Processing math: 100%
MENU

We use cookies to improve your experience with our site.

2023 lmpact Factor: 4.6 (Q1)
2023 CiteScore:7.8 (Q1)
Articles in press have been peer-reviewed and accepted, which are not yet edited and assigned to volumes/issues, but are citable by Digital Object Identifier (DOI).
Ji Yun-kai, Liu Chang-ling, Zhang Zhun, Meng Qing-guo, Liu Le-le, Zhang Yong-chao, Wu Neng-you. 2022. Experimental study on characteristics of pore water conversion during methane hydrates formation in unsaturated sand. China Geology, 5(2), 276‒284. doi: 10.31035/cg2022013.
Citation: Ji Yun-kai, Liu Chang-ling, Zhang Zhun, Meng Qing-guo, Liu Le-le, Zhang Yong-chao, Wu Neng-you. 2022. Experimental study on characteristics of pore water conversion during methane hydrates formation in unsaturated sand. China Geology, 5(2), 276‒284. doi: 10.31035/cg2022013.

Experimental study on characteristics of pore water conversion during methane hydrates formation in unsaturated sand

  • Corresponding author: Chang-ling Liu, qdliuchangling@163.com
  • Received Date: 12 October 2021
  • Accepted Date: 09 March 2022
  • Available Online: 29 March 2022
  • Understanding the pore water conversion characteristics during hydrate formation in porous media is important to study the accumulation mechanism of marine gas hydrate. In this study, low-field NMR was used to study the pore water conversion characteristics during methane hydrate formation in unsaturated sand samples. Results show that the signal intensity of T2 distribution isn’t affected by sediment type and pore pressure, but is affected by temperature. The increase in the pressure of hydrogen-containing gas can cause the increase in the signal intensity of T2 distribution. The heterogeneity of pore structure is aggravated due to the hydrate formation in porous media. The water conversion rate fluctuates during the hydrate formation. The sand size affects the water conversion ratio and rate by affecting the specific surface of sand in unsaturated porous media. For the fine sand sample, the large specific surface causes a large gas-water contact area resulting in a higher water conversion rate, but causes a large water-sand contact area resulting in a low water conversion ratio (Cw=96.2%). The clay can reduce the water conversion rate and ratio, especially montmorillonite (Cw=95.8%). The crystal layer of montmorillonite affects the pore water conversion characteristics by hindering the conversion of interlayer water.
  • Natural gas hydrates (NGHs) are solid crystalline substances formed by gas molecules and water molecules under low temperatures and high pressures (Sloan E and Koh C, 2008). The main component of gas molecules in NGHs is methane with a proportion up to about 90% (Zhan L et al., 2018). NGH is a kind of clean resources with great application prospect due to its wide distribution, high energy intensity and high inventory. The resource potential of NGHs is huge in China, especially in the clayey-silt reservoir in the South China Sea (Li JF et al., 2018). In 2020, China completed the second NGHs production test in the South China Sea, setting world records for daily gas production and total gas production (Ye JL et al., 2020). It effectively promoted the industrialization of NGHs. However, in order to realize the industrialized exploitation of NGH, it is urgent to accurately evaluate the hydrate reservoir resources and build the industrialized hydrate reserve bases.

    Hydrate saturation is the ratio of the hydrate volume to the pore volume in the hydrate-bearing sediments. It is an important parameter to evaluate the hydrate reservoir resources. And it affects the physical properties of hydrate reservoirs such as acoustic wave velocity (Bu QT et al., 2017), resistivity (Priegnitz M et al., 2015; Zhao EM et al., 2021), mechanical properties (Li YL et al., 2021; Wu P et al., 2021), and permeability (Hou J et al., 2018; Li G et al., 2019). Therefore, the accurate acquisition of hydrate saturation will directly affect the accuracy of hydrate resources evaluation and the rationality of exploitation plan establishment. The excess-gas method is the most common method of gas hydrate formation in porous media in the laboratory (Yin ZY et al., 2018). In this method, hydrate saturation is usually determined by assuming that all the pore water is converted into hydrate. However, the pore water in the sediments cannot be completely converted into hydrates in the laboratory, namely, the pore water conversion ratio, that is the ratio of the content of pore water converted into hydrate to the total content of pore water, is less than 100%. Linga P et al. (2012) conducted the hydrate formation experiment in a bed column filled with silica sand particles. They found that the hydrate formation rate and water conversion ratio in the bed column were greater than that in the stirred vessel. Mekala P et al. (2014) studied the methane hydrate formation kinetics in Toyoura sand. It was found that the presence of salt significantly affected the kinetics of methane hydrate formation, which resulted in a 6 time decrease in the water conversion ratio and a significant decrease in hydrate formation rate under the same conditions. Siangsai A et al. (2015) studied on the effect of activated carbon particle size on methane hydrate formation. They found that the average water conversion ratio was in the range of 75.5%–96.5%. It follows that hydrate saturation may be overestimated calculated by the assumption that pore water is completely converted into hydrate. Therefore, understanding the characteristics of pore water conversion in the process of hydrate formation in sediments is the key to accurately obtaining hydrate saturation. It is of great significance for understanding the accumulation law of hydrate and evaluating the hydrate reservoir resources.

    The low-field NMR technology (LF-NMR) can accurately monitor the variation of pore water content and its distribution characteristics during hydrate formation in the sediments. It has been widely used in hydrate community (Ji YK et al., 2019; Kleinberg R et al., 2003; Minagawa H et al., 2008; Zhang YC et al., 2021a). Ge XM et al. (2018) monitored the variation of transverse relaxation time (T2) distribution in the formation and dissociation processes of methane hydrate in sedimentary rocks. The behaviors and habits of methane hydrate formation and dissociation are analyzed in samples with different pore size distribution. Chen HL et al. (2017) investigated the behaviors of CO2 hydrate formation and dissociation in sand by LF-NMR. They analyzed the changes in phase saturation and hydrate growth habits in pores. Kuang YM et al. (2020) observed the formation and dissociation of tetrahydrofuran hydrate in porous media by LF-NMR. It was found that the relative free water and bound water consumption can affect the permeability of hydrate-bearing sediments during hydrate phase transition. Liu LL et al. (2021) combined measurements of LF-NMR and X-ray computed tomography to investigate transverse surface relaxivity in quarzitic sands containing xenon hydrate. Zhang YC et al. (2021b) analyzed some implications of interface evolution for the NMR-based property predictions. So far, researchers have mainly used LF-NMR to study the changes in pore structure and physical parameters of sediments in the process of hydrate formation and dissociation. However, few studies focused on characteristics of pore water conversion in the process of hydrate formation in porous media.

    In view of the fact that LF-NMR can accurately analyze the pore water conversion characteristics in sediments, this study uses a self-designed LF-NMR monitoring apparatus to carry out the methane hydrate formation experiment in unsaturated sand samples based on the influence factors analysis on T2 distribution. The effects of sand size and clay on the pore water conversion ratio and rate are analyzed. And the law of pore water conversion is revealed. It has important theoretical significance and application value for accurately obtaining hydrate saturation and evaluating hydrate reservoir resources.

    The schematic diagram of the self-designed LF-NMR monitoring apparatus for gas hydrate is show in Fig. 1. The experimental apparatus consists of the LF-NMR system, the high-pressure reactor, the gas injecting system, the temperature and pressure control system, and the data acquisition system. It was designed for the gas hydrate formation in porous media and real-time LF-NMR measurement.

    Figure  1.  Schematic diagram of the experimental apparatus

    The LF-NMR system (Suzhou Niumag Analytical Instrument Corp., China) mainly consists of a permanent magnet, a probe coil and a radio frequency (RF) module. It can measure T2 of hydrogen atoms (1H) in the pore fluids. The 1H Larmor frequency is 21.3 MHz. The magnetic field strength is 0.5 Tesla. The magnetic field temperature is 32℃. The homogeneity of the magnetic field is 20 ppm for a length of 60 mm. The diameter of the probe coil is 70 mm.

    Fig. 2 shows the schematic diagram of the high-pressure reactor which is the core component of the apparatus. It is mainly made of polyetheretherketone (PEEK) which has no effect on the magnetic field and RF field (Zhang Z et al., 2021). The size of sample chamber is Φ30 mm×H60 mm. Fluorinert (FC-40, 3M Company, USA) is used as the confining liquid and the isolating liquid. Fluorinert, which has no hydrogen, cannot be detected in the LF-NMR measurement. And its low dielectric property can minimize the RF losses (Ersland G et al., 2010).

    Figure  2.  Schematic diagram of the high-pressure reactor.

    The gas injecting system consists of a methane gas bottle, a gas booster pump, a hold-up vessel, a pressure reducing value, and an air compressor. It can maintain the constant-pressure supply of methane gas to the high-pressure reactor in the process of methane hydrate formation in sand samples.

    The temperature and pressure control system consists of a confining pressure tracking pump, a circulator, a low-temperature thermostat, and a circulating thermostat. The confining pressure tracking pump is used to maintain a constant difference between the confining pressure and the pore pressure to prevent the sample chamber from rupturing. The circulator and the low-temperature thermostat are used to maintain the sample chamber in a constant low-temperature environment. The circulating thermostat is used to prevent the low-temperature confining liquid from affecting the temperature of the permanent magnet by circulating the isolating liquid.

    The data acquisition system mainly consists of a computer, a pressure transducer, a thermocouple, and an A/D module. The pressure transducer (Senex, China) with an accuracy of 0.25% is used to collect pore pressure of the sample chamber. The thermocouple (PT100) with an accuracy of ±0.1°C is used to acquire temperature of the sample chamber.

    Methane gas (99.99%) and deionized water were used to form methane hydrate in sand samples. Table 1 shows the key parameters of sand samples.

    Table  1.  The key parameters of sand samples
    No.Diameter of silica sand/μmClay typeWeight ratio of silica sand to clay
    1250‒500//
    275‒150//
    3250‒500Kaolinite17:1
    4250‒500Montmorillonite17:1
     | Show Table
    DownLoad: CSV

    In this study, the methane hydrate formation was carried out using the excess-gas method in unsaturated sand samples, the same as the method used in the previous experiments (Ji YK et al., 2019). The pore pressure (6 MPa) and temperature (2°C) of sample chamber were kept constant during the methane hydrate formation process. LF-NMR measurements were conducted under in situ conditions. The details of the experimental procedure are as follows:

    (i) Based on the volume of the sample chamber, the density of the sand sample, and the porosity of the sand sample (0=45%), the sand weight was calculated. Based on the water saturation (Sw=50%), the water weight was calculated. After mixing the sand and water thoroughly, the unsaturated sand was placed in a closed environment for 1 day to mix well.

    (ii) The unsaturated sand was loaded into the high-pressure reactor. The parameters of T2 distribution measurement were determined.

    (iii) The dry sand was loaded into the high-pressure reactor. The spin-echo-train decay curve of the dry sand sample was measured. It was used to remove substrate signal in the T2 distribution inversion.

    (iv) The standard samples, in which a certain amount of H2O was added to D2O, were cooled to maintain their temperature at 2°C in a thermostat. The quantitative calibration experiment of the signal intensity of T2 distribution was conducted using standard samples.

    (v) The unsaturated sand was loaded into the high-pressure reactor. The difference between the confining pressure and the pore pressure was maintained at 0.5 MPa by the confining pressure tracking pump. The temperature of the sample chamber was maintained at 2°C by the circulator and the low-temperature thermostat.

    (vi) Methane gas was injected slowly to increase the pore pressure to 6 MPa by the gas injecting system. The pore pressure was maintained at 6 MPa during the methane hydrate formation process by the gas injecting system. T2 distribution measurements were conducted until the end of methane hydrate formation.

    In this study, the T2 distribution measurement is carried out by the Car-Purcell-Meiboom-Gill (CPMG) pulse sequence. A 90° RF pulse is followed by a long series of 180° RF pulses in the entire CPMG pulse sequence. In this study, the waiting time is 5000 ms. The number of spin-echo train is 18000. The echo spacing (TE) is 0.169 ms. The first CPMG echo cannot be detected until 0.169 ms after the pulse sequence initiation. T2 of 1H in the solid phase, such as methane hydrate, is less than TE. Therefore, it cannot be detected in the LF-NMR system because the complete relaxation occurs before the detection of the first echo at t= TE (Kleinberg R et al., 2003).

    Based on the NMR relaxation mechanisms of pore fluids in the sediments, the T2 of fluids in porous media involves three independent relaxation mechanisms: the bulk relaxation, the surface relaxation and the diffusion relaxation (Coates G et al., 1999; Kenyon W, 1997). The T2 of fluids in porous media is calculated by Eq. (1).

    1T2=1T2B+1T2S+1T2D

    (1)

    where T2B, T2S and T2D are the bulk, surface and diffusion transverse relaxation time. 1/T2B and 1/T2D can be effectively neglected in many porous media (Ge XM et al., 2017; Grunewald E and Knight R, 2011). The expression of T2S is given by Eq. (2).

    1T2S=ρ2(S/SVV)pore

    (2)

    where ρ2 is the T2 surface relaxivity. (S/V)pore is the ratio of pore surface to volume which is related to pore shape factor (Fs) and pore radius (r), as shown in Eq. (3).

    (SV)pore=FSr

    (3)

    Based on Eq. (2) and Eq. (3), the T2 of fluids in porous media is calculated by Eq. (4).

    1T2ρ2FSr

    (4)

    Eq. (4) shows that the T2 of pore fluids is proportional to r. Therefore, T2 distribution obtained by the LF-NMR measurement can be used to evaluate the pore structure.

    The signals from the pores add linearly in porous media with broad distribution of pore size (Kleinberg R et al., 2003). Thus, the observed spin-echo-train decay is shown in Eq. (5).

    M(t)=imietT2i=imieρ2FSrit

    (5)

    where M is the amplitude of the transverse magnetization. mi is the magnitude of the transverse magnetization of the ith pore at t= 0. It is proportional to the water volume of the ith pore. The sum of the coefficient mi is proportional to the water volume of the sample. T2 distribution is derived from Eq. (5). The total signal intensity in T2 distribution is proportional to the sum of the coefficient mi. Therefore, water content in the sand sample can be obtained from the T2 distribution. It is the theoretical foundation to calculate water content in the Section 3.3.

    Fig. 3 shows the relationship between the total signal intensity of T2 distribution and water weight in the pure water, the sand sample (No.2 in Table 1) and the montmorillonite-containing sand sample (No.4 in Table 1). It can be seen that the water weight and the total signal intensity of T2 distribution are in the same straight line for different samples. Therefore, the signal intensity of T2 distribution is not affected by the sediment in the experimental range. LF-NMR can be used to quantitatively characterize the relationship between the signal intensity of T2 distribution and the water content, and accurately measure the content of pore water in sediments.

    Figure  3.  The relationship between the total signal intensity of T2 distribution and water weight.

    Fig. 4 shows the relationship between the normalization signal intensity of T2 distribution and temperature. The signal intensity of T2 distribution decreases with the increase of temperature for different samples. Based on Curie’s Law, the magnetization magnitude is proportional to the inverse of the absolute temperature (Cowan B, 1997). Thus, the increase of temperature results in the decrease of magnetization magnitude. It causes the decrease of signal intensity of T2 distribution. Fig. 5 shows T2 distributions of pure water at different temperature. It can be seen that T2 distribution shifts to the right with the increase of temperature. The reason is that the decrease of fluid viscosity leads to the increase of T2 with the increase of temperature. Therefore, the temperature of sample needs to keep constant during the experiment. In the quantitative calibration experiment of the signal intensity of T2 distribution, the temperature of standard samples should be the same as the temperature in hydrate formation experiment. In addition, the decrease of magnet temperature will cause the increase of magnetic field strength. In order to avoid the influence of temperature on the magnet, the isolation measure between the confining liquid and the magnet during the experiment is required to prevent the low-temperature confining liquid from affecting on the temperature of magnet, as shown in Fig. 2.

    Figure  4.  The relationship between the normalization signal intensity of T2 distribution and temperature.
    Figure  5.  T2 distributions of pure water at different temperature.

    Fig. 6 shows the relationship between the normalization signal intensity of T2 distribution and pore pressure. Increasing pore pressure by injecting nitrogen has no effect on the signal intensity of T2 distribution. However, increasing pore pressure by injecting methane results in the increase of signal intensity of T2 distribution. It shows a good linear relationship between the signal intensity of T2 distribution and pore pressure in the methane case. The signal intensity increased by 9.5% when the pore pressure rose from atmospheric pressure to 6 MPa by injecting methane gas. This indicates that the T2 of 1H in methane can be detected. According to the equation of state, the mole number of methane is proportional to the pressure under isothermal and isopyknic conditions. That is, the number of 1H is proportional to the pore pressure. It leads to a linear increase of signal intensity of T2 distribution with the increase of pore pressure.

    Figure  6.  The relationship between the normalization signal intensity of T2 distribution and pore pressure.

    According to the above analysis, when calculating water content based on T2 distribution, the influence of methane on T2 distribution should be considered, especially under high pore pressure condition. In this section, the fine sand sample (No.2 in Table 1) is taken as an example to introduce the calculation method of water content in the process of hydrate formation considering the influence of methane on T2 distribution under high pore pressure condition.

    The relationship between the water weight and the total signal intensity of T2 distribution, which is obtained in the quantitative calibration experiment, is shown in Fig. 7. The calibration equation for water content is shown in Eq. (6).

    Figure  7.  The relationship between water weight and the total signal intensity of T2 distribution.

    mw=aIw

    (6)

    where mw is the water weight. Iw is the total signal intensity of water in T2 distribution. a is the undetermined coefficient. a=7.95544×10−5 in the experiment. The correlation coefficient of calibration equation is 0.99984. This indicates that the total signal intensity of T2 distribution has a good correlation with the water weight.

    Based on the total signal intensity of T2 distribution before methane is injected into the sand sample, the water weight of sand sample is calculated by Eq. (6). The initial water saturation of sand sample is calculated by Eq. (7).

    Sw=mw/mwρwρwVϕ0×100%

    (7)

    where Sw is the water saturation. ρw is the water density. V is the apparent volume of sand sample. ϕ0is the porosity of sand sample. The initial water saturation of sand sample calculated by Eq. (7) is 50.02%. The initial water saturation of sand sample calculated by the weight method is 50.00%. This indicates that the error of the water weight calculated by T2 distribution is very small.

    According to the influencing factors analysis, the mole number of methane is proportional to the total signal intensity of methane in T2 distribution. Due to the constant temperature and pressure during the experiment, the total signal intensity of methane in T2 distribution is proportional to the methane volume, as shown in Eq. (8).

    Vg=bIg

    (8)

    where Vg is the methane volume. Ig is the total signal intensity of methane in T2 distribution. b is the undetermined coefficient. The initial methane volume in the sand sample is Vg0=VVw0 = 19.099.55 = 9.54cm3where Vw0 is the initial water volume. When the pore pressure rose from atmospheric pressure to 6 MPa before the methane hydrate formation, the total signal intensity of T2 distribution increased from 120006.34 to 131462.95. Thus, the undetermined coefficient b=9.54131462.95120006.34=8.32707×104cm3.

    Based on Eq. (6) and Eq. (8), the water weight during the hydrate formation can be calculated by Eq. (9).

    mw=a(IVgb)

    (9)

    The molecular structure of methane hydrate is CH4·6H2O (Sloan E and Koh C, 2008). The weight of hydrate formed in the sand sample can be calculated by Eq. (10).

    mh=mw0mw18×6(16+18×6)=1.14815(mw0mw)

    (10)

    The methane volume in the sand sample can be calculated by Eq. (11).

    Vg=Vϕ0mwρwmhρh

    (11)

    Based on Eq. (9), Eq. (10) and Eq. (11), the water weight during the hydrate formation can be obtained. The water conversion ratio during the hydrate formation can be calculated by Eq. (12).

    Cw=mw0mwmw0×100%

    (12)

    The water conversion rate during the hydrate formation can be calculated by Eq. (13).

    Rw=mw,t1mw,t2t2t1

    (13)

    In this study, when Rw is less than 1×10−5 g/h, it is considered that the hydrate formation in the sand sample has ended. The water conversion ratio at this time is regarded as the final water conversion ratio.

    Fig. 8 shows that the variation of T2 distribution during methane hydrate formation in the fine sand sample (No.2 in Table 1). The T2 distribution of sand sample shows one peak before the hydrate formation. T2 distribution gradually shows two peaks in the process of hydrate formation. Based on the assumption of normal distribution model, the T2 distribution is a unimodal distribution for the homogeneous porous media. As for the porous media with strong heterogeneity, the T2 distribution shows two or three peaks. It indicates that the pore distribution of sand sample is homogeneous before the hydrate formation. Hydrate formation causes the heterogeneity of pore distribution in the sand sample. This phenomenon may be caused by the randomness of the nucleation position of hydrate. Hydrate will be preferentially formed at a certain position in the sand sample. The large pores were smashed into several smaller pores isolated by hydrate clusters upon hydrate formation (Liu YZ et al., 2021). It results in the reduction of pore size. The size of other pores in which hydrate is not formed or formed later is larger. It leads to the heterogeneous distribution of pore size in the sand sample.

    Figure  8.  The variation of the T2 distribution during methane hydrate formation in the fine sand sample.

    The left end of T2 distribution tends to shift to the left in the process of hydrate formation, as shown in Fig. 8. T2 of pore water is affected not only by rock surface but also by hydrate surface due to the hydrate formation. It leads to the increase of T2 surface relaxivity part of the pore water in Eq. (6), which in turn causes T2 to decrease. Meanwhile, the right end of T2 distribution tends to shift to the right in the process of hydrate formation, as shown in Fig. 8. The hydrate formation may reduce the influence of rock surface and increase the influence of hydrate surface on T2 of part of the pore water. Because the T2 surface relaxivity of hydrate is one order of magnitude smaller than that of rock (Kleinberg R et al., 2003), the T2 surface relaxivity in Eq. (6) decreases, which in turn causes T2 to increase. In addition, changes in the initial pore structure will lead to changes in T2. Due to the limitations of the experiment, it is impossible to make a related analysis in this study. It will be further studied in the later stage.

    Fig. 9 shows that the variation of the water conversion ratio and rate during methane hydrate formation in the fine sand sample (No.2 in Table 1). The inset figure shows the variation of the water conversion ratio and rate from t=0 h to t=6 h. The water conversion rate is faster in the early stage of hydrate formation. In the first 3 hours, more than 80% of pore water is converted into hydrate. Due to the lower Gibbs free energy of nucleation and the higher concentration of methane molecules and water molecules at the methane-water interface (Sloan E and Koh C, 2008), the methane hydrate film is usually formed quickly at the methane-water interface (Ji YK et al., 2021; Tohidi B et al., 2001). The hydrate film prevents the direct contact of water and methane gas. The increase of the thickness of hydrate film renders it more difficult for gas or water molecules to diffuse through the hydrate film. It leads to the water conversion rate to slow down. The water conversion rate fluctuates and decreases with the formation of hydrate, as shown in the inset figure of Fig. 9. Hydrate nucleation in porous media is random. The induction time is different at different positions in porous media. It can cause the fluctuation of water conversion rate. In addition, Liang HY et al. (2021) proposed a new growth pattern of hydrate protrusions on the outer surface of the hydrate shell, which can also explain the characteristics of water conversion in this work. As the hydrate shell thickened, the mode of mass transfer of water through the hydrate shell from diffusion shifted to permeation, resulting in the increase of effective diffusion coefficient of interstitial water. It can cause the secondary increase of water conversion rate.

    Figure  9.  The variation of the water conversion ratio and rate during methane hydrate formation in the fine sand sample.

    Fig. 10 shows that the variation of the water conversion ratio during methane hydrate formation in the coarse sand sample (No.1 in Table 1) and the fine sand sample (No.2 in Table 1). The time when the water conversion ratio reaches 80% for the fine sand sample (t2) is less than the time for the coarse sand sample (t1). Therefore, the pore water conversion rate is faster in the fine sand sample. The reason is that the specific surface of fine sand sample is larger. It causes the larger contact area between water and methane. Larger contact area between water and methane results in a larger pore water conversion rate. The final water conversion ratio of the fine sand sample (Cw=96.2%) is lower than that of the coarse sand sample (Cw=99.7%). Due to the smaller pore size in the fine sand sample, the capillary force has a greater hindering effect on hydrate formation as the hydrate is formed. Under the same gas driving force, the pore pressure in the fine sand sample is easier to be lower than the equilibrium pressure for the hydrate phase. It causes the lower water conversion ratio in the fine sand sample. Rouquerol J et al. (1988) and Handa Y et al. (1992) pointed out that there was a layer of water molecules with a thickness of 0.8 Å on the solid surface, which could not form ice or hydrate due to the large potential energy of water molecules near the solid surface. The interface interaction between solid and water has an important influence on the crystallization process of water (Jia JL et al., 2013). Therefore, the larger specific surface results in the lower the water conversion ratio in the fine sand sample.

    Figure  10.  The variation of the water conversion ratio during methane hydrate formation in the coarse and fine sand sample.

    Fig. 11 shows that the variation of the water conversion ratio during methane hydrate formation in the coarse sand sample (No.1 in Table 1), the kaolinite-containing sand sample (No.3 in Table 1) and the montmorillonite-containing sand sample (No.4 in Table 1). The time when the water conversion ratio reaches 80% for the coarse sand sample (t1) is the shortest, followed by the time for the kaolinite-containing sand sample (t3), and the time for the montmorillonite-containing sand sample (t4) is the longest. Therefore, the clay can inhibit the hydrate formation in sand samples, especially montmorillonite. Both sides of the crystal layer of montmorillonite are composed of oxygen atoms. The force between the crystal layers is intermolecular force. The hydrogen bonds don’t exist between the crystal layers. It causes water molecules to be easy to enter the crystal layers. In addition, the crystal surface combines with a large number of exchangeable cations due to a large number of lattice substitution in montmorillonite. When water molecules enter the crystal layers, the diffuse electric double layers are formed due to the dissociation of exchangeable cations. The negatively charged crystal surfaces repel each other and expand. It causes water molecules to be easier to enter the crystal layers. Therefore, a large number of water molecules exist between the crystal layers. The water molecules between the crystal layers can form hydrate with methane molecules (Yan KF et al., 2019). However, methane molecules are restricted to contact with water molecules due to the existence of the crystal layers. It results in a significant decrease in the water conversion ratio and rate in the montmorillonite-containing sand sample. However, the hydrogen bonds exist between the crystal layers of kaolinite. It causes water molecules to be difficult to enter the crystal layers. And the crystal surface combined with few exchangeable cations. Compared with montmorillonite, there is no crystal layer that hinders the contact between methane molecules and water molecules. Therefore, the water conversion rate is faster and the water conversion ratio is higher in the kaolinite-containing sand sample. According to the above analysis, the addition of kaolinite will increase the specific surface of sand sample. It makes more water be affected by the solid surface and fail to form hydrates. It results that the water conversion ratio of the kaolinite-containing sand sample is lower than that of the coarse sand sample.

    Figure  11.  The variation of the water conversion ratio during methane hydrate formation in the coarse and clay-containing sand sample.

    In this study, the low-field NMR technology was used to study the effects of sand size and clay on characteristics of pore water conversion during methane hydrate formation in unsaturated porous media. The main conclusions are as follows:

    (i) T2 distribution is not affected by sediment type and pore pressure within the experimental range, but is significantly affected by temperature. The increase of the pressure of hydrogen-containing gas, such as methane, can increase the hydrogen content to cause the increase of signal intensity of T2 distribution.

    (ii) Hydrate formation aggravates the heterogeneity of pore distribution in the sand sample. The influence of the T2 surface relaxivity of hydrate on T2 of pore water should be considered in T2 distribution analysis.

    (iii) The randomness of the hydrate nucleation and the mode of mass transfer of water through the hydrate shell from diffusion shifted to permeation can cause the fluctuation of water conversion rate.

    (iv) The specific surface of sand sample affects the water conversion ratio and rate. The fine sand sample has a larger specific surface. It causes a larger gas-water contact area to lead to a larger water conversion rate. At the same time, it causes a larger water-sand contact area to lead to a lower water conversion ratio.

    (v) The clay can reduce the pore water conversion ratio and rate, especially montmorillonite. In addition to the influence of the specific surface of sand samples, the crystal layers of montmorillonite can hinder the water conversion.

    Yun-kai Ji and Chang-ling Liu conceived of the presented idea. Chang-ling Liu and Neng-you Wu supervised the findings of this work and were in charge of overall direction and planning. Yun-kai Ji and Zhun Zhang carried out the experiment. Yun-kai Ji, Qing-guo Meng, Le-le Liu and Yong-chao Zhang contributed to the interpretation of the results. All authors provided critical feedback and contributed to the final manuscript.

    The authors declare no conflicts of interest.

    The authors greatly appreciate the financial support of the National Natural Science Foundation of China (41876051 and 41872136), the China Postdoctoral Science Foundation (2021M701815) and the Postdoctoral Innovative Talents Support Program in Shandong Province (SDBX2021015).

  • Bu QT, Hu GW, Ye YG, Liu CL, Li CF, Wang JS. 2017. Experimental study on 2-D acoustic characteristics and hydrate distribution in sand. Geophysical Journal International, 211(2), 990–1004. doi: 10.1093/gji/ggx351
    Chen HL, Wei WF, Tian HH, Wei HZ. 2017. NMR relaxation response of CO2 hydrate formation and dissociation in sand. Acta Physico-Chimica Sinica, 33(08), 1599–1604. doi: 10.3866/PKU.WHXB201704194
    Coates G, Xiao LZ, Prammer M. 1999. NMR Logging: Principles And Applications. Houston, Halliburton Energy Services, 35‒50.
    Cowan B. 1997. Nuclear Magnetic Resonance and Relaxation. Cambridge, Cambridge University Press. 83‒123
    Ersland G, Husebø J, Graue A, Baldwin B, Howard J, Stevens J. 2010. Measuring gas hydrate formation and exchange with CO2 in Bentheim sandstone using MRI tomography. Chemical Engineering Journal, 158(1), 25–31. doi: 10.1016/j.cej.2008.12.028
    Ge XM, Fan YR, Liu JY, Zhang L, Han YJ, Xing DH. 2017. An improved method for permeability estimation of the bioclastic limestone reservoir based on NMR data. Journal of Magnetic Resonance, 283, 96–109. doi: 10.1016/j.jmr.2017.09.004
    Ge XM, Liu JY, Fan YR, Xing DH, Deng SG, Cai JC. 2018. Laboratory investigation into the formation and dissociation process of gas hydrate by low-field NMR technique. Journal of Geophysical Research:Solid Earth, 123(5), 3339–3346. doi: 10.1029/2017JB014705
    Grunewald E, Knight R. 2011. The effect of pore size and magnetic susceptibility on the surface NMR relaxation parameter T2. Near Surface Geophysics, 2(9), 169–178. doi: 10.3997/1873-0604.2010062
    Handa Y, Stupin D. 1992. Thermodynamic properties and dissociation characteristics of methane and propane hydrates in 70-ANG-radius silica gel pores. The Journal of Physical Chemistry, 96(21), 8599–8603. doi: 10.1021/j100200a071
    Hou J, Ji YK, Zhou K, Liu YG, Wei B. 2018. Effect of hydrate on permeability in porous media: Pore-scale micro-simulation. International Journal of Heat and Mass Transfer, 126, 416–424. doi: 10.1016/j.ijheatmasstransfer.2018.05.156
    Ji YK, Hou J, Cui GD, Lu N, Zhao EM, Liu YL, Du QJ. 2019. Experimental study on methane hydrate formation in a partially saturated sandstone using low-field NMR technique. Fuel, 251, 82–90. doi: 10.1016/j.fuel.2019.04.021
    Ji YK, Hou J, Zhao EM, Liu CL, Guo TK, Liu YL, Wei B, Bai YJ. 2021. Pore-scale study on methane hydrate formation and dissociation in a heterogeneous micromodel. Journal of Natural Gas Science and Engineering, 95, 104230. doi: 10.1016/j.jngse.2021.104230
    Jia JL, Zhang Y, Li G, Chen ZY, Li XS, Wu HJ. 2013. Experimental study on formation behavior of methane hydrate in sea mud from South China Sea. Geoscience, 27(06), 1373–1378. doi: 10.3969/j.issn.1000-8527.2013.06.014
    Kenyon W. 1997. Petrophysical principles of applications of NMR logging. Log Analyst, 38(2), 21–40.
    Kleinberg R, Flaum C, Straley C, Brewer P, Malby G, Peltzer E, Friederich G, Yesinowski J. 2003. Seafloor nuclear magnetic resonance assay of methane hydrate in sediment and rock. Journal of Geophysical Research:Solid Earth, 108(B3), 2137. doi: 10.1029/2001JB000919
    Kleinberg R, Flaum C, Griffin D, Brewer P, Malby G, Peltzer E, Yesinowski J. 2003. Deep sea NMR: Methane hydrate growth habit in porous media and its relationship to hydraulic permeability, deposit accumulation, and submarine slope stability. Journal of Geophysical Research:Solid Earth, 108(B10), 2508. doi: 10.1029/2003JB002389
    Kuang YM, Zhang LX, Song YC, Yang L, Zhao JF. 2020. Quantitative determination of pore‐structure change and permeability estimation under hydrate phase transition by NMR. AIChE Journal, 66, e16859. doi: 10.1002/aic.16859
    Li G, Li XS, Lv QN, Zhang Y. 2019. Permeability measurements of quartz sands with methane hydrate. Chemical Engineering Science, 193, 1–5. doi: 10.1016/j.ces.2018.08.055
    Li JF, Ye JL, Qin XW, Qiu HJ, Wu NY, Lu HL, Xie WW, Lu JA, Peng F, Xu ZQ, Lu C, Kuang ZG, Wei JG, Liang QY, Lu HF, Kou BB. 2018. The first offshore natural gas hydrate production test in South China Sea. China Geology, 1(1), 5–16. doi: 10.31035/cg2018003
    Li YL, Dong L, Wu NY, Nouri A, Liao HL, Chen Q, Sun JY, Liu CL. 2021. Influences of hydrate layered distribution patterns on triaxial shearing characteristics of hydrate-bearing sediments. Engineering Geology, 294, 106375. doi: 10.1016/j.enggeo.2021.106375
    Liang HY, Guan DW, Yang L, Zhang LX, Song YC, Zhao JF. 2021. Multi-scale characterization of shell thickness and effective volume fraction during gas hydrates formation: A kinetic study. Chemical Engineering Journal, 424, 130360. doi: 10.1016/j.cej.2021.130360
    Linga P, Daraboina N, Ripmeester J, Englezos P. 2012. Enhanced rate of gas hydrate formation in a fixed bed column filled with sand compared to a stirred vessel. Chemical Engineering Science, 68(1), 617–623. doi: 10.1016/j.ces.2011.10.030
    Liu LL, Zhang Z, Liu CL, Wu NY, Ning FL, Zhang YC, Meng QG, Li CF. 2021. Nuclear magnetic resonance transverse surface relaxivity in quartzitic sands containing gas hydrate. Energy & Fuels, 35(7), 6144–6152. doi: 10.1021/acs.energyfuels.1c00225
    Liu YZ, Zhang LX, Yang L, Dong HS, Zhao JF, Song YC. 2021. Behaviors of CO2 hydrate formation in the presence of acid-dissolvable organic matters. Environmental Science & Technology, 55(9), 6206–6213. doi: 10.1021/acs.est.0c06407
    Mekala P, Babu P, Sangwai J, Linga P. 2014. Formation and dissociation kinetics of methane hydrates in seawater and silica sand. Energy & Fuels, 28(4), 2708–2716. doi: 10.1021/ef402445k
    Minagawa H, Nishikawa Y, Ikeda I, Miyazaki K, Takahara N, Sakamoto Y, Komai T, Narita H. 2008. Characterization of sand sediment by pore size distribution and permeability using proton nuclear magnetic resonance measurement. Journal of Geophysical Research, 113(B7). doi: 10.1029/2007JB005403.
    Priegnitz M, Thaler J, Spangenberg E, Schicks J, Schrötter J, Abendroth S. 2015. Characterizing electrical properties and permeability changes of hydrate bearing sediments using ERT data. Geophysical Journal International, 202(3), 1599–1612. doi: 10.1093/gji/ggv245
    Siangsai A, Rangsunvigit P, Kitiyanan B, Kulprathipanja S, Linga P. 2015. Investigation on the roles of activated carbon particle sizes on methane hydrate formation and dissociation. Chemical Engineering Science, 126, 383–389. doi: 10.1016/j.ces.2014.12.047
    Sloan E, Koh C. 2008. Clathrate Hydrates of Natural Gases. Boca Raton, CRC Press, 156-168.
    Tohidi B, Anderson R, Clennell M, Burgass R, Biderkab A. 2001. Visual observation of gas-hydrate formation and dissociation in synthetic porous media by means of glass micromodels. Geology, 29(9), 867. doi: 10.1130/0091-7613(2001)029<0867:VOOGHF>2.0.CO;2
    Rouquerol J, Rodriguez-Reinoso F, Sing K, Unger KK. 1988. Characterization of Porous Solids. Amsterdam, Elsevier Science, 439–447.
    Wu Pe, Li YH, Sun X, Liu WG, Song YC. 2021. Mechanical characteristics of hydrate-bearing sediment: A review. Energy & Fuels, 35(2), 1041–1057. doi: 10.1021/acs.energyfuels.0c03995
    Yan KF, Li XS, Chen ZY, Zhang Y, Xu CG, Xia ZM. 2019. Methane hydrate formation and dissociation behaviors in montmorillonite. Chinese Journal of Chemical Engineering, 27(5), 1212–1218. doi: 10.1016/j.cjche.2018.11.025
    Ye JL, Qin XW, Xie WW, Lu HL, Ma BJ, Qiu HJ, Liang JQ, Lu JA, Kuang ZG, Lu C, Liang QY, Wei SP, Yu YJ, Liu CS, Li B, Shen KX, Shi HX, Lu QP, Li J, Kou BB, Song G, Li B, Zhang HE, Lu HF, Ma C, Dong YF, Bian H. 2020. The second natural gas hydrate production test in the South China Sea. China Geology, 3(2), 197–209. doi: 10.31035/cg2020043
    Yin ZY, Moridis G, Tan H, Linga P. 2018. Numerical analysis of experimental studies of methane hydrate formation in a sandy porous medium. Applied Energy, 220, 681–704. doi: 10.1016/j.apenergy.2018.03.075
    Zhan L, Wang Y, Li XS. 2018. Experimental study on characteristics of methane hydrate formation and dissociation in porous medium with different particle sizes using depressurization. Fuel, 230, 37–44. doi: 10.1016/j.fuel.2018.05.008
    Zhang YC, Liu LL, Wang DG, Chen PF, Zhang Z, Meng QG, Liu CL. 2021a. Application of low-field nuclear magnetic resonance (LFNMR) in characterizing the dissociation of gas hydrate in a porous media. Energy & Fuels, 35(3), 2174–2182. doi: 10.1021/acs.energyfuels.0c03855
    Zhang YC, Liu LL, Wang DG, Zhang Z, Li CF, Meng QG, Liu CL. 2021b. The interface evolution during methane hydrate dissociation within quartz sands and its implications to the permeability prediction based on NMR data. Marine and Petroleum Geology, 129, 105065. doi: 10.1016/j.marpetgeo.2021.105065
    Zhang Z, Liu LL, Li CF, Liu CL, Ning FL, Liu ZC, Meng QG. 2021. A testing assembly for combination measurements on gas hydrate-bearing sediments using X-ray computed tomography and low-field nuclear magnetic resonance. Review of Scientific Instruments, 92(8), 85108. doi: 10.1063/5.0040858
    Zhao EM, Hou J, Du QJ, Liu YG, Ji YK, Bai YJ. 2021. Numerical modeling of gas production from methane hydrate deposits using low-frequency electrical heating assisted depressurization method. Fuel, 290, 120075. doi: 10.1016/j.fuel.2020.120075

Catalog

    Figures(11)  /  Tables(1)

    Metrics
    PDF Downloads(167) Abstract views(7252) HTML views(3377) Cited by(18)

    /

    DownLoad:  Full-Size Img  PowerPoint
    Return
    Return
    x Close Forever Close