Evaluation on the gas production potential of different lithological hydrate accumulations in marine environment

Energy 91 (2015) 782e798

Contents lists available at ScienceDirect

Energy journal homepage: www.elsevier.com/locate/energy

Evaluation on the gas production potential of different lithological hydrate accumulations in marine environment Li Huang a, b, Zheng Su a, Neng-You Wu a, c, * a

Key Laboratory of Gas Hydrate, Guangzhou Institute of Energy Conversion, Chinese Academy of Sciences, Guangzhou 510640, China University of Chinese Academy of Sciences, Beijing 100049, China c Qingdao Instiute of Marine Geology, China Geological Survey, Qingdao 266071, China b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 29 November 2014 Received in revised form 9 July 2015 Accepted 17 August 2015 Available online 22 September 2015

In this work, the marine hydrate deposits were classified into CHR (clay reservoir), SIHR (siltstone reservoir) and SHR (sandstone reservoir) according to the grain sizes of the sediments. Based on field measurements and proper estimations in Shenhu area, the gas production potentials of these different lithological hydrate reservoirs were numerically studied through Tough þ Hydrate. The simulation results reveal that SHR can provide the most desirable gas production potential with the following superior features: (a) a burst gas release at initial stage of the production; (b) the highest average gas release rate 1.7  103 ST (standard temperature) m3/d; (c) the highest total release gas volume 1.7  107 ST m3, (d) the highest gas-to-water volume ratio with an average value of 9.04. However, the evolution of the spatial distributions of the characteristic parameters indicates that the gas production in SHR has met some challengeable problems in both technology and environment aspects. On the other hand, compared with CHR, SIHR shows a higher gas release rate and cumulative volume but a worse gas-to-water volume ratio during the entire production period. In addition, the evolution of the salinity spatial distribution indicates an unsatisfactory impact on the environment in the later stage of production for SIHR. © 2015 Elsevier Ltd. All rights reserved.

Keywords: Hydrate Shenhu Grain size Different lithological deposits Gas production potential

1. Introduction Gas hydrate, as one of the most promising alternative energy sources in 21 century, is clathrate, ice-like compounds in which the gas molecules occupy the cage structure composed of hydrogenbonded water molecules [1e3]. In nature, the gas hydrate only occurs in sediments of slopes at the outer continental margins and beneath the arctic permafrost as these places feature the low temperature and high pressure that favor the formation of gas hydrate [4,5]. As a research hot spot, the marine gas hydrate is attracting more and more attention, not only because of the substantial gases it contains (the amount of methane in oceanic hydrates is approximately two orders of magnitude greater than that in permafrost [6]), but also because it may cause submarine geohazards and global climate change under improper operations [7]. However, the key utilization factor for this new energy resource lies in the technical and economic feasibility of gas production from

* Corresponding author. No.62, Fuzhou Rd, Qingdao, Shandong, China. Tel.: þ86 13922774239. E-mail address: [email protected] (N.-Y. Wu). http://dx.doi.org/10.1016/j.energy.2015.08.092 0360-5442/© 2015 Elsevier Ltd. All rights reserved.

the deposits. Nowadays, the prevailing production methods includes: (I) depressurization in which the reservoir pressure is decreased to below the hydration pressure under the given reservoir temperature [8,9]; (II) thermal stimulation in which the reservoir temperature is raised above the hydration temperature under the given reservoir pressure [10,11]; (III) inhibitor injection which shifts the pressure-temperature equilibrium conditions to make the hydrates unstable [12]; and (IV) the exchange of methane molecules in the hydrate structure for carbon dioxide molecules [13e15]. Among them, depressurization is accepted as the most promising method for gas production in Class 2 hydrate deposits [16,17]. Although the application of additional thermal stimulation can enhance the gas production, the effect is limited [11]. Significant hydrate deposits are confirmed to exist in Shenhu area, SCS (South China Sea), but the proliferation studies on the gas production potential mainly focus on the production designs, while few studies pay attention to the effects of the geological characteristics in the deposits [18e22]. In fact, preliminary studies have verified that the geological accumulation settings play an important role in the formation and the dissociation of the hydrate in a given area [23e25]. Moreover, Boswell et al. proposed that the most promising and accessible hydrate is the sand-dominated reservoir

L. Huang et al. / Energy 91 (2015) 782e798

Nomenclature

Vr

C CHR I K kQ kQRD kQRW Mw

X

P Qr RGW r, z rw rmax S SHR SIHR t T

specific heat (J/kg/K) clay hydrate reservoir production potential intrinsic permeability (m2) composite thermal conductivity (W/m/K) “dry” thermal conductivity (W/m/K) “wet” thermal conductivity (W/m/K) cumulative volume of water produced from the well (m3) pressure (Pa) volumetric rate of CH4 released from of dissociation (ST m3/s) gas-to-water ratio (ST m3 of gas per kg water) cylindrical coordinates (m) well radius (m) maximum radius of the simulation domain (m) phase saturation sandstone hydrate reservoir siltstone hydrate reservoir time (days) temperature ( C)

in the “Hydrate Energy Pyramid” theory [26,27]. However, they didn't provide any numerical evidence for the details. In this paper, the gas production potentials for different lithological hydrate reservoirs in Shenhu area were investigated by means of numerical simulations. The code we employed is Tough þ Hydrate, which is developed by the Lawrence Berkeley National Laboratory [28]. By using an equilibrium and kinetic hydrate formation and dissociation model, the code can simulate the non-isothermal gas release, phase behavior as well as fluid and heat flow under conditions typical for natural CH4 hydrate deposits in complex geological media [28]. Through the comparison of simulation results, the numerical evidences for the demonstration of “Hydrate Energy Pyramid” as well as the reasons for the selection of the exploitation targets in South China Sea can be provided. 2. Hydrates in research area 2.1. Depositional setting The South China Sea is bordered to the west by Vietnam, to the north by the southeastern mainland and Taiwan of China, to the east by the Philippines and to the south by Malaysia and Indonesia. Its northern slope consists of three oil and gas bearing basins: Southeast Hainan Basin, Pearl River Mouth Basin and Southwest Taiwan Basin (Fig. 1(a)) [29]. Shenhu area, as a most promising location where the hydrate forms, is tectonically located in the Pearl River Mouth Basin [30]. The Pearl River Mouth Basin is in the tectonic subsidence process since the middle of Miocene, which is conducive for the formation of gas hydrate along with the high sedimentation rate [31]. During the GMGS-1 drilling expedition in 2007, the hydrate samples were detected at SH2, SH3 and SH7 within 200 m below the seafloor at a water depth of up to 1500 m, as shown in Fig. 1(b) [32,33]. The hydrate-containing sediment layers were 10e25 m thick with a maximum saturation value of 48% [32]. The laboratory analysis of the core samples reveals that the types of hydrate bearing sediments are distinctive. For instance, the sediments in SH2 are

783

cumulative volume of CH4 released from dissociation (ST m3/s) mass fraction (kg/kg) discretization along the z-axis (m) grain density (kg/m3)

Dz rR

Greek symbols van Genuchten exponent porosity

l F

Subscripts and superscripts 0 denotes initial state A aqueous phase B HBL base cap capillary G gas phase i salinity irA irreducible aqueous phase irG irreducible gas n permeability reduction exponent nG gas permeability reduction exponent w well

predominantly silts with sand content less than 12% of the sample. The sediments in SH7 are coarser and dominated by sand [34]. The differences of the grain size distributions are clearly shown in Fig. 2 [34].

2.2. Classification of gas hydrate reservoirs In the petroleum industry, the reservoirs are normally divided into 4 classes according to the grain sizes of the sediments, namely the clay reservoir, siltstone reservoir, sandstone reservoir and conglomerate reservoir. Their respective harmonic mean grain sizes are shown in Table 1. As a result of long-distance transport and weathering, however, the sediments in marine are generally fine particles, this phenomena is also confirmed in Shenhu area [34]. Therefore, only three types of hydrate deposits are developed in that area, which are CHR (clay hydrate reservoir), SIHR (siltstone hydrate reservoir) and SHR (sandstone hydrate reservoir).

2.3. Permeabilities of the hydrate reservoirs 2.3.1. The estimations of permeability: type 1 On the basis of the closely graded ball filling theory, the sediments in a hydrate reservoir are simplified to be composed of uniform spherical particles. Then the permeability of the system can be estimated by the empirical KozenyeCarman equation [35]:

K1 ¼

kF3 tS2

(1)

where k is the Kozeny constant related to the shape of capillary cross-section, t is the ratio of the trace length of fluid flow to the length of rock, normally the ratio of the Kozeny constant to t is 0.23. And F represents the porosity of the reservoir, S is the specific surface whose datum is referenced from the rock volume. For the random accumulation of the sediment particles, there are [35].

784

L. Huang et al. / Energy 91 (2015) 782e798

Fig. 1. (a) Sketch map of the research area in South China Sea [30]; (b) The locations of the drilled sites in Shenhu [33].

Fig. 2. Grain size distributions in the hydrate-bearing sediments from SH2 and SH7 [34].

F ¼ 0:36; S ¼

1:896 3:792 ¼ r d

(2)

where r and d are the average radius and diameter of the sediment particles, respectively. Thus the KozenyeCarman equation can be described as: 2

K1 ¼

2 kF3 d ¼ 0:23  0:363  ¼ 7:463e4 d tS2 3:7922

(3)

2

here the permeability is proportional to d . 2.3.2. The estimations of permeability: type 2 As another derivation type of the KozenyeCarman equation, the permeability is given by (4) [36]:

Table 1 Classification of the hydrate reservoirs in terms of grain size. Reservoir types

Grain size (d)

Average (d)

Clay Siltstone Sandstone Conglomerate

<0.005 mm 0.005e0.05 mm 0.05e2 mm >2 mm

0.003 mm 0.009 mm 0.098 mm 2.5 mm

K2 ¼

F3 180ð1  FÞ2

2

d

(4)

where F is the porosity, d is the average grain size of the sediments. 2 In the equation the permeability K not only is proportional to d , but also has some hybrid relationship with the porosity. The two types of the estimations of permeability are totally distinctive in equation forms. By comparing the calculation results of the two equations in Table 2 (in type 2, the porosity is assumed to be 36%), we can see that the errors for the two permeabilities are both 15.21%, which can't be ignored and obviously vary with the value of the porosity of reservoir. Thus in the later simulation, we employed the Type 2 estimation method for the properties of the reservoirs in Shenhu area.

Table 2 The permeability estimations of the different lithological hydrate reservoirs. Hydrate deposit types

d (mm)

K1 (m2 )

K2 (m2 )

Errors (%)

CHR SIHR SHR

0.003 0.009 0.098

6:7167 e15 6:0450 e14 7:1675 e12

5.6953 e15 5.1258 e14 6.0775 e12

15.21 15.21 15.21

L. Huang et al. / Energy 91 (2015) 782e798

The geological system in our study references the early studies [11,20]. The HBL between a permeable overburden and underburden with the same thickness of 20 m is 40 m thick. This is considered and confirmed to be enough for thermal transfer and pressure transmission as well as fluid flow for a 30-year production. A very thin confining boundary is hypothetical both at the top and bottom of the geometry system to prevent the vertical mass and heat exchange, especially the yielded gaseous effusion and brine water invasion. While a vertical well of rw ¼ 0.1 m in analogy to the research of the Ulleung Basin is applied at r ¼ 0 m [37]. The production interval (10 m) is located in the middle of the HBL, enabling the upper and lower layer of the HBL self-sealing layers to remain the products in the HBL. The research domain along the r axis is within the range of 80 me80 m, but only the positive part is simulated in this work because of the symmetry of the system, as shown in Fig. 3. 3.2. System properties In the absence of field measurement data, some system properties are obtained from the published literature, listed in Table 3. The sediment porosity of the system is 38% [19]. And the corresponding permeabilities of the different lithological hydrate reservoirs can be calculated by using the second type of empirical KozenyeCarman equation (as discussed early), the results are shown in Table 4. As the drilling results and geochemical studies show that the content of CH4 in hydrate deposits in Shenhu area is above 96%, the gas component is deemed to be 100% CH4 in our

Parameters

Value

Hydrate zone thickness Water zone thickness (overburden & underburden) Initial pressure PB (at base of HBL) Initial temperature TB (at base of HBL) Gas composition Initial saturations in the HBL Water salinity Intrinsic permeability Kr ¼ Kz Grain density rR Porosity F Dry thermal conductivity kQRD Wet thermal conductivity kQRW

40 m 20 m

Composite thermal conductivity model Capillary pressure model

14.58 MPa 10.76  C 100% CH4 SH ¼ 0.3, SA ¼ 0.7 0.03 Refer to Table 4. 2750 kg/m3 0.38 1.0 W/(m K) 3.1 W/(m K)
1 
1 kQC ¼ kQRD þ SA2 þ SH2 kQRW  kQRD Þ þ FSl kQl =

3.1. Geometry system and well design

Table 3 Hydrate deposits properties and initial conditions in Shenhu area.

=

3. Numerical models and simulation approach

785

i1l h 1 Pcap ¼ P0 ðS* Þ =l  1 A sirA Þ s* ¼ ðsð1s irA Þ SirA 0.19 L 0.45 P0 0.1 MPa  n Relative permeability model A SirA krA ¼ S1S irA  n G SirG krG ¼ S1S irA OPM model 3.572 0.02 0.20

N SirG SirA

research. The initial saturation of the hydrate and the water salinity in hydrate accumulation are 30% and 0.03 respectively.

3.3. Domain discretization and simulation specifics As the significance of the vicinity of the well has been emphasized in the gas production, the hybrid discretization is employed in the simulation. The radial direction r is divided into 50 grids, of which the first three grids are 0.2 m followed by an exponential increase. The axial direction z is divided into 112 grids, of which the height of the HBL and the adjacent domain (i.e., the lower overburden and upper underburden) is 0.5 m. And a wider discretization of 1 m is conducted remote of the HBL along the z axis, and even wider segments with a height of 2 m are applied near the boundaries. Consequently, the total 50  112 ¼ 5600 gridblocks are generated in (r, z), of which 5390 are active (the remaining being boundary cells). The uppermost and lowermost boundaries are inactive gridblocks with constant temperature and pressure. Such nonuniform discretization is available beneficial for the accurate description of the spatial distribution of the geological properties and the production results.

Table 4 The corresponding permeabilities of the hydrate reservoirs in this study (the porosity is 38%). Hydrate deposit types Average grain size (mm) Intrinsic permeability K(m2 )

Fig. 3. The schematic of hydrate deposit simulated in this study [11].

CHR SIHR SHR

0.003 0.009 0.098

7:1374 e15 6:4236 e14 7:6164 e12

786

L. Huang et al. / Energy 91 (2015) 782e798

Fig. 4. Volumetric rates of gas (Qr) released from diverse lithological hydrate reservoirs.

3.4. Initial conditions and disturbance description The base of the HBL in our study is at an average depth of 1410 m under the oceanic surface, and the initial pressure at that depth can be computed as 14.58 MPa by assuming a hydrostatic distribution of the pressure in the sediments beneath the seafloor [37]. On the hydrate P-T equilibrium curve, the equilibrium hydration temperature of the base can be achieved. However, as for the effective and easy disturbance, the initial temperature TB is initialized as slightly lower than the equilibrium temperature, as shown in Table 3. Knowing that the local geothermal gradient is 0.047  C/m [21], we can determine the temperatures at the top and the bottom boundaries. Eventually the remaining temperature profile can be obtained by a short simulation. In order to destabilize the stable system, the depressurization of Pw ¼ 3 MPa is imposed in the perforated interval for the recovery of the released gas during the production process, as depicted in Fig. 3.

driving force formed in the coarse sediments, resulting in the severe hydrate dissociation in the domain. Regardless of the decreases, the Qr in the coarse SHR is remarkably faster than that in SIHR, exceeding the average Qr of 500 m3/d (ST (standard temperature)) in CHR within the initial 100 days. This shows a favorable performance in the coarse sediments with a high intrinsic permeability (according to the KeC equation). After 100 days, continuous declines of Qr are observed in both SHR and CHR, while the Qr in SIHR exhibits a slight increase until t ¼ 640 days. This is mainly because of the mild permeability increase resulted from the hydrate consumption in the SIHR production, which also occurs in CHR and SHR but leads to different phenomena. For CHR, the lower permeability would not be affected much because of the relatively weak hydrate reaction. While for SHR, the initial depletion of hydrate accounts for the majority of reservoir, thus the continuous dissociation is not strong enough to increase

4. Results and discussions 4.1. Evolution of the productions 4.1.1. Gas production Fig. 4 shows the volumetric rate evolution of the gas (Qr) released from the diverse lithological hydrate reservoirs. Because the pressure differential between the well and the HBL will decrease as the production progresses, the overall trends of these three rate curves show a decline behavior from the initiation. And this trend is consistent with the experimental and simulation results of the hydrate reservoir decomposition by constant-pressure depressurization in our lab [38]. The Qr in SHR shows a burst in the very inception of the production process and a subsequent gradual decrease afterwards, which also matches with the experiment of the gas production during the hydrate decomposition in sand matrix [39e41]. Nevertheless, the gas released rates in CHR and SIHR show different degrees of decrease during the entire period. This may be attributed to the rapid response to the large

Fig. 5. Cumulative volume of gas (Vr) released from diverse lithological hydrate reservoirs.

L. Huang et al. / Energy 91 (2015) 782e798

787

Fig. 6. Cumulative volume of water (Mw) produced from the well of diverse lithological hydrate reservoirs.

the permeability. During this period, the increasing Qr in SIHR is identical to the decreasing Qr in SHR with a value of 1.7  103 m3/ d (which is also the average value of Qr in SHR) at about t ¼ 550 days. After t ¼ 640 days, Qr of SIHR turns to decline till the end of production (30 years). At approximately t ¼ 1080 days, the gas release rate in SHR begins to decline sharply, at which time a slighter increase trend appears in the Qr of CHR due to the permeability increase. This also happens in SIHR at t ¼ 100 days. However, the performance emerges in the later stage of the production is even better than in the initial days in CHR, which shows

an opposite result to that in SIHR. At the end of the 30-year production, the Qr in these two reservoirs are 180 m3/d and 600 m3/ d respectively, remaining about 86.3% and 27.9% of hydrate respectively in the reservoirs. Although Qr in SHR exhibits a sharp decline at the end of the hydrate dissociation (at approximately t ¼ 1080 days), it produces the maximum cumulative gas volume (1.7  107 ST m3), as shown in Fig. 5. This is 5.7 times higher than the estimation of the gas production of annual 1.0  106 m3 on the basis of the experimental result in Zhou's research [39]. However, if the hydrate saturation

Fig. 7. Gas-to-water ratio (RGW) produced from the well of diverse lithological hydrate reservoirs.

788

L. Huang et al. / Energy 91 (2015) 782e798

discrepancy between Shenhu (38%) of this study and his study (11%) was taken into consideration, the estimation of the cumulative volume should be considerable. It is evident that the gas production potential in this reservoir is more desirable than that in CHR or SIHR because of the higher cumulative volume of gas (Vr) during the entire production period. Additionally, when comparing the production between CHR and SIHR, SIHR manifests more advantages, such as a consistent higher Qr and Vr, and less hydrate remaining in the reservoir. The different sediment type with distinctive grain size is the reason for the gas production performance variation, which can be regarded as different magnitude of the intrinsic permeability in these diverse lithological hydrate reservoirs. 4.1.2. Water production Fig. 6 shows the cumulative volume of water (Mw) produced from the well in these diverse lithological hydrate reservoirs. Continuing uptrend can be observed in the three curves, indicating continuous water production accompanying by the gas release over the duration of production. It is evident that the Mw of SHR turns to be exceeded by the Mw of SIHR at t ¼ 550 days, which is coincide with the change of Qr in these two reservoirs (as mentioned in Fig. 4). And since no hydrate dissociates in SHR from that time, the Mw in the latter stage of production just represents the cumulative volume of brine water which exists in situ at that location.

The gas production advantage in SIHR is eroded because of the consistently higher Mw during the entire production period, resulting in a total MW ¼ 7.55  106 m3 of water. Thus the gas-towater ratio RGW ¼ VP/MW should be additionally employed as a relative criterion while evaluating the gas production potential. 4.1.3. Gas-to-water ratio To a better result than expected, the conclusion that the gas production potential in SHR is most desirable among these three hydrate reservoir types is supported by the RGW pattern in Fig. 7. The RGW in SHR with an average value of 9.04 is higher than that in CHR or SIHR during the entire period prior to the cessation of the hydrate dissociation (at approximately t ¼ 1080 days), of which a local maximum RGW of approximately 45 is reached at t ¼ 150 days. Meanwhile, the RGW in CHR with an average value of 7.16 always exceeds that in SIHR with an average value of 4.64 throughout the entire production duration (30 years), and the difference value between the two curves grows larger gradually. This proves that the production potential of the CHR is more promising in the relative criterion, especially at the later stage of the production, which coincides with the higher Qr in the later stage in CHR (as discussed in Fig. 4). These observations both in absolute criterion (Qr, Vr, Mw) and relative criterion (RGW) all demonstrate the absolute superiority of SHR in depressurization production. Additionally, the comparison

Fig. 8. Spatial distribution evolution of hydrate saturation SH (unit: 100%) in CHR during the gas production.

L. Huang et al. / Energy 91 (2015) 782e798

between SIHR and CHR indicates that the gas production potential in SIHR is more favorable in the absolute criterion but undesirable in the relative criterion. Since the depressurization is accepted as the most promising method in the gas production from Class 2 hydrate deposits, it may come to a corollary that the sequence of the gas production potential in marine hydrate reservoirs is ISHR > ISIHR > ICHR in the absolute criterion, conversely ISHR > ICHR > ISIHR in the relative criterion, no matter which decomposition method is employed. However, it has not been fully studied yet. On the other hand, the hydrate exploitation concerns some relevant operational and environmental matters besides the gas and water production results. In order to obtain the most available drilling target, the natural properties should also be investigated in the alternatives.

4.2. Spatial distribution of characteristic parameters The spatial distributions of natural properties over the production duration can reveal the differences of some individual aspects: I) the evolution of hydrate saturation SH provides the effects of the depressurization on the hydrate destruction in these diverse

789

lithological hydrate deposits to differential extents, II) the evolution of gas saturation SG reveals the corresponding complexity of gas recovery, i.e., production difficulty, III) the evolutions of the salinity Xi and temperature T reveal the significant environmental impacts. Furthermore, all these spatial distributions evolutions of the characteristic parameters can better explain the gas and water production performances discussed previously. 4.2.1. CHR (Clay hydrate reservoirs) 4.2.1.1. Spatial distributions of SH and SG. Fig. 8 describes the spatial distribution evolution of hydrate saturation (SH) in CHR over time. The regions of r < 40 m are clearly investigated because of the limited extent of hydrate dissociation during the production period. In the initial 2 years of simulation, the hydrate-free zone is only near the well and the dissociation interface represents a good symmetry at the axis of z ¼ - 40 m (i.e., location of the center of the well), as shown in Fig. 8(aec). Moreover, the dissociation front is receding to the internal HBL in this period. As time advances, the symmetrical dissociation interface is destructed, reaching to a more intense reaction interface at the top of HBL where depressurization is employed. This is distinctively different from what always happen in the depressurization-induced hydrate

Fig. 9. Spatial distribution evolution of gas saturation SG (unit: 100%) in CHR during the gas production.

790

L. Huang et al. / Energy 91 (2015) 782e798

dissociation [35], which may result from the configuration of the well design. At the same time, some secondary hydrate obviously emerges near the upper dissociation interface in Fig. 8(d), resulting from the gas flow from the dissociation front to the production well when encountering the available conditions (especially the temperature condition, which can be demonstrated in its spatial distribution in Fig. 10). As production progresses, the position of the emerged secondary hydrate gradually shifted downward to the lower hydrate interface, reaching a maximum amount at the bottom of the HBL at the end of the period (30 years). It is because the existing secondary hydrate formed before builds a barrier which impedes the gas flow, resulting in the accumulation of the released gas in the reservoir and enhancing the formation of the secondary hydrate. The SG distributions at the same time points of Fig. 8(aec) in Fig. 9 indicate that the released gas accumulates in the hydrate consumption regions circled by the dissociation interface and rises towards the well during the production period. Additionally, it is clear that the released gas is also accumulated at the bottom of the HBL at a low level of approximately SG ¼ 0.03. As time progresses, the SG distribution pattern evolves into a tadpole-like shape. This is because of the continuous dissociation along the lower hydrate interface plus the gas flow obstruction imposed by the formed

secondary hydrate barrier, which keeps the released gas at that location. The high SG near the lower hydrate interface in Fig. 9(eef) makes the formation of the secondary hydrate possible, which is consistent with the SH distribution in Fig. 8. As time advances, the length of the “tadpole tail” becomes shorter and shorter (Fig. 9(eei)), companying with the declining dissociation rate of the hydrate in CHR during the production period (Fig. 4). 4.2.1.2. Spatial distribution of T. Fig. 10 represents the temperature spatial distribution evolution in CHR over time. During the first 1 year of production, the lower T can be discerned in the zone where the hydrate is exhausted, declining towards the dissociation front due to the endothermic hydrate dissociation reaction. But thereafter, the lower T in this area begins to increase which may be attributed to the heat supply from the surroundings. It is noted that this temperature increase would also account for the loci of the formation of the secondary hydrate in Fig. 8. What's more, the intense dissociation reactions at the top of the HBL (described in Fig. 8) result in the continuous T decline there, as shown in Fig. 10(cef). For the continuous gas recovery, the heat will be brought from the reservoir to the well, which will cause T increase in the region between the secondary hydrate obstruction and the lower HBL interface (Fig. 10(fei)). This would also explain the

Fig. 10. Spatial distribution evolution of temperature T (unit:  C) in CHR during the gas production.

L. Huang et al. / Energy 91 (2015) 782e798

reason why the secondary hydrate only occurs along the dissociation front with the possible low temperature there.

4.2.1.3. Spatial distribution of Xi. As a significant effect of the hydrate decomposition on the environment, the spatial distribution evolution of the salt concentration in the aqueous phase Xi is investigated in Fig. 11. The intense dissociation activities would result in mass fresh water release and the decline of water salinity, which indicates that the salinity change can be a symbol of the extent of dissociation reactions. According to the results in Fig. 8(aeb), the low Xi should appear in the column area near the well, which indeed occurs in the Xi spatial distributions in Fig. 11(aeb). As time advances, the pronounced activities emerged at the upper and lower hydrate interface result in the gradual decline of Xi at these two regions, as shown in Fig. 11(cef). As a consequence, a high salinity belt is remarkably identified at the elevation of the central of the well. Actually, this phenomenon is a combined effect of the intense dissociation reactions and the gravity of water during the drainage. The maximum Xi occurs at the tadpole-tail, which is caused by the water consumption during the formation of the secondary hydrate, as depicted in Fig. 11(feh). And as the continuous drainage during the production, the maximum Xi fades in the end finally.

791

4.2.2. SIHR (Siltstone hydrate reservoirs) 4.2.2.1. Spatial distributions of SH and SG. Figs. 12 and 13 show the evolutions of the SH and SG spatial distributions in SIHR during the production process. Unlike the patterns in Fig. 8, the hydrate indicates a more pronounced disturbance near the top of the HBL in addition to the vicinity of the well at the initial stage. This is because that the same depressurization operation in coarse sediments of the hydrate reservoir leads to a more desirable pressure drop, contributing to the stronger hydrate destruction in the patterns. Notwithstanding these figures of the same certain times in Figs. 8 and 12 are totally different, it is also evident that the pattern in Fig. 12(d) is much the same as that in Fig. 8(i). It seems that the hydrate activities in SIHR are the delayed appearances of that in CHR. Similar to and even more obvious than that in Fig. 8, the secondary hydrate of SIHR firstly occurs upon the dissociation front at t ¼ 1800 days but gradually declines as the time advances. Obviously, the more intense dissociation in SIHR results in less hydrate remaining at the end of the production period when compared with that in CHR. The SG spatial distributions in Fig. 13 indicate that the gas accumulation zone at early stage (Fig. 13(aed)) is exactly the same as the shortening “tadpole tail” pattern in the later production stage in CHR. However, it clearly shows that the gas accumulation

Fig. 11. Spatial distribution evolution of salinity Xi (unit: 100%) in CHR during the gas production.

792

L. Huang et al. / Energy 91 (2015) 782e798

area narrows to be quite limited since t ¼ 3600 days in Fig. 13(e), which maintains till the end of the production. It implies that the released gas is produced from the well instantly rather than accumulating in the system. This is attributed to the intense hydrate dissociation (demonstrated in Fig. 11) and the better recovery performance from the hydrate reservoir with a desirable permeability in the coarse sediments, with less gas accumulation remaining in the deposit. 4.2.2.2. Spatial distribution of T. Fig. 14 shows the temperature spatial distribution of the SIHR during the production. The decreasing T is remarkable along the upper hydrate formation interface because of the intense dissociation reaction at the initial stage of production, as shown in Fig. 14(aee). While, as the gas continuously flow to the well for recovery, the temperature immediate vicinity of the well rises from the beginning (Fig. 14(aei)). This phenomenon occurs at the lower hydrate interface as well. The whole domain shows a more pronounced cooling than in CHR (Fig. 10). 4.2.2.3. Spatial distribution of Xi. The corresponding Xi distribution evolution in SIHR during the production period is described in Fig. 15. The lower Xi close to the top of the HBL in Fig. 15(aec) can be explained as the dilution of the massive water released from the

intense dissociation, which is similar to the results in Fig. 11(def). This further confirms the “delayed appearances” in SH distribution evolution mentioned before. However, as the production advances, the Xi beneath the overburden gradually evolves to a low level due to the gravity of the water and the high permeability of the underburden (Fig. 15(eei)). It is noted that Xi in the underburden remains in the low level ultimately, indicating that the depressurization in SIHR in marine would have an unsatisfactory effect on the environment in the later stage of production. 4.2.3. SHR (Sandstone hydrate reservoirs) 4.2.3.1. Spatial distribution of SH and SG. According to the SH spatial distribution evolution of SHR, the interruption of gas release in SHR at about t ¼ 1080 days depicted in Fig. 4 can be explained by the exhaustion of hydrate in Fig. 16(d). And this is why our investigation of the reservoir properties evolutions is about the whole deposit within 3 years. The high SH region below the overburden base is clearly identified in Fig. 16(a), indicating more pronounced reaction at the HBL base. This is quite different to that in CHR or SIHR which shows more intense hydrate dissociation at the upper hydrate interface. The reason for the difference is that the good connectivity in the coarse sediments of high intrinsic permeability allows more desirable pressure drop in the deeper depth where there are the higher pressure conditions. Thus more intense hydrate reactions

Fig. 12. Spatial distribution evolution of hydrate saturation SH (unit: 100%) in SIHR during the gas production.

L. Huang et al. / Energy 91 (2015) 782e798

takes place there. However, it is noted that even the highest SH is only 6%, which validates the observation that the majority of hydrate has been already dissociated in the 180 days’ production. As time progresses, the dissociation interface recedes from the well. It is especially unusual that the SH at the HBL base increases at t ¼ 720 days in Fig. 16(c), which indicates the formation of secondary hydrate at the base. Thereafter, the secondary hydrate continues to be dissociated, and is exhausted at t ¼ 3 years. Due to the rapid dissociation of hydrate in SHR, vast released gas accumulates in the deposit, as shown in the SG spatial distribution in Fig. 17. It is noted that the released gas distributes lamellarly and rises toward the upper boundary in the production process, which is never happened in CHR or SIHR. This is because of the high permeability of the overburden that allows the gas to flow upward by the buoyancy. For this reason, the significance of the existence of near-impermeable boundary is accentuated. In the absence of a confining layer, the released gas would escape to the surface, making the gas production of SHR more challenging for the technology and environmental protection. 4.2.3.2. Spatial distribution of T. The T spatial distributions in SHR throughout the production process are shown in Fig. 18. As expected, the temperature within the dissociation zone declines because of the endothermal reaction of the hydrate decomposition.

793

This phenomenon is in good agreement with the observations of temperature variation in Zhou and Haligva's experiments of the hydrate decomposition via depressurization in the quartz sand matrix [39,42]. It is obvious that the lower T emerges in the middle of the HBL in the deposit, which is exactly distinctive from that in CHR or SIHR. As time advances, the area of lower temperature narrows progressively companying the hydrate dissociation, which is analogous to the later stage of the production in SIHR (Fig. 14(eei)). Although the hydrate is exhausted before t ¼ 3 years, the pattern in Fig. 18 (d) indicates that the thermal recovery of the reservoir has not completed or has not even started. 4.2.3.3. Spatial distribution of Xi. Consistent with the SH distribution patterns, the lower Xi appears at the bottom of the HBL due to the intense hydrate dissociation, as shown in Fig. 19. It coincides with the results of the SG and T distribution evolutions of SHR as well. However, even in the more coarse-grained sediments, the uncontrollable environment impact in SIHR does not occur in the Xi spatial distribution in SHR. This may be attributed to the fact that when compared with SIHR, the permeability in SHR reaches a higher level that, rather than accumulating in the deposit by gravity, is more conducive to the water drainage. For this advantage, the production in SHR should be taken into possible consideration.

Fig. 13. Spatial distribution evolution of gas saturation SG (unit: 100%) in SIHR during the gas production.

794

L. Huang et al. / Energy 91 (2015) 782e798

4.3. Comparison with the world's first offshore hydrate production test in Japan The world's first and also the only marine hydrate production test to date was conducted in SHR of Daini Atsumi Knoll, northern slope of Nankai Trough, Japan on March 12 in 2013 [43,44]. This test was terminated on the sixth day due to the large amount of sand production, resulting in a cumulative volume of produced gas equals to 119,500 m3 [45]. Based on the integrated geology and geophysics results from the extensive geophysical logging and pressure coring at the target site AT1 [44,46e48], the gas production performance was estimated through depressurization at a constant pressure 3 MPa by a vertical well, which is the field operations in the production test [46]. The simulated results were compared to the field data obtained during the hydrate production test, shown in Fig. 20. It can be found that the volumetric rate of the produced gas QP in field observation is very close to that in the numerical simulation. And this greatly proves the validation of numerical code employed in this study. But for the volumetric rate of the produced water QW, the field production is much lower than numerical simulation. As a matter of fact, the average QP and QW in the field observation are 20,000 ST m3/d and 200 m3/d, respectively, which leads to a mole ratio of the produced water to methane of nearly 6. And it is of interest to find that this mole ratio

just equals to the hydrate decomposition performance (in the methane hydrate system, the hydrate decomposition follows the equation CH4$6H2O ¼ CH4 þ 6H2O [49]), which means that no insitu interstitial water was produced in the field test. However, this is unlikely to occur in the practical productions and it could be attributed to some separation operation before the products were transferred to the surface. Fig. 20 also shows the comparison of the gas production performance in SHR between our work and field test on Daini Atsumi Knoll. As can be observed in the figure that QP and QW of the field test are both lower than that in SHR of Shenhu owing to the variance of the geologic features of the hydrate reservoirs. For instance, the hydrate concentrated interval in the field production test is nearly 60 m while the thickness of the HBL in our work is 40 m, the porosity of the hydrate-bearing sediments in field production test is estimated as 40e60% according to the core samples while in Shenhu it is 38% [47,48]. Other than these, knowledge of the related geologic features and the production performance of these two hydrate reservoirs is listed in Table 5 [45,47]. As both in the sandstone hydrate reservoirs, the average QP of these reservoirs over the production duration is of the same magnitude, which are 2.0  104 ST m3/d and 8  104 ST m3/d, respectively. Moreover, the cumulative volume of 120,000 ST m3 can be comparable to the cumulative volume of 518,000 ST m3 in SHR when considering the

Fig. 14. Spatial distribution evolution of temperature T (unit:  C) in SIHR during the gas production.

L. Huang et al. / Energy 91 (2015) 782e798

Fig. 15. Spatial distribution evolution of salinity Xi (unit: 100%) in SIHR during the gas production.

Fig. 16. Evolution of hydrate saturation SH (unit: 100%) spatial distribution in SHR during the gas production.

795

796

L. Huang et al. / Energy 91 (2015) 782e798

Fig. 17. Evolution of gas saturation SG (unit: 100%) spatial distribution in SHR during the gas production.

Fig. 18. Evolution of temperature T (unit:  C) spatial distribution in SHR during the gas production.

discrepancy of the geologic features between these two field hydrate reservoirs. Thus, the evaluation results can be regarded as reasonable and considerable to guide the field production in the marine hydrate accumulations in the future. 5. Conclusions In this work, we numerically investigated the gas production potential from the different lithological hydrate reservoirs most likely to occur in Shenhu area, South China Sea by depressurization through a vertical well. These are CHR, SIHR and SHR, and their permeability is 7.137 e15 m2, 6.4236 e14 m2 and 7.6164 e12 m2

respectively according to the KozenyeCarman empirical equation, in which the porosity is 38%. Based on the research results, the following conclusions can be drawn: (1) Among the different lithological hydrate reservoirs, SHR shows the most desirable gas production potential with the highest average Qr (1.7  103 ST m3/d), the highest total Vr (1.7  107 ST m3) and the highest average RGW (9.04) over a production duration of 3 years. (2) During the production process, continuous decline of Qr and RGW and increase of Vr and Mw can be observed in both SIHR and CHR. The SIHR shows a better gas production

L. Huang et al. / Energy 91 (2015) 782e798

797

Fig. 19. Evolution of salinity Xi (unit: 100%) spatial distribution in SHR during the gas production.

Fig. 20. Comparison of volumetric rate of the produced gas and water of SHR between Shenhu and the field data in the first production test on Daini Atsumi Knoll.

Table 5 Comparison of hydrate reservoirs and the production performance between SHR in Shenhu and field offshore test on Daini Atsumi Knoll in the duration of the six days production. Hydrate deposit features

SHR in Shenhu

Field offshore production test

Hydrate zone thickness Initial hydrate saturation Porosity Initial pressure of HBL Initial pressure of HBL Average of Qp Cumulative volume of produced gas

40 m 30% 38% 14.58 MPa 10.76  C 8  104 ST m3/d 518,000 ST m3

60 m 50e80% 40e60% 13.63 MPa 14.25  C 2.0  104 ST m3/d 120,000 ST m3

performance than CHR does, accompanying with a worse gas-to-water ratio which is 7.16 on average. (3) We may conclude that the sequence of the gas production potential in marine hydrate reservoirs is ISHR > ISIHR > ICHR in the absolute criterion, conversely ISHR > ICHR > ISIHR in the relative criterion. However, no matter which decomposition method is employed, further study remains necessary for this issue. (4) The spatial distribution evolution of the characteristic parameters indicates that the hydrate dissociation occurs more intensively at the upper hydrate interface of both CHR and SIHR. To some extent, the gas production of CHR is a delayed pattern of the SIHR, which shows an unsatisfactory effect on the environment in the later stage of production. (5) The severe dissociation activities in SHR appear at the lower hydrate interface. And the highlighted existence of near-

798

L. Huang et al. / Energy 91 (2015) 782e798

impermeable boundary brings its production more technological and environmental problems. Acknowledgments This research is supported by the National Natural Science Foundation of China (No.41276057), Key Arrangement Programs of the Chinese Academy of Sciences (KGZD-EW-301) and China Geological Survey (No.GZH2012006003). The authors would like to acknowledge the support with sincere gratitude. References [1] Sloan Jr ED, Koh C. Clathrate hydrates of natural gases. 3th ed. CRC Press; 2008. [2] Makogon YF, Holditch SA, Makogon TY. Natural gas-hydrates d a potential energy source for the 21st century. J Petroleum Sci Eng 2007;56(1):14e31. [3] Makogon YF. Natural gas hydrates d a promising source of energy. J Nat Gas Sci Eng 2010;2(1):49e59. [4] Kvenvolden KA. Gas hydrates d geological perspective and global change. Rev Geophys 1993;31(2):173e87. [5] Lee SY, Holder GD. Methane hydrates potential as a future energy source. Fuel Process Technol 2001;71(1):181e6. [6] Moridis GJ, Sloan ED. Gas production potential of disperse low-saturation hydrate accumulations in oceanic sediments. Energy Convers Manag 2007;48(6):1834e49. [7] Kvenvolden KA. Potential effects of gas hydrate on human welfare. Proc Natl Acad Sci 1999;96(7):3420e6. [8] Zhao JF, Yu T, Song Y, Liu D, Liu WG, Liu Y, et al. Numerical simulation of gas production from hydrate deposits using a single vertical well by depressurization in the Qilian Mountain permafrost, Qinghai-Tibet Plateau, China. Energy 2013;52:308e19. [9] Jiang XX, Li SX, Zhang LN. Sensitivity analysis of gas production from class I hydrate reservoir by depressurization. Energy 2012;39(1):281e5. [10] Kamath VA, Godbole SP. Evaluation of hot-brine stimulation technique for gas production from natural gas hydrates. J Petroleum Technol 1987;39(11): 1379e88. [11] Su Z, Huang L, Wu NY, Yang SX. Effect of thermal stimulation on gas production from hydrate deposits in Shenhu area of the South China Sea. Sci China Earth Sci 2013;56(4):601e10. [12] Li G, Li XS, Tang LG, Zhang Y. Experimental investigation of production behavior of methane hydrate under ethylene glycol injection in unconsolidated sediment. Energy & Fuels 2007;21(6):3388e93. [13] Yuan Q, Sun CY, Yang X, Ma PC, Ma ZW, Liu B, et al. Recovery of methane from hydrate reservoir with gaseous carbon dioxide using a three-dimensional middle-size reactor. Energy 2012;40(1):47e58. [14] Marcelle-De Silva J, Dawe R. Towards commercial gas production from hydrate deposits. Energies 2011;4(12):215e38. [15] Xu CG, Cai J, Lin FH, Chen ZY, Li XS. Raman analysis on methane production from natural gas hydrate by carbon dioxide-methane replacement. Energy 2015;79:111e6. [16] Moridis GJ, Kowalsky M. Gas production from unconfined class 2 oceanic hydrate accumulations. In: Economic geology of natural gas hydrate. Netherlands: Springer; 2006. p. 249e66. [17] Lee JY, Ryu BJ, Yun TS, Lee J, Cho GC. Review on the gas hydrate development and production as a new energy resource. KSCE J Civ Eng 2011;15(4):689e96. [18] Li G, Moridis GJ, Zhang KN, Li XS. The use of huff and puff method in a single horizontal well in gas production from marine gas hydrate deposits in the Shenhu Area of South China Sea. J Petroleum Sci Eng 2011;77(1):49e68. [19] Su Z, Cao YC, Wu NY, He Y. Numerical analysis on gas production efficiency from hydrate deposits by thermal stimulation: application to the Shenhu Area, South China Sea. Energies 2011;4(12):294e313. [20] Su Z, Moridis GJ, Zhang KN, Wu NY. A huff-and-puff production of gas hydrate deposits in Shenhu area of South China Sea through a vertical well. J Petroleum Sci Eng 2012;86:54e61. [21] Su Z, He Y, Wu NY, Zhang K, Moridis GJ. Evaluation on gas production potential from laminar hydrate deposits in Shenhu Area of South China Sea through depressurization using vertical wells. J Petroleum Sci Eng 2012;86: 87e98. [22] Li G, Moridis GJ, Zhang KN, Li XS. Evaluation of gas production potential from marine gas hydrate deposits in Shenhu Area of South China Sea. Energy & Fuels 2010;24(11):6018e33. [23] Seol J, Lee H. Natural gas hydrate as a potential energy resource: from occurrence to production. Korean J Chem Eng 2013;30(4):771e86. [24] Wu NY, Huang L, Su Z, Yang SX, Wang HB, Liang JQ, et al. A study of geological evaluation indicators for the exploitation potential of marine natural gas hydrates: theory and methodology. Nat Gas Ind 2013;33(7):11e7.

[25] Bhade P, Phirani J. Gas production from layered methane hydrate reservoirs. Energy 2015;82:686e96. [26] Boswell R, Collett TS. The gas hydrates resource pyramid. Fire in the ice e methane hydrate newsletter. National Energy Technology Laboratory, U.S Department of Energy; 2006. [27] Boswell R, Collett TS. Current perspectives on gas hydrate resources. Energy & Environ Sci 2011;4(4):1206e15. [28] Moridis GJ, Kowalsky MB, Pruess K. TOUGHþ HYDRATE v1. 0 user's manual: a code for the simulation of system behavior in hydrate-bearing geologic media. Berkeley, CA, U.S.A: Lawrence Berkeley National Laboratory; March, 2008. [29] Wu NY, Zhang GX, Liang JQ, Su Z, Wu DD, Lu HL, et al. Progress of gas hydrate research in northern South China Sea. Adv New Renew Energy 2013;1:80e94. [30] Wu NY, Zhang HQ, Yang SX, Zhang GX, Liang JQ, Lu JA, et al. Gas hydrate system of Shenhu area, Northern South China Sea: geochemical results. J Geol Res 2011;2011:1e10. [31] Yang SX, Zhang HQ, Wu NY, Su X, Schultheiss P, Holland M, et al. High concentration hydrate in disseminated forms obtained in Shenhu area, North Slope of South China Sea. In: Proceeding of the 6th International Conference on Gas Hydrates. British Columbia, Canada: Vancouver; July 6-10, 2008. [32] Zhang HQ, Yang SX, Wu NY, Su X, Holland M, Schultheiss P, et al. Successful and surprising results for China's first gas hydrate drilling expedition. Fire in the ice e methane hydrate newsletter. National Energy Technology Laboratory, U.S Department of Energy; 2007. [33] Wu SG, Dong DD, Yang SX, Zhang GX, Wang ZJ, Li QP, et al. Genetic model of the hydrate system in the fine grain sediments in the northern continental slope of South China Sea. Chin J Geophys 2009;52(7):1849e57 [Chinese Edition]. [34] Liu CL, Ye YG, Meng QG, He XL, Lu HL, Zhang J, et al. The characteristics of gas hydrates recovered from Shenhu area in the South China Sea. Mar Geol 2012;307:22e7. [35] Kong X. Higher seepage mechanics. Hefei. China University of Science and Technology Press; 1999. [36] Bear J. Dynamics of fluids in porous media. New York: Eisevier; 1972. [37] Moridis GJ, Reagan MT, Kim SJ, Seol Y, Zhang K. Evaluation of the gas production potential of marine hydrate deposits in the Ulleung Basin of the Korean East Sea. SPE J 2009:759e81. [38] Li G, Li B, Li XS, Zhang Y, Wang Y. Experimental and numerical studies on gas production from methane hydrate in porous media by depressurization in pilot-scale hydrate simulator. Energy & Fuels 2012;26:6300e10. [39] Zhou Y, Castaldi MJ, Yegulalp TM. Experimental investigation of methane gas production from methane hydrate. Industrial Eng Chem Res 2009;48(6): 3142e9. [40] Li XS, Yang B, Zhang Y, Li G, Duan LP, Wang Y, et al. Experimental investigation into gas production from methane hydrate in sediment by depressurization in a novel pilot-scale hydrate simulator. Appl Energy 2012;93:722e32. [41] Zhao JF, Zhu ZH, Song YC, Liu WG, Zhang Y, Wang DY. Analyzing the process of gas production for natural gas hydrate using depressurization. Appl Energy 2015;142:125e34. [42] Haligva C, Linga P, Ripmeester JA, Englezos P. Recovery of methane from a variable-volume bed of silica sand/hydrate by depressurization. Energy & Fules 2010;24:2947e55. [43] Japan completes first offshore methane hydrate production test e methane successfully produced from deepwater hydrate layers Fire in the ice e methane hydrate newsletter. National Energy Technology Laboratory, U.S Department of Energy; 2013. 13 (2). [44] Fujii TM, Nakamizu M, Tsuji Y, Namikawa T, Okui T, Kawasaki M, et al. Methane-hydrate occurrence and saturation confirmed from core samples, eastern Nankai Trough, Japan. In: Collett T, Johnson A, Knapp C, Boswell R, editors. Natural gas hydrates e energy resource potential and associated geologic hazards; 2009. p. 385e400. AAPG Memoir 89. [45] Matsuzawa M, Terao Y, Hay B, Wingstrom L, Duncan M, Ayling L. A completion system application for the world's first marine hydrate production test. In: Proceeding of offshore technology conference; May 5-8, 2014. Houston, Texas, USA. [46] Yamamoto K, Terao Y, Fujii T, Ikawa T, Seki M, Matsuzawa M, et al. Operational overview of the first offshore production test of methane hydrates in the eastern Nankai Trough. In: Proceeding of offshore technology conference; May 5-8, 2014. Houston, Texas, USA. [47] Fujii T, Suzuki K, Takayama T, Tamaki M, Komatsu Y, Konno Y, et al. Geological setting and characterization of a methane hydrate reservoir distributed at the first offshore production test site on the Daini-Atsumi Knoll in the eastern Nankai Trough, Japan. Mar Petroleum Geol 2015. http://dx.doi.org/10.1016/ j.marpetgeo.2015.02.037. [48] Suzuki K, Schultheiss P, Nakatsuka Y, Ito T, Egawa K, Holland M, et al. Physical properties and sedimentological features of hydrate-bearing samples recovered from the first gas hydrate production test site on Daini-Atsumi Knoll around eastern Nankai Trough. Mar Petroleum Geol 2015. http://dx.doi.org/ 10.1016/j.marpetgeo.2015.02.037. [49] Moridis GJ, Reagan MT, Boyle KL, Zhang KN. Evaluation of the gas production potential of some particularly challenging types of oceanic hydrate deposits. Transp Porous Media 2011;90:269e99.