langmuir Pressure Depth vs

Gas Content Langmuir Volume Permeability

Data Points/Observations (n) 51 51 37 37 13

Correlation Coefficient (R) 0.78053 0.65841 0.75580 0.83800 0.66000
Regression Coefficient (R2) 0.60923 0.43350 0.57123 0.70225 0.43560
P-value Intercept 0.00000 0.00630 0.00000 0.00000 0.00039
X Variable 0.00000 0.00000 0.00000 0.00000 0.01410
Coefficients Intercept 4.45808 2.16358 18.37318 13212.98200 1.04453
X Variable − 0.00260 0.00620 0.00740 − 170.69671 − 0.00081
Standard Error Intercept 0.22250 0.75842 0.83707 944.62680 0.20818
X Variable 0.00030 0.00101 0.00108 18.78778 0.00028
95% CB lower level Intercept 4.01094 0.63950 16.67380 11295.29000 0.58633
X Variable − 0.00320 0.00420 0.00520 − 208.83800 − 0.00140
95% CB upper level Intercept 4.90522 3.68770 20.07250 15130.68000 1.50270
X Variable − 0.00200 0.00820 0.00960 − 132.55600 − 0.00020
90% CB lower level Intercept 4.08500 0.89210 – – –
X Variable − 0.00310 0.00450 – – –
90% CB upper level Intercept 4.83110 3.43510 – – –
X Variable − 0.00200 0.00790 – – –
80% CB lower level Intercept – – – – 0.76070
X Variable – – – – − 0.00100
80% CB upper level Intercept – – – – 1.32840
X Variable – – – – − 0.00040
70% CB lower level Intercept – – 17.49300 12219.22000 –
X Variable – – 0.00630 − 190.46200 –
70% CB upper level Intercept – – 19.25380 14206.75000 –
X Variable – – 0.00854 − 150.93200 –

CB=Confidence Band.

Fig. 4. Relation between depth and gas content for the Raniganj coals.

as P > 0.05, In this paper, authors have attempted to capture the sparse
Fig. 3. Relation between depth and moisture content for the Raniganj coals.
distribution with low and high case trends with depth from the trend line
equation of lower and upper limits of the 90% confidence band
increases with depth. Faiz et al. (2007) observed an increasing trend in respectively (Fig. 4, Table 2). The trend line equations for the calculation
gas content up to the depth of 600 m and a slightly decreasing trend in of mid, high, and low gas content are as follows (Eqs. (4)–(6)):
gas content for the further increase of depth up to 900 m for the Sydney
basin coals. A recent study by Kędzior (2019) found a good positive Gas Content(mid) = 0.0062 × Depth + 2.1636; (4)
relationship between gas content and depth for the Silesian basin, polish
coal. The previous study by the authors (Mohanty et al., 2018) has Gas Content (high) = 0.0079 × Depth + 3.4351; (5)
shown an increasing trend of gas content with depth. Although the
relationship found between depth and gas content for the studied coal is Gas Content (low) = 0.0045 × Depth + 0.8921 (6)
moderate (R2 = 0.43, linear), the relationship is statistically significant The depth related to gas contents taken on daf basis is plotted for

6
S. Chattaraj et al. Journal of Natural Gas Science and Engineering 92 (2021) 103969

establish the low, mid, and high case trends of Langmuir volume for the
seams penetrated through the borehole BH#B (Fig. 5). Authors have
found a statistically significant moderate positive linear trend for both
studied Raniganj (R2 = 0.57, P < 0.05) coals. Mid values of the Langmuir
volume for the corresponding depth have been derived from the
trendline equation of the actual data and the equation is as follows (Eq.
(7)):
Langmuir Volume (mid) = 0.0074 × Depth + 18.373 (7)
Based on the data trend, a 70% confidence band was chosen for the
depth vs Langmuir volume plot (Table 2) and the equations generated
from the regression analysis are as follows (Eqs. (8) and (9)).
Langmuir Volume (high) = 0.0085 × Depth + 19.2358 (8)

Langmuir Volume (low) = 0.0063 × Depth + 17.4926 (9)

For the model compatibility, it has been multiplied by the coal
density to convert the Langmuir volume into a cc/cc unit.

3.3.2. Langmuir pressure

Fig. 5. Relation between depth and Langmuir volume for the Raniganj coals.
The globally established relations between methane adsorption ca­
pacities (in terms of Langmuir parameters) with reservoir parameters
determining low, mid, and high cases. Subsequently, the in situ estimates are neither very conclusive nor definite. Most of the explanations are
of gas content calculated using ash and moisture values of the corre­ based on the results of experiments performed under different condi­
sponding samples are used as input to the GEM model. The relationship tions for different basins (Chattaraj et al., 2016). Zhao et al. (2001); Pini
established is shown in Fig. 4. et al. (2010); Zhang and Tao (2011) and Lin et al. (2012) have concluded
that VL decreases but PL increases with increasing temperature. How­
3.3. Adsorption isotherms ever, on the other hand, Chen et al. (1995) and Lin et al. (2014) found
that both VL and PL decrease with increasing temperature, which is
Depth vs Langmuir volume (VL) and temperature vs Langmuir similar to the present study (Fig. 6). On the contrary, Wang et al. (2012)
pressure (PL) are plotted to calculate the low, mid and high values of the and Hao et al. (2014) established that both VL and PL increase with
VL and PL. As both Langmuir volume and Langmuir pressure are deter­ increasing temperature. A strong negative, statistically significant rela­
mined through the adsorption experiment, the same confidence interval tionship has been found between the reservoir temperature and Lang­
has been used for both the parameters. muir pressure (R2 = 0.71, P < 0.05). Equations for lower and upper
limits of 70% confidence band equation used for generating the low and
3.3.1. Langmuir Volume high Langmuir pressure value for each seam and the equations are as
Moisture decreases and carbon increases with the increase in matu­ follows (Eqs. (10) and (11)).
rity corresponding to the depth. The adsorption of methane in coal is Langmuir Pressure (high) = − 150.932 × Temperature + 14206.75 (10)
mainly micropore dominant. With the increase in coal maturity
aromaticity increases, which leads to the conversion of macropores and and
mesopores into micropores, thereby increase in the adsorption capacity
(Liu, 2010; Bandopadhyay and Mohanty, 2014 ; Guo and Guo, 2018). Langmuir Pressure (low) = − 190.46 × Temperature + 12219.22 (11)
The Langmuir volume (on daf basis) is plotted against depth to
3.4. Methane saturation in coal

Coal is said to be saturated if the amount of gas held within its matrix
is the same as the capacity of coal to adsorb the gas at the same PT
condition. The level of undersaturation increases as the measured gas
content falls below the adsorption isotherm curve. At any pressure, the
maximum gas content the coal can hold is given by (Eq. (12)):
VL P
[GC]max = (12)
PL + P
Coal saturation is described as the ratio of [GC]measured to [GC]max .
Mathematically (Eq. (13)),
[GC]measured
Saturation = (13)
[GC]max

The less the saturation level, more the dewatering is needed to

reduce the reservoir pressure to the critical desorption pressure when
the gas desorption starts from the coal. However, despite the highly
under saturated nature of Rocky Mountain basin coals, it is one of the
major CBM producing basins in the world (Hanson et al., 1996).
In Raniganj Coalfield, the observed coal saturation has a wide range
Fig. 6. Relation between reservoir temperature and Langmuir pressure for of variation – ranging from 27% to 100% (Table 1, Fig. 7). The coal
Raniganj coals. saturation data are tabulated in Table 1. The available data on coal

7
S. Chattaraj et al. Journal of Natural Gas Science and Engineering 92 (2021) 103969

Fig. 7. Coal saturation shown with respect to the adsorption isotherm for the (a) seam R–I (b) seam R–II (c) seam R–III (d) seam R–IV (e) seam R–V (f) seam R–VI (g)
seam L-2 (h) seam L-2A (i) seam R–VII (j) seam R–VIII (k) seam R-IX of the Raniganj Coalfield.

saturation suggests that the majority of the observed saturation values time it is pulled out of the hole. The pressure/temperature data thus
lie in the range below 50%. Based on the range of variation on coal recorded during injection and fall-off periods goes into the pressure
saturation, the high case is described with gas content which makes coal transient analysis. The interpretation of injection/fall-off test (ΔP vs Δt,
)
100% saturated (Fig. 7 (h), Table 1). The low case on coal saturation is
and t ∂∂Pt vsΔt gives vital information on permeability, skin and reservoir
described with the low case gas content trend with depth.
boundaries. Since typical CBM reservoirs are initially saturated with
water, the injection/fall-off tests ensure that the flow within the reser­
3.5. Permeability voir remains single-phase flow, and thus conventional pressure transient
analysis techniques can be applied to estimate permeability and skin of
Coal porosity and permeability are the dominant factors controlling the CBM reservoirs.
methane storage and production (Shi et al., 2014; Jena et al., 2018; In coalbed methane reservoirs, the permeability decreases with
(Mohanty, 2020)). Coal permeability is a function of multiple parame­ depth. As overburden load increases with depth, effective stress in­
ters such as geological setting, depth, coal rank, development of natural creases, thereby decreasing the cleat aperture width as the joints and
fractures, etc. in the coal reservoir (Fu et al., 2001; Chen and Zhang, fractures are compacted. The degree of closure of cleat aperture with
2007; Meng et al., 2010; Liu et al., 2012). The in situ permeability of the depth depends upon the cleat compressibility. In other words, the degree
coal seams was determined through pressure transient analysis of of permeability loss with depth is a function of cleat compressibility in a
injection/fall-off test conducted in coreholes by the CBM operators in normal stress environment. Liu et al. (2016) found lower permeability
Raniganj Coalfield. The target coal formation is isolated with inflatable from the well test in the coals from a greater depth than the shallower
packers and water is injected using pumps to increase the formation ones in the southern Qinshui Basin, China. A modeling study by Olufemi
pressure which is then allowed to fall off. The packer assembly contains et al. (2004) has found that relative permeability is the most influencing
downhole shut-in equipment to shut the well after the completion of parameter for gas production for the ECBM project, especially at the
injection period. The downhole shut-in arrangement is useful in mini­ early stages. Pan and Connell (2012) presented the behavior of the ab­
mizing the test period dominated by wellbore storage effects. The solute permeability of coal is central to a range of gas migration pro­
memory gauges run with packer assembly to record the pressure and cesses of methane. Injection/fall-off tests were carried out in six
temperature downhole from the time the packer is run-in-hole to the

8
S. Chattaraj et al. Journal of Natural Gas Science and Engineering 92 (2021) 103969

Fig. 7. (continued).

Table 3
In situ permeability of Raniganj coal seams with the depth.
Depth Seam In situ Permeability (md)

336.50 R–VIII 7.120

418.90 R–VII 5.220
447.90 R-L2A 4.220
472.10 R-L2 3.250
661.00 R–VI 2.040
868.00 R–VI 3.980
624.70 R–V 6.412
903.00 R–V 0.790
769.20 R–IV 3.403
999.00 R–IV 3.650
854.10 R–III 1.913
915.20 R–II 2.276
1006.60 R–I 1.010

boreholes. The measured values of in situ permeability lie between 0.790

and 7.120 mD (Table 3). These permeability values are representative of
average in situ permeability at reservoir scale manifesting the combined
effect of coal matrix and fracture networks. Detailed regression analysis Fig. 8. Variation in in situ permeability with the depth.
reveals a statistically significant moderate decreasing exponential trend
for permeability (k) with depth having a regression coefficient value of According to the data trend, 80% confidence upper and lower band
0.44 (exponential fit). Mid values of the in situ permeability of the cor­ were chosen for the depth vs in situ permeability plot for the determi­
responding depth has been derived from the trendline equation of the nation of the high and low value of permeability respectively (Table 2)
actual data and the equation is as follows (Eq. (14)): and the equations generated from the regression analysis are as follows
(Eqs. (15) and (16)).
Permeability (mid) = 1.1080E + 01e− 1.866677E− 03×Depth
(14)

9
S. Chattaraj et al. Journal of Natural Gas Science and Engineering 92 (2021) 103969

Table 4 useful in quantifying the value of constant (C). The constant (C) value
Cleat spacing according to field-scale value. can be adjusted to achieve a reasonable value of cleat porosity, from the
Cleat Spacing (mm) above equation, which is adequate to produce the same level of water
production that is observed in pilot wells.
Low Value Mid Value High Value
In Raniganj Coalfield, the cleat spacing measured during the field
22 27 35 study varies between 22 and 33 mm. Based on this variation of the cleat
spacing low, mid and high value of the porosity and permeability rela­
tion is generated. The values are tabulated in Table 4.
Permeability (high) = 21.2997e− 0.0010×Depth
(15)
In this paper, based on the coal geometry established from field
studies, authors have used below porosity-permeability equation (in
Permeability (low) = 5.7630e− 0.0030×Depth
(16)
field units) (Eq. (19)):
The data permeability data used for the depth correlation is pre­ [ ]
sented below (Fig. 8): k
(19)
1
φ= 3
4150a2
3.6. Porosity-permeability relationship In the above equation, permeability (k) is in mD (milli Darcy) and
cleat spacing (aa) in mm (millimetres).
Fracture porosity value was calculated from the porosity- Using above porosity-permeability relationship, cleat permeability
permeability relationship developed by Reiss (1980) and modified by values have been transformed into cleat porosity value. Depending upon
Robertson (2005). Reiss (1980) developed a conceptual model for the the variation of cleat spacing, low, mid, and high case porosity-
determination of cleat permeability from the cleat spacing and porosity permeability relations have been generated (Fig. 9).
data by assuming a matrix-block geometry described as a bundle of
vertical matchsticks guided by uniaxial stress system. Later, Robertson
(2005) has modified the relationship by assuming coal block as cubic 3.7. Rock compressibility
geometry under biaxial or hydrostatic confining pressures. The studies
have revealed that both conceptualizations – matchstick and cubical Generally, the laboratory measurement of rock compressibility in
arrangements – lead to cubic exponent relationship of porosity (ϕ) to CBM reservoir is limited and mostly, CBM operators determine this
permeability (k). Mathematically, it can be expressed as (Eq. (17)): value through iterative productivity matching techniques when
adequate production history is available. In this paper, the authors have
k = Cϕ3 a2 (17)
used an innovative approach for determining the range of rock
where, a is the cleat spacing; constant (C) depends upon the arrange­ compressibility by analyzing the data generated for permeability with
ment of coal matrix being considered. respect to depth. It is assumed that coals at most depths underwent
Given the fact that permeability of coal seam is a measurable similar cooking processes resulting in similar initial cleat dimensions.
quantity with well tests and cleat spacing may be known from outcrop Subsequently, the cleat closure happens with burial leading to decrease
studies, the above relationship can be used to determine cleat porosity in permeability. It is globally observed in CBM reservoirs that perme­
(Eq. (18)) i.e., ability decreases logarithmically with depth (Bandyopadhyay et al.,
[ ] 2020).
k 13 Effective stress (σe ) at any depth can be expressed as the difference
ϕ= (18)
Ca2 between overburden stress (σob ) and pore pressure (Pp ). Mathematically,
The value of cleat porosity, for a known permeability and cleat σ e = σob − Pp (20)
spacing, obtained from the above relation is dependent upon the value
of constant (C) being used, thus depends upon the arrangement being Assuming lithostatic gradient of 1 psi/ft and hydrostatic gradient of
considered. Considering that the coal is a highly heterogeneous reser­ 0.433 psi/ft, the above expression reduces to
voir, any assumption on a uniform arrangement will have its drawbacks. σ e = σob − Pp = (1 − 0.433) × Depth(ft) = 0.567 × Depth(ft) (21)
Therefore, the determination of constant (C) is a challenge. However,
the availability of water production data from pilot wells, if any, is Cleat compressibility is defined as fractional change in porosity with
change in stress
1 ∂ϕ
cf = − (22)
ϕ ∂σe

∫φ2 σe )2
(∫
dφ
= Cf ∂σ e (23)
φ
φ1 (σ e )2

φ2 { }
= exp − 0.567cf (d2 − d1 ) (24)
φ1

where depth d1 and d2 are in feet.

Assuming that Permeability is proportional to the cube of porosity,

Table 5
Rock compressibility estimation.
Depth Permeability (mD) [derived from Rock Compressibility (per psi)
(m) mid-case equation Eq (14)] [calculated from Eq (26)]

400 3.1 1.75E-04

1000 1.7
Fig. 9. Porosity-permeability relationship for the Raniganj coals.

10
S. Chattaraj et al. Journal of Natural Gas Science and Engineering 92 (2021) 103969

Table 6 relative permeability by injecting the gas and water simultaneously at a

Assumptions for Corey’s model. constant rate is tedious. Corey (1954) had derived an empirical equation
Parameters Values for the calculation of relative permeability for the conventional reservoir
system. The equation is as follows (Eqs. (27)–(29))
Irreducible water saturation, Siw 20%
Critical gas saturation, Sgc 5% [ ]n
krg = (krg )max 1 − S∗w g (27)
Max. relative permeability for water, Krwmax 1.0
Max. relative permeability for gas, Krgmax 1.0 ( )n
Exponent of the relative permeability curve shape for gas, ng Mid High Low krw = (krw )max S∗w w (28)
2 − 3.5 2.5
Exponent of the relative permeability curve shape for water, Mid High Low Sw − Siw
nw 2 − 3.5 2.5 S∗w = (29)
1 − Siw

where, krg = Relative permeability to gas at the irreducible water satu­

and that the reduction in porosity due to stress increase follows the
normal compressibility formulation: ration, dimensionless; S*w = Normalized water saturation, dimensionless;
n = Normalized gas saturation exponent, dimensionless; Sw = Water
′
( )13
k2 φ {
= 2 = exp − 0.567cf (d2 − d1 )
}
(25) saturation, fraction, Siw = Irreducible water saturation; ng = Exponent
k1 φ1 governing the relative permeability curve shape for gas; nw = Exponent
[( )13 ] governing the relative permeability curve shape for water.
In CBM, if cleat aperture size is large, gas and water phases can flow
Ln kk21
in the cleat without significantly interfering with each other. This is
cf = − (26)
0.567(d2 − d1 ) corresponding to the case where ng = nw = 1. the resulting relative
permeability curves are straight lines. In the absence of experimental
The above equation suggests that from any given straight section of
data, fracture relative permeability is usually assumed linear functions
the log perm vs depth plot, an effective compressibility can be calculated
of wetting phase saturation. On the other hand, when cleat apertures are
independent of knowing the porosity. In order to get an estimate of rock
very small, wall roughness and tortuosity can affect fluid flow. In this
compressibility, the permeability value has been calculated at two
case, it is reasonable to assume that two or more flowing phases may
arbitrarily selected depths, which are then used in Eq (26) to determine
interfere with one another as if they were confined to the pore space of
the rock compressibility value. Table 5 summarizes the process. The
an intergranular porous medium. The resulting fracture relative
approach used above indicates that the rock compressibility is in the
permeability curves will be nonlinear functions of wetting phase satu­
order of 10− 04.
ration.
In Raniganj Coalfield, there is no experimental data available on
3.8. Relative permeability relative permeability curves. In this regard, the authors have attempted
to capture the uncertainty in relative permeability by defining a range
The best way to establish a representative relative permeability on the variation of relative permeability curves. The low, mid, and high
curve for CBM reservoirs is through history matching. However, due to cases were generated by varying the value of ng and nw from 1 to 3 with
non-availability of the adequate field data, the history matching exercise ng as the mid-value of 2 (Fig. 10). This variation in ng and nw captures a
is beyond the scope of the present study. Hence, the relative perme­ wide range of distribution and uncertainties that prevail in CBM
ability was modeled using the Corey’s correlation (Corey, 1954). The reservoirs.
relative permeability curve was generated using parameters used in The impact of connate or irreducible water saturation on the pro­
Corey’s equation. The lists of assumptions for Corey’s model are out­ duction behaviour of CBM well was also investigated. However, it was
lined in Table 6. observed that it has a minimal impact during the very late stage of
CBM production involves the dewatering of the coal seam until the production. The observation is as expected because it takes considerable
reservoir pressure reaches the gas saturation level. Hence, the interplay
between water and gas occurs within the coal cleats, which affects the
Table 7
mobility of both water and gas, making relative permeability as one of
Model description and assumptions.
the main influencing parameters of CBM production. Measurement of
Parameters Description/Assumption

Grid system Cartesian grid system with I × J × K = 31 × 31 ×

12
Grid dimension Each grid in x-direction = 18.35 m
Each grid in y-direction = 18.35 m
Layers in the model 12
Fracture porosity (ф) Derived from the porosity-permeability
relationship
Initial reservoir pressure, psi Derived from pressure vs. depth correlation
Gas content, m3/t Derived from depth vs. gas content correlation
Langmuir volume, m3/t Derived from depth vs VL correlation
Langmuir pressure, psi Derived from temperature vs PL correlation
Rock compressibility, psi-1 1.75E-04
Coal desorption time, days Derived from the canister desorption analysis
Gas composition 100% CH4
Permeability, mD Derived from depth vs perm correlation
Perm anisotropy PERMX=PERMY (no perm anisotropy)
Anisotropy permeability ratio, 0.1
Kv/Kh
Relative permeability Curves generated using Corey’s method
Initial water saturation, Swi 100%
Reservoir temperature Derived from depth
Coal density, g/cc 1.4 (average value of density measured for all the
Fig. 10. Relative permeability curve for low, mid and high cases using Cor­
samples)
ey’s model.

11
S. Chattaraj et al. Journal of Natural Gas Science and Engineering 92 (2021) 103969

Table 8
Input parameters for model run.
Sample No. Seam Thickness, Depth, Langmuir Volume (VL, ar, cc/gm) Langmuir Pressure (PL, kPa) Fracture Pressure Sorption Time

m m Mid High Low Mid High Low (Pf, kPa) (τ, days)

BH#B/02 R-X 2.3 254.12 15.9 16.9 14.9 6791 8948 4634 2489 2.99
BH#B/03 R-IX 1.3 313.79 13.4 14.2 12.4 6485 8686 4284 3074 2.16
BH#B/08 R–VIII 4.0 428.75 16.9 18.1 15.7 5896 8182 3611 4200 3.28
BH#B/11 R–VII 2.8 505.69 15.9 17.4 14.4 6009 8279 3740 4954 2.8
BH#B/L2A L-2A 1.3 537.5 16.8 18.0 15.5 5339 7705 2974 5265 3.66
BH#B/12 L–2 3.4 566.12 16.7 18.3 15.1 5873 8162 3584 5545 2.62
BH#B/15 R–VI 1.9 668.11 15.9 17.3 14.7 4671 7133 2209 6544 5.2
BH#B/18 R–V 1.7 756.69 18.0 19.8 16.2 4729 7183 2276 7412 5.06
BH#B/20 R–IV 5.6 916.17 21.2 23.4 19.0 4251 6774 1729 8974 4.15
BH#B/22 R–III 0.9 1019.88 20.9 23.3 18.6 3722 6321 1124 9990 2.51
BH#B/24 R–II 1.5 1118.1 19.4 21.7 17.4 3398 6043 753 10,952 3.52
BH#B/27 R–I 5.3 1162.68 20.0 22.3 17.7 3159 5838 480 11,389 4.27

Sample No. Seam Gas Content (ar, cc/gm) Fracture Permeability, mD Relative Permeability Skin

Mid High Low Mid High Low Mid High Low Mid High Low

BH#B/02 R-X 2.60 3.81 1.41 6.89 16.5 2.689

BH#B/03 R-IX 3.04 4.40 1.69 6.17 15.6 2.248
BH#B/08 R–VIII 3.18 4.53 1.84 4.98 13.9 1.592
BH#B/11 R–VII 3.66 5.18 2.18 4.31 12.8 1.264
BH#B/L2A L-2A 4.11 5.80 2.46 4.06 12.4 1.149 ng ¼ 2 ng ¼ 1 ng ¼ 3 ¡1 ¡3.5 2.5
BH#B/12 L–2 4.02 5.66 2.42 3.85 12.1 1.055 nw ¼ 2 nw ¼ 1 nw ¼ 3
BH#B/15 R–VI 5.06 97.06 3.11 3.18 10.9 0.777
BH#B/18 R–V 4.99 6.93 3.10 2.70 9.99 0.595
BH#B/20 R–IV 5.69 7.84 3.61 2.00 8.52 0.369
BH#B/22 R–III 6.88 9.41 4.40 1.65 7.68 0.270
BH#B/24 R–II 6.26 8.56 4.03 1.37 6.96 0.201
BH#B/27 R–I 7.24 9.87 4.68 1.26 6.66 0.176

Fig. 11. (a) 3D view of the borehole locations along with the seam disposition and (b) 3D view of the depth-wise seam sequence in the modeled borehole.

production years for coal reservoirs to realize its irreducible water

saturation. Therefore, in this study, the uncertainty in relative perme­
ability is captured by defining a range of exponents only.

3.9. Sorption time

Coalbed methane (CBM) reservoirs are naturally fractured forma­

tions, comprising both permeable cleats and matrix blocks. The inter­
action between cleats and matrix describes the rate of flow from a matrix
element into the cleat system in response to a methane concentration
gradient. Gas diffusivity in the matrix is represented by a function of
shape factor, gas desorption time, and reservoir pressure. Time taken to
desorb 63.2% of the total gas, generally reported in hours and days is
known as sorption time (τ). It is an important input for many coal gas
Fig. 12. Grid pattern of the modeled reservoir. reservoir simulators to observe the well performance. It is also used to
calculate 90–95% desorption time and it depends upon the several
reservoir properties. The sorption time has been calculated from the in

12
S. Chattaraj et al. Journal of Natural Gas Science and Engineering 92 (2021) 103969

Fig. 13. (a) Range of overall gas production rate from 12 seams of Raniganj Coalfield obtained through the CMG-GEM simulator (b) Simulated gas rate of 10 cases
closest to the actual gas production rate.

Fig. 14. (a). Range of overall water production rate from 12 seams of Raniganj Coalfield obtained through the CMG-GEM simulator. (b) Simulated water rate of 10
cases closest to the actual water production rate.

Fig. 15. Range of overall cumulative (a) gas (b) water production from 12 seams of Raniganj Coalfield obtained through the CMG-GEM simulator along with the
actual gas and water production plots, respectively.

situ gas content data. The CBM simulator GEM models the matrix-cleat 4. Reservoir modeling and simulation
interactions by incorporation of sorption time, which represents the
combined effect of shape factor and diffusion co-efficient. Chen et al. The coal seams are modeled as a dual-porosity medium with time-
(2013) suggested that desorption time plays a key role during the early dependent, non-linear desorption of methane from the coal matrix
stage of production, but diminishes when the adsorption phase takes described by isotherms relating to matrix gas content and pressure. The
over the dominant role. For the studied Raniganj coals, the sorption time Cartesian model was 31 £ 31 blocks each with a length of 18.35 m to
varies from 2.16 to 5.2 days (Table 8). represent an 80 acres drainage area. GEM software features a range of
dual porosity and dual permeability techniques capable of modeling
both coal and shale gas reservoirs. GEM includes options for gas sorption
in the matrix, gas diffusion through the matrix, two-phase flow through

13
S. Chattaraj et al. Journal of Natural Gas Science and Engineering 92 (2021) 103969

( )
βc kx Ax krg
Tgxi− 1,i = Tgxi,i+1 =
μg Bg △x i
= Gas − phase transmissibility between grid block i ± 1 and i

( )
βc kx Ax krw
Twxi− 1,i = Twxi,i+1 =
μg Bg △x i

= Water phase transmissibility between grid block i ± 1 and i

qm = Source term for the ith block = Flow rate of gas desorbing from coal
matrix and coming into cleat system.
In the matrix system, the gas is present in the adsorbed state which
desorbs into the cleat system when pressure depletes. Therefore, the coal
matrix acts as source for the cleat system. The flow in the matrix is
governed by the Fick’s law which relates the flow rate of gas with the gas
concentration gradient.
According to Fick’s law (Eq. (32)),
Fig. 16. Effects of the parameters on estimated ultimate recovery (EUR) from ( )
wells drilled in Raniganj Coalfield. qm = Vm σDc Cm − Cf (32)

the natural fracture system. Hence, GEM reservoir simulator has been where, Vm = Matrix Volume, m3 ; σ= ShapeFactor, m− 2 Dc =
used for this study. The twelve layers in the model represent the twelve Diffusion ​ m2
coefficient, day ; Cm = Average ​ Matrix ​ Gas​ Concentration, m3 ofm​ coal;,
3

seams (Fig. 11). The model doesn’t include the interburden layers be­
Cf = Equilibrium​ Matrix ​ Gas​ Concentration, m3 ofm​ coal
3

tween the coal seams. The model layout describes the producing coal
Desorption Time is given by (Eq. (33)),
seams as stacked layers – which is consistent with the existing comple­
tion strategy in Raniganj block CBM wells wherein the wells are 1
τ= (33)
completed in coal seams only with no perforation in the interlayers. The σDc
twelve seams are stacked on each other and individual seam properties
Hence Eq. (32) can be written as (Eq. (34)),
have been taken into consideration for the input of CMG-GEM simulator.
As only the targeted (all 12 seams) seams have been perforated, inter­ Vm ( )
qm = Cm − Cf (34)
layer properties between the two seams have not been considered. τ
However, production profiles of BH#B bore well shown for both gas and Cm and Cf are evaluated by the reservoir simulators at different time
water (Figs. 13–16) are combined from all the seams. The model is set up steps according to the Langmuir equation (Eqs. (35) and (36))
to represent a single well producing under commingled completion
( )n+1
strategy. Model description and assumptions are described in Table 7. VL Pg
The basic equations which govern the flow through grids of a CBM Cf = ( )n+1 (35)
PL + Pg
reservoir simulation model are – 1) mass conservation equation, 2)
momentum conservation equation in the form of Darcy’s law in the coal ( )n
VL Pg
cleat system, 3) Fick’s law driving gas flow from coal matrix to cleat Cm = ( )n (36)
PL + Pg
system, 4) Real gas law equation, and 5) pore volume compressibility
equation defining cleat porosity changes with pressure. ( )n+1
CBM reservoirs, being a dual-porosity system, are conceptualized in Pg = Pressure of gas phase in a grid at the current time level
the form of a matrix-cleat system arrangement. The basic mass balance ( )n
equation applies in the cleat system where the matrix system acts as an Pg = Pressure of gas phase in a grid at the previous time level
additional source for the cleat system. The example is illustrated below The model set up with the above reservoir and well parameters are
for a 1D reservoir discretized in grid blocks with grid dimensions Δx and run with two operating constraints – 1) water rate of 100 m3/day as a
Δy in the x and y direction respectively (Fig. 12). primary constraint and 2) minimum flowing bottom-hole pressure of
The application of the mass balance equation for the ith block in the 200 kPa as a secondary constraint. The water constraint of 100 m3/day
cleat system gives the flow equations for gas and water phase as below. in the model is mainly driven by the average water production observed
Reservoir simulators apply the mass balance equation in time steps for in the currently existing CBM wells in the area. During this study, 243
every grid. In the example below, the mass balance equation is written numbers of simulation runs were made to study the influence of several
for ith block in the cleat system for the flow of fluids during the time parameters on the performance of CBM wells based on input data
between (n+1)th and nth time step (Eqs. (30) and (31)). summarized in Table 8.
[( )n+1
[( )] [( )] Vb i φSg
Tgxi− 1,i Pgi− 1 − Pgi + Tgxi+1,i Pgi+1 − Pgi − qg + qm =
αc △t Bg i 5. Results and discussion
( )n ]
φSg
− A total of five (05) key uncertainty parameters were considered to
Bg i evaluate the production characteristics of the CBM wells in Raniganj
(30) Coalfield – 1) adsorption isotherm, 2) gas content, 3) permeability, 4)
[( )n+1 ( )n ] relative permeability, and 5) skin. The basis for the high and low cases of
Twxi− 1,i [(Pwi− 1 − Pwi )] + Twxi+1,i [(Pwi+1 − Pwi )] − qw =
Vbi ϕSw
−
ϕSw each parameter is explained above in section 3 on Reservoir Charac­
αc △t Bw i Bw i terization. Single well models were set up in CMG-GEM simulator to
(31) represent every possible scenario on these five key uncertainty param­
eters (Table 12). A high and low case on each parameter causes the
where, possibility of the existence of three (03) scenarios on each parameter –

14
S. Chattaraj et al. Journal of Natural Gas Science and Engineering 92 (2021) 103969

low, mid and high. Therefore, there exists a possibility of a total of 243 Table 10
(3 × 3 × 3 × 3 × 3 = 243) scenarios with a combination of all the un­ Case matrix for 10 closest simulation runs to the actual production from Well-1.
certainty parameters. The case matrix containing all possible 243 sce­ Isotherm Gas Permeability Relative Skin
narios is presented in Table 12. The low, mid, and high cases on each Content Permeability
parameter are represented by nomenclature (− 1), (0), and (1) Case-150 0 1 0 0 1
respectively. Case-153 0 1 0 1 1
The single well models on each of 243 cases were run in prediction Case-155 0 1 1 − 1 0
mode in CMG-GEM for a production period of 25 years. The model re­ Case-157 0 1 1 0 − 1
Case-160 0 1 1 1 − 1
sults from 243 cases were extracted and presented in the form of gas Case-162 0 1 1 1 1
rate, water rate, and cumulative gas and cumulative water in Case-228 1 1 0 − 1 1
Figs. 13–16. Therefore, the average monthly production was calculated Case-231 1 1 0 0 1
for the first 25 years to explain the temporal characteristics and trends Case-234 1 1 0 1 1
Case-235 1 1 1 − 1 − 1
for CBM wells of Raniganj Coalfield to guarantee high reliability of
comparisons over long time scales. Experimental design, as used in
reservoir simulation, is foremost a sampling process to find a subset of 5.2. Cumulative gas and water production
input variables that can still yield an accurate view of the behaviour of
the output variables. In principle, the parameter matrix is composed of The grey colour shaded area shows the range of gas production
all data that is input into the modeling process (such as gas content, VL, variation obtained from all the 243 cases (Fig. 15a). The uncertainty
PL, permeability, relative permeability, skin etc.). The output, say ulti­ range described by the key parameters has the potential to impact the
mate recovery is then analyzed from the combination cases. The key peak cumulative gas vary from nearly 0.11 to 82.83 MMSCM. Maximum
strength from the application of the methodology is that probability cumulative gas production has been achieved for case-243 and is min­
distributions of the input parameters can be translated in probability imum is for case 1, i.e. overall high case and overall low case respec­
distributions of output parameters to quantify the impact of uncertainty tively. However average cumulative production is 12.33 MMSCM and
on the expected outcomes. 4.18MMSCM for the overall mid-case (Case-122) for the total simulated
well life of 25 years.
The shaded area shows the range of water production variation ob­
5.1. Gas rate and water rate
tained from all the 243 cases (Fig. 15b). The uncertainty range described
by the key parameters has the potential to impact the peak cumulative
The shaded area shows the range of gas production variation ob­
water production varying from 0.007 to 0.11 MMSCM.
tained from all the 243 cases (Fig. 13a). The uncertainty range described
by the key parameters has the potential to impact the peak gas rate to
vary from nearly 2 to 56,201 m3/day. Among all the cases, the maximum 5.3. Effects of factors on gas production
gas rate has been achieved for the overall high case (Case-243) during
the 25th month from the initial date. However, the minimum value of gas CBM productivity is dependent on the combined effect of several
rate shows that gas production starts since the very first day (may be a parameters such as depositional history, rank, gas content, permeability,
very negligible amount) as the dewatering starts in all the cases. etc. (Kaiser et al., 1994). Among these, the individual effect on cumu­
The volume of water contained in the coalbed depends upon the cleat lative gas production of five parameters viz, Langmuir parameters
porosity. It maintains the reservoir pressure and holds the methane gas (Volume and Pressure), gas content, permeability, porosity and relative
as an adsorbed layer within the coal; and finally, during the production permeability on the cumulative gas production has been studied. In
stage, additional formation water may enter into the coal seam. Hence, order to study the influences of these factors on the overall CBM pro­
water plays an important role in CBM storage as well as CBM production duction, CMG-GEM software was run to conduct sensitivity analysis. The
(Su et al., 2005). The water production from all 243 cases varies between cases and their corresponding EUR value estimated from the model are
0 and 129.84 m3/day. The data of water rate are plotted in Fig. 14a, tabulated below (Table 9).
where the grey colour shaded area represents the total range of variation From Fig. 16 it is evident that gas content is the most sensitive factor
in water rate for all the 243 cases. From the data table, it can be inferred controlling CBM productivity, which is similar to the previous study by
that there may remain a possibility where only the gas will be produced, Zou et al. (2010) on the Qinshui Basin coal, China.
no dewatering is required. Figs. 13b and Fig. 14b show the 10 profiles
each of gas rate and water rate, respectively, those are closer to the
actual production data. The input data for Langmuir volume varies from
18.7 to 32.8 cc/cc, Langmuir pressure 3159–8948 KPa, gas content
5.33–13.82 cc/cc and permeability 1.26–16.5mD for the 10 closest
cases.

Table 9
Cases to show the effect of individual parameters on the estimated EUR values.
Case Description Case Number EUR (MMSCM)

Overall Mid Case Case-122(0,0,0,0,0) 04.17768

Isotherm High Case Case-203(1,0,0,0,0) 06.54977
Isotherm Low Case Case-41(-1,0,0,0,0) 02.23751
Gas Content High Case Case-149(0,1,0,0,0) 35.23122
Gas Content Low Case Case-95(0,-1,0,0,0) 00.90237
Permeability High Case Case-131(0,0,1,0,0) 06.37962
Permeability Low Case Case-113(0,0,-1,0,0) 02.45166
Relative Permeability High Case Case-125(0,0,0,1,0) 04.72088
Relative Permeability Low Case Case-119(0,0,0,-1,0) 03.57377
Skin High Case Case-123(0,0,0,0,1) 05.05537
Skin Low Case Case-121(0,0,0,0,-1) 03.49464
Fig. 17. Observed monthly water and gas production from Well-1.

15
S. Chattaraj et al. Journal of Natural Gas Science and Engineering 92 (2021) 103969

Table 11
Summary results of history matching of six closest cases.
Seam Case - Case - Case - Case - Case - Case -
1 2 3 4 5 6

Permeability, R-X 10.6 8.6 8.1 8.6 8.6 8.6

mD R-IX 9.5 7.7 7.9 7.7 7.7 7.7
R–VIII 7.7 6.2 5.7 6.2 6.2 6.2
R–VII 6.6 5.4 5.4 5.4 5.4 5.4
L-2A 6.2 5.1 5.3 5.1 5.1 5.1
L–2 5.9 4.8 5 4.8 4.8 4.8
R–VI 4.9 4.0 4.2 4.0 4.0 4.0
R–V 4.2 3.4 3 3.4 3.4 3.4
R–IV 3.1 2.5 2 2.5 2.5 2.5
R–III 2.5 2.1 1.6 2.1 2.1 2.1
R–II 2.1 1.7 1.2 1.7 1.7 1.7
R–I 1.9 1.6 1 1.6 1.6 1.6
Gas Content R-X 4.1 4.6 4.1 4.1 4.1 3.3
(m3/t) R-IX 4.3 5.0 4.3 4.3 4.3 3.4
R–VIII 5.2 6.1 5.5 5.2 5.2 4.4
R–VII 5.8 6.5 6.2 5.8 5.8 4.9
L-2A 7.7 7.7 8.0 7.7 7.7 6.4
Fig. 18. Well-1: Observed gas production vs Model gas production. The model L–2 7.5 7.9 7.9 7.5 7.5 6.3
is run at the observed water production from Well-1 and history match is R–VI 9.0 10.2 9.3 9.0 9.0 7.5
explored with comparison on model gas rate vs observed gas rate. This figure R–V 9.0 10.2 9.5 9.0 9.0 7.6
compares six (06) cases of the closest match out of several runs. R–IV 10.3 11.0 10.8 10.3 10.3 8.7
R–III 11.9 11.2 12.4 11.9 11.9 9.9
R–II 10.9 11.5 11.3 10.9 10.9 9.0
R–I 12.3 12.0 12.7 12.3 12.3 10.2
Drainage Area, 70 75 80 76 76 80
(Acres)
Skin − 1.5 − 1.5 − 1.5 − 2.0 − 2.5 − 1.5

5.4. Validation of model data

The production data from one well i.e.,Well-1, located in Raniganj

Coalfield has been considered to validate the uncertainty range estab­
lished through the present investigation. The monthly gas and water
production from Well-1 is shown in Fig. 17. The Well-1 has more than
seven (7) years of production history and presents a valid case to draw
essential insights about coal reservoir properties. The production data
from Well-1 was thus used to perform a history matching exercise to
quantify the important coal properties like cleat permeability and gas
content. Several runs were done in CMG-GEM in the course of history
matching. The model was run with the observed water rate constraint
and the match was investigated upon the gas rates from Well-1. Out of
multiple realizations run, the six cases with a close match are presented
Fig. 19. Well-1 - Observed water production vs Model water production. The
in Fig. 18 and Fig. 19. While performing history matching, it was
model is run at the observed water production from Well-1 and thus model
water rate matches very well with the observed water rate. observed that the impact of coal permeability, gas content, and relative
permeability on gas production is more than other parameters. Although
several runs were done varying all parameters, the six cases presented in
Figs. 18 and 19 that closely match the observed production from Well-1
The Tornado plot (Fig. 16) shows the impact of key parameters on
correspond to a variation in coal permeability and gas content with base
the estimated ultimate recovery (EUR). The key finding from this
case assumption on Langmuir isotherm. The cleat porosity is related to
parametric study is that gas content significantly impacts the recovery
coal permeability in the model according to Eq (19). In order to check
factor in the Raniganj Coalfield. Along with gas content, isotherms and
the impact of the drainage area, the size of the model was also varied
permeability have also a significant impact on EUR. The data on
while history matching exercise. It is important to note that the drainage
permeability provides its values to lie within the range defined by low
area in the six (6) cases of the closest match was found to fall in the range
case and high case depth correlations presented in Fig. 8. On the other
from 70 acres to 80 acres – which is closer to the development strategy
hand, the uncertainty range on gas content has a wide variation with the
employed by the operators in the area. Table 11 summarizes the results
possibility of coal saturation to lie between 27% and 100%. Table 1 il­
of history match for the six (6) closest cases.
lustrates that Raniganj coals are having a wide range of saturation dis­
It is important to note that the values of permeability and gas content
tribution with a huge upside potential going towards 100% saturated
obtained from the history matching lie in the band described by low and
case and low downside potential with further undersaturation. In this
high cases on these parameters. The gas rate from Well-1 is highlighted
study, the range of uncertainty related to upside potential on gas content
in Fig. 13a. It is to be noted that the model runs (from 243 cases) shown
is captured by considering 100% saturation in the high case. Table 10
in Fig. 13a are anchored upon the assumption that there is no downtime
representing the case matrix for 10 closest simulation runs to the actual
in the well production. However, the observed production from Well-1
production from Well-1 also conforms that gas content, permeability
has some downtime effects leading to a slight extension of time to
and isotherm are the key drivers.
peak gas rate. The water rate from Well-1 is highlighted in Fig. 14a. A
good degree of agreement exist between model and actual data.

16
S. Chattaraj et al. Journal of Natural Gas Science and Engineering 92 (2021) 103969

Table 12
Design matrix for reservoir simulation runs.
Case AI GC k kr S Case AI GC k kr S Case AI GC k kr S

1 − 1 − 1 − 1 − 1 − 1 41 − 1 0 0 0 0 81 − 1 1 1 1 1
2 − 1 − 1 − 1 − 1 0 42 − 1 0 0 0 1 82 0 − 1 − 1 − 1 − 1
3 − 1 − 1 − 1 − 1 1 43 − 1 0 0 1 − 1 83 0 − 1 − 1 − 1 0
4 − 1 − 1 − 1 0 − 1 44 − 1 0 0 1 0 84 0 − 1 − 1 − 1 1
5 − 1 − 1 − 1 0 0 45 − 1 0 0 1 1 85 0 − 1 − 1 0 − 1
6 − 1 − 1 − 1 0 1 46 − 1 0 1 − 1 − 1 86 0 − 1 − 1 0 0
7 − 1 − 1 − 1 1 − 1 47 − 1 0 1 − 1 0 87 0 − 1 − 1 0 1
8 − 1 − 1 − 1 1 0 48 − 1 0 1 − 1 1 88 0 − 1 − 1 1 − 1
9 − 1 − 1 − 1 1 1 49 − 1 0 1 0 − 1 89 0 − 1 − 1 1 0
10 − 1 − 1 0 − 1 − 1 50 − 1 0 1 0 0 90 0 − 1 − 1 1 1
11 − 1 − 1 0 − 1 0 51 − 1 0 1 0 1 91 0 − 1 0 − 1 − 1
12 − 1 − 1 0 − 1 1 52 − 1 0 1 1 − 1 92 0 − 1 0 − 1 0
13 − 1 − 1 0 0 − 1 53 − 1 0 1 1 0 93 0 − 1 0 − 1 1
14 − 1 − 1 0 0 0 54 − 1 0 1 1 1 94 0 − 1 0 0 − 1
15 − 1 − 1 0 0 1 55 − 1 1 − 1 − 1 − 1 95 0 − 1 0 0 0
16 − 1 − 1 0 1 − 1 56 − 1 1 − 1 − 1 0 96 0 − 1 0 0 1
17 − 1 − 1 0 1 0 57 − 1 1 − 1 − 1 1 97 0 − 1 0 1 − 1
18 − 1 − 1 0 1 1 58 − 1 1 − 1 0 − 1 98 0 − 1 0 1 0
19 − 1 − 1 1 − 1 − 1 59 − 1 1 − 1 0 0 99 0 − 1 0 1 1
20 − 1 − 1 1 − 1 0 60 − 1 1 − 1 0 1 100 0 − 1 1 − 1 − 1
21 − 1 − 1 1 − 1 1 61 − 1 1 − 1 1 − 1 101 0 − 1 1 − 1 0
22 − 1 − 1 1 0 − 1 62 − 1 1 − 1 1 0 102 0 − 1 1 − 1 1
23 − 1 − 1 1 0 0 63 − 1 1 − 1 1 1 103 0 − 1 1 0 − 1
24 − 1 − 1 1 0 1 64 − 1 1 0 − 1 − 1 104 0 − 1 1 0 0
25 − 1 − 1 1 1 − 1 65 − 1 1 0 − 1 0 105 0 − 1 1 0 1
26 − 1 − 1 1 1 0 66 − 1 1 0 − 1 1 106 0 − 1 1 1 − 1
27 − 1 − 1 1 1 1 67 − 1 1 0 0 − 1 107 0 − 1 1 1 0
28 − 1 0 − 1 − 1 − 1 68 − 1 1 0 0 0 108 0 − 1 1 1 1
29 − 1 0 − 1 − 1 0 69 − 1 1 0 0 1 109 0 0 − 1 − 1 − 1
30 − 1 0 − 1 − 1 1 70 − 1 1 0 1 − 1 110 0 0 − 1 − 1 0
31 − 1 0 − 1 0 − 1 71 − 1 1 0 1 0 111 0 0 − 1 − 1 1
32 − 1 0 − 1 0 0 72 − 1 1 0 1 1 112 0 0 − 1 0 − 1
33 − 1 0 − 1 0 1 73 − 1 1 1 − 1 − 1 113 0 0 − 1 0 0
34 − 1 0 − 1 1 − 1 74 − 1 1 1 − 1 0 114 0 0 − 1 0 1
35 − 1 0 − 1 1 0 75 − 1 1 1 − 1 1 115 0 0 − 1 1 − 1
36 − 1 0 − 1 1 1 76 − 1 1 1 0 − 1 116 0 0 − 1 1 0
37 − 1 0 0 − 1 − 1 77 − 1 1 1 0 0 117 0 0 − 1 1 1
38 − 1 0 0 − 1 0 78 − 1 1 1 0 1 118 0 0 0 − 1 − 1
39 − 1 0 0 − 1 1 79 − 1 1 1 1 − 1 119 0 0 0 − 1 0
40 − 1 0 0 0 − 1 80 − 1 1 1 1 0 120 0 0 0 − 1 1
121 0 0 0 0 − 1 162 0 1 1 1 1 203 1 0 0 0 0
122 0 0 0 0 0 163 1 − 1 − 1 − 1 − 1 204 1 0 0 0 1
123 0 0 0 0 1 164 1 − 1 − 1 − 1 0 205 1 0 0 1 − 1
124 0 0 0 1 − 1 165 1 − 1 − 1 − 1 1 206 1 0 0 1 0
125 0 0 0 1 0 166 1 − 1 − 1 0 − 1 207 1 0 0 1 1
126 0 0 0 1 1 167 1 − 1 − 1 0 0 208 1 0 1 − 1 − 1
127 0 0 1 − 1 − 1 168 1 − 1 − 1 0 1 209 1 0 1 − 1 0
128 0 0 1 − 1 0 169 1 − 1 − 1 1 − 1 210 1 0 1 − 1 1
129 0 0 1 − 1 1 170 1 − 1 − 1 1 0 211 1 0 1 0 − 1
130 0 0 1 0 − 1 171 1 − 1 − 1 1 1 212 1 0 1 0 0
131 0 0 1 0 0 172 1 − 1 0 − 1 − 1 213 1 0 1 0 1
132 0 0 1 0 1 173 1 − 1 0 − 1 0 214 1 0 1 1 − 1
133 0 0 1 1 − 1 174 1 − 1 0 − 1 1 215 1 0 1 1 0
134 0 0 1 1 0 175 1 − 1 0 0 − 1 216 1 0 1 1 1
135 0 0 1 1 1 176 1 − 1 0 0 0 217 1 1 − 1 − 1 − 1
136 0 1 − 1 − 1 − 1 177 1 − 1 0 0 1 218 1 1 − 1 − 1 0
137 0 1 − 1 − 1 0 178 1 − 1 0 1 − 1 219 1 1 − 1 − 1 1
138 0 1 − 1 − 1 1 179 1 − 1 0 1 0 220 1 1 − 1 0 − 1
139 0 1 − 1 0 − 1 180 1 − 1 0 1 1 221 1 1 − 1 0 0
140 0 1 − 1 0 0 181 1 − 1 1 − 1 − 1 222 1 1 − 1 0 1
141 0 1 − 1 0 1 182 1 − 1 1 − 1 0 223 1 1 − 1 1 − 1
142 0 1 − 1 1 − 1 183 1 − 1 1 − 1 1 224 1 1 − 1 1 0
143 0 1 − 1 1 0 184 1 − 1 1 0 − 1 225 1 1 − 1 1 1
144 0 1 − 1 1 1 185 1 − 1 1 0 0 226 1 1 0 − 1 − 1
145 0 1 0 − 1 − 1 186 1 − 1 1 0 1 227 1 1 0 − 1 0
146 0 1 0 − 1 0 187 1 − 1 1 1 − 1 228 1 1 0 − 1 1
147 0 1 0 − 1 1 188 1 − 1 1 1 0 229 1 1 0 0 − 1
148 0 1 0 0 − 1 189 1 − 1 1 1 1 230 1 1 0 0 0
149 0 1 0 0 0 190 1 0 − 1 − 1 − 1 231 1 1 0 0 1
150 0 1 0 0 1 191 1 0 − 1 − 1 0 232 1 1 0 1 − 1
151 0 1 0 1 − 1 192 1 0 − 1 − 1 1 233 1 1 0 1 0
152 0 1 0 1 0 193 1 0 − 1 0 − 1 234 1 1 0 1 1
153 0 1 0 1 1 194 1 0 − 1 0 0 235 1 1 1 − 1 − 1
154 0 1 1 − 1 − 1 195 1 0 − 1 0 1 236 1 1 1 − 1 0
(continued on next page)

17
S. Chattaraj et al. Journal of Natural Gas Science and Engineering 92 (2021) 103969

Table 12 (continued )
Case AI GC k kr S Case AI GC k kr S Case AI GC k kr S

155 0 1 1 − 1 0 196 1 0 − 1 1 − 1 237 1 1 1 − 1 1

156 0 1 1 − 1 1 197 1 0 − 1 1 0 238 1 1 1 0 − 1
157 0 1 1 0 − 1 198 1 0 − 1 1 1 239 1 1 1 0 0
158 0 1 1 0 0 199 1 0 0 − 1 − 1 240 1 1 1 0 1
159 0 1 1 0 1 200 1 0 0 − 1 0 241 1 1 1 1 − 1
160 0 1 1 1 − 1 201 1 0 0 − 1 1 242 1 1 1 1 0
161 0 1 1 1 0 202 1 0 0 0 − 1 243 1 1 1 1 1
AI = Adsorption Isotherm; GC = Gas Content; k = permeability; kr = relative permeability; S=Skin

6. Conclusions Credit author statement

The paper depicts the regional variation of key reservoir parameters Sujoy Chattaraj: Laboratory work, simulation and writing, Rajeev
of Raniganj Coalfield. The paper also demonstrates the efficacy of the Upadhyay: Conceptualization, simulation, analysis/interpretation and
analysis-based understanding to define the range of uncertainties asso­ manuscript preparation, Debadutta Mohanty: Conceptualization, over­
ciated with the subsurface parameters and identifying the key perfor­ all supervision, analysis/interpretation, manuscript preparation and
mance indicators of CBM wells in Raniganj Coalfield. The CBM reservoir communication, Gopinath Halder: Writing, reviewing and editing, Tar­
simulation studies have been deployed to quantify the effect of key keshwar Kumar: Writing, reviewing and editing.
uncertainty parameters and describe a range of possible production
profiles from CBM wells of Raniganj Coalfield under a commingled
strategy of 12 stacked seams. The studies presented in this paper find its Declaration of competing interest
application and usefulness in the following ways:
The authors declare that they have no known competing financial
➢ For the first time attempt has been made to translate the core scale interests or personal relationships that could have appeared to influence
reservoir characterization to regional-scale reservoir evaluation. the work reported in this paper.
➢ Cleat compressibility measurement on small core gives unrepresen­
tative values for application over large-scale coal reservoir. This Acknowledgements
paper presents a more scientific approach to infer cleat compress­
ibility from the slope of the straight line log (k) versus the depth plot. DM is thankful to SERB and CSIR-CIMFR for financial support
It is important to note that the approach suggested in this paper will through Grant No. SR/S4/ES- 591/2011 and Grant No. MLP-82/2019-
work well when the effective stress changes primarily as a function of 20, respectively. SC is acknowledging CSIR for funding fellowship
depth, which is appropriate for normally stressed regimes. through Grant No. 31/22 (0028)/2017-EMR-I. Thanks are due to the
➢ The measurement of cleat porosity in the laboratory is also difficult Director, CSIR-Central Institute of Mining and Fuel Research, India for
as the coal samples are too small and brittle to adequately capture his kind permission to publish the paper. The authors are thankful to the
the field-scale cleat network. In this paper, the authors have pre­ anonymous reviewers for their valuable suggestions to improve the
sented a method of estimating cleat porosity in Raniganj Coalfield manuscript.
using the knowledge of cleat spacing and cleat permeability. The
method is useful as its application can give a porosity distribution in References
the CBM reservoir similar to permeability, which otherwise would be
difficult. ASTM D3172, 2013. Standard Practice for Proximate Analysis of Coal and Coke. ASTM
➢ In Raniganj Coalfield, the experimental values of Langmuir volume International, West Conshohocken, PA, p. 2. https://doi.org/10.1520/D3172.
ASTM D3176, 1979. Standard Method for Ultimate Analysis of Coal and Coke [C, H, S, N,
varies between 17.4 and 29.8 cc/gm (daf); Langmuir pressure O, Ash], 409–412. ASTM International, West Conshohocken, PA, p. 4. https://doi.
2451–7827 kPa; gas content 2.07–13.03 cc/gm (daf); and cleat org/10.1520/D3176-15.
permeability 0.79–7.12 mD. Bandopadhyay, A.K., Mohanty, D., 2014. Variation in hydrogen content of vitrinite
concentrates with rank advance. Fuel 134, 220–225.
➢ The uncertainty range on gas content has a wide variation with the
Bandyopadhyay, K., Mallik, J., Shajahan, R., Agarwal, N., 2020. Closing the gap between
possibility of coal saturation to lie between 27% and 100%. Raniganj analytical and measured coal permeability 281, 118752, 1-13.
coals are having a wide range of saturation distribution with a huge Bao, Y., Ju, Y., Yin, Z., Xiong, J., Qi, Y., 2020. Influence of reservoir properties on the
methane adsorption capacity and fractal features of coal and shale in the upper
upside potential going towards 100% saturated case and low
Permian coal measures of the south Sichuan coalfield, China. Energy Explor. Exploit.
downside potential with further undersaturation. 38 (1), 57–78.
➢ From the present investigation, it is inferred that there are three Bertard, C., Bruyet, B., Gunther, J., 1970. Determination of desorbable gas concentration
major reservoir properties where associated uncertainties can influ­ of coal (direct method). Int. J. Rock Mech. Min. Sci. 7, 43–65. https://doi.org/
10.1016/0148-9062(70)90027-6.
ence the production behavior considerably – 1) gas content, 2) shape Chattaraj, S., Mohanty, D., Kumar, T., Halder, G., 2016. Thermodynamics, kinetics and
of isotherm, and 3) coal permeability. The upside and downside modeling of sorption behavior of coalbed methane – a review. Journal of
potential concerning these three reservoir properties is much more Unconventional Oil and Gas Resources 16, 14–33.
Chattaraj, S., Mohanty, D., Kumar, T., Halder, G., Mishra, K., 2019. Comparative study
than the other ones. on sorption characteristics of coal seams from Barakar and Raniganj formations of
➢ The study presented in this paper finds its application in supporting Damodar Valley Basin, India. Int. J. Coal Geol. 212, 103202. https://doi.org/
the CBM operators of Raniganj Coalfield with business decisions 10.1016/j.coal.2019.05.009, 1-19.
Chen, J.G., Zhang, J.F., 2007. Systematic control effect of structures on permeability of
during the early exploration stage – when the unavailability of high coal rank reservoirs. Nat. Gas Geosci. 18 (1), 134–136.
appraisal (or pilot) data presents a major bottleneck in quantifying Chen, C.G., Xiao, X.H., Zhang, D.J., 1995. The dependence of temperature to the
the range of peak gas rate and time to peak gas. Production profiles of adsorption of methane on anthracite coal and its char. Coal Conversion 18 (3),
88–92.
both gas and water show a wide range of variation up to 56,201 m3/
Chen, Z., Liu, J., Kabir, A., Wang, J.G., Pan, Z.J., 2013. Impact of various parameters on
day and up to 129.84m3/day, respectively. the production of coalbed methane. SPE J. 18, 910–923. https://doi.org/10.2118/
162722-PA.
Chen, Y., Liu, D., Yao, Y., Cai, Y., Chen, L., 2015. Dynamic permeability change during
coalbed methane production and its controlling factors. J. Nat. Gas Sci. Eng. 25,
335–346.

18
S. Chattaraj et al. Journal of Natural Gas Science and Engineering 92 (2021) 103969

CMG, 2003. Computer Modeling Group, Inc., “Tutorial: Building, Running, and Liu, T., Liu, S., Liu, B., Fu, X., Zhu, C., Yang, W., Zhao, Y., 2020. Stress response during
Analyzing Coalbed Methane Model Using Builder and GEM”. in-situ gas depletion and its impact on permeability and stability of CBM reservoir.
Corey, A.T., 1954. The Interrelation between gas and oil relative permeabilities. Prod. Fuel 266, 117083.
Mon. 19 (1), 38–41. Lv, Y., Tang, D., Xu, H., Luo, H., 2012. Production characteristics and the key factors in
Cui, X., Bustin, R., 2006. Controls of coal fabric on coalbed gas production and high-rank coalbed methane fields: a case study on the Fanzhuang Block, Southern
compositional shift in both field production and canister desorption tests. SPE J. Qinshui Basin, China. Int. J. Coal Geol. 96 (97), 93–108.
111–119. No. SPE-89035. Meng, Z.P., Tian, Y.D., Li, G.F., 2010. Characteristics of in-situ stress field in Southern
DGH report, 2019. Directorate General of Hydrocarbons. Website. http://www.dghindia. Qinshui Basin and its research significance. J. China Coal Soc. 35 (6), 975–981.
org/index.php/page?pageId=63&name=Hydrocarbon%20Scenario. Meng, Y., Wang, J.Y., Li, Z., Zhang, J., 2018. An improved productivity model in coal
Faiz, M., Saghafi, A., Sherwood, N., Wang, I., 2007. The influence of petrological reservoir and its application during coalbed methane production. J. Nat. Gas Sci.
properties and burial history on coal seam methane reservoir characterization, Eng. 49, 342–351.
Sydney Basin, Australia. Int. J. Coal Geol. 70 (1–3), 193–208. https://doi.org/ Mohanty, D., 2011. Geologic and genetic aspects of coal seam methane (chapter – 4). In:
10.1016/j.coal.2006.02.012. Singh, A.K., Mohanty, D. (Eds.), First Indo-US Workshop on Coal Mine Methane,
Fu, X.H., Qin, Y., Li, G.Z., 2001. An analysis on the principal control factor of coal CIMFR Dhanbad, 17-20 October 2011. Cygnus, pp. 45–60.
reservoir permeability in central and southern Qinshui basin, Shanxi. J. Geomech 7 Mohanty, D., 2020. Coal seam gas - evaluation, extraction and environmental issues
(1), 45–52. (chapter 7). In: Kundu, S.N., Nawaz, M. (Eds.), Sustainable Energy and Environment:
Fu, X.H., Qin, Y., Wang, G.G.X., Rudolph, V., 2009. Evaluation of gas content of coalbed An Earth System Approach. Apple Academic Press, pp. 173–200 [ISBN:
methane reservoirs with the aid of geophysical logging technology. Fuel 88, 9781771887632 (Hard Cover); 9780429430107 (eBook)].
2269–2277. Mohanty, D., Chattaraj, S., Kumar, T., Halder, G., 2017. Sorption characteristics of coals
Gee, E.R., 1932. The geology and coal resources of the Raniganj Coalfield; Geol. Surv. from Tirap (west) colliery, NE India. In: Proc. International Conference on
India Memoir 61, pp343. Challenges and Prospects of Petroleum Production and Processing Industries (PEDJP
Ghosh, S.C., 2002. The Raniganj coal basin: an example of an Indian Gondwana rift. 2017), IIT (ISM), Dhanbad, 12-14 January 2017, p. 10.
Sediment. Geol. 147 (1–2), 155–176. https://doi.org/10.1016/S0037-0738(01) Mohanty, D., Chattaraj, S., Singh, A.K., 2018. Influence of coal composition and maturity
00195-6. on methane storage capacity of coals of Raniganj coalfield, India. Int. J. Coal Geol.
Guo, D., Guo, X., 2018. The influence factors for gas adsorption with different ranks of 196, 1–18.
coals. Adsorpt. Sci. Technol. 36 (3–4), 904–918. Olufemi, O., Ertekin, T., Smith, D.H., Bromhal, G., Sams, W.N., Jikich, S., 2004. Carbon
Hanson, W.B., Duhrkoph, D.J., Waller, S.F., 1996. Geotechnical uncertainty risk. In: dioxide sequestration in coal seams: a parametric study and development of a
Coalbed Gas Exploration. Geol. Soc. of Am. Annu. Meeting. Abstr. with Program, practical prediction/screening Tool using neuro-simulation. In: SPE Annual
Denver, Co, p. A-41. Technical Conference and Exhibition. https://doi.org/10.2118/90055-ms.
Hao, S.X., Chu, W., Jiang, Q., Yu, X., 2014. Methane adsorption characteristics on coal Pan, Z., Connell, L.D., 2012. Modelling permeability for coal reservoirs: a review of
surface above critical temperature through Dubinin–Astakhov model and Langmuir analytical models and testing data. Int. J. Coal Geol. 92, 1–44.
model. Colloid. Surface. Physicochem. Eng. Aspect. 444, 104–113. Pan, Z., Wood, D.A., 2015. Virtual special issue: coalbed methane (CBM) exploration,
Jalali, J., Mohaghegh, Sh.D., Gaskari, G., 2010. Coalbed methane reservoir simulation reservoir characterisation, production, and modelling: a collection of published
and uncertainty analysis with artificial neural networks: sharif university of research (2009-2015). J. Nat. Gas Sci. Eng. 26, 1491–1494.
technology. Scientia Iranica Chemistry and Chemical Engineering 17, 65–76. Pini, R., Ottiger, S., Burlini, L., Storti, G., Mazzotti, M., 2010. Sorption of carbon dioxide,
Jena, G.K., Mohanty, D., Chattaraj, S., 2018. Investigation on microfractures in coals of methane and nitrogen in dry coals at high pressure and moderate temperature.
Sohagpur coalfield: implication for CBM reservoir evaluation. In: ‘Technological International Journal of Greenhouse Gas Control 4 (1), 90–101.
Advancements and Emerging Mining Methods (TAEMM-2018)’ organized by Reiss, L.H., 1980. The Reservoir Engineering Aspects of Fractured Formations. Gulf
Institute of Engineers in Association with CSIR-CIMFR at Dhanbad during August 24- publishing company, Houston, Texas, ISBN 0-87201-303-0.
25, 2018, pp. 323–331. Robertson, E.P., 2005. Measurement and Modeling of Sorption Induced Strain and
Kaiser, W.R., Hamilton, D.S., Scott, A.R., Tyler, R., Finley, R.J., 1994. Geological and Permeability Changes in Coal. Ph.D. Dissertation. Colorado School of Mines, Golden,
hydrological controls on the producibility of coalbed methane. J. Geol. Soc. 151, Colorado.
417–420. Shi, J.Q., Durucan, S., 2008. Modeling mixed-gas adsorption and diffusion in coalbed
Kang, J., Fu, X., Gao, L., Liang, S., 2018. Production profile characteristics of large dip reservoirs, SPE 114197. In: 2008 SPE Unconventional Reservoirs Conference Held in
angle coal reservoir and its impact on coalbed methane production: a case study on Keystone. Colorado, USA.
the Fukang west block, southern Junggar Basin, China. J. Petrol. Sci. Eng. 99–114, Shi, J.Q., Pan, Z., Durucan, S., 2014. Analytical models for coal permeability changes
2018. during coalbed methane recovery: model comparison and performance evaluation.
Karacan, C.Ö., 2008. Evaluation of the relative importance of coal bed reservoir Int. J. Coal Geol. 136, 17–24.
parameters for prediction of methane inflow rates during mining of longwall Singh, A.K., Mohanty, D., 2015. CO2 sequestration potential of Indian coalfields (chapter
development entries. Comput. Geosci. 34, 1093–1114. – 9). In: Goel, M., Sudhakar, M., Shahi, R.V. (Eds.), Carbon Capture, Storage and
Karacan, C.Ö., Drobnaik, A., Mastalerz, M., 2014. Coalbed reservoir simulation with Utilization: a Possible Climate Change Solution for Energy Industry. The Energy
geostatistical property realizations for simultaneous multi-well production history Research Institute (TERI), pp. 133–147.
matching: a case study from Illinois Basin, Indiana, USA. Int. J. Coal Geol. 131, Singh, P.K., Singh, V.K., Rajak, P.K., Singh, M.P., Naik, A.S., Raju, S.V., Mohanty, D.,
71–89. 2016. Eocene lignites from Cambay basin, Western India: an excellent source of
Karimi, K., 2005. Coal Bed Methane Reservoir Simulation Studies. ME thesis submitted at Hydrocarbon. Geoscience Frontiers 7–5, 811–819.
School of Petroleum Engineering, The University of New South Wales, p. pp149. Su, X., Lin, X., Liu, S., Zhao, M., Song, Y., 2005. Geology of coalbed methane reservoirs in
Kędzior, S., 2019. Distribution of methane contents and coal rank in the profiles of deep the southeast Qinshui Basin of China. Int. J. Coal Geol. 62, 197–210.
boreholes in the Upper Silesian Coal Basin, Poland. Int. J. Coal Geol. 202, 190–208. Suggate, R.P., 1974. Coal ranks in relation to depth and temperature in Australian and
Keles, C., Vasilikou, F., Ripepi, N., Agioutantis, Z., Karmis, M., 2019. Sensitivity analysis New Zealand oil and gas wells. N. Z. J. Geol. Geophys. 17 (1), 149–167. https://doi.
of reservoir conditions and gas production mechanism in deep coal seams in org/10.1080/00288306.1974.10427995.
Buchanan County, Virginia. Simulat. Model. Pract. Theor. 94, 31–42. Van Krevelen, D.W., 1993. Coal; Typology-Chemistry-Physics-Constitution, third ed.
Kissell, F.N., McCulloch, C.M., Elder, C.H., 1973. The Direct Method of determining Elsevier, Amsterdam, p. 979. Coal).
methane content of coalbeds for ventilation design. US Bur. Mines, Rep. Invest. Wang, G., Cheng, W.M., Pan, G., 2012. Influence of temperature on coal’s adsorbing
7767, 17. capability. J. Saf. Environ. 12 (5), 231–234.
Kou, Z., Wang, H., 2020. Transient pressure analysis of a multiple fractured well in a Wang, H., Kou, Z., Guo, J., Chen, Z., 2021. A semi-analytical model for the transient
stress-sensitive coal seam gas reservoir. Energies 13, 1–20. https://doi.org/10.3390/ pressure behaviors of a multiple fractured well in a coal seam gas reservoir. J. Petrol.
en13153849. Sci. Eng. 198, 108159. https://doi.org/10.1016/j.petrol.2020.108159.
Kumar, S., Boral, P., Choudhury, N., Ghosh, G., Mohanty, D., 2006. The natural fairways Warren, J.E., Root, P.J., 1963. The behavior of naturally fractured reservoirs. Soc. Pet.
and their implication to CBM recovery: a perspective view. In: National Conference Eng. J., Sept. 245–255.
on GTFM-2006, Indian School of Mines, Dhanbad, 2nd-4th November 2006, Wei, Q., Li, X., Hu, B., Zhang, X., Zhang, J., He, Y., Zhang, Y., Zhu, W., 2019. Reservoir
pp. 77–83. characteristics and coalbed methane resource evaluation of deep buried coals: a case
Lin, Y.B., Ma, D.M., Liu, Y.H., 2012. Experiment of the influence of temperature on study of the No.13–1 coal seam from the Panji Deep Area in Huainan Coalfield,
coalbed methane adsorption. Coal Geol. Explor. 40 (6), 24–28. Southern North China. J. Petrol. Sci. Eng. 867–884, 2019.
Lin, H.F., Yao, F., Li, S.G., 2014. Experiment of coal’s temperature and water content Wu, Y., Pan, Z., Zhang, D., Lu, Z., Connell, L.D., 2018. Evaluation of gas production from
influencing methane absorption quality. Coal Ming Technology 19 (3), 9–12. multiple coal seams: a simulation study and economics. International Journal of
Liu, P., 2010. A Research on the Methane Adsorption/desorption in the Mid and High Mining Science and Technology 28, 359–371.
Rank Bituminous. The Master Degree Thesis of Xi. An university of science and Xu, H., Tang, D., Zhao, J., Tao, S., Li, S., Fang, Y., 2015. Geologic controls of the
technology, China (in Chinese with English abstract). production of coalbed methane in the Hancheng area, southeastern Ordos Basin.
Liu, H.H., Sang, S.X., Wang, G.G.X., Li, Y., Li, M., Liu, S., 2012. Evaluation of the J. Nat. Gas Sci. Eng. 26, 156–162.
synergetic gas-enrichment and higher-permeability regions for coalbed methane Zhang, X., Tao, Y.Q., 2011. Experiment study on gas isothermal adsorption from coal
recovery with a fuzzy model. Energy 39 (1), 426–439. under different temperature conditions. Coal Engineering 4, 87–89.
Liu, H., Sang, S., Xue, J., Wang, G., Xu, H., Ren, B., Liu, C., Liu, S., 2016. Characteristics Zhao, Z.G., Tang, X.Y., Zhang, G.M., 2001. Experiment and significance of isothermal
of an in situ stress field and its control on coal fractures and coal permeability in the adsorption of coal on methane under higher temperature. Coal Geol. Explor. 23 (4),
Gucheng block, southern Qinshui Basin, China. J. Nat. Gas Sci. Eng. 36 (B), 130–131.
1130–1139. https://doi.org/10.1016/j.jngse.2016.03.024.

19
S. Chattaraj et al. Journal of Natural Gas Science and Engineering 92 (2021) 103969

Zhou, F., Guan, Z., 2016. Uncertainty in estimation of coalbed methane resources by Zuber, M.D., Olszewski, A.J., 1992. The impact of errors in measurement of coalbed
geological modelling. J. Nat. Gas Sci. Eng. 33, 988–1001. https://doi.org/10.1016/j. methane reservoir properties on well production forecasts. In: Proceedings of the
jngse.2016.04.017. 62nd SPE Annual Technical Conference and Exhibition. Paper No. 24908,
Zou, M., Wei, C., Pan, H., Seyas, S.K., Cao, J., 2010. Productivity of coalbed methane Washington, DC.
wells in southern of Qinshui Basin. Min. Sci. Technol. 20, 765–769. Zuber, M.D., Olszewski, A.J., 1993. Coalbed methane production forecasting:
Zou, M., Lv, X., Huang, Z., Wei, S., Zhang, M., Sun, C., 2018. Modeling and prediction for measurement accuracy required for key reservoir properties. In: Proceedings of the
gas production during coalbed methane drainage based on two indirect reservoir 1993 International Coalbed Methane Symposium, Birmingham, Alabama,
parameters. Energy Explor. Exploit. 1–14. https://doi.org/10.1177/ pp. 549–559.
0144598718777105, 0(0).

20