Stimulation Effect Evaluation of Boundary Sealing and Reservoir Fracturing on Offshore Challenging Gas Hydrates

Abstract

Depressurization combined with thermal stimulation based on injection-production well patterns is considered promising for gas hydrate development. Nevertheless, its direct application to Shenhu challenging hydrates may be problematic due to the presence of low reservoir permeability and permeable boundaries. The present study proposes to improve the development potential of Shenhu hydrate by reservoir reconstruction, including boundary sealing and reservoir fracturing, and numerically investigates the production performance. The results showed that water intrusion, hot loss, and gas leakage can be effectively addressed by boundary sealing. Nevertheless, it cannot enhance productivity as thermal decomposition gas accumulated around the injection well. Conversely, reservoir fracturing can significantly improve extraction efficiency as substantial amounts of hydrates dissociate along the fractures, and the gas can be well recovered through the fractures. However, reservoir fracturing was not conducive to water control and energy utilization as it induced more severe water flooding and gas leakage. Under the synergistic effect of the two, there was no methane leakage, and the gas production rate increased with increasing fracture conductivity, while the gas-to-water ratio and energy ratio presented the opposite trend. To obtain a favorable production performance, a fracture with a conductivity of 1–10 D·cm was recommended. Therefore, the combination of boundary sealing and reservoir fracturing makes it feasible for safe and efficient extraction of offshore challenging hydrate under the injection-production mode.

1. Introduction

Natural gas hydrate (NGH) is a crystalline compound in which guest gas molecules (mainly methane) are encapsulated in a cage of water molecules [1]. Due to the wide distribution and huge reserves, with a carbon content of about 1/3 of the global total [2,3], NGH is widely recognized as a promising new clean energy resource that will play a significant role in future energy supply [4].

At present, Japan, China, the European Union, as well as other countries and regions, have been vigorously advancing the commercial development and utilization of hydrates [5]. However, more than 97% of NGH is hosted in deep-sea shallow sediments, facing enormous challenges for development due to the complex marine environment, unique occurrence conditions, and a range of potential geo-environmental issues [6]. The Shenhu Area of the South China Sea is currently one of the most active NGH trial production sites in the world, where NGH is endowed in low-permeability clayey silty sediments without impermeable boundaries [7,8,9]. According to the classification of hydrates by Moridis et al. (2011), this falls under the category of challenging hydrates due to the great extraction difficulty [10]. Furthermore, the two rounds of trial production of Shenhu hydrates have proven the feasibility of recovering gas from this deposit [11,12]. Especially in the second trial production, horizontal wells were used for productivity enhancement, with 86.14 × 10⁴ m³ of gas produced within 30 days [12]. Nevertheless, it is still far below the accepted commercial productivity of 5–50 m³/d [13,14]. Consequently, enhancing the productivity of challenging hydrates has become a hot topic.

The feasible NGH extraction methods include depressurization, thermal stimulation, gas displacement (mainly CO₂), and inhibitor injection [15]. Among them, the combination of depressurization and thermal stimulation is generally considered highly promising as it can provide sufficient heat for hydrate decomposition [16,17,18,19]. Particularly, the injection-production mode based on a multi-well shows tantalizing potential because of the efficient mass and heat transfer and has been proposed for the commercial development of NGH by Japan [20]. However, this development mode may be problematic if applied directly to challenging hydrates. The most pronounced distinction is that the permeability of the clayey silty deposits in the Shenhu Area (mostly less than 10 mD) is much lower than the sandy deposits in the Nankai Trough (hundreds to thousands of mD), which is unfavorable for temperature–pressure transfer and fluid flow. Li et al. (2010) and Su et al. (2012) found that thermal stimulation was confined to a limited area [21,22]. Feng et al. (2015) and Jin et al. (2016) revealed that the low permeability cuts off the inter-well hydraulic connection, resulting in the thermal decomposition gas being hard to recover [23,24]. Additionally, the absence of an impermeable boundary results in significant heat fluid losses, which are highly detrimental to energy utilization [25]. Most importantly, thermal decomposition gas tends to escape to the cover layers driven by the high injection pressure, significantly increasing the geo-environmental risk [26]. These issues need to be well addressed in challenging hydrate development under the injection-production mode.

Reservoir fracturing (RF) is an important strategy for unconventional oil and gas development [27,28,29]. Because of this, RF has been proposed to improve the injection-production performance of challenging hydrates. Furthermore, the fracability of hydrate deposits and the stimulation potential have been initially evaluated [30,31,32,33,34,35,36]. Ju et al. (2020) found that RF improved the hot water wave area, thereby facilitating hydrate decomposition [34]. Zhong et al. (2011, 2022) discovered that RF greatly increased the mass and heat transfer between injection and production wells, and was expected to reach commercial productivity [13,37]. However, the hot water loss and gas leakage induced by the open boundary were not considered. According to our previous assessment, uncontrolled hydrate decomposition caused by RF can significantly increase the risk of gas leakage [38]. Against this background, the current study attempts to address the negative effects induced by RF and permeable boundaries.

The permeable cover layer of the NGH deposits is similar to the boundary conditions of gas reservoirs containing edge or bottom water. The primary solution technique is boundary sealing (BS), that is, the use of sealing and plugging materials (e.g., cement) to create impermeable artificial barriers that prevent water intrusion. Yue et al. (2012) developed a flow model for horizontal wells in bottom-water-driven reservoirs with impermeable barriers, demonstrating that the barrier can enhance the critical rate and delay water breakthrough [39]. Liu et al. (2023) constructed an improved artificial barrier by injecting oil-soluble resin as the selective water shut-off agent during the crack generation phase [40]. Bai et al. (2024) reviewed the macroscopic effects and microscopic mechanisms of gel agents during the “injection, migration, plugging, and stabilization” process [41]. Similarly, this technology can be integrated into hydrate development. Our previous research has demonstrated that the hydraulic connection between the reservoir and caprock can be severed by the BS, thereby addressing issues including injected hot loss, boundary water intrusion, and methane leakage [42]. However, the low reservoir permeability hindered the migration of thermal decomposition gas to production wells, resulting in no contribution to productivity during the first three years of hot water injection. Given all the considerations above, the combination of BS and RF may be a novel and promising development mode for offshore challenging hydrates.

This study intends to improve the development potential of challenging hydrates under the injection-production mode through reservoir reconstruction, including BS and RF. The production performance, including NGH decomposition, gas–water production, energy utilization, as well as hot loss and gas leakage, was numerically analyzed, with Shenhu challenging hydrates as the geological background. Furthermore, the synergistic effects between BS and RF were thoroughly evaluated. The novel findings of this study could provide valuable insights for challenging hydrate extraction.

2. Modeling

2.1. Development Plan

The development plan is designed based on the geological characteristics of hydrates at the Shenhu SH2 site, as shown in Figure 1. Geological surveys indicate that the hydrate deposits at the Shenhu SH2 site belong to Class 2 [43,44]. Furthermore, both the overlayer (OL) and underlayer (UL) are permeable with the same permeability as the hydrate-bearing layer (HBL) of 10 mD. Consequently, this represents a classic challenging hydrate characterized by the absence of impermeable boundaries and low reservoir permeability [10]. In the development plan, a five-spot horizontal well pattern with a rectangular distribution was adopted, including one hot water injection well located in the middle of the reservoir and four production wells distributed at the top and bottom of the HBL. Previous studies have demonstrated that this type of well deployment is efficient for gas recovery [45,46]. One feasible reservoir reconstruction scheme is given as follows: firstly, a high-conductivity artificial fracture is created in the middle of the HBL by RF through the injection well; then, impermeable artificial barriers are formed at the top and bottom of the HBL by injecting sealing materials (e.g., cement) through the production wells, thereby sealing open boundaries.

Figure 1. Proposed development plan for the NGH reservoir at the Shenhu SH2 site, wherein the injection and production horizontal wells were arranged along the X-axis; all horizontal sections were drilled along the Y-axis; and the Z-axis represents the vertical direction.

2.2. Geometric Modeling

The HBL thickness at Shenhu SH2 is reported to be 0–44 m [43], and 40 m is taken here. Several researchers have demonstrated that a 30 m thick cover layer is sufficient for the simulation of reservoir temperature–pressure evolution [47,48,49]. While considering the permeable characteristics in this case, the thickness of the OL and UL was set to be 80 m. The horizontal well spacing between the injection and production wells was 100 m, assuming that the reservoir was homogeneous along the Y-axis, and thus only one unit length of the horizontal well section was modelled. As a result, the model size is 200 m (X) × 1 m (Y) × 200 m (Z).

The fracture width and half-length were set as 0.01 and 1 m, with a conductivity of 1–50 D·cm. The 1 m-thick layer at the top and bottom of the HBL was modeled as the barrier, and the permeability was set as 0 to simulate the sealing effect. In the X direction, the grids were refined from 1 m to 0.2 m toward the wellbore, resulting in 203 subdivided meshes. In the Y direction, only one grid was subdivided. In the Z-direction, the grids were refined from the model boundary toward the fracture to 0.01 m, resulting in 106 subdivided meshes. As a result, the total number of meshes is 21,518.

2.3. Simulation Code

Tough + Hydrate (T + H, v1.0), developed by Lawrence Berkeley, was used here because of its reliability in describing phase transitions, heat exchange, and fluid flow in complex media. Furthermore, its accuracy has been fully tested on both experimental and mine scales [50,51,52], which ensures the accuracy of the simulation results.

2.4. Initial and Boundary Conditions

Reservoir pressure and temperature were initialized using hydrostatic pressure and geothermal gradients due to the permeable cover layers. Dirichlet boundary conditions were imposed at the top and bottom model boundaries, as sufficient cover layers have been considered. Model parameters are presented in Table 1. Notably, to ensure the reliability of simulation results, the selected reservoir parameters in the model, including porosity, permeability, and hydrate saturation, as well as the properties of the OL and UL, were consistent with the real geological conditions of Shenhu SH2 Site.

Table 1. Model parameters.

Reservoir Parameters	Value and Unit
Sediment and seawater densities	2650 kg/m3, 1024 kg/m3
Sediment porosity	0.38
Sediment permeability	10 mD
Hydrate saturation	0.55
Sediment specific heat	1000 J/(kg·K)
Dry and wet thermal conductivity of sediments	1 W/(m·K), 3.1 W/(m·K)
Initial average reservoir pressure	14.70 MPa
Seafloor temperature	3.60 °C
Thermal gradient	50 °C/km
Injection and production parameters	Value & Unit
Production pressure	4.50 MPa
Injection pressures	20 MPa
Hot water temperature	60 °C
Constitutive parameters	Value & Unit
Capillary pressure $P_{cap}$	$P_{cap}=-P_{thr}{\left[{({\displaystyle \frac{S_A-S_{irA}}{S_{mxA}-S_{irA}}})}^{-1/\lambda }-1\right]}^{1-\lambda }$
Gas phase relative permeability $k_{rG}$	$k_{rG}={\left({\displaystyle \frac{S_G-S_{irG}}{1-S_{irA}}}\right)}^{n_G}$
Aqueous phase relative permeability $k_{rA}$	$k_{rA}={\left({\displaystyle \frac{S_A-S_{irA}}{1-S_{irA}}}\right)}^n$
$P_{thr}$	1 × 105 Pa
$\lambda$	0.45
$S_{mxA}$	1.00
$n$	3.50
$n_G$	3.50
$S_{irA}$	0.30
$S_{irG}$	0.03

2.5. Simulation Scheme

In our simulations, the production behaviors under various BS cases including sealing OL (BS_O), UL (BS_U), or both (BS_OU) were investigated; subsequently, various RF cases were analyzed, including C_f of 1 (RF₁), 5 (RF₅), 10 (RF₁₀), and 50 (RF₅₀) D·cm, respectively; and finally, the synergistic effect between RF and BS was evaluated. Notably, the unreconstructed reservoir serves as a base case for comparative analyses. Consequently, the final simulation scheme in Table 2 was developed.

Table 2. Simulation scheme.

Cases	Sealed Layer	Cf, D·cm	Cases	Sealed Layer	Cf, D·cm
Base case	None	None	RF10	None	10
BSOU	OL, UL	None	RF50	None	50
BSO	OL	None	BSOU + RF1	OL, UL	1
BSU	UL	None	BSOU + RF5	OL, UL	5
RF1	None	1	BSOU + RF10	OL, UL	10
RF5	None	5	BSOU + RF50	OL, UL	50

3. Results and Discussion

3.1. Production Performance Under BS

3.1.1. Reservoir Physical Field Distribution

The reservoir physical field distribution at 1800 days in the base, BS_OU, BS_O, and BS_U cases is shown in Figure 2.

Figure 2. Reservoir pressure (P), temperature (T), hydrate saturation (S_h), and gas saturation (S_g) in the base, BS_OU, BS_O, and BS_U cases.

The high injection pressure extended to the cover layer in the base case (Figure 2a), and considerable injected hot water flowed into the OL and UL (Figure 2b). Furthermore, the propagation of low production pressures was suppressed, similar to the boundary water intrusion phenomenon in the development of conventional gas reservoirs containing edge or bottom water. This interlayer permeability difference is detrimental to hydrate dissociation and gas recovery. It can be observed that the thermal decomposition gas escaped into cover layers (Figure 2c) and a significant amount of NGH was reserved (Figure 2d). In the BS_OU case, the high injection pressure can be successfully maintained in the HBL, and the low production pressure was also efficiently propagated, creating a satisfactory pressure gradient between the injection and production wells (Figure 2e). In addition, there was no hot water loss (Figure 2f) and gas escape (Figure 2g). Regrettably, gas was hard to transport to the production wells due to the low reservoir permeability. This causes a large amount of gas that is trapped in the vicinity of the injection wells (Figure 2g), thus inhibiting the injection and diffusion of hot water (Figure 2f). As a result, the NGH decomposition efficiency (Figure 2h) was lower relative to the base case. In the BS_O case, high injection pressure maintenance failed (Figure 2i). The mechanism is that the injected hot water penetrated through the bottom of the HBL and migrated downward into the permeable UL, causing partial thermal decomposition gas to escape into the UL (Figure 2j–l). In the BS_U case, high injection pressure primarily propagated towards the permeable OL (Figure 2m). This results in hot water breaking through the top of the HBL and thermal decomposition gas escaping into the permeable OL (Figure 2n–p). Thus, partial sealing does not fully prevent hot water loss and gas escape.

3.1.2. Hydrate Dissociation Behavior

The variation of NGH dissociation rate (Q_h, gas released volume per day) and NGH decomposition ratio (R_hd, defined by the ratio of decomposed hydrate mass to the in situ hydrate mass) in the base, BS_OU, BS_O, and BS_U cases is shown in Figure 3. The variation of Q_h in the base case presents four different phases. In the first phase (~330 days), Q_h increased greatly with injected hot water spreading and low production pressure propagation and peaked at 113 m³/d. Subsequently (~1000 days), Q_h decreased sharply, with a minimum value of 50 m³/d. Mechanisms contributing to this phenomenon include a significant amount of hot water leaked into UL and OL, the slower and slower forward advance of the thermal decomposition leading edge, as well as the boundary water intrusion that inhibits the propagation of low production pressures. In the third phase (~2340 days), inter-well interaction gradually intensified with NGH decomposition, and therefore Q_h increased slowly and peaked at 71 m³/d. After that, Q_h dropped sharply as there was less and less hydrate remaining in the reservoir.

Figure 3. Variation of Q_h and V_hd in the base, BS_OU, BS_O, and BS_U cases.

Unlike the base case, the initial Q_h in the BS_OU case decreased rapidly with a minimum value of 45 m³/d at 327 days. This is because the gas accumulated around the injection well suppressed the injection and diffusion of hot water. Then, a piston-like displacement gradually formed under a stable inter-well pressure gradient, and thus Q_h stabilized at 55 m³/d. Afterward, Q_h began to rise at about 1800 days and peaked at 98 m³/d at 3090 days. This is because the gas is recovered by the production wells, which promotes the injection and diffusion of hot water, thus increasing the hydrate decomposition efficiency. After that, Q_h decayed dramatically as the inter-well NGH was nearly completely decomposed. The variation of Q_h in BS_O and BS_U cases is similar to the base case. In addition, the NGH decomposition cycle T_h for the base, BS_U, and BS_O cases is 4130–4530 days, suggesting that sealing the UL or OL alone has a limited effect on hydrate decomposition. Nevertheless, simultaneously sealing both of them decreased the initial Q_h, but the mid- and post-Q_h were more substantial, resulting in the shortest T_h of 3800 days.

3.1.3. Gas Production Behavior

The variation of gas production rate Q_g and cumulative gas production V_g in the base, BS_OU, BS_O, and BS_U cases is presented in Figure 4.

Figure 4. Variation of Q_g and V_g in the base, BS_OU, BS_O, and BS_U cases.

As shown in Figure 4, the variation of Q_g presented three phases in all cases. Firstly, there was a low gas production phase, where Q_g stabilized after a slight decrease, and the produced gas was contributed by depressurized decomposed gas. Furthermore, Q_g in the BS_OU case is the maximum, about 20 m³/d, followed by those in BS_O and BS_U cases, about 15 m³/d, and the minimum is 12 m³/d in the base case. The second phase was entered at approximately 1000 days, where Q_g increased significantly as more and more thermal decomposition gas was recovered by production wells. Affected by the evolutions of Q_h, the peak values of Q_g in the BS_OU case are the maximum (118 m³/d), followed by 100 m³/d in the base case, 88 m³/d in the BS_O case, and 86 m³/d in the BS_U case. Ultimately, there is a gas production depletion phase due to the insufficient gas supply. Additionally, the final V_g within 5760 days in all cases was consistent, and boundary conditions barely affected the gas recovery cycles T_g, as they were about 4700–4900 days.

3.1.4. Water Production Behavior

Figure 5 shows the variation of water production rate Q_w and cumulative water production V_w with gas recovery ratio R_gr (defined by the ratio of V_g to total gas available in the reservoir) in the base, BS_OU, BS_O, and BS_U cases. It can be seen that Q_w exhibits three phases in all cases, including an initial slight rise, followed by a brief decline, and another rapid rise. The mechanism is that more and more injected water was recovered by the production wells as inter-well interactions increased. This is similar to the water flooding phenomenon in water-driven reservoir development, since open boundaries could induce severe water intrusion. The Q_w in the base case is the highest, followed by the BS_O, BS_U, and BS_OU cases, respectively. Specifically, the V_w at R_gr of 0.7 (acceptable for water-driven development of China’s offshore low-permeability gas reservoirs) in the base case is 3.13 × 10⁴ m³, which is 5.89, 1.38, and 1.61 times that of BS_O, BS_U, and BS_OU cases, respectively. Consequently, boundary intrusion can be well addressed by simultaneously sealing the upper and lower boundaries, which is an effective means of water control.

Figure 5. Variation of Q_w and V_w in the base, BS_OU, BS_O, and BS_U cases.

Figure 6 shows the variation of gas-to-water ratio R_gw (defined by the ratio of V_g to V_w) in the base, BS_OU, BS_O, and BS_U cases. It can be seen that R_gw in the base case is the lowest (<5), and its change presents three stages of fall-rise-fall, which correspond to the initial stage of low gas and low water production, the middle stage of increasing water and gas production, and the later stage of low gas and high water production, respectively. Compared to the base, the evolution of R_gw in BS_O and BS_U cases is similar, and its enhancement is not significant. Encouragingly, R_gw in the BS_OU case is satisfactory, reaching 32 at an R_gr of 0.7, 5.87, 4.25, and 3.63 times that of the base, BS_O, and BS_U cases, respectively. Therefore, gas–water production behavior can be improved by sealing OL and UL.

Figure 6. Variation of R_gw in the base, BS_OU, BS_O, and BS_U cases.

3.2. Injection-Production Performance Under Reservoir Fracturing

3.2.1. Reservoir Physical Field Distribution

Figure 7 shows the reservoir physical field distribution at 720 days in the RF₁, RF₅, RF₁₀, and RF₅₀ cases. It can be seen that as C_f increases from 1 to 50 D·cm, the high injection pressure propagates faster (Figure 7a–d) and a wider wave of hot water in the horizontal direction is observed (Figure 7e–h). This improves the reservoir heating efficiency, thus promoting hydrate decomposition (Figure 7i–l). Furthermore, more hydrate dissociates along the fracture surface as C_f increases, which leads to a shift in hydrate decomposition mode from piston-type to volumetric-type. However, increasing C_f could not avoid gas leakage (Figure 7m–p), which is consistent with previous studies. It should be specifically pointed out that the current model assumes a constant C_f value throughout the production process. In reality, given the weakly cemented nature of NGH deposits, fractures may buckle, become plugged with sand proppant, or interact with natural heterogeneities. This may impact long-term forecasts. Consequently, the variation pattern of C_f as well as its influence needs to be further investigated.

Figure 7. Reservoir pressure (P), temperature (T), hydrate saturation (S_h), and gas saturation (S_g) in the RF₁, RF₅, RF₁₀, and RF₅₀ cases.

3.2.2. Hydrate Dissociation Behavior

Figure 8 shows the variation of Q_h and R_hd in the base, RF₁, RF₅, RF₁₀, and RF₅₀ cases. As can be seen, the variation of Q_h in the RF cases is similar to that in the base case. The difference is that the initial Q_h increases with the increase in C_f. Specifically, the peak value increases from 130 to 317 m³/d as C_f increases from 1 to 50 D·cm, which is 1.16–2.83 times higher than that in the base case. The mechanism is that a greater C_f is more favorable for hot water migration, allowing more hydrate to dissociate along the fracture. Additionally, at greater C_f, the leading edge of hydrate decomposition moves away from the fracture faster, which leads to a higher post-peak Q_h decay rate. The T_h reduces from 3700 to 2170 days as C_f increases from 1 to 50 D·cm, a reduction of 430–1960 days compared to the base case. As a result, increasing C_f is helpful for hydrate decomposition.

Figure 8. Variation of Q_h and V_hd in the base, RF₁, RF₅, RF₁₀, and RF₅₀ cases.

3.2.3. Gas Production Behavior

Figure 9 shows the variation of Q_g and V_g in the base, RF₁, RF₅, RF₁₀, and RF₅₀ cases.

Figure 9. Variation of Q_g and V_g in the base, RF₁, RF₅, RF₁₀, and RF₅₀ cases.

With C_f increasing, the initial low productivity phase is greatly shortened, and the corresponding time for Q_h to peak is earlier. Specifically, this time is shortened from 2160 days to 400 days as C_f increases from 1 to 50 D·cm, a reduction of 260–1790 days relative to the base case. The mechanism is that the gas flow rate inside the fracture is higher with a greater C_f, and therefore, thermal decomposition gas can be extracted more rapidly from production wells. However, the peak value of Q_g does not always increase as C_f increases, with a maximum value of 161 m³/d occurring at C_f of 10 D·cm. This is attributed to the fact that increasing C_F simultaneously increases the hot water injection rate, thereby reducing the gas-phase relative permeability in the fracture. Additionally, it is observed that the V_g within 5760 days decreases from 24.13 to 22.12 × 10⁴ m³ as C_F increases from 1 to 50 D·cm, a reduction of 0.18–2.18 × 10⁴ m³ compared to the base case. This suggests that gas leakage occurs in the RF case and that the amount of leaked gas increases with increasing C_f. Thus, increasing C_f can substantially increase the gas recovery efficiency, but also induce more serious gas leakage.

3.2.4. Water Production Behavior

Figure 10 shows the variation of Q_w and V_w in the base, RF₁, RF₅, RF₁₀, and RF₅₀ cases. At the early production stage (R_gr < 0.1), the influence of C_f on Q_w is not significant. Interestingly, Q_w and V_w do not always increase with increasing C_f. Specifically, when R_gr is 0.7, V_w decreases from 2.84 to 1.94 × 10⁴ m³ as C_f increases from 1 to 5 D·cm, a reduction of 0.29–1.19 × 10⁴ m³ compared to the base case. In contrast, V_w increases from 1.94 to 3.60 as C_f increases from 5 to 50 D·cm, indicating that increasing C_f in this interval is more favorable for water production compared to gas production. Consequently, properly controlling the C_f value of 1–10 D·cm is conducive to water control.

Figure 10. Variation of Q_w and V_w in the base, RF₁, RF₅, RF₁₀, and RF₅₀ cases.

Figure 11 shows the variation of R_gw in the base, RF₁, RF₅, RF₁₀, and RF₅₀ cases. The R_gr corresponding to the peak value of R_gw reduced from 0.72 to 0.20 as C_f increases from 1 to 50 D·cm, a reduction of 0.08–0.60 relative to the base case. The reason is that increasing C_f can significantly improve Q_g, but also induce premature water flooding. The R_gw at R_gr of 0.7 increases first and then decreases with increasing C_f, and the maximum value of 8.82 appears at C_F of 5 D·cm, 1.61 times that in the base case. As a result, RF can improve the gas and water production behavior, but needs to control C_f reasonably.

Figure 11. Variation of R_gw in the base, RF₁, RF₅, RF₁₀, and RF₅₀ cases.

3.3. Production Performance Under BS and RF

3.3.1. Reservoir Physical Field Distribution

Figure 12 shows the reservoir physical field distribution at 1800 days in the BS_OU + RF₁, BS_OU + RF₅, BS_OU + RF₁₀, and BS_OU + RF₅₀ cases. The high injection pressure only propagated in the HBL, and ideal pressure gradients between the injection and production wells were generated (Figure 12a–d). Similar to the RF cases in Figure 6, the hot water spread area increases as C_F increases (Figure 12e–h), and therefore, more hydrates are decomposed along the fracture (Figure 12i–l). The difference is that there is no hot water loss (Figure 12e–h) and gas leakage (Figure 12i–l). As a result, BS can improve the utilization efficiency of hot water while effectively addressing gas leakage induced by RF.

Figure 12. Reservoir pressure (P), temperature (T), hydrate saturation (S_h), and gas distribution (S_g) in the BS_OU + RF₁, BS_OU + RF₅, BS_OU + RF₁₀, and BS_OU + RF₅₀ cases.

3.3.2. Hydrate Dissociation Behavior

The variation of Q_h and V_hd in the base, BS_OU + RF₁, BS_OU + RF₅, BS_OU + RF₁₀, and BS_OU + RF₅₀ cases is shown in Figure 13. As can be seen, the peak value of Q_h increases from 122 to 204 m³/d as C_f increases from 1 to 50 D·cm, which is 1.72–2.87 times that of the base case. Additionally, the T_h reduced from 3370 to 2380 days as C_f increased from 1 to 50 D·cm, a reduction of 760–1750 days relative to the base case. As a result, increasing C_f in BS + RF cases is still an effective method for promoting hydrate decomposition.

Figure 13. Variation of Q_h and R_hd in the base, BS_OU + RF₁, BS_OU + RF₅, BS_OU + RF₁₀, and BS_OU + RF₅₀ cases.

3.3.3. Gas Production Behavior

Figure 14 shows the variation of Q_g and V_g in the base, BS_OU + RF₁, BS_OU + RF₅, BS_OU + RF₁₀, and BS_OU + RF₅₀ cases. The Q_g in the BS + RF cases is much higher than that in the base case. Furthermore, as C_F increases from 1 to 50 D·cm, the peak value of Q_g increases from 142 to 186 m³/d, and the T_gr is shortened from 4210 to 3720 days, indicating that the extraction efficiency is greatly improved by increasing C_f. The final V_g within 5760 days in all cases is consistent, indicating there is no gas leakage. As a result, BS can greatly improve gas recovery efficiency while addressing gas leakage induced by RF.

Figure 14. Variation of Q_g and V_g in the base, BS_OU + RF₁, BS_OU + RF₅, BS_OU + RF₁₀, and BS_OU + RF₅₀ cases.

3.3.4. Water Production Behavior

The variation of Q_w and V_w in the base, BS_OU + RF₁, BS_OU + RF₅, BS_OU + RF₁₀, and BS_OU + RF₅₀ cases is shown in Figure 15. The Q_w in C_f of 1–10 D·cm cases is much lower than that in the base case. For C_f of 50 D·cm, Q_w is higher than that of the base case in the mid and late periods. In addition, Q_w and V_w increase with increasing C_f. At R_gr of 0.7, V_w increases from 0.55 to 1.61 × 10⁴ m³ as C_f increases from 1 to 50 D·cm, accounting for 17–51% of the base case. Thus, significant reductions in water production are achieved.

Figure 15. Variation of Q_w and V_w in the base, BS_OU + RF₁, BS_OU + RF₅, BS_OU + RF₁₀, and BS_OU + RF₅₀ cases.

Figure 16 shows the variation of R_gw in the base, BS_OU + RF₁, BS_OU + RF₅, BS_OU + RF₁₀, and BS_OU + RF₅₀ cases. The R_gw in BS + RF cases is much higher than in the base case, and it decreases with increasing C_f. Specifically, the R_gw at R_gr of 0.7 decreases from 30 to 10 as C_F increases from 1 to 50 D·cm, 1.84–5.53 times higher than the base case. Therefore, combining BS with RF greatly improves gas–water production performance.

Figure 16. Variation of R_gw in the base, BS_OU + RF₁, BS_OU + RF₅, BS_OU + RF₁₀, and BS_OU + RF₅₀ cases.

4. Synergistic Effects of BS and RF

From the analysis of the above results, it is found that the hot water loss and boundary water intrusion can be addressed by BS; RF can significantly improve extraction efficiency, but also exacerbate the gas leakage, whereas the combination of the two seems to result in a satisfactory production behavior. Here, the synergistic effects of BS and RF on gas and water production are analyzed with the average gas production rate Q_ag (Figure 17a) and R_gw (Figure 17b) at R_gr of 0.7. In addition, for the multi-well injection development mode, energy ratio ER, defined by V_g·C_g/[C_w·M_i·(T_i–T₀)], is a crucial parameter to measure the economy and feasibility of exploitation [53]. As a result, the ER at R_gr of 0.7 is employed as a key indicator for synergistic effect analysis (Figure 17c). Notably, the ER used here did not account for the energy required to create artificial barriers and perform reservoir fracturing. Given the significant challenges in reconstructing hydrate deposits, more comprehensive economic feasibility assessments need to be conducted based on systems analysis methods before field application [54].

Figure 17. The Q_ag (a), R_gw (b), and ER (c) at R_gr of 0.7 in all simulation cases.

As can be observed from Figure 17a, the Q_ag in the BS_O, BS_U, and BS_OU cases is slightly lower than that in the base case. The mechanism involves the thermal decomposition gas accumulating around the injection well, which suppresses hot water injection and thereby reduces the thermal stimulation area. This indicates that boundary sealing alone does not enhance productivity. Surprisingly, Q_ag can be significantly increased by RF, especially at C_f ≥ 5 D·cm. Furthermore, the synergistic effect of BS and RF shifts with changes in C_f. That is, they exhibit positive synergy at C_f of 1 and 50 D·cm, and negative synergy at C_f of 5 and 10 D·cm. However, this synergistic effect is not significant as the variation interval of Q_ag is only [−9, 11]. This indicates that in fractured reservoirs, the effect of BS on productivity is very limited. Furthermore, under the synergistic effect of BS and RF, Q_ag increases significantly with the increase in C_f, specifically from 64 to 99 m³/d as C increases from 1 to 50 D·cm, an enhancement of 10–45 m³/d relative to the BS_OU case. Given that BS can address RF-induced methane leakage and has a limited impact on Q_ag, it is suggested here to increase productivity by adjusting C_f.

As can be seen from Figure 17b, relative to the base case, there is very limited improvement in R_gw by RF alone or sealing one open boundary. However, simultaneously sealing the upper and lower boundaries obtains the highest R_gw of 32, 6.40 times that of the base case. At different C_f values, the R_gw after boundary sealing is all enhanced, which demonstrates the excellent water control ability of the BS. However, the increase in C_f weakens the water control ability of BS as a considerable amount of injected water induces severe water flooding. Nevertheless, the final water yield under the synergistic effect of BS and RF is still much lower than that of the base case.

As can be seen from Figure 17c, the ER of the base case is 1.06, indicating that it is not economically feasible to develop challenging hydrates directly using the multi-well injection-production mode. After the application of BS, the ER reaches 1.28, 1.30, and 3.76 for the BS_O, BS_U, and BS_OU cases, respectively, indicating that only the simultaneous sealing of the upper and lower boundaries could significantly improve ER because hot water loss is effectively addressed. The ER values in the RF case are all lower than 1.0, indicating that the application of RF alone is still not economically feasible. In addition, ER decreases significantly with increasing C_f under the synergistic effect of BS and RF. However, they are still greater than 1.0, 1.18–3.41 times that of the base case.

To summarize, BS can effectively address boundary water intrusion and hot water loss, thereby substantially increasing R_gw and ER, but barely affecting Q_ag, whereas RF can significantly increase Q_ag, but also induce severe gas leakage and decrease R_gw and ER. Nevertheless, with the synergistic effect of the two, safe extraction and the relatively ideal production behavior can be obtained by regulating the C_f.

5. Conclusions

The present research numerically investigated the injection-production performance under various BS and/or RF cases and thoroughly analyzed their synergistic stimulation mechanisms. The following conclusions can be drawn.

• BS can effectively address boundary water intrusion, hot loss, and methane leakage, thereby reducing water production and enabling the full utilization of injection energy. However, it cannot speed up hydrate decomposition and gas production. Additionally, partial sealing does not fully prevent hot water loss and gas escape.
• RF is very effective in facilitating hydrate decomposition and gas recovery, and the extraction efficiency improves greatly with increasing C_f. However, the final ER is lower than 1.0 due to the severe water flooding induced by the fracture, making it unfeasible for economic extraction. Additionally, RF exacerbates the risk of gas leakage, which poses a great challenge for the safe exploitation of marine hydrates.
• There is no methane leakage under the synergistic effect of RF and BS. Furthermore, Q_h and Q_g improve significantly as C_f increases, whereas R_gw and ER show the opposite trend. Therefore, rational regulation of C_f value is essential to obtain satisfactory multi-well injection-production behavior, with 1–10 D·cm recommended.
• BS is an indispensable technology for the economic and safe extraction of offshore challenging hydrates under the multi-well injection-production mode. Although RF is helpful for hydrate decomposition and gas production, its induced water flooding and methane leakage should not be neglected in future applications.

The current simulations indicate that the combination of BS and RS can significantly enhance the performance of the injection-production well pattern. Nevertheless, it should be noted that the fracture and barrier settings in our model are relatively ideal to maximize the desired effect, and the injection-production parameters employed also need further evaluation and optimization. Additionally, a homogeneous model without interlayers was employed because the hydrate distribution remains unclear. Consequently, follow-up research should fully account for these issues to advance and support engineering implementation.