Numerical Evaluation of a Novel Development Mode for Challenging Oceanic Gas Hydrates Considering Methane Leakage

Abstract

The exploitation of challenging oceanic gas hydrate reservoirs with low permeability and permeable boundary layers faces the challenges of methane leakage and low production. Considering this aspect, a novel five-spot injection–production system combined with hydraulic fracturing was proposed. In particular, the potential of this development mode, including hydrate dissociation, gas production, and gas capture, was evaluated in comparison with a three-spot injection–production system. The results showed that increasing the fracture conductivity cannot prevent CH₄ leakage in the three-spot, and the leakage accounted for 5.6% of the total gas production, even at the maximum fracture conductivity of 40 D·cm. Additionally, the leakage amount increased as the well spacing increased, and the leakage accounted for 36.7% of the total gas production when the well spacing was 140 m. However, the proposed development mode completely addressed CH₄ leakage and significantly increased gas production. The average gas production rate reached 142 m³/d per unit length of the horizontal section, which was expected to reach the commercial threshold. The variance analysis indicated that optimal plans for the challenging hydrates in the Shenhu area were well spacing of 100–120 m and fracture conductivity greater than 20 D·cm.

1. Introduction

Natural gas hydrate (NGH) is a cage-structured crystalline solid formed by natural gas (primarily methane) and water under high pressure and low temperature [1]. The global natural gas reserves in NGH are approximately 2.5–120 × 10¹⁵ stm³, nearly twice the total carbon content of fossil fuels [2,3,4]; thus, NGH is generally considered to play a crucial role in future energy supply and climate management [5,6].

NGH is contended to be a challenging natural gas reservoir because its development faces complex phase transition, heat transfer, multiphase flow, formation deformation, and many other challenges, compared with conventional natural gas development [6,7,8,9,10]. According to the reservoir properties and occurrence environment, the challenging NGH reservoirs are characterized by low initial reservoir temperature (CH-k), lack of impermeable boundary layers (CH-b), high hydrate stability (CH-s), and low reservoir permeability (Ch-k) [11]. A vast majority of global oceanic NGH reservoirs have both low reservoir permeability and permeable boundary layers (CH-kb) [12,13], which is the main type of oceanic NGH and the primary target for commercial development.

The premise of efficient gas recovery from NGH reservoirs is to obtain a considerable hydrate decomposition rate. However, the depressurization method, which is generally considered to be the most promising, exhibits a low hydrate decomposition rate and poor gas production in CH-kb [11,14,15]. This is caused by two mechanisms: first, the low reservoir permeability and permeable boundaries inhibit pressure drop transmission and fluid flow [16,17,18]; second, insufficient reservoir sensible heat further reduced the hydrate decomposition capacity [19,20]. Therefore, exploring the efficient development mode of CH-kb has recently become an urgent topic.

Depressurization combined with thermal fluid injection could provide extra heat for hydrate dissociation, which is a promising improved scheme for NGH development [21,22]. In particular, the injection–production system with multiple wells, such as two-spot [23] and three-spot (TS) well patterns [24], significantly increases drainage area and enhances gas recovery, and has been envisaged by Japan for the commercial development of NGH [25]. However, its direct applicability in CH-kb is problematic because the gas released by hydrate thermal decomposition is difficult to pass through the inter-well hydrate undissociated zones where hydrates fill in pores and thereby form barriers to gas production. At present, using hydraulic fracturing (HF), a key stimulation technology in the commercial development of shale and tight gas, to create high conductivity artificial fractures between the injection well (well_i) and production well (well_p) is a promising solution. Several numerical simulations confirmed that fractures significantly enhanced hydrate decomposition and gas production capacity, wherein hydraulic fracturing combined with a three-spot well pattern (HF + TS) was used [26,27,28]. However, existing studies have not yet focused on the risk of methane leakage in hydraulic fractured CH-kb. When the hydrate decomposition front breaks through to the permeable boundaries, the gas released by hydrate thermal decomposition escapes into boundary layers under the drive of high injection pressure [29]. This is unfavorable for gas recovery and potentially pollutes the oceanic and atmospheric environment. Recently, methane released by oceanic hydrate dissociation is considered to play an important role in the global carbon cycle, as it is leaking through permeable seabed or natural fractures [30,31,32,33,34]. Furthermore, the massive release of methane induced by NGH development in the future may affect the atmospheric CH₄ budget and exacerbate the greenhouse effect [35,36,37]; therefore, there is an urgent need to evaluate the risk of methane leakage in hydraulic fractured CH-kb.

In our previous studies, we successfully addressed methane leakage in the non-fractured CH-kb using a novel five-spot well pattern, wherein two capture wells (well_c) were arranged at the boundary layers to recover the escaped gas [29]. However, because the low reservoir permeability weakens the inter-well interaction, the spacing between injection and production wells is limited to 50 m. Small space well pattern significantly increases the drilling cost, reduces economic benefits, and is not conducive to the commercial development of CH-kb. Given this, a novelty development mode combining hydraulic fracturing and the five-spot injection-production system (HF + FS) is proposed here to address the aforementioned challenges. In this study, the production potential of the proposed development mode (HF + FS) for CH-kb at the SH2 site in the Shenhu area, South China Sea, was evaluated by comparing it with HF + TS. Specifically, the impacts of fracture conductivity (F_c) and well spacing (W_s) on hydrate decomposition, gas production, methane escape, and gas recovery under the two development modes were thoroughly discussed, and ideal development plans were obtained through variance analysis (ANOVA). These findings provide valuable references for the safe and efficient development of CH-kb.

2. Modeling

2.1. Development Plan Design

The implementation of HF + FS includes the following steps: deploying the five-spot well pattern, performing hydraulic fracturing between the well_i and well_p, and production with depressurization combined with thermal fluid injection.

As shown in Figure 1, all wells are horizontal and extend along the Y direction. The well_i and well_c are arranged by drilling a three-branch horizontal well, wherein the middle section is used for injection, and the upper and lower sections are used for gas capture. Additionally, well_i and well_p are connected through a horizontal artificial fracture created by hydraulic fracturing, and its width is 10 mm [38,39]. Due to the lack of engineering cases, the injection pressure and temperature in existing simulations are 16–22 MPa and 30–90 °C, respectively [21,22,23,24,26,27,28,29]. For comparison, we adopted the same injection–production scheme as that in our previous study [29], that is, the injection temperature and pressure were 60 °C and 16 MPa, and the bottom hole pressures of well_p, upper well_c, and lower well_c were 4, 10, and 11 MPa, respectively. Additionally, geomechanical issues are not considered in the model, because current research mainly focuses on CH₄ migration and hydrate decomposition behavior under the two development modes.

2.2. Generalization of the Fracture

In this study, the fracture is generalized as porous media with high conductivity, and the multi-flow in the fracture is described as a continuous flow that follows Darcy’s law [40]. Therefore, the two-phase seepage in the matrix and fracture that considers the gravity and capillary forces can be given by Ref. [41]:

$$F^l=-k\frac{k_r^l}{{\mu }^l}\left(\nabla P^l-{\rho }^{l}g\right),$$

(1)

$$F_i^g=-k\frac{k_r^g{\rho }^g{C}_i^g}{{\mu }^g}\left(1+\frac{\omega }{P^g}\right)\left(\nabla P^g-{\rho }^{g}g\right),$$

(2)

where $F^l$ and $F_i^g$ are the fluxes of the liquid and gas phases, respectively, wherein l, g, and i represent the liquid phase, gas phase, and gas component, respectively; $k$ is the absolute permeability of the matrix, m²; $k_r^l$ and $k_r^g$ are the relative permeabilities; ${\mu }^l$ and ${\mu }^g$ are the fluid viscosities, Pa·s; $P^l$ and $P^g$ are the fluid pressures, Pa; ${\rho }^l$ and ${\rho }^g$ are the fluid densities, kg/m³; $g$ is the gravitational acceleration, m/s²; $C_i^g$ is the mass friction; and $\omega$ is the Klinkenberg factor, Pa.

The $k_r^l$ and $k_r^g$ are expressed as [41]:

$$k_r^l={\left(\frac{S^l-S_{ir}^l}{1-S_{ir}^l}\right)}^{n^l},$$

(3)

$$k_r^g={\left(\frac{S^g-S_{ir}^g}{1-S_{ir}^g}\right)}^{n^g},$$

(4)

where $S^l$ and $S^g$ are the water and gas saturation, $S_{ir}^l$ and $S_{ir}^g$ are the irreducible saturation of liquid and gas, $n^l$ and $n^g$ are the stone indexes of liquid and gas.

The value of $P^l$ can be obtained by Ref. [42]:

$$P^l=P^g+P_c,$$

(5)

where $P_c$ is the capillary pressure, Pa, and is expressed as [43]:

$$P_c=-P_t{\left[{\left(\frac{S^l-S_{ir}^l}{1-S_{ir}^l}\right)}^{-1/\vartheta }-1\right]}^{1-\vartheta },$$

(6)

where $P_t$ is threshold pressure, Pa; and $\vartheta$ is the pore structure index.

2.3. Numerical Code

To simulate the production behavior of CH-kb during depressurization and thermal fluid injection, the hydrate reaction based on thermodynamic conditions, mass conservation, energy conservation, and the flux of possible components and phases should be considered. The details of these governing equations have been given in previous studies [23,41]. Several simulators have been developed to solve these governing equations, such as Tough + Hydrate (T + H), and MH21-Hydrate [38]. T + H is developed by Lawrence Berkeley Laboratory and is popular in simulating field-scale hydrate reservoir development [41]. Three forms of heat exchange, heat conduction, heat convection, and heat radiation, are considered in T + H. Heat sources include hydrate reaction heat, phase change latent heat, the sensible heat of the reservoir itself, etc. The fluid flow in T + H follows Darcy’s law, and the convection and molecular diffusion of dissolved gases and inhibitors in water are considered. The code includes four components, including water, CH₄, CH₄ hydrate, and inhibitor, and four phases, including gas, liquid, ice, and hydrate. Hydrate decomposition is based on the equilibrium model or kinetic model [44]. Therefore, T + H is powerful in modeling non-isothermal hydrate decomposition, phase behavior, heat exchange, and fluid flow in hydrate reservoirs. Furthermore, the reliability of this code has been verified both at the laboratory and field scales [45,46]; thus, T + H is employed in this study.

2.4. Initial Conditions and Model Parameters

Post-drilling data at the SH2 site indicates that hydrates occur in argillaceous silt sediment below 185 mbsf (meter below seafloor), with thickness, porosity, and hydrate saturation of 0–40 m, 0.33–0.48, and 26–48%, respectively (40 m, 0.38, and 44%, respectively, in this study) [47]. Core testing revealed that the absolute permeabilities of the overburden, hydrate-bearing layer (HBL), and underburden were 10 mD [48]. Assuming that the reservoir properties are homogeneous along the Y direction, horizontal wells with long production sections are used for gas recovery and fluid injection, thus the thickness of the model in the Y direction could be set to 1 m, as shown in Figure 1.

The reservoir temperature and geothermal gradients were 14.87 °C and 0.047 °C/m, respectively [49]; therefore, the domain temperature can be obtained. The pore water is assumed to be interconnected due to the permeable boundary layers, and the pressure distribution was derived from the reservoir pressure of 14.97 MPa and pressure gradient of 0.01 MPa/m [49]. The temperature and pressure at the upper and lower boundaries of the model are considered constant because HBL is covered by seawater and underlain by enough thick sediment. Based on this, the initial equilibrium state of the reservoir is obtained (Figure 2). The values of other model parameters are listed in

Table 1. Model parameters.

Parameters	Value
Formation thickness (Overlayer, HBL, and underlayer)	30, 40, 30 m [26]
Intrinsic permeability of wells	1 × 10−9 m2 (=1000 D) [11]
Intrinsic porosity of wells	1 [11]
Initial water saturation (Overlayer, HBL, and underlayer)	1, 0.56, 1
Porewater salinity (Overlayer, HBL, and underlayer)	0.03
Gas composition	CH4 100%
Dry thermal conductivity (All deposits)	1 W/m/K
Wet thermal conductivity (All deposits)	3.10 W/m/K
Grain density (All deposits)	2600 kg/m3
Grain-specific heat (All deposits)	1000 J·K/kg
Threshold pressure $P_t$	1 × 105 Pa [27]
pore structure index $\vartheta$	0.45
Index for liquid phase $n^l$	3.50
Index for gas phase $n^g$	3.50
Irreducible aqueous saturation $S_{ir}^l$	0.30
Irreducible gas saturation $S_{ir}^g$	0.03

.

2.5. Simulation Scenarios

The fracture conductivity depends on the stress condition, proppant performance, etc., mostly in dozens of D cm [34,35,36,37]. To obtain an ideal development plan, four levels of fracture conductivity are set at 5, 10, 20, and 40 D·cm, and four levels are set at 80, 100, 120, and 140 m for well spacing, as presented in

Table 2. Simulation scheme.

Scenario Number	Development Mode	Ws (m)	Fc (D·cm)	Scenario Number	Development Mode	Ws (m)	Fc (D·cm)
01#	TS	100	_	13#	HF + FS	80	10
02#	HF + TS	100	5	14#	HF + FS	100	10
03#	HF + TS	100	10	15#	HF + FS	120	10
04#	HF + TS	100	20	16#	HF + FS	140	10
05#	HF + TS	100	40	17#	HF + FS	80	20
06#	HF + TS	80	20	18#	HF + FS	100	20
07#	HF + TS	120	20	19#	HF + FS	120	20
08#	HF + TS	140	20	20#	HF + FS	140	20
09#	HF + FS	80	5	21#	HF + FS	80	40
10#	HF + FS	100	5	22#	HF + FS	100	40
11#	HF + FS	120	5	23#	HF + FS	120	40
12#	HF + FS	140	5	24#	HF + FS	140	40

.

3. Results

3.1. Comparison of Different Development Modes

Based on the simulation results of scenarios 01, 04, and 18#, we compared the production potential of different development modes (TS, HF + TS, HF + FS) for CH-kb.

3.1.1. Spatial Evolution of the Physical Properties

1.

Reservoir temperature evolution

Figure 3 depicts the reservoir temperature evolution in different development modes. In TS, the high-temperature (higher than 30 °C) zone initially spreads circularly around well_i (Figure 3a,b). At 1800 d, the thermal decomposition front advanced to the boundaries (Figure 3c), and hot water flowed into the boundary layers, resulting in non-uniform heat diffusion (Figure 3c). Compared to that in TS, the temperature evolution in HF + TS differed considerably. First, hot water mainly flowed into the fracture, resulting in a faster horizontal heat diffusion rate. Second, the hot water injection was strengthened by the fracture, leading to a larger heating area. At 360 d, the horizontal and vertical heating radii are 68 and 9 m, respectively (Figure 3d), which are higher than that at 300 d (7 m) in TS. At 600 d, the high temperature around the well_p indicates that hot water has flowed into it (Figure 3e). This is a phenomenon known as water-flooding in the petroleum industry that dramatically increases the water yield, and the time of occurrence is T_f. At 1800 d, the vertical heating radius reached 44 m (Figure 3f). In HF + FS, vertical heat diffused faster at 600 d (Figure 3h). This is because the low pressure of well_c induces more hot water to migrate vertically. Moreover, since the well_c prevents hot water from passing through the permeable boundaries, the vertical heating radius is still 20 m at 1800 d (Figure 2i). Therefore, the heating efficiency is significantly enhanced by the fracture, and hot water infiltration into the boundary layer is avoided by well_c.

2.

Reservoir pressure evolution

Figure 4 shows the reservoir pressure evolution in different development modes. In TS, two low-pressure (≤13 MPa) zones and a high-pressure (≥15 MPa) zone are formed around well_p and well_i (Figure 4a–c), respectively. As the pressure transmission and fluid injection are inhibited by the low reservoir permeability, both the low- and high-pressure zones exhibit a slow diffusion rate. In HF + TS, the high-pressure zone diffuses faster along the fracture (Figure 4d). Moreover, the radius of the low-pressure zone is 48 m at 1800 d (Figure 4f), which is 2.09 times that in TS. At 1800 d, the radius of the high-pressure zone shrinks to 15 m due to the pressure relief (Figure 4f). In HF + FS, the low-pressure transmission of well_p is the same as that in HF + TS, and two low-pressure zones are also formed around well_c, which intersect with the low-pressure zones formed by well_p (Figure 4i). This indicates that the gas is well recovered, thereby relieving the reservoir pressure.

3.

Hydrate saturation evolution

Figure 5 shows the hydrate saturation evolution in different development modes. In TS, the depressurization decomposition zone and thermal decomposition zone expand circularly with a slow expansion rate (Figure 5a,b). At 1800 d, the decomposition radii of the thermal stimulation and depressurization are 13 and 21 m (Figure 5c), respectively, indicating that thermal stimulation has a better hydrate decomposition efficiency. Moreover, the gas leaking into the overburden recombines with free water and re-form hydrates (Figure 5c). In HF + TS, hydrates dissociate rapidly along the fracture, while more hydrates are generated in the overburden (Figure 5d–f). In HF + FS, the hydrate saturation evolution is similar to that in HF + TS, but no hydrate in the overburden (Figure 5g–i). Therefore, the arrangement of well_c effectively avoids the secondary formation of hydrate in the overburden.

4.

Gas saturation evolution

Figure 6 displays the gas saturation evolution in different development modes. In TS, gas saturation evolution is similar to that of hydrate saturation. The gas released by thermal stimulation accumulated at the decomposition front, and can not be produced (Figure 6a–c). Thus, there is no synergy effect between well_i and well_p, and thermally decomposed gas migrates into the boundary layer (Figure 6c). In HF + TS, the gas released by thermal stimulation migrates to the well_p through the fracture (Figure 6d–f). However, more gas leaked into the boundary layers (Figure 6d–f), and the farthest vertical escape position was 46 m away from welli (Figure 6f), which was higher than that in TS (35 m). In HF + FS, the gas saturation evolution is similar to that in HF + TS, but there is no methane in the boundary layers (Figure 6g–i). Therefore, methane leakage is addressed by well_c.

3.1.2. Hydrate Dissociation Behavior

Figure 7 shows the variation of the CH₄ released rate (Q_r) and hydrate dissociation ratio (D_r) in different development modes, wherein D_r is the ratio between the cumulative decomposed hydrate mass (M_d, kg) and the original hydrate mass (M_o, kg) in the reservoir, and is expressed as:

$${\mathrm{D}}_{\mathrm{r}}=\frac{{\mathrm{M}}_{\mathrm{d}}}{{\mathrm{M}}_{\mathrm{o}}},$$

(7)

In TS, Q_r is minimum and stable at approximately 50 m³/d. In HF + TS, the variation of Q_r exhibited three stages. In the first stage (0–50 d), the stable state of the hydrates was broken instantaneously; thus, the initial Q_r value was the highest. In the second stage (50–450 d), Q_r stabilized around 200 m³/d. In the third stage (450–1800 d), Q_r gradually decreases. This is caused by two factors: first, hot water bursts into the well_p along the fracture, which reduces the heating efficiency. Second, the hydrate decomposition front gradually moves away from the fracture. In HF + FS, the evolution of Q_r was similar to that in HF + TS, but it was higher in the first two stages and declined faster in the third stage. This is because the presence of the well_c further strengthened the initial injection, while the subsequent water-flooded further reduced the heating efficiency. The final D_r in TS, HF + TS, and HF + FS is 31%, 93%, and 95%, respectively. This indicates that the final D_r is improved by the fracture, whereas it is barely affected by the existence of well_c.

3.1.3. Gas Production Behavior

Figure 8 displays the variation of the gas production rate (Q_p) and cumulative gas production (V_p) in different development modes. In TS, Q_p is minimum and totally contributed by depressurization. In HF + TS, affected by Q_r, Q_p also presents three stages. In the first stage (0–85 d), Q_p decreases rapidly. Subsequently (85–374 d), gas migrates to well_p; thus, Q_p increases and reaches its peak of 143 m³/d, which is 5.72 times that in TS. In the third stage (374–1800 d), the decrease of Q_r leads to a decrease in Q_p. For HF + FS, the variation of Q_p is similar to that in HF + TS. Attributed to the gas recovery by well_c, the maximum Q_p reaches 203 m³/d, which is 1.43 times that in HF + TS. The final V_p in HF + FS is 20.75 × 10⁴ m³, which is 5.47 and 1.18 times of TS and HF + TS, respectively. Therefore, well_c and the fracture synergistically enhance the production.

3.1.4. Methane Leakage, Gas Capture, and Recovery Efficiency

The volume of methane leakage (V_l, m³) consists of two parts: the volume of free gas in the boundary layers (${\mathrm{V}}_{\mathrm{l}}^{\mathrm{f}}$, m³) and the volume of gas trapped in re-formed hydrates (${\mathrm{V}}_{\mathrm{l}}^{\mathrm{h}}$, m³), and is expressed as:

$${\mathrm{V}}_{\mathrm{l}}={\mathrm{V}}_{\mathrm{l}}^{\mathrm{f}}+{\mathrm{V}}_{\mathrm{l}}^{\mathrm{h}},$$

(8)

To characterize the methane capture ability of well_c, the cumulative gas production of well_c was defined as the volume of captured gas (V_c). Recovery efficiency (R_e) is defined as the ratio between cumulative gas production (V_p, m³) and the volume of gas released by hydrate decomposition (V_r, m³), and is expressed as:

$${\mathrm{R}}_{\mathrm{e}}=\frac{{\mathrm{V}}_{\mathrm{p}}}{{\mathrm{V}}_{\mathrm{r}}},$$

(9)

Figure 9 depicts V_l in TS and HF + TS, V_c in HF + FS, and R_e in different development modes at 1800 d. The V_l in TS and HF + TS are 15,080 and 18,389 m³, respectively, indicating that CH₄ leakage is aggravated by the fracture. The V_c in HF + FS is 40,750 m³ and there is no CH₄ leakage, which is consistent with the gas distribution shown in Figure 7i. Moreover, V_c is higher than V_l, indicating that well_c also enhances V_p. The R_e in HF + FS is 0.93, an increase of 0.14 and 0.42 relative to TS and HF + TS, respectively. Therefore, HF + FS is a better development mode for CH-kb. It is important to note that 7% of the gas was trapped in the reservoir at 1800 d (Figure 6i). To avoid methane leakage, depressurized production can be maintained until free gas is completely produced.

3.2. The Effect of Fracture Conductivity

Scenarios 02–05, 10, 14, 18, and 22# simulated the production behavior of the two development modes (HF + TS, HF + FS) with different fracture conductivity at the well spacing of 100 m. By comparing these simulation results, the effect of fracture conductivity on production behavior is analyzed.

3.2.1. Hydrate Dissociation Behavior

Figure 10 depicts the variation of Q_r in HF + TS and HF + FS at different F_c. In HF + TS, Q_r varies gently around 90 m³/d when F_c is 5 D·cm, which is similar to that in TS, indicating that an F_c value of 5 D·cm is insufficient for gas migration. When F_c is 10 D·cm, T_f is 1000 d. For F_c of 20 and 40 D·cm, T_f further increases to 400 and 70 d, respectively. Additionally, Q_r at 1800 d is lower when F_c ≥ 20 D·cm than that when F_c ≤10 D·cm. This is attributed to the inadequacy of remaining hydrates for the gas supply when F_c ≥ 20 D·cm. In HF + FS, Q_r is higher at the early stage of production and lower at the later stage of production. The final D_r values in HF + TS are 0.51, 0.82, 0.93, and 0.96 for F_c of 5, 10, 20, and 40 D·cm, respectively (Figure 11), indicating that D_r increases as F_c increases, but the increased rate gradually decreases. The final D_r in HF + FS and HF + TS are almost the same, thus the existence of well_c has little effect on hydrate decomposition.

3.2.2. Gas Production Behavior

Figure 12 depicts the variation of Q_p in HF + TS and HF + FS at different F_c. In HF + TS, Q_p is lower than 50 m³/d and gradually decreases when the F_c is 5 D·cm. When F_c is 10 D·cm, Q_p shows a downward trend in the first 750 d as the gas released by thermal stimulation does not migrate to the well_p. Subsequently, Q_p increases and peaks at 85 m³/d. When F_c is 20 and 40 D·cm, the gas migration speed is faster; thus, Q_p starts to rise at 100 and 45 d, showing peak values of 143 and 232 m³/d, respectively. In HF + FS, Q_p is higher and shows a period of increase even at a low F_c value. Figure 13 shows that the final V_p values in HF + TS and HF + FS are 5.64 and 10.52, 12.93 and 17.56, 17.46 and 20.75, and 19.94 and 22.26 × 10⁴ m³ for F_c values of 5, 10, 20, and 40 D·cm, respectively. That is, the final V_p increases with increasing F_c. Compared to HF + TS, the final V_p values in HF + FS increased by 86%, 35%, 18%, and 11%, respectively, indicating that the contribution of well_c to V_p is greater at a lower F_c value.

3.2.3. Methane Leakage, Gas Capture, and Recovery Efficiency

Figure 14 displays the variation of V_l, V_c, and R_e with F_c. Notably, V_l and V_c initially increase and subsequently decrease with the increase in F_c. This can be explained by the impacts of F_c on hydrate decomposition and gas migration. In particular, increasing F_c in the range of 5–10 D·cm significantly improved the hydrate decomposition efficiency, while the gas migration capacity was insufficient, thereby exacerbating the CH₄ leakage. When F_c > 20 D·cm, the improvement of the gas migration capacity is more significant than the increase in Q_r; thus, more gas was produced by well_p, which decreased the V_l and V_c values. However, a leakage volume of 11,236 m³ still exists at the highest F_c, indicating that the reduction of V_l by increasing the F_c is unsatisfactory. Additionally, V_c was higher than V_l, indicating that all escaped gas was recovered by well_c.

R_e increases with increasing F_c while the increased rate decreases gradually. Additionally, the R_e values in HF + FS increased by 0.31, 0.21, 0.14, and 0.10 compared to those in HF + TS at F_c of 5, 10, 20, and 40 D·cm, respectively. This indicated that the contribution of well_c to R_e decreases as F_c increases.

3.3. The Effect of Well Spacing

Based on the simulation results of scenarios 04, 06–08, and 17–20#, we compared the impacts of W_s on the production potential of HF + TS and HF + FS.

3.3.1. Hydrate Dissociation Behavior

Figure 15 displays the variation of Q_r in HF + TS and HF + FS at different W_s. In the first 450 days, Q_r decreased as W_s increased. This is attributed to the lower pressure gradient in the fracture, which reduces the fluid flow velocity. Under the same W_s, the Q_r in HF + FS in the first 450 days was higher than that in HF + TS, but it dropped sharply after 450 days. Figure 16 shows that the final D_r values in HF + TS and HF + FS are 0.98 and 0.98 (W_s = 80 m), 0.93 and 0.95 (W_s = 100 m), 0.83 and 0.87 (W_s = 120 m), and 0.70 and 0.72 (W_s = 140 m), respectively. This indicated that the D_r decreases as W_s increases, whereas the existence of well_c negligibly affected the final D_r.

3.3.2. Gas Production Behavior

Figure 17 shows the variation of Q_p in HF + TS and HF + FS at different W_s. When W_s is 80 m, Q_p increased the fastest at the initial production stage, while it declines the most rapidly after 450 days. This is because when W_s is too small, production wells are flooded earlier and there is not enough hydrate to maintain long-term production. In HF + FS, the variation in Q_p presents three phases. However, in HF + TS, the pressure gradient in the fracture was insufficient for gas migration when W_s ≥120 m, thus, Q_p is lower and there is no rising phase. The final V_p values in HF + FS were 18.62, 20.75, 21.19, and 18.70 × 10⁴ m³ at W_s values of 80, 100, 120, and 140 m, respectively. These values were 1.05, 1.19, 1.47, and 1.67 times those in HF + TS (Figure 18), indicating that the enhancement of the V_p increases as W_s increases. Additionally, the reasonable W_s should be 100–120 m because Vp under this condition is more attractive

3.3.3. Methane Leakage, Gas Capture, and Recovery Efficiency

Figure 19 depicts the variation of V_l, V_c, and R_e with W_s. Notably, V_l and V_c increase with increasing W_s, indicating that the methane leakage is aggravated after increasing the W_s. Furthermore, V_c is higher than V_l even at W_s of 140 m, indicating no CH₄ leakage in HF + FS. The R_e values in HF + TS and HF + FS are 0.91 and 1, 0.79 and 0.93, 0.6 and 0.86, and 0.54 and 0.79 at W_s values of 80, 100, 120, and 140 m, respectively. Thus, with increasing W_s, more gas is trapped, and the contribution of well_c to R_e is greater.

3.4. Variance Analysis

Analysis of variance (ANOVA) can examine the significance of the difference between the means of the factors [50]. In this study, a two-factor mixed variance model is established through the “aov” function in the “R language” (https://www.r-project.org/ (accessed on 18 June 2022)). The V_p, V_c, and R_e values of 300, 600, 900, 1200, 1500, and 1800 d in HF + FS are considered evaluation indicators, and F_c and W_s are the independent variables. Thus, the significance of F_c and W_s are obtained, as shown in Figure 20. Additionally, the main effect and interaction effect of F_c and W_s were visualized, as shown in Figure 21.

3.4.1. Gas Production

As shown in Figure 20a, the significance of F_c to V_p is highest in the entire production period, and the significance level of W_s is lower and gradually decreases as the production progresses. Thus, F_c has a more significant impact on V_p. Moreover, F_c has a positive effect on V_p, whereas W_s has a negative effect on it, except at W_s of 80 m (Figure 21a–c). Thus, it is necessary to further improve F_c to overcome the negative effects of W_s to enhance production. For the current simulations, the average Q_p per unit horizontal section length is higher than 100 m³/d when F_c ≥ 20 D·cm, and a horizontal section length of 600 m can reach the commercial threshold of 60,000 m³/d [25,48,49].

3.4.2. Gas Capture

Figure 20b exhibits the significance level of F_c to V_c is highest at 300 d, which subsequently becomes insignificant at 900 d. The significance level of W_s increases gradually and reaches the maximum at 1200 d. Thus, the initial and final V_c values are mainly controlled by F_c and W_s, respectively. As shown in Figure 21d, F_c has a positive effect on V_c at 300 d, indicating that the rapid decomposition of hydrate caused by the fracture increased the risk of CH₄ leakage. After 900 d (Figure 21e,f), W_s exhibited a positive effect on V_c. Furthermore, the effect of F_c on V_c changed from positive to negative as the production progresses. This is because the thermal decomposition front gradually moves away from well_c, and more gas is produced when F_c is higher. Therefore, a higher F_c reduces V_c, while the necessity of deploying well_c increases under larger W_s.

3.4.3. Recovery Efficiency

As shown in Figure 20c, both F_c and W_s are insignificant to R_e at the initial stage, and their significance levels were enhanced as the production progresses. Moreover, the significance level of W_s is higher than that of F_s, indicating that W_s have a more significant impact on R_e. Additionally, the negative effect of W_s on R_e and the positive effect of F_c on R_e become increasingly clear as the production progresses (Figure 21g–i). Therefore, a higher R_e can be obtained with a smaller W_s and a greater F_c, e.g., R_e values were higher than 0.85 for W_s ≤ 120 m and F_c ≥ 20 D·cm.

4. Discussion

The previous study has obtained the optimal development plan of FS for non-hydraulically fractured reservoirs [29]. To evaluate the production potential of the proposed development mode in this study, the development indicators of the two are compared, as shown in Figure 22. The fracture conductivity and well spacing of the proposed development mode used for comparison here are 40 D·cm and 120 m, respectively. It can be seen that the well spacing is 50 m in a non-hydraulically fractured reservoir, which is only 41.7% of that in a hydraulically fractured reservoir. The cumulative gas production in hydraulically fractured reservoir is 243.2 × 10³m³, which is 3.1 times of non-hydraulically fractured reservoir. The volume of captured gas increases from 22.9 to 43.1 × 10³m³, indicating that there is a greater risk of gas escape in hydraulically fractured reservoir. Additionally, the recovery efficiency increases by 6.3% after fracturing, and the horizontal well section length required to reach the commercial gas production level is reduced from 1362 m to 444 m. Therefore, the proposed development mode is helpful to reduce drilling costs and improve the economics of challenging hydrate development.

It is important to note that the present simulations were conducted under an ideal fracture morphology as field trials of hydraulic fracturing in hydrate reservoirs have not yet been carried out. In fact, for weakly consolidated hydrate reservoirs, it is generally considered that hydraulic fracturing is challenging, and the fracture propagation patterns are complex due to the inhomogeneous hydrates in sediments [51,52,53]. Thus, it is necessary to further evaluate the effect of fracturing morphology on reservoir stimulation and methane leakage. Additionally, various injection-production well patterns, such as two-, three-, five-, seven-, and nine-spot, have been designed in conventional oil and gas development, and there is an optimal matching relationship between fracture distribution and injection–production system [54,55,56]. Unlike conventional oil and gas, the design of hydrate development mode needs to consider the methane leakage due to the lack of impermeable boundaries. Furthermore, the methane escape behavior is affected by reservoir-cover conditions and injection–production parameters. Consequently, to safely and efficiently recover methane from challenging hydrates, the applicability of different development modes to challenging hydrates with various occurrence conditions should be further explored.

5. Conclusions

In this study, we compared the production potential of the three-spot and five-spot well patterns for hydraulically fractured CH-kb reservoirs. The effects of fracture conductivity and well spacing on methane leakage and gas production were investigated. The main conclusions are as follows.

• For the three-spot well pattern, there was considerable methane leakage into the permeable boundary layers. The leakage volume increases as the well spacing increases and reaches 46,341 m³ when the well spacing is 140 m. The impact of fracture conductivity on methane leakage shows two opposite stages: when the fracture conductivity is lower than 20 D·cm, the leakage volume increases as the fracture conductivity increases; when the fracture conductivity is higher than 20 D·cm, the opposite is true. However, enhancing the fracture conductivity can not prevent methane leakage, as a leakage volume of 11,236 m³ was observed at the greatest fracture conductivity of 40 D·cm.
• The proposed five-spot well pattern combined with hydraulic fracturing could realize the safe development of CH-kb under a large well spacing. Even at the well spacing of 140 m, there is still no methane leakage. Additionally, increasing the well spacing increased the risk of methane leakage. Consequently, deploying capture wells is crucial for expanding safe well spacing.
• Compared with the three-spot, the five-spot further increases the gas production and recovery and is expected to meet the commercial production threshold of hydrates. In the current simulations, the five-spot maximizes total gas production and recovery efficiency by 87 and 31%, respectively, compared to the three-spot. The average gas production rate reached 142 m³/d per unit horizontal section length, and a horizontal section length of 425 m was estimated to meet the commercial threshold.
• The development mode of hydraulic fracturing combined with the five-spot well pattern enhances the production potential of CH-kb while addressing methane leakage, which is safe and efficient for the development of low-permeable hydrate reservoirs with permeable boundary layers. In particular, a comprehensive analysis of gas production, methane capture, and gas recovery indicates that the CH-kb hydrate reservoir at the SH2 site in the Shenhu area is suitable for a well spacing of 100–120 m and fracture conductivity greater than 20 D·cm.