Geomechanically Sustainable Gas Hydrate Production Using a 3D Geological Model in the Ulleung Basin of the Korean East Sea

Abstract

Although various simulation studies on gas hydrate production have been conducted, a single vertical well in the cylindrical system has been adopted in most research. However, this system has a limited ability to predict commercial production in gas hydrate reservoirs. In order to facilitate commercial production, a field-scale reservoir model with a multi-well system must be constructed using geological data, such as seismic data, well logging data, core data, etc. The depressurization method is regarded as a practical production strategy because it has high levels of production efficiency and economical effectiveness. However, this method can lead to subsidence due to the increased effective stress. In this work, we studied a production simulation strategy for commercial gas hydrate production. A three-dimensional geological model with a realistic field scale is constructed using seismic and well logging data from the Ulleung Basin of the Korean East Sea. All of the grids are refined in the I and J direction, and the grids near the production well are very small to consider realistic hydrate dissociation. The cyclic depressurization method is adopted for the increase in the geomechanical stability, rather than the non-cyclic depressurization method. Various case studies are conducted with alternating bottomhole pressures for the primary and secondary depressurization stages over 100 days. Geomechanical stability is significantly enhanced, while cumulative gas production is relatively less reduced or nearly maintained. In particular, all cases of the cumulative gas production at 6 MPa during the secondary depressurization stage are similar to the non-cyclic case, while the geomechanical stabilities of those cases are restored. This study is thought to have contributed to the development of technology for commercial gas hydrate production with a geomechanical stability study using a reservoir-scale model with a multi-well system.

1. Introduction

Gas hydrate is a crystalline compound that occurs when water forms a cage-like structure physically trapped guest gas molecules [1]. Research of gas hydrate has been broadly conducted in various scientific and industrial fields, such as energy recovery, CO₂ capture and storage, gas separation, water desalination, gas storage and transport [2]. Gas hydrate is stable under high-pressure and low-temperature conditions [3]. Accordingly, gas hydrate deposits are found in permafrost, marine sediments and deep lakes, where the pressure is over 0.6 MPa and temperature is less than 300 K at the same time [4]. As represented in Figure 1, the phase boundary and geothermal gradient are used to define the hydrate stable region in permafrost and marine sediments [5]. In Figure 1, gas hydrate exists in the yellow shaded zone enclosed by a geothermal gradient (red dash line) and phase boundary (purple line). The difference between the two conditions is that the gas hydrate stable zone (GHSZ) is below the sea floor in the case of marine sediments. Gas hydrate in marine sediments exists 1200~1500 m below sea level, while the occurrence depth is 200~1100 m below sea level in permafrost.

The map in Figure 2 indicates that natural gas hydrate has been recovered and is inferred to present in various deep marine and permafrost region [6]. Through deep coring projects or shallow seabed coring operations, most of the recovered gas hydrate samples were acquired. Inferred gas hydrate occurrences are sites where bottom simulating reflectors (BSRs) have been identified using available seismic data. Site names in this map show gas hydrate drilling projects and expeditions.

Marine sediments occupied more than 95% of global gas hydrate resources and have many possibilities as a next energy resource. As shown in Figure 3, the worldwide resource of natural gas hydrate is enormously large. It is estimated to be 3000 trillion cubic meters (TCMs), which is much bigger than those of conventional gas (30 TCMs), early unconventional resources (30 TCMs), such as tight gas and shallow shales, and emerging unconventional resources (300 TCMs), such as CBM and deep shales [7]. In going down this pyramid, in-place resource volume and dependence on technology increase, while resource quality, concentration and recoverability decrease. Nowadays, we can only commercially produce conventional resources and early/emerging unconventional resources. Although many field production tests have been conducted, for instance, the Messoyakha gas field in Russia, the Malik site in Canada, Alaska’s North Slope in the USA, the Nankai Trough in Japan and the northern South China Sea in China, there is no commercial production site for gas hydrate [8].

In order to produce gas hydrate, gas hydrate in a stable state should be made to be in an unstable state by external action, and this process is called dissociation. Dissociation is related to the thermodynamic property of natural gas hydrate. There are several production methods for gas hydrate using hydrate dissociation, such as the depressurization method, thermal stimulation method and gas/inhibitor injection method. Various pilot tests for gas hydrate production have been conducted in the hydrate reservoirs using several production methods. A pilot test for gas hydrate production using thermal stimulation (hot water circulation) was conducted at the Mackenzie Delta in the Canadian Arctic in 2002 [9]. However, the production volume of gas hydrate was only 468 m³ over 5 days, even with the large amount of energy consumption [10]. The gas injection method using CO₂ and N₂ was utilized at Ignik Sikumi, Alaska, and the results represented that the injection rate was reduced by the hydrate formation of the injected gas in the sediments [11]. However, there are few experimental results showing the efficiency improvement using new techniques such as separate injection and production well [12]. The inhibitor injection method has many disadvantages such as rapid reduction in effectiveness by dilution, the cost of chemical inhibitors and the possibility of precipitation [13]. In both the thermal stimulation and the inhibitor injection methods, the direction of the injected fluid is opposite to the flow direction of dissociated natural gas from hydrates in the reservoir. This phenomenon will bring about low levels of production recovery [14]. The depressurization method has many advantages for the commercial production from a hydrate reservoir due to the high levels of production efficiency and technical/economical effectiveness [14,15,16]. In particular, this method has been applied successfully in various pilot tests for gas hydrate production, such as in the Mallik of Canada, the Nankai Trough of Japan and the South China Sea of China [17,18,19].

In spite of the successful pilot test using the depressurization method, there are no commercial production sites for gas hydrate in the world. The main two reasons are that the production rate did not satisfy a minimum gas production rate and also the vertical subsidence problem. The first problem is related to the dissociation of gas hydrate. This problem leads to relatively lower gas production than in general gas reservoirs. The second problem is that most hydrate-bearing sediments (HBS) consist of unconsolidated porous rocks [20]. Accordingly, there is a possibility of a vertical subsidence problem in the depressurization method, which is lowering the reservoir pressure [21,22,23].

In this study, simulation studies are conducted using the cyclic depressurization method with a multi-well system for commercial gas hydrate production in the Ulleung Basin of the Korean East Sea. Generally, most simulation studies have used a single production well in a cylindrical system. Multiple production well systems in a geology model at field scale can increase the gas production rate to satisfy the economical production rate. Accordingly, a 3D, realistic geological model with a reservoir scale is constructed using seismic data and well logging data of the Ulleung Basin. In order to describe realistic gas hydrate production, a geological model with a multi-well system is reduced into a small-scale model, and grids near the multiple production wells are refined. The cyclic depressurization method, which uses alternating depressurization and shut-in periods, is proposed for the increasing of the recovery factor [24]. On the other hand, Lee et al. adopted this model for its geomechanical stability, and its results represented a successful outcome with little loss of gas production in the cylindrical system with a single vertical well [25]. We also adopt this method for its geomechanically sustainable production. Various case studies are conducted with alternating bottomhole pressures of the primary and secondary depressurization stages over 100 days. Geomechanical stability is significantly enhanced, while cumulative gas production is relatively less reduced or nearly maintained. This study is thought to have contributed to the development of technology for commercial gas hydrate production via a geomechanical stability study using a reservoir scale model with a multi-well system.

2. Geology of the Ulleung Basin and Geological Model Construction

2.1. Geology of the Ulleung Basin

The study area is located in Ulleung Basin, the East Sea. The East Sea occupies the region between Korea and the Japan island arc in the northwestern margin of the Pacific (Figure 4a). In the aspect of plate tectonics, the East Sea is located on the eastern margin of the Eurasian Plate [26,27]. Physiographically, the East Sea is composed of three deep basins (Japan, Yamato, and Ulleung Basin) and three highs (Oki Bank, Korea Plateau, and Yamato Ridge). These deep basins are interpreted as back-arc basins induced by the subduction of the tectonic plate along the Japan island arc [28,29,30]. The Ulleung Basin is at the southernmost part among three deep basins in the East Sea (Figure 4a). The basin is bounded by the Korean Strait to the south, the Korean Peninsula to the west, the Korean Plateau to the north and the Oki bank to the east. The continental shelf of the Ulleung Basin ranges about 200~300 m in water depth. It is deepened through the continental slope basinward up to 2500 m in water depth. The continental slope dipping to the deep basin floor is steeper on the western slope (up to 10°) than the southern slope (1~2°).

The basin evolution of the East Sea was probably initiated before the late Oligocene with crustal thinning and seafloor spreading in the Japan basin [31,32]. Subsequently, the rifting and spreading of these deep basins in the East Sea continued until the Middle Miocene, until the extensional stress regime was changed. In the early Middle Miocene, the tectonic regime inverted to a compressional regime and precursors indicating back-arc closing occurred in Tsushima Island. After that, the basin was in the closing stage, and uplift and compressional fault or fold structures of Neogene strata were predominantly formed around the Ulleung Basin. The Neogene strata filling the basin during the Miocene and Pliocene epochs include widespread mass transport deposits on the slope area and extensive turbidite and hemipelagic sediment on the basin floor [33].

The Ulleung Basin is a well-known site for gas hydrate research [34,35,36,37,38]. The occurrence of gas hydrate in sedimentary succession is inferred from geophysical indicators, such as BSR (bottom simulating reflectors), enhanced reflectors, acoustic blanking and seismic chimneys [34,36,38]. The existence of natural gas hydrate in the Ulleung Basin has been confirmed by two drilling expeditions in UBGH-1 and UBGH-2 [34,36]. The gas hydrate in the Ulleung Basin is found in three different occurrence types: (1) the pore-filling type in porous sediment layers, (2) the fracture-filling type in seismic chimney structures and (3) the disseminated type in silt [39]. In addition, it is confirmed that the gas hydrate stability zone (GHSZ) in the basin floor of the Ulleung Basin is formed within a sedimentary succession less than 200 m below the seafloor [36]. Sediments consisting of the GHSZ in the slope area are characterized by vertically stacked mass transport deposits, whereas sediments of GHSZ in the base of the slope and basin floor are mainly composed of deep-water turbidite sand layers and hemipelagic settling muds.

2.2. Geological Model Construction

The study area is located in the northwestern corner of the basin floor in the Ulleung Basin, including drilling site UBGH2-6. Water depth in this area is about 2156 m below the sea surface (Figure 4b,c). Topographic variation is relatively gentle and smooth. The sedimentary succession shown in the vertical seismic profile is characterized by vertically stacked chaotic lens-shaped bodies in the lower part and well-stratified parallel reflections in the upper part (Figure 5). The chaotic seismic facies in the lower part is interpreted as a stacked mass transport deposit. This unit is gently tilted from site UBGH2-6 in the southwest direction. The upper seismic facies, seen as high-continuity and well-stratified parallel reflections, show turbidite and hemipelagic sediment [36]. BSR appears at about 3.15 s in two-way travel time, cross-cutting the upper well-stratified reflections with negative polarity to that of the seafloor (Figure 5). The presence of gas hydrate in this area is confirmed by drilling at the UBGH2-6 site. Sedimentary cores and LWD (logging while drilling) log responses indicate that the gas hydrate deposits are mainly concentrated between 110~155 mbsf (meter below seafloor) and calculated gas hydrate saturations in this high-resistivity interval show more than 18%. Thus, the GHSZ in the study area is characterized by turbidite and hemipelagic sediment succession above the BSR.

On the basis of the seismic interpretation, including changes in internal seismic facies and the identification of regional key horizons, the sedimentary success is subdivided into four seismic units: FA, FB, FC and FD (Figure 5). FA is the lowermost unit overlain by inferred BSR. The overall internal seismic facies of FA shows the high amplitude and parallel reflections with high continuity. In this unit, a dim spot that abruptly decreased the amplitude is observed between 140 and 153 mbsf (Figure 5). The sedimentary core recovered in the drilling site UBGH2-6 shows that FA is correlated to the interval in which the frequency and individual thickness of turbidite sand layers increase and to the high-resistivity response zone interpreted as a highly saturated gas hydrate-bearing zone [36]. The overlying units (FB, FC and FD) are characterized by relatively low-amplitude reflections with high continuity. These units are bounded by regionally well-stratified key horizons. These units are correlated to the interval in which the frequency and individual thickness of turbidite sand layers are decreased. Moreover, in the study area, two normal faults are observed at the southwestern part of site UBGH2-6 within 1 km in distance (Figure 5 and Figure 6). These faults in the vertical seismic profile cut the seismic units, including FA, FB, FC and FD, showing that these faults have been recently activated. The observed BSR in the seismic profile is cross-cutting these faults (Figure 5).

The geological model is built for the production simulation of gas hydrate in the Ulleung Basin. The model area is defined as a square of 5 km by 5 km in size centered on the site UBGH2-6, as shown in Figure 7. For the structural frame of this geological model, the present study uses the surfaces binding each seismic unit and two deep-seated faults (Figure 7). Zones within the model are consequently set as four zones, and the lowermost boundary is set as inferred BSR (Figure 7). In the gridding step, which laterally divides the modeling area, the lateral size (I and J direction) is defined as 100 m by 100 m, considering the bin size of the 3D seismic data and the lateral extension of deep turbidite, which are input data confining the horizontal resolution.

To reflect the geological information in the reservoir model, this study defines HBS and upscales the properties related with gas hydrate, such as gas hydrate saturation and porosity. Herein, the HBS is defined as a porous sedimentary layer including pore-filling gas hydrate. With regard to gas hydrate occurrence, the sedimentary succession highly concentrated with gas hydrate at site UBGH2-6 approximately ranges from 140 to 153 mbsf. This occurrence zone is correlated with the dimming zone in the FA unit. These individual HBS can be selected by using well log and core analysis. The cross-plot between the gas hydrate saturation and grain size shows linear correlation, and the gas hydrate-bearing porous layer has more than 18.8% in saturation [40]. On the basis of the cut-off value, we roughly identified HBS within vertical resolution of the calculated saturation. Compared with the sedimentary cores, we selected 14 HBS within the GHSZ. Thus, the reservoir model in this study is composed of 14 sand layers and 13 shale ones, as shown in Figure 8 [40]. The range of hydrate saturation is 38.8~86.2%. The overburden thickness is 140 m and the HBS thickness is 13 m.

3. Reservoir Simulation Approach

3.1. Reservoir Simulator

In the simulation of gas hydrate production, many conditions, such as multiple components, phase transition and thermal and geomechanical response, should be considered, unlike in general oil and gas production simulation. Accordingly, the numerical simulator is of great significance. There are various numerical simulators for gas hydrate simulation, for example, TOUGH + HYDRATE of LBNL, STOMP-HYD of PNNL, MH-21 of NIAIST, HydrateResSim of LBNL and STARS of CMG [41]. These models have confirmed their accuracy and suitability through comparative studies by many researchers [42,43,44]. STARS, which is a commercial reservoir simulator with a thermal and geomechanical model, can describe hydrate reservoirs. This simulator includes the Kim–Bishnoi kinetic parameters that can describe heat of dissociation and hydrate thermodynamic stability, which is the core mechanism for hydrate simulation [45]. In this study, STARS 2009 version is utilized as the simulator of gas hydrate simulation to consider the geomechanical response and reservoir model on a large scale.

3.2. Reservoir Model

The reservoir model is set up using the developed geological model. Figure 9 shows the reservoir model in this study. The initial model is discretized into 50 × 47 × 58 = 136,300 gridblocks. This number of gridblocks is not too large for the general black oil simulation. However, the grid size in the I and J direction is 100 m in the original system. In the large grid size system, hydrate dissociation for the production of gas hydrate is not easily acquired because it is hard to lower the grid pressure. Most hydrate reservoir simulation studies using cylindrical systems have adopted very small grid size systems, especially near the production well. Accordingly, this model is cropped to 10 × 10 × 58 in order to prevent the number of grids becoming too large after the refinement of the reservoir model, with an example shown by the red line rectangle in Figure 9. Then, all of the grids are refined in the I and J direction. In particular, the grids near the production wells are very small, less than 1 m, to consider hydrate dissociation. The number of total grids in this system is 106 × 106 × 58 = 651,688. Due to the generally low permeability of the system and the relative abundance of clays, the well spacing is restricted to 500 m [46]. Therefore, four vertical production wells are located with 500 m well spacing.

Table 1. Initial conditions and properties.

Parameter	Value
Overburden thickness (m)	140
Underburden thickness (m)	157
Layer thicknesses and porosities	As in Figure 8
Hydrate saturation in HBS	As in Figure 8
Initial pressure at top layer (MPa)	22.261
Initial temperature at top layer (°C)	0.482
Dry thermal conductivity (W/m/K)	1
Wet thermal conductivity (W/m/K)	1.45
Bottomhole pressure (MPa)	9
Intrinsic permeability (m2)	Overburden	Sand	Mud interlayer	Underburden
	2.00 × 10−18	1.78 × 10−13	2.00 × 10−16	2.00 × 10−19
Porosity (%)	67	45	67	67
GH saturation (%)	0	38.8~86.2	0	0
Bulk density (kg/m3)	2620	2650	2640	2660
Young’s modulus (MPa)	14	40 (at Sh = 0)	18	20
		1400 (at Sh = 1)
Poisson’s ratio	0.35	0.25	0.35	0.35
Cohesion (MPa)	0.030	0.035	0.030	0.040
Rock compressibility (1/Pa)	1.0 × 10−8

represents the input parameters used in this study. Sand porosity is 45% and mud porosity is 67%. The intrinsic permeability of sand is 1.78 × 10⁻¹³ and that of mud is 2.00 × 10⁻¹³. The thickness, porosities and hydrate saturation of the HBS are shown in Figure 8. In particular, we use different values of Young’s modulus according to the hydrate saturation. The reservoir model of this study uses experimental data, such as the relative permeability model and the dynamic permeability model, which is a function of the solid (hydrate), as illustrated in Figure 10 and Figure 11 [3,40]. The experimental results of relative permeability were validated with the results of X-ray CT (computerized tomography), and it showed good matching results [40]. In addition, even though the intrinsic permeability was different with each sample, the permeability reduction trends with increasing hydrate saturation were similar for all the samples [3].

3.3. Scenario of the Cyclic Depressurization Simulation

The cyclic depressurization simulation is composed of primary and secondary depressurization stages. These two stages are alternately repeated within the cycle until production time is finished. Each production stage has a bottomhole pressure and production time. In this simulation, we change the bottomhole pressure of the primary and secondary depressurization stage, while the production time of both stages is fixed. The bottomhole pressure of the primary depressurization stage is changed from 6 MPa to 12 MPa and that of the secondary depressurization stage is 12 MPa, 16 MPa and the shut-in case. The production time of the primary and secondary depressurization stages is 8 and 2 days, respectively. The base case assumes that the bottomhole pressure of the primary depressurization stage is 9 MPa, compared to 16 MPa for the secondary depressurization stage. The non-cyclic depressurization case is also represented in order to analyze an effect of the cyclic depressurization method, and we use the values of the base case. Production time is set at 100 days.

4. Numerical Simulation

4.1. Non-Cyclic Depressurization Case

The bottomhole pressures are varied (6, 9 and 12 MPa) in the non-cyclic depressurization case. The lower bottomhole pressure case causes a high gas production rate and cumulative gas production (Figure 12). The field gas production rate of the 6 and 9 MPa cases steadily maintains a certain level, excluding the 6 MPa case (Figure 12a). The field gas production rate of all cases consistently increases and that of the 12 MPa case is very low compared with other cases. Field cumulative gas production is between 2.71 × 10³ m³ (in the case of 12 MPa) and 9.69 × 10⁵ m³ (in the case of 6 MPa) (Figure 12b). The cumulative gas production of the 12 MPa case represents a very low value, near zero. Therefore, that case is excluded in the simulation study of the cyclic depressurization simulation.

Figure 13 represents the I-K section of vertical subsidence in the case of 9 MPa according to the production time (25, 50, 75 and 100 days) in the well No. 1 area. Overall, vertical subsidence is represented in the upper side of the HBS and overburden, while uplift is shown in the bottom of the HBS and underburden. Additionally, there are large values of the subsidence and uplift near the production well. In particular, the upper side of the HBS has the highest value of the vertical subsidence. The boundary of the subsidence and uplift is between the 22nd and 23rd layer of the HBS, which has 27 layers. In other words, subsidence has an enormous effect, more so than uplift. In addition, the subsidence value is −0.79 m and uplift value is 0.24 m at 100 days.

Figure 14 represents the vertical subsidence at the upper side of the HBS, which has the highest value of the vertical subsidence. In all cases, the amount of vertical displacement increases as production time goes by. The amount of vertical displacement records a sharp reduction and shows a convergence to a certain level in all cases. In the case of low bottomhole pressure, there is larger subsidence than other cases because the high gas production rate leads to the low pore pressure. The vertical displacement is from −0.30 m (in the case of 12 MPa) to −1.58 m (in the case of 6 MPa). For sustainable and stable gas hydrate production, the vertical subsidence problem should be solved.

4.2. Cyclic Depressurization Case: 9 MPa in the Primary Depressurization Stage

The production conditions during the secondary depressurization stage are changed, with 12 and 16 MPa of bottomhole pressure, including the shut-in case, while the bottomhole pressure of the primary depressurization stage is fixed at 9 MPa. As shown in Figure 15a, the field gas production rate of all the cases represents an overall increasing trend. However, there are very different gas production rates during the primary depressurizations stage, although those have the same production conditions (9 MPa). The gas production rates of the shut-in case and the 16 MPa case are larger than that of the non-cyclic case, but that of the 12 MPa case is smaller than that of the non-cyclic case. The reason is that gas saturation and permeability is reduced by the regeneration of the gas hydrate near the well. In the cases of shut-in and 20 MPa, field gas production rates during the secondary depressurization stage are nearly zero. On the other hand, the 12 MPa case shows a higher field gas production rate than other case. The field gas production rate is continuously increased, and it is 451 m³/day after 100 days. Each cumulative field gas production is between 4.41 × 10⁴ m³ and 5.48 × 10⁴ m³, while it is 5.48 × 10⁴ m³ in the non-cycle case, as shown in Figure 15b. In particular, there are no differences in the cumulative gas production with the shut-in case and non-cyclic case, although no gas is produced during the secondary depressurization stage in the case of shut-in. Additionally, the slight difference of that with the 16 MPa case and non-cyclic case is represented, and the difference is only 6.8%. The reason may be the geothermal heat supply from the overburden, underburden and interlayer of shale during the shut-in period. However, the cumulative gas production of the 12 MPa case is 4.41 × 10⁴ m³, and this case has a large difference with other cases because of the relatively low gas production during the primary depressurization stage.

The vertical subsidence results of well No. 1 at the upper side of the HBS are represented in Figure 16. The vertical subsidence in all cases continuously increases according to the time. Additionally, in all cases, the amount of vertical displacement increases during the primary depressurization stage, while it decreases during the secondary depressurization stage. The range of final vertical displacement is from −0.74 m to −0.71 m, while the shut-in case represents −0.79 m. Geomechanical stabilities in all cases are increased in both the primary and secondary depressurization stage. In particular, the stability of the secondary depressurization stage is more restored than that of the primary depressurization stage. The reason is that pore pressure is increased during the secondary depressurization stage because of the relatively low level of gas production. In addition, the vertical subsidence of the non-cyclic case has a strong geomechanical effect result for the same reason as in the previous sentence. The difference in vertical subsidence between the shut-in case and non-cyclic case is just 9.9%. This value is very high compared with the difference in the cumulative gas production. Accordingly, geomechanical stability increases significantly using the cyclic depressurization method, especially in the case of the shut-in.

4.3. Cyclic Depressurization Case: 6 MPa in the Primary Depressurization Stage

Production conditions during the secondary depressurization stage are changed, such as in the 9 MPa case (with the 12 MPa, 16 MPa and shut-in cases). Field gas production rates of this case study are more than those of the 9 MPa case during the primary and secondary depressurization stages because we utilize a low bottomhole pressure (6 MPa) condition (Figure 17a). In addition, the rates are continuously increased over 60 days and field gas production rates are more than 20,000 m³ at that time. In all cases, the field gas production rates during the primary depressurization stage are higher than those of the non-cyclic case. In the secondary depressurization stage, there is less gas production than in the primary depressurization stage, even in the case of 12 MPa. As shown in Figure 17b, cumulative gas production is between 9.23 × 10⁵ m³ (in the case of 12 MPa) and 9.52 × 10⁵ m³ (in the case of 6 MPa), while it is 9.69 × 10⁵ m³ in the case of shut-in. There are few differences in cumulative gas production between the non-cyclic case and other cases. In particular, the difference between the non-cyclic case and shut-in case is just 1.8%. From this study, it is found that the cyclic depressurization method has a strong advantage in terms of the production rate.

Figure 18 represents the vertical subsidence of well No. 1 at the upper side of the HBS. The amount of vertical displacement increases according to the production time. In this case, geomechanical stability also worsens according to the time, while it is restored during the secondary depressurization stage. Geomechanical stability is increased in both the primary and secondary depressurization stages using cyclic depressurization compared to the non-cyclic case. The results of the 16 MP case and shut-in case are very similar, as are the results of the field gas production rates. The range of final vertical displacement is from −1.44 m to −1.42 m, while the shut-in case represents −1.60 m. These values are higher than that of the 9 MPa case study because of the high field gas production rate. In particular, geomechanical stability is enhanced by 11.3%, while the difference in the cumulative field gas production rate is very small. It is shown that the cyclic depressurization method is a powerful production method for geomechanical stability without a decrease in the gas production rate.

5. Conclusions

In this study, a numerical simulation was conducted using a 3D geological model with a multi-well system in order to develop technology for commercial gas hydrate production. Additionally, the cyclic depressurization method was adopted in order to secure the geomechanical stability. A 3D geological model of the Ulleung Basin of the Korean East Sea was built using the seismic data, well logging data and experimental data. By employing the STARS program of CMG, which has a hydrate kinetic model and a geomechanics module, we obtained reliable and stable results. We conducted simulation studies with various production strategies over 100 days. By applying the cyclic depressurization method, the amounts of vertical subsidence were significantly reduced, while cumulative gas production was relatively less reduced or nearly maintained. In particular, in the 6 MPa case during the secondary depressurization stage, there were very few differences with the non-cyclic case. This study is thought to have contributed to commercial gas hydrate production with a geomechanical stability study conducted using a reservoir-scale model and a multi-well system.