As a vital future energy resource, achieving high-efficiency exploitation of natural gas hydrates (NGHs) still faces challenges, and depressurization combined with other enhancement technologies, such as reservoir stimulation, may be the optimal path. Unlike previous studies that mainly focused on the impact of boundary sealing on single-well or complex-structure well types, this work systematically compared the production performance of two well-net modes under boundary sealing conditions. Based on China’s first offshore NGH trial production, and a numerical simulation method combined with J index (mainly affected by well type) was used to systematically compare the short- and long-term yield-increasing effects of five-spot wells (FSW) and cluster horizontal wells (CHW) for depressurization of Class 1 hydrate reservoirs with boundary sealing. The results indicate that both types of wells have better productivity performance due to the low-permeability barrier formed by boundary sealing to suppress water invasion and concentrate pressure energy to decompose hydrates, but their performance differs over time. The five-spot wells showed a more substantial overall improvement. Compared with the base case, with the boundary sealing, the cumulative gas production (Vg) of the five-spot wells and cluster horizontal wells increased by 169.8% and 155.1%, respectively, and the gas-to-water ratio (Rgw) increased by 680.6% and 409.3%, respectively. Although the cluster horizontal wells performed well in the first 120 days, the five-spot wells with boundary sealing performed well after 120 days and achieved a higher J index of 9.5 after 720 days. The results indicate that cluster horizontal wells demonstrate higher short-term gas production efficiency, whereas five-spot wells offer long-term development potential. The optimal engineering decision should therefore be based on whether the project’s core strategic objective is short-term pilot verification or long-term development. These findings provide a theoretical reference for multi-well development strategies in Class 1 hydrate reservoirs under boundary sealing conditions.
Graphical Abstract
boundary sealingcluster horizontal wellfive-spot wellsnatural gas hydratenumerical simulationThe author(s) declared that financial support was received for this work and/or its publication. This research was supported by the National Key R&D Program of China (No.2023YFC2808700), Guangdong MEPP Fund (No.GDOE(2019)A39).section-at-acceptanceEconomic GeologyIntroduction
Natural gas hydrates (NGHs) are high-energy-density crystalline substances formed by water and gas molecules (mainly methane) in low-temperature and high-pressure environments, widely present in deep-sea sediments and terrestrial permafrost worldwide (Sloan, 2003; Boswell, 2009; Chong et al., 2016; Boswell and Collett, 2011; Moridis et al., 2011). As a potential alternative energy with abundant reserves, its commercial extraction still poses challenges. Although China’s offshore NGH trial production has successfully validated the effectiveness of the depressurization and increased the daily gas production to 2.87 × 104 m3, there is still a huge gap from the commercial threshold of 5 × 105 m3/d (Li et al., 2018; Ye et al., 2020; Wu et al., 2021). This indicates that relying solely on traditional depressurization methods makes it difficult to achieve economic exploitation, and more efficient development strategies must be explored.
The current main approaches of increasing production include: using complex structured wells to expand the drainage area; Applying thermal stimulation to improve dissociation rate; And implementing reservoir stimulation, such as hydraulic fracturing, near-wellbore stimulation, and boundary sealing, to change the local or overall permeability of the reservoir, etc (Wu et al., 2021; Huang et al., 2023). Recent advancements have enriched the spectrum of enhanced depressurization strategies. These include combining depressurization with CO2 replacement and developing novel thermodynamic and kinetic inhibitors to manage hydrate formation more efficiently and sustainably (Zhang et al., 2024; Zhang et al., 2025a; Zhang et al., 2025b). Among the above technologies, boundary sealing can optimize the pressure field and fluid migration in the reservoir. The principle is to build low-permeability artificial barriers in the reservoir through new materials such as cement, gel, polymer, and nano particles to form an effective vertical seal, thus inhibiting the invasion of water, and significantly improving the pressure reduction depressurization (Zhu et al., 2021; Sun et al., 2022; Sun Z. et al., 2019). The essence of this technology is similar to the widely used water plugging practices in conventional oil and gas reservoirs (Hua et al., 2013; Zhao et al., 2018; Yue et al., 2015).
In recent years, the increasing production mechanism and engineering applicability of boundary sealing in NGH development have gradually been revealed, and a series of numerical simulation studies have verified its effectiveness from different perspectives (Zhao et al., 2020; Wang et al., 2024; Li et al., 2021; Guo et al., 2024; Lv et al., 2022; Nie et al., 2024; Qin et al., 2025). Zhao et al. (2020) proposed that an artificial barrier could be built in the permeable layer through gel injection, and the simulation results showed that the hydrate dissociation rate increased from 8.9% to 45.4%. Li et al. (2021) further introduced the hydraulic fracturing-assisted boundary sealing method and found that when sealing lengths exceed the fracturing radial length, this approach boosts gas production by 93.2% and reduces water production by 62.9%, reflecting the significant advantages of this combination technology in controlling water and increasing gas production. Lv et al. (2022) evaluated the adaptability of artificial barriers for vertical well depressurization, and the results indicated a 20.88% reduction in water production alongside an improved gas-to-water ratio. Guo et al. (2024) explored the collaborative mechanism of network fracturing and boundary sealing, demonstrating that this combined method can effectively prevent the water invasion and optimize the pressure propagation path, resulting in an increase of 11.1 times in gas production and 13.3 times in gas-to-water ratio. In terms of parameter optimization, Qin et al. (2025) further elucidated the impact of sealing layer performance on mining efficiency and proposed optimized sealing parameters for vertical wells. Ning et al. (2022) analyzed the efficiency enhancement mechanism of multilateral wells under boundary sealing. These research results clearly outline the evolution path of boundary sealing technology from mechanism exploration to parameter optimization.
Although these studies have demonstrated the potential of boundary sealing, current research is still mainly limited to single vertical or single complex structure well configurations. Under boundary sealing conditions, there is still a lack of systematic comparative analysis of production performance between different well-net patterns (specifically, five-point wells and cluster horizontal wells). Therefore, this work focuses on two representative well-net patterns. To enable a direct and fair comparison of their intrinsic performance, boundary sealing is introduced and maintained as an idealized engineering condition throughout the simulation. A 3D geological model is established, and the differential production response of the two well-net types is systematically compared and analyzed under this unified premise of effective water control.
MethodologyGeological background
The SHSC4 well is located in the Baiyun Sag of the Pearl River Mouth Basin (Figure 1), and is the target of China’s first offshore NGH trial production in 2017 in the Shenhu area, which provides the stratigraphic, petrophysical, and production benchmarks for this work. This site has a water depth of 1266 m, with a seafloor temperature of about 3 °C and a geothermal gradient of 43.653 °C/km (Li et al., 2018; Zhang et al., 2017). The reservoir is vertically stratified into three primary units: the Gas Hydrate-Bearing Layer (GHBL: 201-236 mbsf, 35 m), the Three-Phase Layer (TPL: 236-251 mbsf, 15 m), and the Free Gas Layer (FGL: 251-278 mbsf, 27 m) (Li et al., 2018; Zhang et al., 2017). Key petrophysical properties derived from core and log data: GHBL: Porosity 0.35, permeability 2.9 mD, average hydrate saturation 34%. TPL: Porosity 0.33, permeability 1.5 mD, average hydrate saturation is about 31%. FGL: Porosity 0.32, permeability 7.4 mD, free-gas saturation 7.8%. The lithology is quartz-feldspar dominated with 25%–35% clay minerals (mainly montmorillonite and illite) (Li et al., 2018; Zhang et al., 2017). The reservoir exhibits significant heterogeneity: the FGL permeability is about 5 times that of the TPL vertically (Li et al., 2018; Sun Y. et al., 2019; Ma et al., 2020; Cao et al., 2023; Qin et al., 2020). The SHSC-4 well was completed across the interval (201-271 mbsf) and produced for 60 days, yielding a cumulative gas production of 3.09 × 105 m3 (Li et al., 2018; Sun Y. et al., 2019; Ma et al., 2020; Cao et al., 2023; Qin et al., 2020).
SHSC4 well site. (adapted from Hao et al., 2018. Copyright 2022 American Chemical Society) (Hao et al., 2022).
Map of the South China Sea showing China, Vietnam, and the Philippines with three outlined regions labeled Xisha sea area, Shenhu sea area, and Dongsha sea area; a yellow star marks the 2017 first offshore NGH production test in Shenhu sea area.Simulation code
TOUGH + HYDRATE V1.0 is a professional numerical simulator designed for NGH development (Moridis et al., 2008). It supports various production strategies, including depressurization, thermal stimulation, and inhibitor injection. The code has been rigorously validated through both laboratory experiments and field tests, demonstrating high predictive accuracy and reliability (Su et al., 2012; Yin et al., 2019). In this work, the parallel version pT + H v1.0 and the equilibrium model are employed to simulate depressurization with five-spot wells and cluster horizontal wells, under the assumptions of controllable sand production, validity of Darcy’s law, and non-deformable geological media (Zhang et al., 2008; Kowalsky and Moridis, 2007). The mathematical model is grounded in mass and energy conservation principles, with the governing equations for mass and energy balance given by Equations 1, 2, respectively (Moridis et al., 2008):
Mass conservation equation
ddt∫VnMκdV=∫ΓnFκ·ndΓ+∫VnqκdVwhere Mκ is mass accumulation, Fκ denotes mass flux, and qκ represents source/sink terms.
Energy conservation equation
ddt∫VnMθdV=∫ΓnFθ·ndΓ+∫VnqθdVwhere θ is heat component, is heat flux, Mθ, Fθ, and qθ are correspond to heat accumulation, flux, and source/sink ratio, respectively.
Model and cases design
The schematic diagram is shown in Figure 2. The model domain was discretized with an unstructured grid system. Specifically, the X-Y plane is split into 36 layers along the X-direction and 50 layers along the Y-direction. Furthermore, to enhance the resolution of the region where physical property variations are prominent during exploitation, local grid refinement is implemented within the 2 m vicinity of the wellbore, and the grid spacing in both the X and Y-axes is 1 m, resulting in a total grid count of 5,220 on the X-Y plane. The Z-direction is discretized into 95 layers (Figures 2a,b), leading to a total grid number of 495,900 for the entire model. This grid setup ensures high-precision capture of multiphase flow and heat transfer behaviors, as well as the changes in reservoir physical properties during hydrate exploitation.
Figure consists of two groups of subplots labeled (a) FSW and (b) CHW, each highlighted in yellow boxes. For both groups, left panels show X-Y coordinates with vertical bars and data points, and right panels show Y-Z cross-sections with horizontal stratigraphic boundaries labeled in red as OB, GHBL, TPL, FGL, and UB with thicknesses in meters. FSW panels include blue markers and lines, while CHW panels feature pink bars. Both illustrate spatial and vertical placement of features across a 3D grid.
Four simulation cases were designed, with detailed parameter configurations listed in Table 1. Cases 1A and 1B employ a five-spot wells pattern, including 5 vertical wells, each with a length of 70 m. Case 1A serves as the base case without boundary sealing, while Case 1B incorporates sealing layers to examine the impact of boundary sealing. Cases 2A and 2B adopt the cluster horizontal wells pattern, comprising 3 horizontal wells, each with a length of 150 m. Likewise, Case 2A is free of sealing layers, and Case 2B integrates sealing layers. Overall, these cases are designed to elucidate the optimal well configuration and the effectiveness of sealing layers for gas hydrate exploitation.
Simulation cases design.
Cases
Well-net patterns
Well location
Well quantity
Single well length
Boundary sealing
Case 1A
Five-spot wells
Z = −30 m ∼ −100 m
5
70 m
-
Case 1B
Five-spot wells
Z = −30 m ∼ −100 m
5
70 m
Yes
Case 2A
Cluster horizontal wells
Z = −72.5 m
3
150 m
-
Case 2B
Cluster horizontal wells
Z = −72.5 m
3
150 m
Yes
Comparative framework and the role of the <italic>J</italic> index
This work compared two development strategies for a Class 1 hydrate reservoir under boundary sealing conditions. FSW uses vertical wells for multi-layer joint production on GHBL, TPL, and FGL. In contrast, CHW drive the dissociation of hydrates in GHBL and the recovery of free gases in TPL and FGL through pressure propagation. Although their total wellbore lengths are different (FSW is 350 m, CHW is 450 m), for fairness, we introduced the J index, which normalizes wellbore length, pressure drop, and daily gas production, providing a reasonable measure for comparing the two development strategies. To enable a fair comparison, the productivity index J is introduced (Equation 3), which normalizes gas production by wellbore length and pressure drawdown (Wu et al., 2021):J=Qg/hΔP
Where h represents the total length of the wellbore (m), and ΔP represents the pressure difference (MPa). and Qg represents the daily gas production.
Initial and boundary conditions
This model follows a standard initialization procedure, with the initial pore water pressure calculated using the empirical formula in Equation 4 (Yu et al., 2021; Sun et al., 2017; Yuan et al., 2017):Ppw=Patm+ρswgH+Z×10−6
Within the formula, Ppw stands for porewater pressure (MPa), ρsw is the seawater density (kg/m3), Patm refers to normal atmospheric pressure (MPa), and g denotes gravitational acceleration (m/s2), while H signifies water depth (m) and Z corresponds to sediment depth from the seabed (m).
Subsequently, the initial conditions for the entire formation were solved via the code’s self-balancing function (Figure 3) (Sun et al., 2015). The model boundaries were set as the first type boundary, and the production wellbore was considered as an internal boundary, and the well grids kept a constant pressure difference of 6.5 MPa (Feng et al., 2019). The models and their parameters for describing multiphase flow adopt the pre-validated model parameter settings in this sea area (Sun Y. et al., 2019; Ma et al., 2020; Cao et al., 2023; Qin et al., 2020). Among them, capillary pressure is calculated by the van Genuchten model, and relative permeability is calculated by the Stone model. The parameter settings are detailed in Table 2.
Model’s initial conditions.
Four rectangular contour plots display subsurface properties along Y and depth axes, separated into OB, GHBL, TPL, FGL, and UB layers. Top left shows pressure in pascals. Top right displays temperature in degrees Celsius. Bottom left represents hydrate saturation, and bottom right shows gas saturation. Each plot uses a blue-to-red color scale, with labeled horizons and consistent Z and Y ranges.
Model’s physical properties.
Parameter type
Parameters
Value & unit
Formation Thickness
OB & UB
30 m
GHBL
35 m
TPL
15 m
FGL
27 m
Permeability& Porosity
OB & UB
2.0 mD, 0.30
GHBL
2.9 mD, 0.35
TPL
1.5 mD, 0.33
FGL
7.4 mD, 0.32
Saturation
Hydrate Saturation in GHBL and TPL
Cited from Logging Data
Free Gas Saturation in FGL
Cited from Logging Data
Sealing Layer Properties
Permeability
0.0001 mD
Thickness
1 m
Location
Top of GHBL & Bottom of FGL
Production well
Production Pressure
6.5 MPa
Wellbore Radius
0.1 m
Multiphase flow
Capillary Pressure Model (Sun Y. et al., 2019; Ma et al., 2020; Cao et al., 2023; Qin et al., 2020)
Pcap=‐P0S*‐1/λ‐11‐λ, S*=SA‐SirASmxA‐SirA
SmxA (Maximum aqueous saturation)
1
λ (Capillary pressure exponent)
0.45
P0 (Capillary pressure reference value)
104 Pa
Relative Permeability Model (Sun Y. et al., 2019; Ma et al., 2020; Cao et al., 2023; Qin et al., 2020)
Table 2 provides detailed physical parameters for the model. The formation structure includes 30 m thick overburden (OB) and underburden (UB), a 35 m thick GHBL, a 15 m thick TPL, and a 27 m thick FGL (Yuan et al., 2021). The initial permeability settings for each layer are: 2.0, 2.9, 1.5, and 7.4 mD, respectively; the corresponding porosities are: 0.30, 0.35, 0.33, and 0.32, respectively (Li et al., 2018). Sealing layers with a permeability of 0.0001 mD and a thickness of 1 m are located on the GHBL top and FGL bottom. The wellbore radius is 0.1 m, and the salinity is 3.5%. Initial saturations are obtained from logging curves. Rock properties include a grain density of 2600 kg/m3, a geothermal gradient of 43.653 °C/km. Multiphase flow model parameters are also fully listed (Li et al., 2018).
Model validation
The reliability of the proposed model was assessed by comparing it with actual gas production data from the 2017 Shenhu NGH trial production. Key model parameters include a 70 m vertical well located at the center of the model, with a perforation interval of −201 to −271 mbsf, and a constant production pressure difference of 3 MPa (Qin et al., 2020). The simulated 60-day cumulative gas production is 3.31 × 105 m3, which shows reasonable agreement with the field-observed value of 3.09 × 105 m3, yielding a relative error of +7.1% (Figure 4). The degree of matching between the two reflects the basic applicability of the model. The observed discrepancy is likely caused by near-wellbore hydrate dissociation induced by heat transfer from drilling fluid to the reservoir through high-thermal-conductivity casing during drilling and completion operations, an effect that the current model cannot characterize.
Gas production history fitting.
Scatter plot comparing cumulative production over time for practical production, shown with green triangles, and simulation production, shown with pink squares. Both datasets show a generally linear upward trend, with simulation production slightly exceeding practical production after 30 days. X-axis represents time in days from zero to sixty-five; Y-axis represents cumulative production in ST ten thousand cubic meters from zero to forty.Results and discussionGas and water production
From the gas production rate curve (Qg) and cumulative gas production curve (Vg) in Figures 5a,b, it can be seen that, driven by a large production pressure difference of 6.5 MPa, hydrates rapidly decompose, and a massive high-saturation free gas flows into the TPL wellbore within a short period. The Qg of Case 1A reaches its peak at the beginning, and as a result of the dual impacts of the water invasion and the secondary hydrate generation around the TPL wellbore, Qg gradually decreases slowly. Compared to Case 1A, Case 1B adopts boundary sealing, which effectively reduces the water invasion. The energy of the pressure difference is mainly used to drive the dissociation of hydrates. Therefore, the overall trend of its Qg curve is higher, and the rate of decline in the later stage is also slower. Similarly, the same phenomenon was observed in Cases 2A and 2B. However, a noteworthy detail in Figure 5a is that in the first tens of days, the gas production rate of Case 1B was lower than that of Case 2B. This is mainly because the horizontal section of cluster horizontal wells is completely located at TPL, and high-saturation free gas is rapidly extracted; However, five-spot wells only has the middle section of the wellbore (−65 to −85 m) located at TPL, and the contact area between the wellbore and TPL is relatively small, so its early gas production rate is lower than that of cluster horizontal wells. Compared to cluster horizontal wells, boundary sealing has a more significant enhancement effect in five-spot wells. This is because the top and bottom sections of the vertical wellbore are closer to the upper and lower boundaries, making it easier for water invasion, resulting in rapid water breakthrough; The horizontal well cluster is located in the middle of the TPL, far from the top and bottom water, requiring a longer time for bottom water invasion. After 720 days of production, the Vg of the four cases were 1379.3, 2342.5, 1278.3, and 1982.8 × 104 m3, respectively. Compared with Cases 1A and 2A, Cases 1B and 2B increased by 169.8% and 155.1%, respectively.
Production behavior over 720 days. (a) Evolution of gas production. (b) Evolution of cumulative gas. (c) Evolution of water production. (d) Evolution of cumulative water.
Four line charts display results for four cases (1A–FSW, 1B–FSW, 2A–CHW, 2B–CHW) over seven hundred twenty days. Chart (a) shows Qg decreasing over time; (b) shows Vg increasing, (c) shows Qw generally increasing, and (d) shows Vw rising steadily. Each case is represented by a distinct color and legend for each parameter.
From the water production rate curve (Qw) and cumulative water production curve (Vw) in Figures 5c,d, it can be seen that the Qw curve of Case 1A shows a phenomenon of first decreasing and then rapidly increasing within 30 days. The initial decline in water production rate is attributed to the rapid deceleration of hydrate dissociation, while the sudden increase after 30 days results from the water invasion. Unlike Case 1A, Case 1B employs boundary sealing, which significantly and effectively suppresses the invasion of top and bottom water and greatly reduces water production. Its Qw curve shows a slow downward trend as the weakening driving force and the dissociation of hydrates slow down. Unlike Cases 1A and 1B, the cluster horizontal wells in Case 2A is located in the middle of the TPL, and most of the water in the TPL is bound water (Li et al., 2018; Zhang et al., 2017). Therefore, its Qw curve shows a slow upward trend in the initial stage and then gradually increases, which is attributed to the dissociation of hydrates and the intrusion of bottom water. Case 2B adopted boundary sealing, which further suppressed the invasion of bottom water, and its Qw curve was smaller overall than Case 2A. After 720 days of production, the Vw of the four cases were 351,379, 87,771, 151,345, and 57,369 m3, respectively. The Vw of Case 1B and Case 2B was 24.9% and 37.9% of that of Case 1A and Case 2A, respectively. Obviously, boundary sealing allows for a wider and more uniform distribution of pressure in the reservoir, effectively suppressing the invasion of top and bottom water and significantly improving the productivity of the entire system. In addition, compared with cluster horizontal wells, boundary sealing has a more marked effect on increasing gas production and inhibiting water production in the five-spot wells.
Gas-to-water ratio and <italic>J</italic> index
Figure 6 shows the performance of two key production indicators, gas-to-water ratio (Rgw) and J index. Among them, Rgw (ST m3 of CH4/ST m3 of water) is a key indicator for evaluating the production efficiency, and a higher Rgw indicates that gas production is more economical. From Figure 6A, it can be seen that the Rgw curve of Case 1B is on an upward trend throughout the entire production period, while the Rgw curve of Case 2B is on a downward trend. When production reached 720 days, the Rgw of the four cases were 39.2,266.8,84.4 and 345.5, respectively. Compared with Case 1A and Case 2A, Case 1B and Case 2B have increased by 680.6% and 409.3% respectively, indicating that boundary sealing has a more marked effect on improving the Rgw of the five-spot wells.
Evaluation indicators over 720 days. (a) Evolution of gas-to-water ratio. (b) Evolution of J Index.
Two line charts display results for four cases over 720 days. Panel (a) plots Rgw (methane water ratio) with Case 1A and 1B lower and flatter, and Case 2A and 2B starting higher and declining. Panel (b) plots Jindex with all cases declining rapidly, particularly 2A and 2B, with 1B remaining slightly higher. Legends and axis titles are present.
Generally, the average Qg is the most straightforward indicator for evaluating the production enhancement effect. When production reached 720 days, the Vg of the five-spot wells with boundary sealing is 2342.5 × 104 m3, which is higher than the gas production of the cluster horizontal wells with boundary sealing, which is 1982.8 × 104 m3. But in fact, the total length of the wellbore of the five-spot wells and the cluster horizontal wells is not equal. The overall length of the wellbore for the five-spot wells is 350 m, while that for the cluster horizontal wells is 450 m. Here, the J index serves as a supplementary indicator to assess productivity and is mainly affected by well type. The temporal evolution of the J index within 720 days is shown in Figure 6B, which provides a clear view of the efficiency trade-off between the two well-net development strategies. It can be seen that within a production period of 120 days, the cluster horizontal wells exhibit a markedly higher J index than the five-spot wells, which is due to its wellbore being located in TPL, which allows for efficient early recovery of high-saturation free gas located in TPL. However, after 120 days, the J index of the five-spot wells with boundary sealing has achieved a reversal, which underscores the superior long-term sustainability of the direct multi-layer development strategy under boundary sealing conditions. After 720 days of production, the J index of the four cases was 5.6, 9.5, 4.0, and 6.2, respectively.
Spatial distribution of reservoir physical parametersPore pressure
Figure 7 shows the evolution of pore pressure in four cases over 720 days. It can be observed that there are three commonalities in the pressure field diagrams of Cases 1A and 1 B. Firstly, there is a larger pressure gradient in the upper part of the vertical wellbore owing to the existence of solid gas hydrate in GHBL, which results in a lower initial permeability and therefore a larger pressure gradient. Secondly, the gas expansion effect of TPL and FGL in the lower part results in a relatively smaller pressure gradient. Thirdly, the vertical wellbore located at the center is most affected by the pressure superposition effect, resulting in a relatively large pressure gradient. In addition, it can be observed that there was no significant change in the wellbore pressure of Case 1A throughout the entire production period, which is due to the water invasion, consuming the driving energy and limiting the pressure propagation. Case 1B adopts boundary sealing, effectively suppressing the water invasion, allowing the driving energy to be concentrated on the dissociation of hydrates rather than consumed on the produced water. As production proceeds, the pressure propagation becomes wider. Similarly, Case 2A is also affected by the bottom water invasion, and there is almost no change in pressure propagation throughout the entire production period. Case 2B adopts boundary sealing, and as production progresses, the pressure distribution around the wellbore in the reservoir becomes wider and more uniform. However, compared to the pressure gradient of the five-spot wells, the cluster horizontal wells are smaller, which also proves that boundary sealing has a greater impact on the five-spot wells.
Pore pressure over 720 days.
Grid of sixteen scientific contour plots grouped in four rows and four columns, displaying changes in Pp (Pa) values versus Y (m) and Z (m), color-coded from red to blue, with each subplot labeled by case, time (ranging from 120 to 720 days), and specific layers such as DB, UB, GHEL, TPL, and FGL.Reservoir temperature
Figure 8 shows the evolution of reservoir temperature in four cases over 720 days. It can be observed that under the large pressure difference, hydrates rapidly decompose and are strongly influenced by the Joule-Thomson effect. Both Cases 1A and 1B have developed a low-temperature area near the wellbore, and the vertical wellbore located in the center is affected by the pressure superposition effect, resulting in a relatively large low-temperature area. As production progresses, Case 1A is gradually shrinking in the low-temperature area near the wellbore due to the water invasion. Case 1B adopts boundary sealing, effectively suppressing bottom water invasion. As production progressed, the hydrates continued to decompose, and the low-temperature area became larger. Similarly, in the initial stages, Case 2A also formed a low-temperature area near the wellbore. As production progressed, the low-temperature area became smaller due to the influence of bottom water invasion. Case 2B adopted boundary sealing; the inflow of bottom water was suppressed. As production progressed, the hydrates continued to decompose, and the low-temperature area became larger.
Reservoir temperature over 720 days.
Grid of sixteen temperature contour plots shows subsurface temperature distribution over depth and distance for four cases at four time intervals: 120, 240, 360, and 720 days. Each plot is labeled by case and features color gradients, contour lines, and stratigraphic layer annotations.Hydrate saturation
Figure 9 show the evolution of hydrate saturation for four cases over 720 days. It can be observed that both Cases 1A and 1B generate the secondary hydrates around the TPL wellbore, and the vertical wellbore located in the center was affected by the pressure superposition effect, resulting in a higher degree of hydrate dissociation around the wellbore. Compared to Case 1A, Case 1B adopts boundary sealing, and the energy of the pressure difference is more used for the dissociation. Thus, simultaneously, the saturation of secondary hydrates around the TPL wellbore is lower, and the range of hydrate dissociation is wider. Both Cases 2A and 2B generated a minor quantity of secondary hydrates in the reservoir adjacent to the wellbore root and toe, which is because a massive free gas flowed into the wellbore from the root and toe sections, thereby inducing a strong Joule-Thomson effect. Similarly, Case 2B employs boundary sealing, resulting in a much greater degree and extent of hydrate dissociation compared to Case 2A.
Hydrate saturation over 720 days.
Sixteen-panel scientific visualization shows color-coded cross-sectional diagrams for four cases (1A-FSW, 1B-FSW, 2A-CHW, 2B-CHW) across four time intervals: 120, 240, 360, and 720 days. Each panel includes stratified layers labeled Dk, GHB-L, TPL, and FGL, with a color bar above indicating Shc (mole fraction) from 0.02 to 0.35. Stratigraphic layers are colored blue to red, indicating concentration differences and migration patterns over time for each case. Panel axes are labeled Y (m) and Z (m).
In addition, the formation of secondary hydrates can reduce the permeability of the reservoir around the wellbore, thereby affecting production behavior. In Case 1A, where boundary sealing was not used, water invasion weakened the pressure-driven energy, resulting in a higher saturation of secondary hydrates generated around the TPL wellbore, which limits the free gas flow into the wellbore located at TPL. This is consistent with the decreasing trend of gas production rate and the increasing trend of water production rate (Figures 5a,c); In case 1B, boundary sealing effectively suppressed water invasion, allowing pressure energy to be concentrated on hydrate dissociation. The saturation of secondary hydrates generated around the TPL wellbore was relatively low, and the free gas could flow into the wellbore located at TPL relatively smoothly. At the same time, there were more dissociated gases from GHBL, supporting it to maintain a relatively high gas production rate and a relatively low water production rate (Figures 5a,c); In cases 2A and 2B, the formation of secondary hydrates is limited to the root and toe of the wellbore and has low saturation. Therefore, their differences in gas and water production behavior are mainly controlled by boundary sealing conditions (Figures 5a,c).
Gas saturation
Figure 10 show the evolution of gas saturation for four cases over 720 days. It can be observed that in Case 1A, high-saturation free gas at TPL flows continuously into the wellbore, and the gas saturation at TPL decreases steadily. Moreover, with the dissociation of hydrates in GHBL, the range of low-saturation gas accumulated in the upper reservoir around the wellbore becomes larger. Compared to Case 1A, Case 1B adopts boundary sealing, the degree of hydrate dissociation in GHBL is higher, and the area of low-saturation gas accumulated around the upper wellbore is larger. Similarly, in Case 2A, it was found that high-saturation free gas situated at the TPL continued to flow into the wellbore as production progressed, and the gas saturation around the wellbore decreased steadily. Compared to Case 2A, Case 2B also adopts boundary sealing, and the degree of hydrate dissociation in GHBL is higher, and the area of low-saturation gas accumulated around the upper wellbore is also larger.
Gas saturation over 720 days.
Sixteen-panel scientific figure displaying color-coded contour plots of soil salinity distribution over time for four different cases (Case 1A-FSW, Case 2A-CHW, Case 2B-CHW) at intervals of one hundred twenty, two hundred forty, three hundred sixty, and seven hundred twenty days. Each subplot includes vertical and horizontal axes labeled depth (z in meters) and distance (y in meters), with colored gradients representing salinity concentration (Sg, dimensionless) and a color scale ranging from 0.01 to 0.26. Distinct soil layers and boundaries are annotated.
The evolution of gas saturation distribution is also closely related to gas and water production behavior: in Case 1A without boundary sealing, the combined effect of high saturation secondary hydrates generated around the TPL wellbore and water invasion resulted in a rapid decrease in gas saturation, which is consistent with its gas and water production trend (Figures 5a,c); In Case 1B, due to the use of boundary sealing, the saturation of the secondary hydrate generated around the TPL wellbore is low, and the pressure energy is concentrated for hydrate dissociation. Therefore, the gas saturation and distribution range remain relatively high, and the water production rate is low (Figures 5a,c), further verifying the synergistic effect of boundary sealing in suppressing water invasion and improving gas recovery. Similarly, the same phenomenon can also be observed in Cases 2A and 2B.
Implications and future recommendations
The results of this study indicate that under boundary sealing conditions, the five-spot wells exhibit superior long-term production performance; However, the actual implementation of this strategy still faces multiple challenges. At the engineering level, building an effective sealing layer in deep-sea environments is a major challenge. This requires achieving a wide range and high-precision vertical isolation within the target layers, and ensuring that the injected sealing materials (such as gel, polymer, cement) maintain their sealing performance and mechanical integrity under long-term low-temperature and high-pressure conditions. The relevant construction techniques, long-term maintenance costs, and operational risks need to be comprehensively evaluated. At the simulation level, the geomechanical effects caused by long-term depressurization may lead to reservoir and wellbore instability, as well as sand production and other issues. Neglecting this coupling process may compromise the accuracy of long-term yield forecasting. Additionally, while the overburden and underburden thickness adopted in this model is sufficient to mitigate boundary effects for the investigated production period, studies focusing on longer-term dynamics or complex overburden-reservoir interactions should consider employing intact overlying formations and larger-scale boundary conditions to more accurately capture system-wide behaviors (Ye et al., 2025). Therefore, the conclusion drawn in this study regarding the comparison of short-term and long-term production performance of well-nets is based on the current model framework and specific geological conditions. In order to promote the on-site application of this technology, future work should future work should focus on researching sealing materials and construction techniques, establishing models combined with geomechanics, implementing overlying formations and larger-scale boundary conditions for long-term production forecasting and conducting comprehensive technical and economic feasibility assessments covering the entire lifecycle.
Conclusions
Based on China’s first offshore NGH trial production field data, the numerical approach was employed to comprehensively assess the yield-increasing effect of boundary sealing on Class 1 NGHs with five-spot wells and cluster horizontal wells. The findings can provide theoretical insights for the development of Class 1 NGHs through well-net patterns with boundary sealing. The key conclusions are summarized as follows:
Compared with the single depressurization method, the combination of boundary sealing can significantly improve the recovery rate. This is because the boundary sealing effectively suppresses water production, and the pressure energy is concentrated to drive the dissociation of hydrates, which can achieve more robust and sustainable efficient production of NGHs. Compared to cluster horizontal wells, the vertical wellbore of the five-spot wells is closer to the upper and lower boundaries, making it easier for water invasion, resulting in rapid water breakthrough and inefficient utilization of depressurization energy. Therefore, boundary sealing has a marked enhancement effect on gas production and water production inhibition in five-spot wells. After 720 days of production, the Vg of Cases 1B and 2B were 2342.5 and 1982.8 × 104 m3, respectively, which increased by 169.8% and 155.1% compared to Cases 1A and 2A. The Rgw was 266.8 and 345.5, respectively, which increased by 680.6% and 409.3% compared to Cases 1A and 2A. Within 120 days of production, the performance of the cluster horizontal wells was much better than that of the five-spot wells. However, after 120 days, the performance of the five-spot wells with boundary sealing surpassed that of the cluster horizontal wells. After 720 days of production, the J index of Cases 1B and 2B were 9.5 and 6.2, respectively. These findings reveal a clear strategic implication: under the boundary sealing conditions, cluster horizontal wells provide higher short-term production efficiency, making them suitable for objectives like pilot testing, while five-spot wells offer greater long-term development potential. Therefore, the choice of well-net pattern should be treated as a strategic decision, aligning with the primary objective of the specific project.
Data availability statement
The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding authors.
Author contributions
TW: Conceptualization, Methodology, Software, Writing – original draft. ZC: Formal Analysis, Funding acquisition, Investigation, Writing – review and editing. QL: Resources, Data curation, Visualization, Writing – review and editing. JQ: Data curation, Resources, Visualization, Writing – review and editing. CX: Visualization, Data curation, Resources, Writing – review and editing. JW: Project administration, Supervision, Writing – review and editing.
Conflict of interest
The author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Generative AI statement
The author(s) declared that generative AI was not used in the creation of this manuscript.
Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.
Publisher’s note
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.
ReferencesBoswellR. (2009). Is gas hydrate energy within reach?Science325, 957–958. 10.1126/science.117507419696340BoswellR.CollettT. S. (2011). Current perspectives on gas hydrate resources. Energy Environ. Sci.4, 1206–1215. 10.1039/c0ee00203hCaoX.SunJ.QinF.NingF.MaoP.GuY. (2023). Numerical analysis on gas production performance by using a multilateral well system at the first offshore hydrate production test site in the shenhu area. Energy270, 126690. 10.1016/j.energy.2023.126690ChongZ.YangS.BabuP.LingaP.LiX. (2016). Review of natural gas hydrates as an energy resource: prospects and challenges. Appl. Energy162, 1633–1652. 10.1016/j.apenergy.2014.12.061FengY.ChenL.SuzukiA.KogawaT.OkajimaJ.KomiyaA. (2019). Enhancement of gas production from methane hydrate reservoirs by the combination of hydraulic fracturing and depressurization method. Energy Convers. Manag.184, 194–204. 10.1016/j.enconman.2019.01.050GuoW.ZhongX.ChenC.ZhangP.LiuZ.WangY. (2024). Stimulation effect of network fracturing combined with sealing boundaries on the depressurization development of hydrate reservoir. Energy302, 131752. 10.1016/j.energy.2024.131752HaoY.YangF.WangJ.FanM.LiS.YangS. (2022). Dynamic analysis of exploitation of different types of multilateral wells of a hydrate reservoir in the South China sea. Energy Fuels.36, 6083–6095. 10.1021/acs.energyfuels.2c00493HuaZ.LinM.GuoJ.XuF.LiZ.LiM. (2013). Study on plugging performance of cross-linked polymer microspheres with reservoir pores. J. Petroleum Science Engineering105, 70–75. 10.1016/j.petrol.2013.03.008HuangM.WuL.NingF.WangJ.DouX.ZhangL. (2023). Research progress in natural gas hydrate reservoir stimulation. Nat. Gas. Ind. B10, 114–129. 10.1016/j.ngib.2023.01.005KowalskyM.MoridisG. (2007). Comparison of kinetic and equilibrium reaction models in simulating gas hydrate behavior in porous media. Energy Convers. Manag.48, 1850–1863. 10.1016/j.enconman.2007.01.017LiJ.YeJ.QinX.QiuH.WuN.LuH. (2018). The first offshore natural gas hydrate production test in South China Sea. China Geol.1, 5–16. 10.31035/cg2018003LiS.WuD.WangX.HaoY. (2021). Enhanced gas production from marine hydrate reservoirs by hydraulic fracturing assisted with sealing burdens. Energy232, 120889. 10.1016/j.energy.2021.120889LvT.PanJ.CaiJ.LiR.ShenP. (2022). Adaptability of artificial barrier to depressurization production of marine stratified gas hydrate reservoir. Nat. Gas. Ind.42 (11), 132–140. 10.3787/j.issn.1000-0976.2022.11.014MaX.SunY.LiuB.GuoW.JiaR.LiB. (2020). Numerical study of depressurization and hot water injection for gas hydrate production in China’s first offshore test site. J. Nat. Gas. Sci. Eng.83, 103530. 10.1016/j.jngse.2020.103530MoridisG.KowalskyM.PruessK. (2008). TOUGH+ hydrate V1.0 user’s manual; report LBNL-0149E. Berkeley, CA, USA: Lawrence Berkeley National Laboratory.MoridisG.CollettT.Pooladi-DarvishM.HancockS.SantamarinaC.BoswellR. (2011). Challenges, uncertainties, and issues facing gas production from gas-hydrate deposits. SPE Reservoir Evaluation & Engineering14 (01), 76–112. 10.2118/131792-paNieS.LiuK.XuK.ZhongX.TangS.SongJ. (2024). Numerical study on the stimulation effect of boundary sealing and hot water injection in marine challenging gas hydrate extraction. Sci. Rep.14 (1), 15280. 10.1038/s41598-024-66321-538961272NingF.ChenQ.SunJ.WuX.CuiG.MaoP. (2022). Enhanced gas production of silty clay hydrate reservoirs using multilateral wells and reservoir reformation techniques: numerical simulations. Energy254, 124220. 10.1016/j.energy.2022.124220QinX.LiangQ.YangL.QiuH.XieW.LiangJ. (2020). The response of temperature and pressure of hydrate reservoirs in the first gas hydrate production test in South China Sea. Appl. Energy278, 115649. 10.1016/j.apenergy.2020.115649QinF.SunJ.CaoX.MaoP.ZhangL.LeiG. (2025). Numerical simulation on combined production of hydrate and free gas from silty clay reservoir in the South China Sea by depressurization: formation sealing. Appl. Energy377, 124343. 10.1016/j.apenergy.2024.124343SloanE. (2003). Fundamental principles and applications of natural gas hydrates. Nature426, 353–359. 10.1038/nature0213514628065SuZ.MoridisG.ZhangK.WuN. (2012). A huff-and-puff production of gas hydrate deposits in shenhu area of South China Sea through a vertical well. J. Pet. Sci. Eng.86–87, 54–61. 10.1016/j.petrol.2012.03.02086SunJ.NingF.LiS.ZhangK.LiuT.ZhangL. (2015). Numerical simulation of gas production from hydrate-bearing sediments in the shenhu area by depressurising: the effect of burden permeability. J. Unconv. Oil Gas. Resour.12, 23–33. 10.1016/j.juogr.2015.08.003SunJ.ZhangL.NingF.LeiH.LiuT.HuG. (2017). Production potential and stability of hydrate-bearing sediments at the site GMGS3-W19 in the South China Sea: a preliminary feasibility study. Mar. Petroleum Geol.86, 447–473. 10.1016/j.marpetgeo.2017.05.037SunZ.LiN.JiaS.CuiJ.YuanQ.SunC. (2019). A novel method to enhance methane hydrate exploitation efficiency via forming impermeable overlying CO2 hydrate cap. Appl. Energy240, 842–850. 10.1016/j.apenergy.2019.02.022SunY.MaX.GuoW.JiaR.LiB. (2019). Numerical simulation of the short- and long-term production behavior of the first offshore gas hydrate production test in the South China Sea. J. Pet. Sci. Eng.181, 106196. 10.1016/j.petrol.2019.106196SunY.CaoB.ZhongJ.KanJ.LiR.NiuJ. (2022). Gas production from unsealed hydrate-bearing sediments after reservoir reformation in a large-scale simulator. Fuel308, 121957. 10.1016/j.fuel.2021.121957WangY.ZengY.ZhongX.PanD.ChenC. (2024). Investigation into the effect of permeable boundary sealing on the behavior of hydrate exploitation via depressurization combined with heat injection. Energies17 (20), 5172. 10.3390/en17205172WuN.LiY.WanY.SunJ.HuangL.MaoP. (2021). Prospect of marine natural gas hydrate stimulation theory and technology system. Nat. Gas. Ind. B40, 173–187. 10.1016/j.ngib.2020.08.003YeJ.QinX.XieW.LuH.MaB.QiuH. (2020). The second natural gas hydrate production test in the South China Sea. China Geol.3, 197–209. 10.31035/cg2020043YeH.ChenJ.YaoY.ChenD.WuX.LiD. (2025). Gas recovery from low-permeability muddy silt gas hydrate reservoirs by depressurization coupled with hot water injection: impact of hydro-lock effect. Energy316, 134413. 10.1016/j.energy.2025.134413YinZ.MoridisG.ChongZ.LingaP. (2019). Effectiveness of multi-stage cooling processes in improving the CH4-hydrate saturation uniformity in sandy laboratory samples. Appl. Energy250, 729–747. 10.1016/j.apenergy.2019.05.077YuT.GuanG.WangD.SongY.AbudulaA. (2021). Numerical investigation on the long-term gas production behavior at the 2017 shenhu methane hydrate production-site. Appl. Energy285, 116466. 10.1016/j.apenergy.2021.116466YuanY.XuT.XinX.XiaY. (2017). Multiphase flow behavior of layered methane hydrate reservoir induced by gas production. Geofluids2017, 7851031. 10.1155/2017/7851031YuanY.XuT.JinC.ZhuH.GongY.WangF. (2021). Multiphase flow and mechanical behaviors induced by gas production from clayey-silt hydrate reservoirs using horizontal well. J. Clean. Prod.328, 129578. 10.1016/j.jclepro.2021.129578YueP.ChenX.LiuH.JiaH. (2015). The critical parameters of a horizontal well influenced by a semi-permeable barrier considering thickness in a bottom water reservoir. J. Petroleum Sci. Eng.129, 88–96. 10.1016/j.petrol.2015.02.029ZhangS.MaY.XuZ.ZhangY.LiuX.ZhongX. (2024). Numerical simulation study of natural gas hydrate extraction by depressurization combined with CO2 replacement. Energy303, 131998. 10.1016/j.energy.2024.131998ZhangK.MoridisG.WuY.PruessK. (2008). “A domain decomposition approach for large-scale simulations of flow processes in hydrate-bearing geologic media,” in Proceedings of the 6th International Conference on Gas Hydrates (Vancouver, BC, Canada: ICGH).ZhangW.LiangJ.LuJ.WeiJ.SuP.FangY. (2017). Accumulation features and mechanisms of high-saturation natural gas hydrate in shenhu area,northern South china sea. Pet. Explor. Dev.44, 708–719. 10.1016/s1876-3804(17)30082-4ZhangS.MaY.HuangY.XuZ.LiuX.JiangS. (2025a). Research progress on hydrate replacement mechanism and enhancement methods: a review. Gas Sci. Eng.139, 205631. 10.1016/j.jgsce.2025.205631ZhangS.HuangY.MaY.JiangS.YinS.ZhongX. (2025b). Potassium acetate as an eco-friendly methane hydrate inhibitor: implications for thermodynamics and kinetics. Chem. Eng. J.518, 164560. 10.1016/j.cej.2025.164560ZhaoG.YouQ.TaoJ.GuC.AzizH.MaL. (2018). Preparation and application of a novel phenolic resin dispersed particle gel for in-depth profile control in low permeability reservoirs. J. Petroleum Sci. Eng.161, 703–714. 10.1016/j.petrol.2017.11.070ZhaoE.HouJ.LiuY.JiY.LiuW.LuN. (2020). Enhanced gas production by forming artificial impermeable barriers from unconfined hydrate deposits in shenhu area of South China sea. Energy213, 118826. 10.1016/j.energy.2020.118826ZhuD.PengS.ZhaoS.WeiM.BaiB. (2021). Comprehensive review of sealant materials for leakage remediation technology in geological CO2 capture and storage process. Energy & Fuels35 (6), 4711–4742. 10.1021/acs.energyfuels.0c04416
Edited by:Ebrahim Fathi, West Virginia University, United States