Effects of Impermeable Boundaries on Gas Production from Hydrate Accumulations in the Shenhu Area of the

Based on currently available data from site measurements and the preliminary estimates of the gas production potential from the hydrate accumulations at the SH7 site in the Shenhu Area using the depressurization method with a single horizontal well placed in the middle of the Hydrate-Bearing Layer (HBL), the dependence of production performance on the permeabilities of the overburden (OB) and underburden (UB) layers was investigated in this modeling study. The simulation results indicated that the temperature and the pressure of the HBL were affected by the permeabilities of OB and UB and the effect of depressurization with impermeable OB and UB was significantly stronger than that with permeable boundaries. Considering the percentage of hydrate dissociation, the gas production rate and the gas-to-water ratio, the hydrate deposit with impermeable OB and UB was expected to be the potential gas production target.

Natural gas hydrates (NGH) are solid, nonstoichiometric compounds formed by host water molecules with small guest molecules, such as CH 4 , C 2 H 6 , C 3 H 8 , CO 2 , H 2 S, etc. [1].Natural gas hydrate deposits mainly involve CH 4 and occur in the permafrost and in deep ocean sediments, where the necessary conditions of low temperature and high pressure for hydrate stability exist.Recent seismic explorations and geological researches have shown that natural gas hydrates in these sediments constitute a large natural gas reservoir that invites consideration as a potential strategic energy resource [2][3][4][5].
The Blake Ridge hydrate accumulation offshore the Carolinas [23], in which large volumes of gas hydrate are distributed in vast volumes of fine-grained and relatively undeformed sediment at low (2%-4% average) saturations, has been extensively studied since the °Cean drilling program (ODP) Leg 164 [24], and a preliminary estimate of the gas resource is up to 13.7 × 10 9 t under standard atmospheric conditions [24].Numerical simulations indicate that the hydrate accumulations at Blake Ridge with the free gas reservoir beneath the hydrate-bearing layer may be of economic value for industrial production [25,26].
Recent studies [10,17,18,22] have indicated that, under certain conditions, gas can be produced from natural hydrate deposits at high rates over long periods using vertical wells by means of depressurization and thermal stimulation.Simulation results indicate that horizontal wells [10,19,22] will show significant advantages over the vertical wells in gas production from Class 2 and Class 3 deposits.

Hydrates in the Shenhu Area
Recent studies [27][28][29][30][31][32] have documented the occurrence of the significant gas hydrate deposits in the Shenhu Area of the South China Sea.The Shenhu Area is located in the near southeast of the Shenhu Underwater Sandy Bench in the middle of the north slope of the South China Sea, between the Xisha Trough and the Dongsha Islands (Figure 1).Gas hydrate samples were collected during a recent scientific expedition conducted by the China Geological Survey in the Shenhu Area of the northern South China Sea in May 2007 [28][29][30]32].From site measurements, the heat flow in the Shenhu Area was estimated to range from 74.0 mW/m 2 to 78.0 mW/m 2 .Estimates of the geothermal gradient ranged from 43 °C/km to 67.7 °C/km.The water temperature at the ocean floor T OF is in the 4−5 °C range for water depths exceeding 1000 m [28,30].The drilling results from the hydrate layer in the Shenhu Area, measured from both non-pressure and pressurized cores, indicate that the top of the hydrate layers are located 155-229 m below seafloor (mbsf), and their thickness varies from 10 m to 43 m.These hydrate layers occur at water depths from 1108 m to 1245 m.The sediment porosity ϕ and the in situ salinity X S in the Shenhu Area, measured from the pressure cores, are 33%-48% and 0.0290-0.0315,respectively; the T OF and X S at the seafloor are 3.3-3.7 °C and 0.0328-0.0334,respectively.

Objective and Approach
The main objective of this study is to investigate the effects of the permeabilities of the boundaries, including the overburden (OB) and the underburden (UB), on the gas production potential at the SH7 drilling site of the Shenhu Area with a single horizontal well using the depressurization method.The effects of the permeabilities of OB and UB (k OB and k UB , respectively) on the gas production potential are investigated by the following three cases: (i) k OB = k UB = 75 mD; (ii) k OB = 0 mD, k UB = 75 mD; (iii) k OB = k UB = 0 mD.Case (i) has a similar numerical model, system properties and production method as the reference case in our previous work [10], which is based on the gas hydrate deposit in the South China Sea Shenhu Area, while both Case (ii) and Case (iii) are hypothetical cases.According to the previous classification of the natural hydrate accumulations [8], Case (iii) belongs to Class 3, which involves only an HBL, without underlying mobile fluid zones, and usually bounded by low-permeability OB and UB.This study makes a systematic comparison of the production performance in the cases with different permeabilities of the boundaries.In order to evaluate the production potential of the hydrate deposit, we use two criteria, an absolute criterion and a relative criterion.To satisfy the absolute criterion, a large production potential must be demonstrated, as quatified by an early large cumulative average total gas production rate Q avg (>1.70 × 10 5 ST m 3 /day = 6 MMSCFD), and a large Q avg (>5.68 × 10 4 ST m 3 /day = 2 MMSCFD) over the duration of production (in this work, the duration of more than 80% of the hydrate in the deposit dissociated).The relative criterion is satisfied when the gas-to-water ratio R GW = V P /V W is high, indicating a small cumulative volume of produced water V W relative to V P , where V P is the cumulative volume of the produced CH 4 .R GW provides a measure of the hydrate system response and the effectiveness of dissociation as a gas-producing method.

Method of Production
Previous studies [18,20,21] show that the depressurization-induced dissociation (based on fluid removal) is a promising gas production strategy in the Shenhu Area hydrate deposit because of its simplicity and its technical and economic effectiveness.During the initial stage of production, constant-P production allows a continuous increase in the gas production rate Q PT .This is caused by the continuously increasing k eff in the deposit as S H declines in the course of hydrate dissociation.Furthermore, by setting the well pressure P W > P Q (i.e., the P at the quadruple point), constant-P production provides the necessary bottom pressure control to eliminate the possibility of ice formation and significantly reduce the formation of secondary hydrates.The driving force of the depressurization ∆P W = P 0 − P W = 0.78 P 0 , where P W = 3.0 MPa is the constant well pressure, and P 0 = 13.7 MPa is the initial pressure in the HBL at the elevation of the horizontal well (r W = 0.1 m).

Numerical Simulation Code
For this numerical simulation study, we use both the serial and parallel versions of the TOUGH + HYDRATE code [32,33].This code can model nonisothermal hydration reactions, phase behavior, and flow of fluids and heat under the conditions typical of natural CH 4 -hydrate deposits in complex geologic media.In this code, the impact of the fracturing, grain migration, the volume expansion and the instability in the reservoir are neglected.

Geometry and Domain Discretization
Analysis of hydrate samples from the SH7 site at the Shenhu Area indicates almost pure (99.2%)CH 4 -hydrate in clayey sediments.The system properties and initial conditions that are used in the simulation are shown in Table 1.The wellbore is simulated as a pseudo-porous meium with ϕ = 1.0, k = 5.0 × 10 −9 m 2 (5000 Darcies), and a capillary pressure P cap = 0.The seafloor at this site is at 1108 m depth, and the HBL extends from 155 m to 177 m interval below the seafloor (22 m thick).The domain in all the cases includes an additional 30 m underburden and overburden, which have the same flow properties as the HBL.The boundaries above the overburden and below the underburden are set to be under constant conditions (P and T).   Figure 2 shows the system geometry and configuration of the horizontal well.The geometry and configuration of the hydrate deposit, as well as the corresponding simulation grid, are shown in Figure 3.

Spatial Distributions in Case (i)
In the Case (i), both of the OB and UB are permeable with k OB = k UB = 75 mD.Figures 4-7 show the evolution of the distributions of the hydrate saturation S H , the gas saturation S G , the temperature T and the pressure P over time in the Case (i).White lines in Figure 7    (c) the evolution of the high-S H region (secondary hydrate formation) immediately above the lower dissociation front; (d) the accumulation of gas in the vicinity of the well around the hydrate dissociation interface near the well in the early stage, and then below the lower dissociation interface during the later period.Comparing with the reference case with ∆P W = 0.20 P 0 in the previous study [10], the Case (i) with larger driving force ∆P W = 0.78 P 0 shows some different characteristics: (a) as shown in Figure 4a, the mergence of the lower dissociation interface, the cylindrical interface around the well and the upper dissociation interface occur within approximately t = 60 days in the Case (i); (b) as shown in Figure 4, the hydrate dissociation rate is much higher and almost all the hydrate in the initial HBL are dissociated within 10 years in Case (i), while in the previous reference case, there is still a lot of hydrate undissociated in the HBL after 30 years; (c) as shown in Figure 5, the gas saturation S G , with a maximum of approximately 0.14, is higher in the Case (i) than that in the previous reference case [10] (S Gmax is approximately 0.10), which is caused by the higher hydrate dissocation rate and the more gas released from the hydrate.

Figure 5. Evolution of spatial distribution of S G with permeable boundaries (OB and UB) in Case (i).
As shown in Figure 6, a low-T zone (defined as the zone with T < 12.0 °C) in Figure 6a-e occurs in the vicinity of the cylindrical interface around the well and the upper dissociation interface (Figure 4), which indicates the cooling effects of the endothermic hydrate dissociation reaction.The low-T zone shrinks in Figure 6e,f because of the encroachment by warmer water flooding from the overburden (where there is an inversion of the geothermal gradient, as a result of dissociation-induced cooling in the HBL).The warmer water rising from the UB is clearly depicted by the spatial distribution of T in the vicinity of the lower dissociation interface (Figure 4).All fluids with different source, including the original free water in the HBL, the gas and water released from hydrate dissociation, the water from the overburden, and the warmer water rising from the underburden, converge toward the production well.

Spatial Distributions in Case (ii)
In the Case (ii), the top boundary OB is impermeable with k OB = 0 mD, and the bottom boundary UB is permeable with k UB = 75 mD.The HBL in the Case (ii) is a semi-open system.Figure 8   Under the same depressurization driving force (∆P W = 0.78 P 0 ), there is no fluid flow from the OB to the HBL in the Case (ii) with impermeable OB (k OB = 0 mD), which causes the enhancement of the depressurization effect.The hydrate dissociation rate in the Case (ii) is much higher than that in the Case (i), and almost all the hydrate in the initial HBL is dissociated within 5 years Figure 8f, which is much earlier than that Figure 4f in the Case (i).Furthermore, under the more effective depressurization, the hydrate dissociation occurs all over the HBL simultaneously and the S H decreases in the entire HBL in the Case (ii) (Figure 8).Comparing the S H distributions at the same time points, such as the t = 60 days in Figures 4a and 8a, the t = 2 years in Figures 4d and 8e, and the t = 5 years in Figures 4e and 8f, the above characteristic in Case (ii) is quite different from that in Case (i), in which most of the hydrate dissociation occurs in the vicinity of the dissociation interfaces, including the hydrate dissociation interface around the well, the upper and the lower dissociation interfaces.As shown in Figure 8a-d, the impermeability of the OB causes a "dead zone" of the fluid flow above the production well near the top of the initial HBL during the depressurization from the well, which has a negative effect on the formation and the evolution of the upper dissociation interface.The mergence of the hydrate dissociation interface around the well, as well as the upper and the lower dissociation interfaces, is much later than that in the Case (i) with the permeable OB (Figure 4).

Spatial Distributions in Case (iii)
In Case (iii), both of the top and bottom boundaries OB and UB are impermeable with k OB = k UB = 0 mD. Figure 9 shows the evolution of the S H distribution over time in Case (iii).The time points (a-f) in Figures 9 are t = 60 days, 180 days, 1 year, 2 years, 5 years and 10 years, respectively.Because of symmetry of the well and a no-flow boundary (of fluids and heat) applied at x = 45 m (indicating a well spacing of 90 m), the HBL is a closed system with impermeable upper and lower boundaries (OB and UB) before gas production from hydrate. Figure 9 shows the cylindrical hydrate dissociation interface around the well, the upper and the lower dissociation interfaces near the top and the bottom of the HBL, and the mergence of these interfaces.One of the special characteristics of the S H distribution is that both of the upper and the lower dissociation interfaces are almost horizontal during the entire production process.The reason for this is that the thermodynamic conditions (P and T) along these interfaces are always constant.In a closed system of the initial HBL in Case (iii), the effect of the depressurization is much stronger than that in the open and semi-open system of Case (i) and Case (ii), and the S H is much lower in this case (Figure 9).Furthermore, the depressurization affects the entire HBL obviously and the S H decreases all over the HBL simultaneously (Figure 9), which indicates that the hydrate dissociation occurs in the entire HBL.

Gas and Water Production
Figure 10 shows the evolutions of the volumetric rates of CH 4 production in the gas phase (Q PG ), the total gas production (Q PT ) at the well, and the gas released from the hydrate dissociation (Q R ). Figure 11 shows the evolution of the percentage of the hydrate dissociated in the entire simulated domain.As shown in Figure 10, in all three cases, Q PT , Q PG and Q R decrease with the hydrate dissociation over time.In Case (i) (black line), Q PT is much higher than both Q PG and Q R , which indicates that the source of the majority of the produced gas is CH 4 dissolved in the water, rather than from the free gas phase.The percentage of hydrate dissociated is approximately 96% at t = 3650 days (black line in Figure 11), confirming the results of the existence of the hydrates undissociated in the deposit after 10 years of production shown in Figure 4f.
In Case (ii) (red line), an inflection point on the curve of the Q PT , Q PG and Q R is observed at approximately t = 500 days (Figure 10), when the mergence of the upper and the lower dissociation interfaces occurs Figure 8d.This is caused by the hydrate dissociation between the upper and the lower dissociation interfaces, which connect the upper dissociated zone to the production well, (the undissociated HBL with low k eff plays an important role as a barrier to the downward movement of the gas and water dissociated from the hydrate along the bottom of the OB in the initial HBL).As shown in Figure 8f, there is little hydrate undissociated in the deposit after five years of production, and the percentage of hydrate dissociated is more than 98% at t = 1825 days (red line in Figure 11).As shown in Figure 10, in Case (ii) (red line), after about t = 2200 days, Q PT decreases to approximately 12.0 ST m 3 /day/m of well and then stays constant, and both Q PG and Q R decrease to a very low level (<0.01 ST m 3 /day/m of well), which indicates that the produced gas is from the CH 4 dissolved in the water in the depsoit, rather than from the free gas phase, and that this time point is the end of the hydrate dissociation process.The corresponding hydrate dissociation percentage (red line in Figure 11) approaches 1.0, which indicates that almost all the hydrate in the deposit disappears at that time.
In Case (iii) (green line), Q PT , Q PG and Q R almost coincide until about t = 2200 days (Figure 10), which indicates that the produced gas is CH 4 released from hydrate dissociation and it is in the free gas phase, rather than the gas dissolved in the water.As shown in Figure 10, in Case (iii) (green line), after about t = 2200 days of production, Q R decreases rapidly to lower than 0.01 ST m 3 /day/m of well, and almost all the hydrate in the deposit disappears at that time (the corresponding hydrate dissociation percentage approaches 1.0, green line in Figure 11).Meanwhile, Q PT and Q PG still coincide and decrease over time.As shown in Figure 10, in Case (iii), the cumulative average production rate of total gas Q avg is 31 ST m 3 /day/m of well at time t = 730 days (2 years, the duration of more than 80% of the hydrate in the deposit dissociated), and Q avg = 86 ST m 3 /day/m of well at t = 200 days.The maximum Q avg is 194 ST m 3 /day/m of well at t = 40 days.For a well spacing of 90 m and a horizontal well with the length of 1,000 m, Q avg is 6.20 × 10 4 ST m 3 /day (2.18 MMSCFD) at t = 730 days, and 1.72 × 10 5 ST m 3 /day (6.06 MMSCFD) at t = 200 days, and the maximum Q avg is as high as 3.88 × 10 5 ST m 3 /day (13.67 MMSCFD) at t = 40 days.It is obvious that the large production potential of the hydrate deposits in Case (iii) satisfies the absolute criterion.
As shown in Figure 11, from t = 0 to 800 days, the green line is higher than the red line, indicating that the hydrate dissociation rate in Case (iii) is higher than that in Case (ii) at early times of the production process.The reason for this is the enhancement of the effect of the depressurization in Case (iii) with the upper and the lower impermeable boundaries.With the dissociation of the hydrate in the deposit, the amount of the hydrate undissociated decreases and the temperature of the HBL declines, causing the decrease of the slope of green line in Figure 11.
Figure 12 shows the corresponding water production rates Q W and gas-to-water ratios R GW .In both Case (i) (black line) and Case (ii) (red line), the water production rates Q W increase rapidly to more than 2.0 × 10 4 kg/day/m of well after approximately t = 300 days, and the long-term R GW is prohibitively low (R GW < 10 ST m 3 of CH 4 /m 3 of H 2 O).This indicates that the gas production from the deposits of both Case (i) and (ii) are not economically profitable from the relative criterion point of view.Furthermore, in these two cases, for a well spacing of 90 m and a horizontal well with the length of 1000 m, the water production rate is Q W = 4.0 × 10 7 kg/day (4.0 × 10 4 ton/D), which is unmanageable.As shown in Figure 12, in Case (iii) (green line), both Q W and R GW show quite different profiles: (1) Q W decreases rapidly to lower than 30 kg/day/m of well after approximately t = 200 days.In this case, for a well spacing of 90 m and a horizontal well with the length of 1,000 m, the water production rate is only Q W = 6.0 × 10 4 kg/day (60 ton/D); (2) R GW increases to larger than 190 ST m 3 of CH 4 /m 3 of H 2 O after approximately t = 200 days, which indicates that the relative criterion is satisfied.In a word, it is obvious that the impermeable boundaries showthe best effects on the gas production potential of the hydrate deposits in terms of both the absolute and the relative evaluation criterion.

Summary and Conclusions
In this study, we investigated the effects of the permeabilities of OB and UB on the gas production potential from marine gas hydrate deposits at the SH7 site, in the Shenhu Area of the South China Sea.From the numerical simulation results, the following conclusions are drawn: 1.In Case (i) with permeable upper and lower boundaries, the dissociation is characterized by the following features: the evolution of the hydrate dissociation interface around the well, the upper and the lower dissociation interfaces, the evolution of the secondary hydrate, and the accumulation of gas in the vicinity of the well and below the lower dissociation interface.The relative warmer water rises from the UB and is produced from the well;

Figure 1 .
Figure 1.(a) Location of research area, Shenhu Area, north slope of South China Sea; (b) Bathymetric map of gas hydrate drilling area with sites drilled in the research area.

Figure 2 .
Figure 2. System geometry and configuration of the horizontal well producing from the hydrate deposit at the SH7 site of the Shenhu Area, South China Sea.

Figure 3 .
Figure 3. (a) A schematic of the marine hydrate deposit at the SH7 site of the Shenhu Area, South China Sea; (b) The grid used in the simulations.
indicate the initial position of the top and base of the HBL at z = 11 m and z = −11 m, respectively.The time points (a-f) in Figures 4-7 are t = 60 days, 180 days, 1 year, 2 years, 5 years and 10 years, respectively.Comparison of the spatial distributions of 10 years to the initial HBL provides a measure of the hydrate dissociation profile.

Figure 4 .
Figure 4. Evolution of spatial distribution of S H with permeable boundaries (OB and UB) in Case (i).

Figures 4
Figures 4 and 5 show the evolution of the S H and S G spatial distributions over time in the HBL.The figures show: (a) the evolution of the hydrate dissociation interface around the well, the upper and the lower dissociation interfaces; (b) fast hydrate dissociation rate in the vicinity of the above interfaces;(c) the evolution of the high-S H region (secondary hydrate formation) immediately above the lower dissociation front; (d) the accumulation of gas in the vicinity of the well around the hydrate dissociation interface near the well in the early stage, and then below the lower dissociation interface during the later period.Comparing with the reference case with ∆P W = 0.20 P 0 in the previous study[10], the Case (i) with larger driving force ∆P W = 0.78 P 0 shows some different characteristics: (a) as shown in Figure4a, the mergence of the lower dissociation interface, the cylindrical interface around the well and the upper dissociation interface occur within approximately t = 60 days in the Case (i); (b) as shown in Figure4, the hydrate dissociation rate is much higher and almost all the hydrate in the initial HBL are dissociated within 10 years in Case (i), while in the previous reference case, there is still a lot of hydrate undissociated in the HBL after 30 years; (c) as shown in Figure5, the gas saturation S G , with a maximum of approximately 0.14, is higher in the Case (i) than that in the previous reference case[10] (S Gmax is approximately 0.10), which is caused by the higher hydrate dissocation rate and the more gas released from the hydrate.

Figure 7
shows: (a) the evolution of the pressure gradient around the well over time; (b) the pressure drop in the overburden and underburden; (c) the jagged P distribution and the inflexions near the interface of the dissociated and undissociated zone in the HBL.Of those, (a) and (b) are results of the characteristic of the permeabilities of the HBL and the boundaries, and (c) is a result of the fluid flow caused by gas and water dissociated from the hydrate.

Figure 6 .
Figure 6.Evolution of spatial distribution of T with permeable boundaries (OB and UB) in Case (i).

Figure 7 .
Figure 7. Evolution of spatial distribution of P with permeable boundaries (OB and UB) in Case (i).
shows the evolution of the S H distribution over time in the Case (ii).The time points (a-f) in Figures 8 are t = 60 days, 180 days, 1 year, 555 days, 2 years and 5 years, respectively.

Figure 8 .
Figure 8. Evolution of spatial distribution of S H with impermeable OB and permeable UB in Case (ii).

Figure 9 .
Figure 9. Evolution of spatial distribution of S H with impermeable boundaries (OB and UB) in Case (iii).

Figure 10 .
Figure 10.Effect of the permeability of the boundaries on Q PT , Q PG and Q R .

Figure 11 .
Figure 11.Effect of the permeability of the boundaries on the percentage of hydrate dissociated.

Figure 12 .
Figure 12.Effect of the permeability of the boundaries on Q W and R GW .

3 of
/day/m of well) Time (days)Q W : Case (i) Permeable Boundaries Q W : Case (ii) Impermeable Overburden Q W : Case (iii) Impermeable Boundaries R GW : Case (i) Permeable Boundaries R GW : Case (ii) Impermeable Overburden R GW : Case (iii) Impermeable Boundaries R GW (ST m rate of total CH 4 at time t (ST m 3 /day/m of well) Q W mass rate of aqueous phase production at the well (kg/day/m of well) Q PG volumetric rate of CH 4 production at the well in the gas phase (ST m 3 /day/m of well) Q PT volumetric rate of total CH 4 production at the well (ST m 3 /day/m of well) Q R volumetric rate of CH 4 release from hydrate dissociation (ST m 3 /day/m of well)

Table 1 .
Properties and conditions of the hydrate deposit at site SH7 in the Shenhu area, South China Sea.