Behaviour of methane hydrate bearing sand during monotonic ... · ABSTRACT:This paper, presents the results of methane hydrate (MH) growth habit on the shear behaviour of methane

Indian Geotechnical Conference IGC2016 15-17 December 2016, IIT Madras, Chennai, India

1

Behaviour of methane hydrate bearing sand during monotonic loading:

DEM simulations

J S Vinod Senior Lecturer, Centre for Geomechanics and Railway Engineering, University of Wollongong, NSW-2522 Australia

Masayuki Hyodo Emeritus Professor, Department of Civil Engineering, Yamaguchi University, Japan

BuddhimaIndraratna Distinguished Professor,Centre for Geomechanicsand Railway Engineering, University of Wollongong, NSW-2522 Australia

Roy Miller Former ME student, Centre for Geomechanics and Railway Engineering, University of Wollongong, NSW-2522 Australia

1 INTRODUCTION Methane Hydrates (MH) are crystalline solids consisting of methane molecules surrounded by a cage of interlocking water molecules formed under certain pressure and temperature. It is considered as one of the most feasible sources of energy compared with other known hydrocarbon deposits. The Recent geological survey shows that India has massive deposits of MH deposits estimating around 1890 trillion cubic metres. The deposit on the Krishna Godavari basin is one of the richest and biggest known MH deposit (The Economic times, 2013) in India. However, there are significant challenges both geotechnical and environmental for drilling and production operations of MH. It is understood that the methane is approximately 20 times as effective as greenhouse gas as carbondioxide (e.g. Hyodo et al. 2005; Collet &Dallomore, 2002) and highly unstable (Iwai et al. 2015). A wide range of geotechnical and environmental issues has to be addressed before the production of methane gas from MH safely and without damaging the environmental issues. For example, it has been reported that methane hydrate production may collapse and leads to settlement or landslides on the seabed. The marine substructures are vulnerable to the seabed deformations and no attractive technology is currently available to recover methane economically from

methane hydrate. Until now, large-scale production has not yet commenced as the behaviour of gas hydrates, particularly hydrates of methane, are not fully understood. However, a lot of research on MH as a future energy source has been reported, recently, which shows that countries are eager to develop an effective way to capture MH in order to secure their natural gas stores for the future and also an attempt to combat global warming (e.g. Kvenvolden, 1988; Waite et al., 2009; Collet, 2002; Kvenvolden and Lorenson, 2001; Hyodo et al., 2013;Iwai et al. 2015). A lot of laboratory experiments using triaxial, direct shear and bending tests have been carried out to understand the shear behaviour of methane hydrate- bearing sediment (e.g. Masui et al 2005; Hyodo et al 2002, 2005 & 2013 and Song, 2010). It has been reported that the shear strength, stiffness, and dilation of hydrate-bearing sand is influenced by hydrate saturation, initial confining pressure, and temperature. The stress- strain response of hydrate-bearing soil is affected by the hydrate growth habit (Waite et al. 2009). The microscopic distribution of hydrates in soil is described using three models (1) pore filling: hydrate nucleate on sediment grain boundaries and grow freely into pore spaces without bridging two or more particles. (2) Cementation: Hydrate establishes a bond on the interparticle contact points. (3) load-bearing:

ABSTRACT:This paper, presents the results of methane hydrate (MH) growth habit on the shear behaviour of methane hydrate bearing sand using the Discrete Element Method (DEM). DEM based PFC3D was used to model the two pore growth habitat(i) cementation, bonding of the interparticle contact and (ii) the pore filling, leading to load bearing, behavior approach of MH bearing sand. Monotonic shear loading simulations using DEM have captured, qualitatively, the stress ratio- axial strain behaviour similar to the laboratory experiments. The DEM simulation results highlight that hydrate growth habitat has a significant influence on the shear behaviour of hydrate bearing sand. It was shown that the cementation habit closely captures the variation of peak deviator stress with MH saturation similar to the laboratory experiments. Moreover, the evolution of micro-mechanical parameters (e.g. bond breakage& contact force) during shear loading has been presented and discussed.

KEYWORDS:DEM modelling,monotonic loading, methane hydrate, sand

Behaviour of methane hydrate bearing sand during monotonic loading: DEM simulations

2

Hydrate bridges neighboring grains and contributes mechanical stability to thegranular assembly by becoming part of the loading bearing force chain. Jung et al (2012) report difficulty in controlling hydrate formation, distribution, saturation, and pore habit challenges during laboratory experiments.

Recently, few research studies have been carried to understand the behavior MH bearing sediments using Discrete Element Method (DEM) (Brugada et al.2010; Jung et al 2012; Jian et al, 2013; Vinod et al 2014a; 2014b; Jian et al. 2016). Brugada et al (2010) developed a pore filling DEM model to capture theshearbehaviour of methane hydrate-bearing sand and highlighted that hydrate contribution to the strength of the sediment is purely frictional nature. Jung et al (2012) have concluded based on the DEM simulation that mechanical properties of hydrate-bearing sediments can be expressed as afunction of hydrate saturation, initial porosity, and effective stress. The cementation effect of MH bearing sediments was captured using a simple contact model and DEM simulation highlight that the cohesion/cementation of hydrate-bearing deposit increases with increase in hydrate saturation (Jian et al. 2013). Vinod et al (2014 a &b) investigated the pore filling habit, load bearing, and cementation habit, bonding on the interparticle contact point, has been investigated using DEM simulations.Jiang et al. (2014& 2016) established a bond model(thermo-hydro-mechanical) todescribe the behavior of the bonding MH, in the MH bearing sand environment. The 2D DEM model captured the effect of temperature, pore pressure on the shear behavior of MH bearing sand. In this study,3D DEM simulations using PFC 3D was used to model the MH growth habit on the shear behavior of MH bearing sand.

2 MH BEARING SAND: Laboratory Experiments Hyodo et al. (2005 & 2013) carried out triaxial laboratory experiments on methane hydrate-bearing sand samples. The experimental studies were carried out on acylindrical specimen of 30mm diameter and 50 mm high having an initial porosity of 0.4. All the tests were carried out on Toyora sand as the host sand. A rigorous sample preparation scheme was followed to prepare the MH bearing sand considering the pressure and temperature conditions favorable for MH formation (Hyodo et al. 2013). The host sand was first mixed with predetermined amount of water to achieve the target MH saturation. The moist soil was then placed in a mold with each layer compacted to achieve the required density. This specimen was then subjected to a series of process under specific temperature and pressures as presented in Fig 1. At stage (c) methane

gas was injected and the temperature was reduced to 1C where MH was stable (Fig.1).

Fig. 1: State paths for pressure and temperature to produce MH bearing sand (after Hyodo et al., 2013) A very high pore water pressures were then applied to the sample to consider the condition of the seabed. More details on the sample preparation and testing procedure can be found elsewhere (e.g. Hyodo et al 2013 & 2005). Drained triaxial tests were carried out on MH bearing sand with varying MH saturations. The effect of MH saturation on the shear behavior of MH bearing sand is presented in Fig. 2. MH saturation has a profound influence on the shear behaviour of MH bearing sand. An increase in the initial stiffness and shear strength of sand was observed with increase in the MH saturation. The increase in shear strength with MH saturation may be due to the cementation/bonding of sand particles of sandy sediments beneath the deep ocean floor (Hyodo et al. 2013).

Fig.2: Effect of MH on the stress ratio with axial strain for MH bearing sand (after Vinod et al 2014b)

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

0 2 4 6 8

Str

ess

rati

o, q

/p

Axial strain, (%)

SMH = 53%

= 35%

= 22%

= 0%

3 = 5.0 MPa

Experimental data from Hyodo et al. (2013)

Indian Geotechnical Conference IGC2016 15-17 December 2016, IIT Madras, Chennai, India

3

3 MH BEARING SAND: DEM Simulations In this study, DEM based software PFC3D was used to simulate these laboratory experiments. The cylindrical soil specimens having a diameter =30mm and height = 60mm, similar to the size of the laboratory experiments, was simulated using 7941 spherical particles (host sand). The particle size varies from 0.07 - 1 mm and the samples were prepared at an initial porosity of 0.4. The properties used for MH bearing sand are tabulated in Table 1. A linear force-displacement contact model was used for the simulation program. The particles were generated using radius expansion technique incorporated in PFC3D. The cylindrical lateral wall was assigned a stiffness of 5x107 N/m to create a soft confinement. After generation, the assembly was isotropically compacted to a desired initial confining pressure (fordrained shearing (Fig.3) of the samples. Table 1: Micromechanical Properties used for host sand particles(after Vinod et al. 2014a)

Properties Sand

Normal Stiffness (N/m) 5x108

Shear Stiffness (N/m) 5x108

Particle Size (mm) 0.07-0.1

Density (kg/m3) 2000

Contact friction 0.52

Fig.3: MH bearing host sand (after Vinod et al. 2014a) 3.1 Cementation Habit: The cementation effect of MH was modelled using a parallel bonding model. Parallel bond approximates the physical behaviour of a cement-like substance joining two particles. It can transmit both forces and moments between particles, thus, they may contribute to the resultant force and moment acting on the two bonded particles (Itasca, 2004). The maximum tensile and shear stresses acting

on the periphery of the bond (Itasca, 2004) are given by: (1) (2) Where

max and max are the maximum tensile and

shear stresses acting on the periphery of the bond,

nF and

sF are the normal and shear forces acting on

the bond,

M is the moment acting on the bond, R is the radius of the bond, and 'A and I are the area and moment of inertia of the cross section of the bond. In this study cementation, bond strength, of the MH bearing sand has been varied based on the laboratory experimental investigation (e.g. Hyodo et al. 2009; Jiang et al., 2013, 2014). Therefore, the bond strength of the assembly has been modified to capture the shear behaviour for different percentage of MH saturation (Table 2). A bond shear and normal stiffness of 1x109 N/m and bond radius of 0.2-0.6 m were used for the simulations. A series of drained monotonic tests were carried out on the initial sample (Fig.3) varying the normal and shear bond strength to capture the effect of MH saturation.

Table 2: Micromechanical properties used for DEM simulations (after Vinod et al.2014b)

MH Saturation (%) Normal & Shear Bond

Strength (N/m2) 0 0

22 8.0x106 35 1.2x107 53 1.5x107

The effect of cementation habit on the shear behavior of MH bearing sand is presented in Fig.4. It is evident from the figure that that cementation habit has captured the effect of MH saturation similar to the laboratory experiments. As expected, DEM simulations clearly show that, the stress ratio increases with increase in the MH saturation. However, the reduction in stress ratio after the peak value is mainly due to the breakage of bonds. Fig.5 shows the variation of bond breakage with axial strain. It is evident from the Fig.5 that bond breakage has a significant influence on the shear behaviour of MH bearing sand. A parallel bond break when the magnitude of the tensile normal or shear contact force exceeds the applied normal or shear bond strength. This is clearly captured in Fig. 5, where at

RI

M

A

F n

'max

'maxA

F s

30 mm

60 m

m

SMH= 0%

Behaviour of methane hydrate bearing sand during monotonic loading: DEM simulations

4

low strains all bonds remain intact; however, during shearing CF increases and when CF exceeds the bond strength the bond breaks. The bond breaks non-linearly with axial strain and expected to reach a steady state at large strain (>10%). Further, bond breakage decreases with increase in MH saturation, hence, exhibit higher shear strength at higher MH saturation. Qualitatively, this result is very similar to the laboratory experimental results presented by Hyodo et al. (2013).

Fig.4: Effect of MH on the deviator stress with axial strain for MH bearing sand (after Vinod et al. 2014b)

Fig.5: Variation of percentage of broken bond with axial strain for different MH saturation (after Vinod et al.2014b) 3.2 Pore Filling Habit: The MH particles were randomly generated in the pores of the sand particles using radius expansion technique as described in the earlier section. The MH particles were assumed to be of uniform spherical particles having a size 0.04 mm. The number of particles for different saturation was determined from theinitial void volume. The properties assigned for the MH particles are presented in Table 3. It was observed that the initial porosity slightly decreases with the addition of MH particles.

Table 3: Micromechanical Properties used for MH particles (after Vinod et al. 2014a)

Properties Methane Hydrate

Normal Stiffness (N/m) 1x109

Shear Stiffness (N/m) 1x109

Particle Size (mm) 0.04

Density (kg/m3) 2000

Contact friction 0.58 (SMH = 22%); 0.61 (SMH = 35%); 0.64 (SMH = 53%)

Fig.6 shows samples with hydrate saturation of 35%. A series of drained monotonic triaxial tests were then carried on isotropically consolidated sample at 3= 5MPa.

Fig.6: Initial assembly with MH (after Vinod et al. 2014a)

Fig.7: Effect of MH on the stress ratio with axial strain for MH bearing sand (after Vinod et al. 2014b) The effect of pore-filling habit on the shear behavior of MH bearing sand is presented in Fig 7. It is clear from the figure that pore-filling habit has asignificant influence on the shear behaviour of sand. The stress

0

0.2

0.4

0.6

0.8

1

1.2

1.4

0 1 2 3 4 5

Str

ess

rati

o, q

/p

Axial Strain, (%)

SMH = 53%

= 35%

= 22 %

= 0 %

3 = 5.0 MPa

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0 1 2 3 4 5

Per

cent

age

of B

ond

s B

roke

n

Axial Strain, (%)

SMH =22%

=35 %

=53 %

0

0.4

0.8

1.2

0 2 4 6

Str

ess

Rat

io, q

/p

Axial strain (%)

SMH = 53%=35%= 22 %= 0 %

3 = 5MPa

30 mm

60

mm

SMH= 35%

Indian Geotechnical Conference IGC2016 15-17 December 2016, IIT Madras, Chennai, India

5

ratio increases with the increase in the methane hydrate saturation. The stress ratio shows a peak value at an axial strain of 2% and thereafter remains constant with axial strain. 4 MICROMECHANICAL EXPLANATIONS Fig.8a &b shows the spatial variation of parallel bond with axial strain for SMH=35%. The parallel bond locations are shown by ablack line between two particles for the specimen with MH saturation of 35% at = 0% and = 5% . Fig. 8 (a) shows all the bonds (cementation) formed on the hostsand before shearing, and,Fig.8(b) clearly shows the development of CF leading to the breakage of thebond (cementation). The bond breakage is expected to reach a steady state at large axial strain levels.

Fig.8 Spatial variation of parallel bond with axial strain (a) = 0%; (b) = 5% (after Vinod et al. 2014b)

Fig.9 shows the spatial variation of contact force (CF) chain developed during shear loading for samples SMH= 0% and SMH=35%. The contact force is represented by a line segment connecting the centroid of two contacting particles, and the line width is proportional to contact force magnitude Fig.9 (a) shows the CF developed for MH bearing sand (SMH=0%) samples at = 0.35% and Fig. 8(b) shows the CF distribution for SMH=35% at = 0.35%. During shear loading, CF chains will develop along the major principal stress direction. It is evident from Fig.9b that the MH particles strongly contribute to load bearing along with MH bearing sand during shear loading. It is evident that mean CF (the average value, overall contacts with non-zero normal force) increases with increase in the axial strain. In fact, the increase of CF with MH saturation is directly reflected in the corresponding increase in deviator stress (Fig.7).

5 COMPARISONS WITH LABORATORY EXPERIMENTS The comparison of DEM simulation results with laboratory experiments is presented in Fig.10. It is evident from the figure that cementation habit clearly captures the variation of peak deviator stress close to the laboratory experiments. The pore filling habit shows a close agreement at low MH saturation (SMH< 20%), however, did not capture the increase in deviator stress at high MH saturation (SMH>20%). This result clearly demonstrates the development of a cementation type hydrate growth habit on the Toyoura sand during laboratory experiments

Fig.10. Comparison with Laboratory experiments (after Vinod et al. 2014b)

6 CONCLUSIONS This paper has presented the effect of MH growth habit on the shear behaviour of sand. Two distinct approaches, (i) cementation habit and (ii) pore-filling were simulated to capture the shear response of MH bearing sand. The cementation habit and pore filling habit was modelled by varying the bond strength and contact friction angle respectively. Both the approaches

3

5

7

9

11

13

15

0 20 40 60

Pea

k D

evia

tor

Str

ess

(MP

a)

SMH(%)

Laboratory Experiments (Hyodo et al 2013)

Cementation Habit: DEM simulations

Pore Filling Habit:DEM simulations

Fig.9: Spatial variation of contact force chain during shear loading (a) SMH =0% = 0.35% (b) SMH =35%; =0.35% (after Vinod et al. 2014)

(b) (a)

CF developed in MH particles

(a) (b)

Bon

d b

reakage

Behaviour of methane hydrate bearing sand during monotonic loading: DEM simulations

6

have qualitatively captured the stress ratio – axial strain variation similar to the laboratory experiments. It can be concluded from DEM simulations that the stress ratio significantly increases with increase in MH saturation. Moreover, DEM simulations highlight that the cementation type hydrates growth may exist during shear loading of Toyuora sand in laboratory condition. In cementation habit, breakage of thebond(cementation) decreases with increase in MH saturation whereas in pore filling habit MH particles strongly contribute towards the overall load bearing arrangement. 7 REFERENCES Brugada J., Cheng Y.P., Soga K., &Santamarina J.C. (2010).‘Discrete element modeling of geomechanicalbehaviour of MH soils with pore-filling hydrate distribution’. Granular Matter, 12(5), pp. 517-525. Collet, T.S. and Dallimore, S.R., (2002), ‘Integrated well log and reflection seismic analysis of gas hydrate accumulations on the Richards Island in the Mackenzie Delta’, Canada, CSEG Recorder, 27( 8), pp. 28-40. Collett, T. S. (2002), ‘Energy resource potential of natural gas hydrates’, AAPG Bull., 86, pp. 1971–1992. Hyodo, M, Y. Nakata, N. Yoshimoto, R. Orense, and J. Yoneda, (2009).‘Bonding strength by methane hydrate formed among sand particles’, Proceedings of the 6th International Conference on Micromechanics of Granular Media, Powders and Grains, Golden, Colo, USA, pp. 79–82. Hyodo, M., Nakata, Y., Yoshimoto, N. &Ebinuma, T. (2005), ‘Basic research on the mechanical behaviour of methane hydrate sediments mixture’, Soils &Fou-dations., 45, pp. 75–85. Hyodo, M., Yoneda, J., Yoshimoto, N., and Nakata, Y. (2013).‘Mechanical and dissociation properties of methane hydrate bearing sand in deep seabed’, Soils and Foundation, 53(2),pp. 299-314. Itasca (2004). PFC3D: Particle flow code. User’s guide, version 4.0.,Minneapolis, USA. Iwai, H; Kimoto, S; Akaki, T & Oka, F (2015_, ‘Stability analysis of MethabeHydrate_Bearing soils considering dissociation, Energies, 8, pp. 5381-5412. Jiang, M., He, J., Wang, J., Chareyre, B., and Zhu, F. (2016). ‘DEM Analysis of Geomechanical Properties of Cemented Methane Hydrate–Bearing Soils at Different Temperatures and Pressures’,Int. J. Geomech., 16(3): Jiang, M., Sun, Y. G. and Yang, Q. J. (2013), ‘A simple distinct element modeling of the mechanical behavior of methane hydrate-bearing sediments in deep seabed’, Granular Matter, 15 (2), pp. 209-220. Jiang, M., Zhu, F., Liu, F. and Utili, S. (2014), ‘A bond contact model for methane hydrate-bearing sediments

with interparticle cementation’, International Journal for Numerical and Analytical Methods inGeomechanics.38(17),1823–1854. Jung J.W., Santamarina J.C., Soga K. (2012). ‘Stress-strain response of hydrate-bearing sands: numerical study using discrete element method simulations’. Journal of Geophysical research, 117, pp.1-12. Kvenvolden, K. A. (1988), ‘Methane hydrate - A major reservoir of carbon in the shallow geosphere?’Chemical Geology, 71, pp. 41 - 51 Kvenvolden, K. A., and T. D. Lorenson (2001), ‘The global occurrence of natural gas hydrates, in Natural Gas Hydrates: Occurrence, Distribution, and Detection’, Geophys. Monogr.Ser., Edited C. K. Paull, & W. P. Dillon, Washington, D. C. 124, pp. 3–18. Masui, A., Haneda, H., Ogata, Y. & Aoki, K. (2005), ‘The effect of saturation degree of methane hydrate on the shear strength of synthetic methane hydrate sediments’.5th International Conference on Gas Hydrates, Trondheim, Norway, pp. 657–663. The Economic Times (2013), Tapping massive deposits of ‘fire ice’ methane hydrate can change India’s energy landscape’ Report by Aiyar, S SAhttp://articles.economictimes.indiatimes.com/2013-03-17/news/37787191_1_hydrate-reserves-methane-hydrate-japan-oil-gas Vinod, J S, Hyodo, M, Indraratna, B and Kajiyama, S (2014a) ‘Shear behaviour of methane hydrate bearing sand: DEM simulations’, International Symposium on Geomechanics from Micro to Macro, IS-Cambridge 2014, pp. 355-359. Vinod, J.S. Hyodo, M; Indraratna, B & Miller, R (2014b), ‘DEM Modelling of Methane Hydrate Bearing Sand,Australian Geomechanics Journal, 49(4): pp. 175-182 Waite W.F., Winters W.J.,& Mason D.H. (2004). ‘Methane hydrate formation in partially water-saturated Ottawa sand’. American Mineralogist, 89(8-9), pp. 1202-1207. Waite,W.F., Santamarina, J.C., Cortes, D.D., Dugan, B., Espinoza, D.N., Germaine, J., Jang, J., Jung, J.W., Kneafsey, T., Shin, H.S., Soga, K., Winter, W., Yun, T.S. ‘Physical properties of hydrate bearing soils’.Rev. Geophys. 47. pp. 1-38.