Anisotropy and fracture analysis for coalbed methane reservoir development in Bokaro coalfield, India

The low permeability of coal seams is a constraint in the efficient production of coalbed methane. However, the presence of natural fractures in coal enhances the permeability, and prior knowledge of sub‐surface fractures in coal seams is vital to identify the prospective seam. This paper investigates the anisotropy and identifies fractures by processing the advanced sonic and resistivity image logs to mitigate challenges in the reservoir. Anisotropy is estimated from the difference in the travel time between fast and slow shear waves. The application of Alford's rotation technique determines the fast shear wave polarization angle which is consistent with the fracture orientation along the NE–SW or NW–SE direction in coal seams. Moreover, the crossover of fast and slow shear waves in the slowness versus frequency plot indicates stress‐induced anisotropy that originates from fractures. Besides, drilling‐induced fractures observed along NE–SW in the resistivity image log indicate the maximum horizontal stress direction. Results from this study compare coal seams based on fractures in adopting better operational activities for optimized production and future geomechanical studies in the Bokaro coalfield situated in India.


I N T RO D U C T I O N
Coalbed methane (CBM) is becoming an alternate energy resource over the last decade, and its exploration has many challenges which need proper planning for extracting methane gas (Bell & Bachu, 2003).Fractures are the mechanical rupture along the weakest plane in the rock that occurs when the concentration of regional and internal tectonic stress acting in the rock exceeds the rock's strength.Another type of fracture known as cleat develops in coal.Cleats are small natural fractures that are discontinuous while fractures are continuous.The objective to discriminate between fractures and cleats is to have a comparative analysis in terms of orientation.In * E-mail: Banerjee_Abir@ongc.co.in, abir.phy@gmail.com the resistivity image log, fractures are recognized in sinusoidal lines, whereas cleats appear as parallel to sub-parallel vertical to transverse lines (Laubach et al., 1998;Pitman et al., 2003).In CBM exploration, knowledge of fractures is vital for the planning and optimization of gas production from the reservoir (Bell & Bachu, 2003).Coal has low permeability, and its permeability increases with the presence of natural fractures/cleats (Pitman et al., 2003).However, in the majority of the coal seam, hydraulic fracturing treatment is executed to create secondary fractures to enhance permeability (Banerjee & Chatterjee, 2021a).Cleats are distinguished as face and butt.Face cleats run parallel to maximum horizontal stress direction and in longer in size compared to butt cleat which is orthogonal to face cleat and smaller in size.Figure 1(a) shows a schematic representation of the sub-surface model where the stresses (vertical stress (S v ), maximum horizontal stress (S H ), and minimum horizontal stress (S h )) act orthogonal to each other along the drilled wellbore.The black colour layer represents the coal seam and the core sample of coal is extracted from the well, Figure 1(b) represents the enhanced view of coal's core sample illustrating the development of the face and butt cleats.The face cleats develop along the S H direction and butt cleats are prominent along the S h direction.The orientation of fractures also helps in understanding the flow path mechanisms and production behaviour (Tokhmechi et al., 2009).Fractures are accurately identified from the physical verification of core and processed resistivity image log, but core samples and resistivity image log are not acquired in every well.Therefore, the sonic log has been used to indirectly estimate fractures from anisotropy analysis and its characterization.
Anisotropy is an occurrence where the physical properties vary directionally in a medium.Anisotropy can be of two types: one is the intrinsic type that originates due to the preferred mineral or grain deposition as platy nature of the thin isotropic layer and another is stress-induced anisotropy that appears from the fracture alignment (Wang, 2002).The variation of physical properties in the horizontal or vertical plane with the borehole axis is known as vertical transverse isotropy (VTI), which is observed in shale or thinly bedded intervals (Cheng & Cheng, 1996;Wang, 2002).In azimuthal anisotropy or horizontal transverse isotropy (HTI), the physical properties vary in the azimuthal direction around the wellbore due to the concentration of stress in fractures (Cheng & Cheng, 1996).Pore structure, fabric and stress are Figure 2 A visual representation of a sonic tool in a borehole consisting of a transmitter (T) and receivers (R1-R8).The transmitter emits acoustic pulses that pass through the borehole and formations are recorded in the receivers.The response of emitted transmitter's signal is short and decays rapidly with time whereas the recorded receiver's signal consists of a signal train with increasing magnitude with time.These signal trains are the composite waveform that is received in successive arrival of (i) compressional waves, (ii) shear waves, (iii) Rayleigh waves, (iv) mud waves and (v) Stoneley waves (Banerjee & Chatterjee, 2021b).
the cause of anisotropy.Coal matrix deformation is related to the pore structure and fabric coal resulting from CBM gas adsorption/desorption.The presence of fractures in coal is connected to stress-induced anisotropy.In the petroleum industry, the anisotropy results are widely used for the optimization of reservoir drainage, and the detection of fault and fracture (Xu et al., 2018).Anisotropy and stress/fracture direction in coal seam determines the permeable path for fluid flow (Hower, 2003;Feng et al., 2019).The occurrence of azimuthal anisotropy from fractures in coal was shown by De et al. (1998), Gray et al. (2002) and Vetri et al. (2003) using seismic data.Wang et al. (2019) have shown that shear wave splitting parameters and fracture orientation extracted from the full azimuth multi-component seismic data are frequency dependent.Also, it was observed that the travel time delay of the split shear waves decreases with increasing frequency.An improvement in the recognition of fractures was shown by Li et al. (2019b) using detailed (3D3C) seismic data based on shear-wave splitting which was supported by advanced sonic log processed data.During the advanced sonic log data acquisition, the transmitter in the tool emits an acoustic waveform that passes through the borehole formation and gets converted into various wave components.The difference in velocity of each wave component results in the arrival of compressional (P), shear (S), Rayleigh, mud and Stoneley (St) waves in successive intervals in the receivers.Figure 2 shows a visual representation of a sonic tool in a borehole consisting of a transmitter (T) and receivers (R1-R8).The transmitter emits acoustic pulses that pass through the borehole and formations are recorded in the receivers.The response of emitted transmitter's signal is short and decays rapidly with time, whereas the recorded receiver's signal consists of the signal train with increasing magnitude with time.These signal trains are the composite waveform that is received in successive arrival of (i) compressional wave, (ii) shear wave, (iii) Rayleigh wave, (iv) mud wave, and (v) Stoneley waves (Banerjee & Chatterjee, 2021b).The P, S and St components have different characteristics.P-wave passes through solid, liquid and gas with the same particle motion and direction of propagation.S-wave passes only through a solid, and its particle motion is perpendicular to the direction of wave propagation in an isotropic medium.But, in an anisotropic medium, a phenomenon called birefringence occurs where S-wave splits into fast and slow components with a different group and phase velocities (Sayers, 2009).The last arrival of the St-wave propagates through a solid-liquid interface along the borehole wall and is sensitive to fractures and permeable formation resulting in velocity reduction and amplitude decays (Stoneley, 1924;Tang & Cheng, 1993;Banerjee & Chatterjee, 2021b).Moreover, the properties of P, S and St waves are affected by anisotropy in the formation.In the previous studies, it was established that S-wave is influenced by azimuthal anisotropy, whereas St-wave responds to VTI anisotropy but P-wave is the least affected by anisotropy Figure 5 Diagram representing S-wave splitting in multi-source and multi-receiver acquisition system.S 1 is the direction of fast shear-wave polarization and S 2 is the direction that is orthogonal to S1 and represents the slow shear wave polarization direction.θ is the angle of the S1 and S2 with respect to radial (R) and transverse (T) components.

Figure 6
The adopted flowchart displaying the steps followed in this study using geophysical well log data.(Hornby et al., 1999;Sun et al., 2003).Processing of acoustic cross-multipole array log data distinguishes and characterizes the anisotropy mechanisms through radial acoustic dispersion analysis (slowness with frequency).For instance, four types of mechanisms: (a) homogeneous isotropic, (b) inhomogeneous isotropic, (c) homogeneous anisotropic, and (d) inhomogeneous anisotropic can be identified in the formation from shear wave slowness versus frequency plots using three types of shear waves namely modelled shear wave (black line), fast shear wave (red line) and slow shear wave (blue line), as shown in Figure 3 (Plona et al., 2000;Moreno et al., 2013).No S-wave splitting occurs over frequency in a homogeneous isotropic medium (Fig. 3a).In an inhomogeneous isotropic medium, no S-wave splitting occurs but slowness data are higher compared with the predictive model (Fig. 3b).In a homogeneous anisotropic medium, parallel splitting of fast and slow S-wave slowness is observed; this mechanism is associated with intrinsic anisotropy due to the layering of shale (Fig. 3c).In an inhomogeneous anisotropic medium, fast and slow S-wave slowness is associated with crossover, due to stress-induced anisotropy around the wellbore (Fig. 3d).However, the accuracy of the result depends on the data quality and processing inputs as conditions such as borehole failure or breakouts, ellipticity, and higher relative dip angle between the borehole axis and formation beds are necessary to be considered that give erroneous results (Barbosa et al., 2021).Alford's rotation is the most used technique to separate the fast and show shear wave from the composite waveform.Alford's rotation is also used to determine the fast shear wave polarization angle.Li et al. (2019a) have applied Alford's rotation to predict fractures in coal seams using multi-components seismic data.Alford's rotation was applied to derive fracture azimuth and anisotropic parameters as a quantitative degree of fracture development.In Bokaro coalfield, anisotropy analysis using a sonic log was not documented.Therefore, in this study, the aim is (a) to analyse and characterize anisotropy by processing cross-dipole components of the sonic waveform, (b) to determine fractures orientations and stress direction from the resistivity image log, and (c) to address the outcome in assisting future planning for CBM reservoir development.

S T U DY A R E A A N D DATA S E T
The study is conducted in the west Bokaro field located in the state of Jharkhand in the eastern part of India.The study area includes two wells named 'B1' and 'B2' that are drilled vertically in water-based mud and data are recorded in a centralized tool position.In well B1, conventional geophysical well logs, resistivity image logs and advanced sonic log are available from depth intervals of 100.0-435.0m.In well B2, only the resistivity image log is available from depth interval 200.0-840.0m.Recorded conventional logs include calliper (CALX), gamma-ray (GR), resistivity (MLR), density (ZDNC), neutron porosity in limestone calibrated (CNCF), and photoelectric (PE).The resistivity image and advanced sonic log are processed to obtain the sub-surface fractures and slowness components: P-wave (DTC), S-wave (DTS) and St-wave (DTST).The geological map of the Bokaro coalfield (Fig. 4) illustrates the surface exposure of formation and well's locations B1 and B2 in the western part.The formational sequences from the top of the reservoir are Mahadeva, Panchet, Raniganj, Barren, Barakar and Talchir.The log data are acquired in the Barakar formation of the early Permian age consisting of sandstone, shale and coal (Banerjee & Chatterjee, 2021a).

M AT H E M AT I CA L BAC K G RO U N D
In anisotropy analysis, Alford's rotation processing technique is vital to determine the rotation angle of fast shear (S) waves.Alford's rotation is a four-component rotation technique that waves are processed in monopole mode, whereas the shear wave is processed in the cross-dipole (YY) mode.In the semblance track, the green, sky blue and red colour lines are the travel time picks of compressional, Stoneley and shear wave components respectively, and the colour code of the semblance plot represents the magnitude of coherence.In the projection track, the colour scale represents coherence magnitude from minimum to maximum, and the red colour in the projection log represents the maximum coherence, and it is picked and shown in a black line that estimates the slowness magnitude of the respective wave component.
determines the orientation of the natural coordinate system (Alford, 1986;Li et al., 2019a).Considering a multi-source and multi-receiver system where S-wave splitting occurs.The splitting of the S-wave into fast and slow S-waves may or may not coincide with the natural coordinate system (Thomsen, 1988;Crampin & Lowell, 1991).In Figure 5, S1 and S2 are the directions, and v 1 and v 2 are the velocities of fast and slow S-waves in multiple sources and receiver systems.S a (t) and S(t) is the source matrix in the acquisition and natural coordinate system.S R (t) and S T (t) represent the radial and transverse sources in the acquisition coordinate, and θ is the angle of the S 1 and S 2 with respect to radial (R) and transverse (T) components.The relationship between these components is expressed in the form of the equations using vector rotation matrix R(θ).The equations are as follows: where and Similarly, the reflected signal V(t) is recorded in the acquisition coordinate, that is, expressed as where , and U(t )= U 11 0 0 U 22 represents matrix in the natural coordinate system, U 11 and U 22 is the reflected signal along the S 1 and S 2 directions.
The above equation from the time domain to frequency domain is expressed as where represent the timedelay function of the fast and slow S-wave, δ 1 and δ 2 are the two-way travel time, and f 1 and f 2 are the filter function for the fast and slow S-wave set for the attenuation, geometric spreading and reflection coefficient.By applying rotation in the data matrix in both source and receiver coordinates, the output of the rotated data matrix is expressed as Considering f 1 (ω) = f 2 (ω), equation ( 5) is expressed as where in W RT and V RT , R represents receiver direction and T represents transmitter direction, and = 2z( 1 ).The multiplication factor exp(±iω) in the frequency domain is equivalent to the time shift in the time domain.
Hence, equation ( 6) determines (1) the angle (θ ) between the fast S-wave and acquisition line, (2) the time lag () between fast and slow S-wave, and (3) the separated fast and slow S-wave by the rotation of the acquisition matrix data V(t) into W(t).Therefore, the parameters (θ , ) is computed by rotating the input data matrix by a set of range in angles and time, and the normalization of off-diagonal elements of the rotated matrix is expressed as where N represents the sample number in the search window; k and p are the integers.

M E T H O D S
The adopted flowchart shown in Figure 6 presents the methodology used and the output obtained in this study based on available data.The conventional log parameters distinguish coal from other lithology.The processed resistivity image log yields a brief understanding of the subsurface fracture system and stress direction.Sonic log processing provides the slowness of various wave components and determines anisotropy from fast and slow S-wave travel time and its orientation.Moreover, anisotropy is characterized by fast and slow shear slowness versus frequency plot.

Conventional log analysis
The 05 coal seams named 'V, VI + VII, VIII, IX and X' are identified as objective seams.The seams: V (413.0-420.7 m), VI + VII (385.0-390.5 m), VIII (332.0-336.5 m), IX (264.5-266.5 m) and X (235.0-242.5 m) are distinguished from shale and sandstone based on the geophysical log cut-off criteria tabulated in Table 1 (Banerjee & Chatterjee, 2021a).The borehole conditions influence the log reading, but the calliper log against coal seams shows good borehole condition as no caving or mud formation is observed.

Resistivity image log analysis
The resistivity image log represents the view of a threedimensional wellbore on a two-dimensional surface in a clockwise (North-East-South-West-North) direction.Quality checks and speed correction is applied in the resistivity image log to determine the tool response while recording the data.Further, processing steps include image equalization and normalization to generate static and dynamic images.The colour spectrum in the image represents low to high resistivity in the formation.Among two types of images, in the static image the assigned resistivity colour spectrum remains constant with depth and matches with a particular band of resistivity values consistently throughout the well, whereas the dynamic image creates more visual contrast in the resistivity distributions in the specifically defined depth interval (5 m).Thus, the resistivity distribution would change every 5 m, and thus the mapping of resistivity colour also changes every 5 m to bring out geological features, but there is no link between colour and actual resistivity.The resistivity image log provides crucial evidence to visualize the sub-surface fractures, borehole breakouts (BOs), and drilling-induced fractures (DIFs) which are the indicators of the present-day stress direction (Kingdon et al., 2016;Chatterjee et al., 2017).BOs are parallel to minimum horizontal stress (S h ) and DIFs are parallel to maximum horizontal stress (S H ) direction.Mud is circulated in wells to control the hydrostatic pressure and prevent the wellbore from collapsing.The circulating mud exerts pressure which is known as mud weight (MW).In general, the MW is kept at a slightly higher side than the hydrostatic pressure for wellbore stability.BOs and DIFs can occur if the MW is too high or too low (Tingay et al., 2005).In this study, DIFs are observed as a thin vertical pair of the line at 180°apart in the image log (Niu et al., 2020).Naturally occurring fractures are identified as sinusoidal, while cleats in the coal seam are either observed as small vertical or sub-vertical or transverse lines in the image log.In well B1, dip magnitude and orientation of fractures are determined.Moreover, in well B2, the direction of DIFs is determined.The following expressions compute dip magnitude from the image log in a vertical well.
where h is the height between the crest and trough of the sinusoidal curve, d is the borehole diameter measured from the calliper log and DOI is the depth of investigation of the electrical signal from the tool (1 inch for the tool).

Data acquisition
The advanced sonic tool contains 2 sets of orthogonal transmitters (T x and T y ) and 8 sets of orthogonal receivers (R x and R y ) in the X and Y directions.When the emitted signal from the transmitter (T x ) is recorded by multiple receivers (R x ), then it is known as inline (XX), and when the T x signal is recorded by R y , it is known as cross-dipole (XY).Similarly, another

Data processing (semblance analysis)
The arrival time of the P, S and St waves increases with increasing distance from the transmitter to the receivers in order from R1 to R8.As discussed in Figure 2, the compressional wave arrives first, followed by the shear and Stoneley wave.
For instance, the intra-receivers spacing is 0.15 m and the distance between the transmitter to the nearest receiver is 1.07 m; hence, the slowness of each wave component is computed from the travel time of the signal from the transmitter to the receiver array.The vertical resolution is defined by the size of the receiver array of a tool which implies that 8 receivers with 0.15 m intra-receiver spacing have a vertical resolution of 1.07 m.Consequently, for measuring the true acoustic slowness, the bed must be at least 1.07 m thick.To resolve thin beds, the number of receivers could be reduced for increasing the vertical resolution.However, simply reducing the number of receivers would make the result more susceptible to noise and borehole irregularities.Therefore, the resolution enhancement is based on the multiple-shot processing technique by overlapping sub-arrays across the same depth level.The size of the sub-array is defined by the number of the receiver's parameters.Each of these sub-arrays is processed using a technique called the semblance method that looks for similarity (coherence) in the waveform (Guo et al., 2008).A mean of the projected semblance statistics is calculated to obtain a higher resolution in the projection log.The separation of P, S, St waves from the receiver array waveforms using semblance processing requires time window parameters.Figure 7 2. The waveform is correlated in a specific moving time window over a range of slowness with incremental steps across the travel time axis from a start time to an end time.At each time window position, the signal correlation across the receivers is analysed in the slowness domain.The semblance results are stored in a contour plot in the slowness-time plane in the form of semblance and projec- tion log.The y-and x-axes of the projection log represent the depth and the slowness magnitude.The best-correlated waveform yields maximum coherence indicated by the red colour in the projection log, and continuous slowness log is picked by tracking the coherence maxima for each wave component.The slowness computed from semblance processing represents the average over the length of the receiver array.In shear processing, only YY dipole mode is used as a reference to obtain the semblance and projection plot.Figure 8 represents the semblance and projection plots of the monopole compressional wave, monopole Stoneley wave, and dipole YY component of shear wave are shown in order from left to right from depth interval 370.0-435.0m.The semblance and projection plots of compressional and Stoneley waves are processed in the monopole mode, whereas the shear wave is processed in cross-dipole (YY) mode.In the semblance track, the green, sky blue and red colour lines are the travel time picks of compressional, Stoneley and shear wave components respectively and the colour code of the semblance plot represents the magnitude of coherence.In the projection track, the colour scale represents coherence magnitude from minimum to maximum and the red colour in the projection log represents the maximum coherence, and it is picked and shown in a black line that estimates the slowness magnitude of the respective wave component.

Data analysis
Anisotropy analysis.The anisotropy magnitude and the orientation of the fast shear wave are computed from the crossdipole (XX, XY, YX, YY) processing.The cross-dipole waveform and normalization log (generated during acquisition), tool azimuth log (provides tool orientation), and P, S-wave slowness log (obtained from data processing) are used for anisotropy analysis by processing the cross-dipole components.Alford's rotation technique is used to compute the slowness magnitude from the fast and slow principal waveforms and the orientation of the fast principal waveform in the presence of an anisotropic medium (Alford, 1986).The resulting waveform is processed using a cross-correlation technique called semblance processing for obtaining fast and slow S-wave slowness (Kimball & Marzetta, 1986;Collet et al., 2015).Figure 9 represents the projection and the slowness log of fast/slow shear shown in the black line from 407.0 to 427.0 m.In the projection track, the colour scale represents coherence magnitude from minimum to maximum and the red colour in the projection log represents the maximum coherence and it is picked shown in a black line that estimates the slowness magnitude of fast and slow waves.The anisotropy in coal seams is analysed from three types of indicators: (a) energy anisotropy, (b) travel time (TT) anisotropy and (c) slowness (DTS) anisotropy.Energy anisotropy is determined from the minimum and maximum energy of the cross-dipole components (XY, YX) relative to all four components (XX, XY, YX, YY) at each depth point.The strength of the anisotropy is measured by the difference between maximum with minimum energy.Moreover, the TT and DTS anisotropy are calculated from the differences in travel time and slowness between fast and slow shear slowness and principle waveform.The travel time is obtained from the first arrival picking in slowness waveform and the percentage of DTS anisotropy is determined by dividing the slowness difference by the average shear slowness.Slowness versus frequency plot.Waveform processing recognizes the presence of anisotropy in the formation but the type of anisotropy cannot be ascertained.Therefore, to distinguish the type of anisotropy, slowness versus frequency plot is generated to extract the information.In the slowness versus frequency plot, the slowness-frequency relation is obtained from input waveforms using the matrix pencil method by analysing the dispersive nature of waves (change of slowness with frequency).The steps in obtaining slowness versus frequency plot include: (1) The poles of the mode waves are calculated using matrix pencil algorithm (Ekstrom, 1996), ( 2) sorting of all the obtained modes considering absolute value of complex amplitude and then eliminating false modes and (3) inverting the dispersion slowness of the true modes by weighted spectral semblance (WSS) algorithm and eliminating the aliases (Sinha et al., 2000).Semblance processing determines the slowness in the time domain that favours the low-frequency limit of the wave containing the highest signalto-noise ratio.For a dispersive wave, the strongest signal can occur at frequencies higher than the low-frequency limit, and hence the resulting slowness is too high (Banerjee & Chatterjee, 2021b).The dispersion result can give erroneous if adequate dispersion corrections are not applied.Therefore, initially, a reference model was prepared considering ideal borehole conditions such as borehole diameter, formation P and S-wave velocity, borehole fluid slowness and the density ratio of formation to borehole fluid.Ideal borehole parameters 8.5inch borehole diameter, P and S-wave velocity are used from monopole and YY dipole slowness pick, borehole fluid slowness is 620 μs/m, formation density used from density log, and fluid density is 1050 kg/m 3 is considered.Similarly, the fast/slow shear wave slowness versus frequency relationship is obtained in the borehole.15(a-e).In Figure 15, the black dots are the predicted model under ideal borehole conditions, while the red and blue dots are the fast and slow shear wave observed model.The crossover between fast and slow S-waves from 5-6 kHz in Figure 15 indicates stress-induced anisotropy that has originated due to fractures and its correlation with the resistivity image log confirms a fractured coal seam.Few DIFs are also observed in resistivity images shown with arrow marks in the dynamic image track in well B2 from depth interval 618.0-650.0m.In dynamic track, the drillinginduced fractures (DIFs) are indicated with an arrow mark which is observed as thin parallel lines separated by 180 degrees (Fig. 16a).The stereonet plot of DIFs indicates the S H direction along NE-SW (Fig. 16b).Error analysis measures the uncertainty in the result, that is obtained from the difference in the observed and predicted magnitude of the fast S-wave polarization direction.The root mean square (RMS) error of the fast S-wave polarization angle versus depth for seams 'X, IX, VIII, VI + VII, and V' is shown in Figure 17a, 17b, 17c, 17d, 17e.The average RMS error of 05 coal seams is 5.38°, 5.15°, 4.40°, 3.37°and 4.81°, respectively.The seam-wise description of shear waves crossover depth, fractures/cleats dip, and orientation, energy anisotropy, TT/DTS anisotropy and fast S-wave polarization angle is tabulated in Table 3.

D I S C U S S I O N S
Fractures cause the origin of anisotropy in coal seams that enhances the permeability.The intensity and orientation of fractures help in planning the well's trajectory.The well's trajectory should pass orthogonal to the orientation of the fracture bisecting maximum numbers of fractures.This will result in increased connectivity and optimized production.Also, the hydraulic fracturing treatment is planned according to the fracture system.The hydraulic fracturing treatment generates secondary fracture as the primary fracture is not sufficient enough to have the necessary permeability for the flow system.Therefore, based on the intensity and orientation of fractures, simulation models are prepared considering the volume of the proppant, an extension of the propagation of fluid during treatment, rock mechanical parameters and many more.Knowledge of fracture orientation from fast shear wave azimuth assists in planning oriented perforation and flow path for fluid movement during hydraulic fracturing treatment.In addition stress direction is used for designing drilling trajectory.For instance, the trajectory of the horizontal/inclined well is stable and creates better connectivity as the trajectory bisects maximum numbers of fractures when it passes through the S h direction.The S H direction obtained supports the previous study along N15°E to N35°E by Paul et al. (2017).Seamwise planning for oriented perforation is needed for production optimization.Moreover, stress direction is a vital input for geomechanical application and effective fracturing in the formations (Todorovic-Marinic et al., 2004;Kayama et al., 2021).In an experimental investigation of coal samples under stress conditions, maximum permeability was observed along the strike direction due to anisotropy in stress, Young's modulus, porosity and cleat compressibility (Yan et al., 2019;Li et al., 2021).The present approach gives promising results in the detection of fractures and anisotropy supported by the resistivity image log.The corroboration of finding anisotropy, fast S-wave direction, slowness versus frequency plot, and S H direction with fractures/cleats presence and orientation from resistivity image log suggests the correctness of processing results.The use of only processed sonic logging can estimate fractures based on anisotropy magnitude and slowness versus frequency plot.Also, fracture orientation can be determined from the fast shear wave polarization angle.This analysis helps in operational planning such as the perforation of coal seams and hydraulic fracturing treatment.Though this method can effectively distinguish the fractures in coal, the limitations lie in discriminating the fracture, as the type of fracture could not be identified.The fracture is sensitive to the frequency spectrum as the time delay between fast and slow shear waves changes with variation in frequency (Wang et al., 2019).

C O N C L U S I O N S
1.The determination of anisotropy, characterization of anisotropy, identification of fractures/cleats and their orientation, and determination of S H direction provide numerous inputs to help the operation and production planning.A seam-wise model for oriented perforation can be implemented.

2.
Anisotropy is observed in coal seams from the difference in travel time between fast and slow shear waves.The crossover between fast and slow shear waves in the slowness versus frequency plot indicates stress-induced anisotropy.

Figure 1
Figure 1 (a) A schematic representation of the sub-surface model where the stresses (vertical stress (S v ), maximum horizontal stress (S H ) and minimum horizontal stress (S h )) act orthogonal to each other along the drilled wellbore.The black colour layer represents the coal seam and the core sample of coal is extracted from the well.(b) The enhanced view of coal's core sample illustrates the development of the face and butt cleats.The face cleats develop along the S H direction and butt cleats are prominent along the S h direction.

Figure 3
Figure3Classification of acoustic anisotropy from shear wave slowness versus frequency plots using three types of shear waves namely modelled shear wave (black line), fast shear wave (red line), and slow shear wave (blue line).The four possible types of acoustic anisotropy are (a) homogeneous isotropic: no splitting of shear waves is observed; (b) inhomogeneous isotropic: no shear wave splitting but deviates with modelled shear wave and a shear wave is a function of radius Vs(r) that causes near failure; (c) homogeneous anisotropic: parallel splitting between fast, slow and modelled shear waves, the shear wave is a function of azimuth Vs(θ ); and (d) inhomogeneous anisotropic: a crossover between fast and slow shear wave is observed, where the shear wave is a function of both radius and azimuth Vs(r, θ ) (redrawn afterMoreno et al., 2013).

Figure 4
Figure 4 Index map of India showing the location of the Bokaro field, which is the study area.The geological map of the Bokaro coalfield drawn after Banerjee and Chatterjee (2021a,b) represents the geological formations and well locations B1 and B2.

Figure 7
Figure 7 (a) A schematic diagram of recorded waveform across eight receivers (R1-R8) representing the moveouts in a dashed line.The moveouts of compressional, shear and Stoneley waves are shown.In each waveform, processing parameters such as start and end time, time window and lower and upper slowness limits are analysed.In this figure, it is shown for only the Stoneley wave.(b) The presentation of the actual waveform, where the rectangular box is the picked arrival time of compressional, shear and Stoneley waves.

Figure 8
Figure 8 Semblance and projection plots of monopole compressional wave, monopole Stoneley wave and dipole YY component of shear wave are shown in order from left to right from depth interval 370.0-435.0m.The semblance and projection plots of compressional and Stoneleywaves are processed in monopole mode, whereas the shear wave is processed in the cross-dipole (YY) mode.In the semblance track, the green, sky blue and red colour lines are the travel time picks of compressional, Stoneley and shear wave components respectively, and the colour code of the semblance plot represents the magnitude of coherence.In the projection track, the colour scale represents coherence magnitude from minimum to maximum, and the red colour in the projection log represents the maximum coherence, and it is picked and shown in a black line that estimates the slowness magnitude of the respective wave component.

Figure 9
Figure 9Projection plot of separated fast and slow shear waves.In the projection track, the colour scale represents coherence magnitude from minimum to maximum and the red colour in the projection log represents the maximum coherence, and it is picked shown in the black line that estimates the slowness magnitude of fast and slow waves.
(a) represents a schematic diagram of the recorded waveform across eight receivers (R1-R8) representing the moveouts in dashed lines along with the receiver array waveforms.The moveouts of compressional, shear and Stoneley waves are shown.In each waveform, processing parameters such as start and end time, time window and lower and upper slowness limits are analysed.In this figure, it is shown for only the Stoneley wave. Figure 7(b) illustrates the actual results of the receiver's travel time consisting of P, S and St waves, and the slope of a line drawn through the arrival of the wave component at each receiver is called the moveout.The moveout represents the differences in velocity of the wave components over the distance covered, which implies that a steeper moveout indicates faster propagation of the signal.Considering Figure 7(b), the parameters (time window and slowness limit) used for the semblance processing of P, S and St waves are tabulated in Table

Figure 15
Figure 15 Shear slowness versus frequency plots, the black dots are the predicted shear wave model under ideal borehole conditions, while the red and blue dots are the observed fast and slow shear waves.The crossover between fast and slow shear waves from 5-6 kHz indicates stress-induced anisotropy in fractured depth points at (a) 419.0 m in seam V, (b) 388.5 m in seam VI + VII, (c) 334.8 m in seam VIII, (d) 265.4 m in seam IX and (e) 236.8 m in seam X.

Figure 16
Figure 16 (a) The layout of the resistivity image log in well B2 from depth interval 622.0-628.0m consisting of depth interval in the first track, static image in the second track and dynamic image in the third track.The colour scale shows the increasing trend of resistivity magnitude.In the dynamic track, the drilling-induced fractures (DIFs) are indicated with an arrow mark, which is observed as thin parallel lines separated by 180 degrees.The fourth track is the dip track that shows the dip magnitude and orientation of DIFs.(b) Stereonet plot showing the DIFs are oriented toward the NE-SW direction (pink shade) and the pink cross is the dip magnitude.

Figure 17
Figure 17 The error analysis by comparing error (degree) versus depth plot in (a) seam X, (b) seam IX, (c) seam VIII, (d) seam VI + VII and (e) seam V.3.The depth-wise correlation of anisotropy phenomena with the resistivity image log confirms the origin of anisotropy due to fractures/cleats.The fast shear wave polarization angle is consistent with the fractures/cleats orientation.4. The analysis from the resistivity image log reveals that the NE-SW is the maximum horizontal stress direction and supports the result from the previous study.Besides, the cleats either orient along NE-SW or NW-SE. 5.This work provides vital information for drilling horizontal/inclined wells, the production flow path, the coal seamwise model for oriented perforation and future geomechanical studies.

Table 1
Geophysical log parameters range against coal, sandstone and shale in west Bokaro coalfield

Table 3
Range of parameters (Vp/Vs, Vst), identification of fractures and cleats, its dip and orientation, anisotropy magnitude and fast shear wave orientation in five coal seams Coal seams (depth)