Sage Journals: Discover world-class research

Abstract

This paper presents a study on the impact of well integrity failure on coalbed methane (also known as coal seam gas) production and potential shallow water contamination using numerical simulations with a finite-difference method. Two connection types and 12 cases were simulated: Type A – leakage through cement sheath and Type B – impaired zonal isolation at the aquifer interval. The effect of the distance between the aquifer and the coal seam, drainage area and desorption time on gas and water production was also inspected. Results show that both connection types have strong effects on the gas and water production; the cumulative water and gas production increases with increasing drainage radius; the distance between aquifer and the coal seam has a negative effect on the water production but a negligible effect on the gas production; desorption time, ranging from 5 to 30 days, has a negligible effect on the water and gas production. Connection Type A yields a potential water contamination but connection Type B does not. Gas concentration in the shallow aquifer decreases sharply with an increase of distance away from the producer and the unsafe area are within an area with a radius ranging from approximately 50 m to 90 m away from the producer in this study.

Keywords

Coalbed methane well integrity shallow aquifer water contamination numerical simulation

Introduction

A wellbore is a pathway through which water and coalbed methane (CBM) are produced. It is usually 200–1000 m in depth (APLNG, 2015) for a CBM well. Hence, the CBM well has to penetrate shallower strata, which may contain groundwater used by farmers, before reaching the target horizons (Davies et al., 2014). CBM wells’ completions are similar to conventional gas wells but with modifications incorporated into the completion procedures because of the coals’ unique properties, e.g. friable, a fractured system, swelling and shrinkage, and fines migration (Pike et al., 2008). In the CBM industry, the well completion programs include open-hole completion, open-hole cavitation and cased-hole completion (Pike et al., 2008). The open-hole completion has advantages that no casing is left to obstruct production; there is no damage to the coal face from cementing and there is unobstructed access to the coal face from the wellbore (Pike et al., 2008). The cavity completion is better for coalbeds with thick seams, good permeabilities, extensive cleats, low ash contents, overpressuring and high in situ stresses. However, if the well penetrated multiple thin coal seams and aquifers, a cased-hole completion is better because it offers the option of selective perforation intervals and avoids the connection between aquifer and borehole, etc. (Bellarby, 2009).

Although cased-hole completion is used, the shallow aquifer may still be connected with coal seams from the leakage through cement sheath which may be caused by poor well completion practices, corrosion of steel casing, the deterioration of cement and shear failure or de-bonding of cement sheath (De Andrade et al., 2015; Davies et al., 2014; Feng et al., 2016). Teodoriu and Asamba (2015) reported that the cement permeability is dependent on salt concentration. Connell et al. (2015) investigated the cement degradation at the cement-formation interface using flooding experiments and found that the erosion rate of the cement has a significant relationship with the calcium concentration difference between the inflow and the outflow. In general, the estimated life of a CBM well ranges from 15 to 20 years or longer (CBM Projects, 2014) which will increase the risk of well integrity failures, e.g. corrosion of steel casing and the deterioration of cement, except for the poor well completion practices. Table 1 summarises the mechanisms of wellbore integrity failure before and after production.

Table 1.

Wellbore integrity before and after production.

Before production	Mechanisms of zonal isolation failure
	Formation damage during drilling (caving)
	Casing centralisation (incomplete cementing)
	Inadequate drilling mud removal
	Incomplete cement placement
	Incomplete cement/formation bond
	Cement shrinkage
	Contamination of cement by mud or formation fluids
After production	Mechanical stress/strain: micro-annulus at casing/cement interface; disruption of formation cement bond; fracture formation within cement
After production	Geomechanical attack: corrosion of casing; degradation of casing (carbonation, sulfate and acid attack)

Adapted from McDaniel and Shadravan (2013).

There are environmental concerns related to the potential contamination of shallow aquifers caused by CBM production from well integrity failure (Cheung et al., 2010; CSIRO, 2012a, 2012b; Williams et al., 2012). The concentration limits for methane in water as defined by the US Department of the Interior, Office of Surface Mining are 10 mg/L of safe limit and 28 mg/L of immediate action limit (Water Research Center, 2011). High concentration of methane in water is potentially explosive because methane can outgas from water in 2 minutes which will yield a higher concentration of methane in the air than the flammable limit (5–15 v% for methane in the air; CSIRO, 2012a, 2012b; Williams et al., 2012). Osborn et al. (2011) reported the contamination of drinking water in 51 of 60 drinking-water wells (85%) by methane with methane concentration 17 times higher on average in shallow wells from active drilling and extraction areas than in wells from non-active areas; the measured highest methane concentration is 64 mg/L. The studied area by Osborn et al. is the Catskill and Lockhaven formations that overlie the Marcellus Shale in Pennsylvania and the Genesee Group that overlies the Utica Shale in New York. Osborn et al. (2011) also concluded that there is no evidence for contamination of drinking water from deep saline brines or fracturing fluid. Cheung et al. (2010) reported leakages of methane from coal seams into shallow groundwater wells in Alberta, Canada, by analysing the major ion and isotope geochemistry.

Currently, there are few studies on the impact of well integrity failures on CBM production and shallow aquifer’s contamination using numerical simulation. This paper presents a conceptual study on this topic. Two connection types and 12 cases were simulated: Type A – leakage through cement sheath (Boukhelifa et al., 2005), that is coal seam is connected with aquifer via the cement-formation interface (Agbasimalo and Radonjic, 2014; Gu and Qin, 2010; Ladva et al., 2005; Opedal et al., 2014; Teodoriu et al., 2013) which can be caused by the poor well completion practices or cement deterioration; and Type B – impaired zonal isolation at the aquifer interval, in which the wellbore is connected with aquifer through the middle section of cement and casing in an aquifer interval which may be caused by the corrosion of steel casing (Davies et al., 2014), cement deterioration (Teodoriu et al., 2013) and mechanical damage (Thiercelin et al., 1998). Effect of the distance between the aquifer and coal seam, the drainage area and the desorption time on CBM production and shallow aquifer’s contamination is also inspected.

Numerical simulation models

Mass conservation equations

Computer Modelling Group’s GEM, a finite-difference simulator, is used in this study. The CBM production simulation in this simulator is based on the following assumptions:

There are only two phases, water and gas, in the coal seam.

The temperature is constant.

Reservoir properties are homogeneous.

The gas adsorption/desorption obey Langmuir equation.

The gas is not dissolved in water.

Based on these conditions, the basic mass conservation equations are

\nabla \cdot [\frac{ρ_{g} {kk}_{g}}{μ_{g}} (\nabla p_{gf} - ρ_{g} g \nabla D)] - q_{gf} + q_{gmf} = \frac{\partial}{\partial t} (φ_{f} ρ_{g} S_{gf})

(1)

\nabla \cdot [\frac{ρ_{w} {kk}_{w}}{μ_{w}} (\nabla p_{wf} - ρ_{w} g \nabla D)] - q_{wf} = \frac{\partial}{\partial t} (φ_{f} ρ_{w} S_{wf})

(2)

P_{cgwf} (S_{gf}) = p_{gf} - p_{wf}

(3)

S_{gf} + S_{wf} = 1

(4) where subscripts w, g, f, m and o refer to the water, gas, fracture, matrix and original, respectively. Equations (1) and (2) refer to the mass conservation of gas and water, respectively. k_g and k_w are the relative permeabilities for gas and water phases, respectively; μ is the viscosity; ρ is the density; k is the permeability; g is the gravity constant; q is the gas rate; ϕ is the porosity; S is the saturation; p is the pressure; D is the elevation; t is the time; P_c is the capillary pressure. Equations (1) to (4) were solved numerically in GEM by finite-difference algorithm (CMG User’s Guide, 2012).

Numerical model setup

Figure 1 shows a typical cased hole for a CBM well. In this well, two casings were installed to isolate aquifers from the ground surface and target zones. The true vertical depth at the top of coal seam is 530 m and the coal seam thickness is 10 m, while the true vertical depth at the top of shallow aquifer is varied with its distance to the coal seam, the thickness of the aquifer is 30 m.

Figure 1.

Schematic diagram showing vertical formations and well completion structure.

In order to simulate the impact of well integrity on CBM production and a shallow aquifer, a radial geometry with cell numbers of 40 and 36 in R and theta directions, respectively is used in simulations. Vertically, the grid size is 5 m, 10 m and 2.5 m for the sandstone, shale and coal seam, respectively. Vertical permeabilities were assumed as 1 mD and 200 mD, and porosities were assumed as 2% and 30% for the coal seam and sandstone, respectively. The shale formation is impermeable. A high permeability conduit caused by the failure of well integrity has an equivalent permeability of 100 D.

Figure 1 shows the two connection types which were studied in this study. Type A represents a leakage through cement sheath and Type B represents an impaired zonal isolation at the aquifer interval. For connection Type A, all grids with a number of two (red grids in Figure 2(b)) in radial direction were assigned a vertical permeability of 100 D. Figure 2 shows the built model for case No. S1-1. For connection Type B, integrity failures of cement (between casing and aquifer) and casing at the middle point of the aquifer were assumed. Figure 3 shows the built model for case No. S2-1.

Figure 2.

2D areal grid system (a) and the conduit with high permeability (b) for case No. S1-1. Colour shows the vertical permeability for sandstone in green and the high permeability conduit in red.

Figure 3.

A vertical 2D section shows the horizontal permeability for case No. S2-1.

Table 2 lists all the 12 cases and their parameters and Table 3 provides the basic parameters of all simulations in this study. Figure 4 shows the relative permeability curves used in simulating water and gas flow in the coal seam, sandstone and high permeability conduits. The simulation models were initialised by the gravity-capillary equilibrium. The initial pressure is 5090 kPa at the reference depth of 509 m for both matrix and cleats by assuming a hydrostatical pressure system. A producer is located in the centre and the perforation intervals are shown in Figure 3. This well is produced with a maximum water rate of 200 m³/d and minimum bottom-hole pressure (BHP) of 300 kPa as constraints. All simulations were run for 10 years with a closed boundary which mimics a circular drainage area with a radius of 300 m.

Table 2.

Simulation cases and parameters.

Case no.	Connection type between the aquifer and coal seam	L_AC (m)	R (m)	Desorption time (τ), days
Base case	No	300	300	30
S1-1	Type A	300	500	30
S1-2	Type A	300	400	30
S1-3	Type A	300	300	30
S1-4	Type A	300	200	30
S1-5	Type A	200	300	30
S1-6	Type A	100	300	30
S1-7	Type A	50	300	30
S1-8	Type A	300	300	20
S1-9	Type A	300	300	10
S1-10	Type A	300	300	5
S2-1	Type B	300	500	30

Note: L_AC is the distance between the aquifer and coal seam; R is the drainage radius.

Table 3.

Basic parameters of all simulations.

Parameters	Values
Model	Dual porosity
Shape factor formulation	Gilman–Kazmi
Model geometry	3D radial (cylindrical) with a vertical well at the centre
Grid system	40×36×41 with R_w of 0.1 m and out radius of 500 m, 400 m, 300 m and 200 m (see Table 2)
Sandstone/coal thickness	30 m/10 m
Matrix porosity, ϕ_m	5% (coal)
Cleat porosity, ϕ	30% (sandstone), 2.0% (coal)
Horizontal permeability, k_h	2000 mD (sandstone), 10 mD (coal)
Vertical permeability, k_v	0.1·k_h
Reservoir pressure at the top of coalbed for both matrix and cleats	5090 kPa at depth of 509 m
Langmuir pressure, P_L	5000 kPa
Langmuir volume, V_L	0.5 mole/kg-coal
Porosity and permeability in high conduits	1%, 100 D
Coal density, ρ	1435 kg/m³
Separation condition	15℃ and 101.3 kPa

Figure 4.

Relative permeability for water and gas in the fracture.

Results and discussion

Impact of the aquifer on CBM production

Figure 5 shows the simulated cumulative water and gas production between the base case, in which the coal seam is isolated with an aquifer by a thick mudstone, and the case No. S1-3 in which the coal seam is connected with the aquifer via the connection Type A. The water from the shallow aquifer first flows to the deeper coal seam then flows into the wellbore. Results show that the cumulative gas productions are similar but a higher cumulative water production in case No. S1-3 than that from the base case which has no connection between the aquifer and coal seam. This is because the producer is mostly constrained by BHP of 300 kPa during production which yields a similar cumulative gas production. More water is produced to maintain the same BHP as a higher feeder of water from the aquifer. The incremental water production, about 11,000 m³ at the end of the simulation, is contributed by the shallow aquifer.

Figure 5.

Comparison of water production for models without (base case) and with connection Type A.

Figure 6 shows the comparison of cumulative gas and water production for case Nos. S1-1 through S1-4 which are in different drainage areas. The cumulative gas and water production at the time point of the 3600th day increases linearly with increase in drainage radius (Figure 7). This is because all simulations are simulated at the same BHP and the production reaches pseudo-steady state in the extended period.

Figure 6.

Comparison of gas (a) and water (b) production for models with different drainage areas but with same connection Type A.

Figure 7.

Relationship of drainage radius with cumulative water and gas production at 3600th day.

Figure 8 shows the comparison of cumulative gas and water production for case Nos. S1-3, S1-5, S1-6 and S1-7 which are in different distances between the aquifer and coal. Results show that the distance has no effect on the gas production but a minor effect on the water production because of gravity. The cumulative water production at the end of simulation decreases with the increase of the distance between the aquifer and coal seam (Figure 9). This is because the aquifer in case No. S1-7 has the highest initial pressure which yields a higher water potential and initial water volume in place.

Figure 8.

Comparison of gas (a) and water (b) production for models with different distances between the aquifer and coal seam but in the same connection Type A. Inset: a portion of the figure zoomed.

Figure 9.

The relationship between the cumulative water production at 3600th day with the distance between coal seam and aquifer.

Figure 10 shows the comparison of the cumulative gas and water production for case Nos. S1-3, S1-8, S1-9 and S1-10 which are with different desorption times. Results show that the desorption time has a negligible effect on cumulative gas and water production. This is because the gas desorption time has a small cumulative sensitivity index to water and gas production compared to other parameters, e.g. permeability, coal seam thickness, etc. (Zhou, 2012).

Figure 10.

Comparison of gas (a) and water (b) production for models with different desorption times but the same connection Type A. Inset: a portion of the figure zoomed.

Figure 11 shows the comparison of the cumulative gas and water production rate between case Nos. S1-1 and S2-1 which have different connection types. Results show that the connection Type B yields lower cumulative water and gas production.

Figure 11.

Comparison of gas (a) and water (b) production for models with different connection types.

Methane concentration in the aquifer

In this study, we calculated gas concentration for each grid using equation (5) based on the simulated gas mass density which was calculated by the software using the equations of state (see Appendix 1), gas saturation and water saturation

C_{g} = \frac{ρ_{gr} \cdot S_{g}}{S_{w}}

(5) where ρ_gr is the mass density of methane at reservoir condition in mg/L, S_g and S_w are the gas and water saturations.

Figure 12 shows the variation of the grid-based average methane concentration along the radius of the top layer in the aquifer at the 3600th day. According to the hazard methane concentration limit of 28 mg/L, the safe radius should be greater than 90 m for case No. S1-10 while it is about 50 m for case No. S1-1. The methane concentration decreases sharply from 10,000 mg/L to approximately 10 mg/L within 100 m of the producer. Osborn et al. (2011) reported that the methane concentration also decreases sharply but the safe radius away from the wellbore is about 500 m which is longer than the results from this study. However, this comparison is very rough for the shortage of details about the reservoir properties, well conditions and operations, etc.

Figure 12.

(a) Simulated average methane concentration from top layer grids vs. the radial distance to the producer at the 3600th day. (b) Measured methane concentration against distance to the producer from Osborn et al. (2011).

The connection Type B has no effect on shallow water quality because the gas in the wellbore is extracted directly from the annular space between the tubing and casing. However, if the well is abandoned, the gas in the wellbore may invade into the aquifer by the gravity and the density difference between gas and water.

Conclusions

Based on the numerical simulation of 12 cases, the effects of the connection type, the distance between the aquifer and coal seam, the drainage area and the desorption time on the water and gas production and the methane concentration in the shallow water were studied. Analysing a real field data to verify the methods used in this study need to be done in the future. Note that this is a conceptual study to address the consequence if there is a well integrity issue. But this rarely happens in CBM industry because of the matured well completion technologies. Following are the conclusions drawn based on this study:

Both connection types, Type A – leakage through cement sheath and B – impaired zonal isolation at the aquifer interval, impact the gas and water production.

With the connection Type A, the cumulative water and gas production increases linearly with the increase of drainage radius; the distance between the aquifer and the coal seam has a negative effect on the cumulative water production but a negligible effect on the gas production and the desorption time, ranging from 5 to 30 days, has a negligible effect on the water and gas production.

The Type A connection affects the water quality but Type B does not. The gas concentration decreases sharply with an increase of distance away from the producer and the unsafe radius ranges from approximately 50 m to 90 m in this study.

Footnotes

Acknowledgements

We thank the anonymous reviewer for the constructive comments. We also acknowledge the Computer Management Group for providing the simulator.

Declaration of conflicting interests

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

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work is supported by the National Natural Science Foundation of China (41472127, 41472129, 41502158), Key Coal-Based Technologies R & D Program of Shanxi Province (MQ2014-01) and the Joint Research Fund for Coalbed Methane Gas of Shanxi Province (2013012004).

Appendix

References

Agbasimalo

Radonjic

(2014) Experimental study of the impact of drilling fluid contamination on the integrity of cement-formation interface. Journal of Energy Resources Technology 136(4): 1–5.

APLNG (2015) Available at: www.aplng.com.au/home/what-coal-seam-gas (accessed 28 September 2017).

Bellarby J (2009) Well Completion Design, Volume 56 (Developments in Petroleum Science). 1st ed. Elsevier Science, pp. 45–46. Amsterdam, The Netherlands.

Boukhelifa

Moroni

James

et al. (2005) Evaluation of cement systems for oil and gas well zonal isolation in a full-scale annular geometry. SPE Drilling & Completion 20(1): 44–53.

CBM Projects (2014) Available at: www.bioregionalassessments.gov.au/assessments/12-resource-assessment-maranoa-balonne-condamine-subregion/12321-coal-seam-gas-projects-under-construction (accessed 28 September 2017).

Cheung

Klassen

Mayer

et al. (2010) Major ion and isotope geochemistry of ﬂuids and gases from coalbed methane and shallow groundwater wells in Alberta, Canada. Applied Geochemistry 25: 1307–1329.

CMG User’s Guide (2012) Advanced Compositional Reservoir Simulator. Computer Modelling Group Ltd.

Connell

Down

et al. (2015) An investigation into the integrity of wellbore cement in CO₂ storage wells: Core ﬂooding experiments and simulations. International Journal of Greenhouse Gas Control 37: 424–440.

CSIRO (2012a) Coal seam gas developments predicting impacts. Available at: www.csiro.au/news/coal-seam-gas (accessed 28 September 2017).

10.

CSIRO (2012b) Water in mining and industry [Chapter 10]. In: Prosser I, Wolf L and Littleboy A (eds). Available at: www.publish.csiro.au/?act=view_file&file_id=9780643103283_Chapter_10.pdf (accessed 28 September 2017).

11.

Davies

Almond

Ward

et al. (2014) Oil and gas wells and their integrity: Implications for shale and unconventional resource exploitation. Marine and Petroleum Geology 56: 239–254.

12.

Deremble

Loizzo

Huet

et al. (2011) Stability of a leakage pathway in a cemented annulus. Energy Procedia 4: 5283–5290.

13.

De Andrade J, Sangesland S, Todorovic J, et al. (2015) Cement sheath integrity during thermal cycling: A novel approach for experimental tests of cement systems. In: SPE 173871 paper presented at the SPE Bergen one day seminar, Bergen, Norway, 22 April 2015.

14.

Feng Y, Podnos E and Gray KE (2016) Well integrity analysis: 3D numerical modeling of cement interface debonding. In: ARMA 16-246 paper presented at the 50th US rock mechanics/geomechanics symposium, Houston, Texas, USA, 26–29 June 2016.

15.

Qin

(2010) Experiments on integrated solidification and cementation of the cement-formation interface using the Mud Cake to Agglomerated Cake (MTA) method. Petroleum Exploration and Development 37(2): 226–231.

16.

Ladva

HKJ

Craster

Jones

TGJ

et al. (2005) The cement-to-formation interface in zonal isolation. SPE Drilling & Completion 20(3): 186–197.

17.

McDaniel J and Shadravan A (2013) Zonal isolation critical in developing unconventional resources. The American Oil & Gas Reporter.

18.

Opedal

NvdT

Torsxter

Vralstad

et al. (2014) Potential leakage paths along cement-formation interfaces in wellbores: Implications for CO₂ storage. Energy Procedia 51: 56–64.

19.

Osborn

Vengosh

Warner

et al. (2011) Methane contamination of drinking water accompanying gas-well drilling and hydraulic fracturing. PNAS 108(20): 8172–8176.

20.

Pike W, Laas R, Barbee MA, et al. (2008) Chapter 7: Advances in Coalbed Methane. Houston, TX, USA: Hart Energy Publications.

21.

Redlich

Kwong

JNS

(1949) On the thermodynamics of solutions. V: An equation of state. Fugacities of gaseous solutions. Chemical Reviews 44: 233–244.

22.

Soave

(1972) Equilibrium constants from a modified Redlich–Kwong equation of state. Chemical Engineering Science 27: 1197–1203.

23.

Teodoriu

Asamba

(2015) Experimental study of salt content effect on class G cement properties with application to well integrity. Journal of Natural Gas Science and Engineering 24: 234–329.

24.

Teodoriu

Kosinowski

Amani

et al. (2013) Wellbore integrity and cement failure at HPHT conditions. International Journal of Engineering and Applied Sciences 2(2): 1–13.

25.

Thiercelin M, Baumgarte C and Guillot D (1998) A soil mechanics approach to predict cement sheath behavior. In: SPE 47375 paper presented at the 1998 SPE/ISRM rock mechanics in petroleum engineering conference, Trondheim, Norway, 8–10 July.

26.

Water Research Center (2011) Available at: www.water-research.net/index.php/methane (accessed 28 September 2017).

27.

Williams J, Stubbs T and Milligan A (2012) An analysis of coal seam gas production and natural resources management in Australia: Issues and ways forward. A report prepared for the Australian Council of Environmental Deans and Directors, Joan Williams Scientific Services Pty Ltd, pp. 51–52. Available at: http://aie.org.au/AIE/Documents/Oil_Gas_121114.pdf (accessed 28 September 2017).

28.

Zhou

(2012) History matching and production prediction of a horizontal coalbed methane well. Journal of Petroleum Science and Engineering 96–97: 22–36.

Numerical investigation of the potential contamination of a shallow aquifer in producing coalbed methane