Elsevier

Energy

Volume 119, 15 January 2017, Pages 121-137
Energy

An experimental evaluation of unique CO2 flow behaviour in loosely held fine particles rich sandstone under deep reservoir conditions and influencing factors

https://doi.org/10.1016/j.energy.2016.11.144Get rights and content

Highlights

  • Effective factors on CO2 migration in fine rich deep sandstone formations are discussed.

  • Significant permeability reduction occurs in sandstone due to fine particle migration.

  • Injection pressure has most effect on CO2 injectivity in fine-rich sandstone formations.

  • SEM analysis are conducted to observe the pore structure variations upon CO2 injection.

  • CO2 migration can be ceased at some stages due to pore throat clogging by migrated fine particles.

Abstract

Lack of understanding of CO2 flow behaviour in loosely bonded fine particles (clay and mineral fragments) rich sandstone formations has limited the optimum usage and the operational efficiency of various CO2 injection-related field applications in these formations. A comprehensive experimental study including core flooding tests, XRD and SEM image analysis was therefore conducted precisely to understand the CO2 flow behaviour in sandstone formations rich with loosely bonded clay and detrital particles. 210 mm long sandstone cores obtained from the Marburg Formation, eastern Australia were flooded with CO2 at a range of temperatures (24–54 °C) and confining pressures (10–60 MPa). Pressure developments along the cores were monitored to identify the fluid migration patterns through the samples. According to the results, CO2 permeability in tested sandstone has a high tendency to decrease with increasing injection pressure, depth (confining pressure) and temperature. Increased confining pressure and temperature caused 40–50% and 10–30% reductions in the CO2 permeability. This is because the permeability of fine-rich sandstone is highly affected by fine particle migration associated increased flowing fluid viscosity, pore shrinkage with fine clay particle accumulations and easy compaction of soft clay minerals. Moreover, the closure of micro-cracks under high confining stresses, CO2 adsorption created by clay swelling, the occurrence of electric double layers around clay minerals and a reduced CO2 slip effect are also affect the permeability reduction. Many of these effects were identified in the micro-scale study conducted using SEM image analysis. Interestingly, the injection of CO2 at higher pressures (>6 MPa) caused the pressure development in the sample to be held for a significant time due to the blocking of CO2 flow by the accumulation of transported clay particles in pores. This pressure holding period lasts until sufficient pressure development occurs at the upstream side of the barrier to initiating a fluid flow by breaking that barrier. The findings of the study will be very useful for advances in numerical modelling and analytical equations and worldwide CO2 geosequestration projects in fine-rich sandstone aquifers.

Introduction

Anthropogenic carbon dioxide has become the primary greenhouse gas, with its rapid increase in the atmosphere over the past several decades creating severe issues for human beings, such as rising sea levels, unexpected climate changes, melting of snow covers and lowering of ground water levels [1]. According to current records, CO2 accounts for more than 60% of the world's total greenhouse gas emissions [2], indicating the importance of finding appropriate CO2 emission control techniques. Geo-sequestration of carbon dioxide, the storage of carbon dioxide in deep geological formations, is considered to be one of the most feasible methods due to the widely available CO2 sinks underground and their substantial storage capacities [3], [4]. According to current research, the most preferable geological conditions for safe CO2 storage are at around 600 m–3000 m below the earth's surface [5], where the pressure and temperature increase at gradients of about 22–27 MPa/km and 20–30 °C/km [6], [7], [8] respectively from the earth's surface. The physical state of CO2 depends on pressure and temperature and therefore may vary with changing depth. For example, CO2 density may increase from 150 kg/m3 to 900 kg/m3 at depths between 500 m and 1500 m [9], [10], mainly caused by the change of its physical state from sub-critical gas to super-critical CO2 (when the temperature goes beyond 31.8 °C and pressure goes beyond 7.35 MPa) or gas to liquid under sub-critical temperatures (<31.8 °C). Below super-critical conditions, CO2 may exist as either a gas or a liquid, depending on the relative position in the P-T phase diagram [10]. Since high density is essential for optimum CO2 storage in sedimentary formations, the injection of high-pressure CO2 (either liquid or super-critical) is preferred in geosequestration processes due to the high density values compared to gaseous CO2 [11]. Injected CO2 may exist either as a sub-critical gas, super-critical gas or a sub-critical liquid, depending on the reservoir conditions [8].

CO2 flow behaviour in reservoir rock after injection is very important for successful CO2 geo-sequestration, and the term “permeability” or “intrinsic permeability” can be used to quantify it. Permeability of a porous media is largely controlled by its pore dimensions, principally the pore aperture or width. According to current research, the permeability of deep reservoir rock may vary from few micro-darcies to several thousands of milli-darcies, depending on the rock type, the depth of the reservoir and the geological position of the reservoir [12], [13]. A number of empirical correlations can be used to predict permeability using sample porosity. One such basic correlation can be generally expressed as [14]:k=m2k0ϕ3where, k is permeability, ϕ is porosity, m is hydraulic radius and k0 is a dimensionless constant, which may vary between 2 and 3 [15]. According to Eq. (1), porosity is dependent on the microstructure of the rock matrix and greatly affects the rock mass permeability. Rock porosity more likely changes with confining stress and temperature due to the associated compaction and expansion of mineral grains and the corresponding pore structure modification [16]. According to the Kozeny equation [17], not only porosity (ϕ) but permeability (k) also has a direct relationship with specific surface (S), the ratio between the internal surface area of pores and rock, as follows:k=cS2ϕ3where, c is a geometry-dependent function that can be expressed as follows [18]:c=(4cos(13arccos(ϕ82π31)+43π)+4)1

According to Eq. (2), a greater surface area causes permeability reductions in rock. This is one of the main reasons for clay-rich sandstones having quite low permeability values compared to other sandstones (clay minerals generally have high surface areas). Some researchers have modified Eq. (2) and obtained Eq. (4) [19] by considering the effect of tortuosity (τ): non-straight flow paths due to uneven distribution of pores, on rock mass permeability.k=ϕ32τ(1ϕ)2S2

However, most of permeability predicting equations have been derived for homogeneous material and need to be modified when employing permeability prediction in heterogeneous materials. Moreover, interactions between rock-grains and injection fluid also significantly alter the pore structure of reservoir rock in various ways, such as moisture-adsorbing swelling in clay minerals creating pore reduction [20] and possible chemical interactions between rock minerals and acidic gases like CO2 [17]. The prediction of reservoir rock permeability has become more difficult due to the widely existing heterogeneities in these rocks including the presence of micro-cracks and fractures [21], the existence of authigenic quartz and authigenic clay depositions [22], various pore throat size distributions and low permeable strata (well-compacted clay layers) at different lithification stages [23]. As a result, low permeability may exist in highly porous sandstones [24], because porosity is simply a measure of the pore volume and does not explain pore sizes or their distribution along the sample. However, permeability measures the ability of fluids to pass through the pore structure, and therefore directly relates to the pore characteristics of the medium. For this reason, the existence of any obstacle to the flow, such as aforementioned heterogeneities, may have a significant influence on reservoir flow behaviour or permeability, and as a result permeability may unexpectedly change, even in a highly porous medium. On the other hand, different fluids behave differently when they traverse through rock pores. For example, the permeability obtained using sub-critical CO2 may differ from that obtained for liquid or super-critical CO2, as the movement of gas may be affected by molecular phenomena such as gas slippage [25]. The precise understanding of pore structure and, its behaviour under various pressures and temperatures for various injection fluids are therefore important if an effective injection of gas in many actual field applications is to be achieved. As a result, the estimation of flow behaviour in a selected reservoir based on laboratory experiments/numerical modelling is necessary, in the light of the extensive time and costs associated with field tests.

Current experimental studies of flow behaviour through reservoir rocks can be divided into three major categories: tri-axial tests [24], [26], [27], [28], [29], core flooding tests [30], [31] and batch autoclave tests [32]. Of these, core flooding tests have been widely used by researchers due to their unique advantages over other testing techniques [30], [31], including the ability to measure the pore pressure development over the sample, which is very important in understanding the fluid flow migration patterns along a core sample or in a reservoir. Laboratory core flooding tests conducted to date on gas injection in sandstone have used samples of limited size, with lengths up to 150 mm and an average sample length around 90 mm, and a mean core diameter around 35 mm [33], and only minor consideration has been given to the temperature effect on sandstone permeability in high confinement environments [24]. There is an increasing awareness of the need to experimentally understand rock heterogeneities and their influence on gas permeability, in order to improve the confidence of flow behaviour predictions using numerical simulations. According to drained tri-axial tests conducted by Shukla, et al. [27] in low CO2 injection pressure conditions (<5 MPa), sandstone permeability reduces with increasing confining pressure and injection pressure. However, high injection pressures may cause the sandstone pore structure to be altered, having a direct influence on sandstone permeability. Nasvi, et al. [26] observed reductions in sandstone permeability for a wide range of CO2 injection pressures (6–20 MPa), with increasing injection pressure, and the reduction was more significant when CO2 became denser. According to these researchers, this is related to Klinkenberg's slip flow effect [34]. Contradictory results have also been observed by some researchers [35] where higher permeability for liquid (water) compared to gas (nitrogen and methane) in sandstone. According to their test results, gas permeability in sandstone decreases with increasing injection pressures, and in contrast, water permeability increases with increasing injection pressure. Dong, et al. [29] found a 10–20% porosity reduction in sandstone when confining stress was increased from 3 MPa to 120 MPa and a reduction in permeability accordingly. This is consistent with the findings of Casse and Ramey Jr [36], who also observed a permeability reduction in sandstone regardless of gas or liquid movement with increasing confining pressure. Ghabezloo, et al. [37] showed a more significant influence of pore pressure variation on sandstone permeability compared to confining pressure. According to Gray and Rex [38], swelling of clay minerals and migration of clay particles with the flow cause sample permeability to change over time.

Although there have been a sufficient number of studies to date on rock/soil flow characteristics under low temperatures and pressures, few detailed experimental studies have been reported on CO2 flow behaviour in sandstone with rock heterogeneities under a wide range of temperatures and pressures. However, it is quite difficult to precisely investigate the fluid flow behaviour through rocks under transient conditions using the frequently-used short rock cores (<120 mm), due to the sudden pressure build-up and the corresponding unsteady flow behaviour in sandstone immediately after gas injection. The aim of the present study is therefore to fill the research gaps by using a highly precise core flooding apparatus, which can be used to test rock specimens up to 300 mm in length under high pressure and temperature conditions. A series of core flooding tests was carried out using 210 mm long clay-rich sandstone cores with high amounts of free clay minerals and detrital particles over a range of confining pressures (10–60 MPa), temperatures (24–54 °C) and injection pressures (2–7 MPa). CO2 flow behaviour in sandstone was predicted using the pore pressure development data acquired at intermediate points located along the test sample (at 0, 70, 140 and 210 mm from upstream) over the time. Micro-scale observations were also conducted using scanning electron microscope (SEM) images to understand the effect of pore-scale changes on flow behaviour. This study therefore provides a detailed understanding of CO2 migration in sandstone by studying sandstone flow characteristics (permeability and pore pressure development) and their dependence on factors such as confining pressure, pore pressure and temperature. Moreover, this study promises to improve the quality and confidence of flow simulations in worldwide gas injection projects and this paper discusses subcritical CO2 flow behaviour under injection pressures up to 7 MPa and CO2 flow behaviour under injection pressures above 7 MPa (liquid and super-critical CO2 injection) will be presented in a separate paper.

Section snippets

Sample description

Warwick sandstone, the type of sandstone used in this study, belongs to the early Jurassic age and was obtained from the Marburg formation in the Clarence Moreton Basin, one of the largest basins in Australia that spreads across the New South Wales-Queensland border in eastern Australia [39]. The Warwick sandstone is yellowish brown in appearance, fine-medium size in grains and moderately-cemented type with high clay content. The 300 × 300 × 300 mm blocks obtained from the site were drilled

A micro-scale examination of natural Warwick sandstone to identify the influence of the sandstone's high clay content on its flow performance

According to Eqs. (1), (2), there is a direct relationship between rock porosity and its permeability, and higher permeability conditions can be expected for high porosities of any rock. However, as mentioned earlier, any obstruction to pore geometry or pore surface properties (texture, surface area, angle of pores to the flow direction) may significantly affect rock permeability [45], [50]. A comprehensive pore-scale examination of a porous medium can therefore be employed to understand the

Conclusions

A comprehensive experimental study was carried out to investigate CO2 flow behaviour in clay-rich Warwick sandstone (>20% clay) under sub-critical (gas, liquid) and super-critical CO2 injections. This paper presents the results of gas CO2 injection into a 38 mm diameter and 210 mm long Warwick sandstone core. Flow behaviour under liquid and super-critical CO2 injection will be discussed by a separate paper. For the present study, a sandstone core was flooded with CO2 and N2 using a

Acknowledgements

This research project is funded by the Australian Research Council (ARC) grant number DP120101761 and the authors would like to thank all the Deep Earth Energy Laboratory staff at Monash University, Clayton campus, Australia and the Monash Centre for Electron Microscopy (MCEM), who dedicated their time and energy to bring this experimental series to a successful conclusion.

References (62)

  • L. Liu et al.

    CO2 injection to granite and sandstone in experimental rock/hot water systems

    Energy Convers Manag

    (2003)
  • S. Ghabezloo et al.

    Effective stress law for the permeability of a limestone

    Int J Rock Mech Min Sci

    (2009)
  • G.P.D. De Silva et al.

    Effect of bedding planes, their orientation and clay depositions on effective re-injection of produced brine into clay rich deep sandstone formations: implications for deep earth energy extraction

    Appl Energy

    (2016)
  • J. Baudracco et al.

    Permeability variations in Berea and Vosges sandstone submitted to cyclic temperature percolation of saline fluids

    Geothermics

    (1995)
  • E. Rosenbrand et al.

    The effect of hot water injection on sandstone permeability

    Geothermics

    (2014)
  • J. Ochi et al.

    Permeability decrease in sandstone reservoirs by fluid injection: hydrodynamic and chemical effects

    J Hydrol

    (1998)
  • Y. Yang et al.

    Influence of lithology and compaction on the pore size distribution and modelled permeability of some mudstones from the Norwegian margin

    Mar Pet Geol

    (1998)
  • P.G. Ranjith et al.

    A new triaxial apparatus to study the mechanical and fluid flow aspects of carbon dioxide sequestration in geological formations

    Fuel

    (2011)
  • IPCC

    Climate change 2013

  • M.B. Dusseault et al.
  • W.D. Gunter et al.

    The role of hydrogeological and geochemical trapping in sedimentary basins for secure geological storage of carbon dioxide

    Geol Soc Lond

    (2004)
  • N. Vandenberghe et al.

    The Merksplas-Beerse geothermal well (17W265) and the Dinantian reservoir

    Geol Belg

    (2000)
  • S. Angus et al.

    Thermodynamic tables project: international thermodynamic tables of the fluid state: carbon dioxide

    (1973)
  • S. Bachu

    Screening and ranking of sedimentary basins for sequestration of CO2 in geological media in response to climate change

    Environ Geol

    (2003)
  • C.E. Neuzil

    How permeable are clays and shales?

    Water Resour Res

    (1994)
  • W.F. Brace

    A note on permeability changes in geologic material due to stress

    Pure Appl Geophys

    (1978)
  • W.F. Brace

    Permeability from resistivity and pore shape

    J Geophys Res

    (1977)
  • J. Park et al.

    Physical property changes of sandstones in Korea derived from the supercritical CO2-sandstone‒groundwater geochemical reaction under CO2 sequestration condition

    Geosci J

    (2015)
  • M.R.J. Wyllie et al.

    Application of the Kozeny equation to consolidated porous media

    Nature

    (1950)
  • J. Mortensen et al.

    The relation among porosity, permeability, and specific surface of chalk from the Gorm field, Danish North Sea

    SPE Reserv Eval Eng

    (1998)
  • A. Costa

    Permeability-porosity relationship: a reexamination of the Kozeny-Carman equation based on a fractal pore-space geometry assumption

    Geophys Res Lett

    (2006)
  • Cited by (32)

    • Particle detachment in reservoir flows by breakage due to induced stresses and drag

      2023, International Journal of Rock Mechanics and Mining Sciences
    • Microscopic remaining oil initiation mechanism and formation damage of CO<inf>2</inf> injection after waterflooding in deep reservoirs

      2022, Energy
      Citation Excerpt :

      With the expansion of oil and gas exploration into deep reservoirs, internationally defined as buried depths of more than 4500 m, it has become the main growth of proved reserves in the early 21st century [6,7]. Injection CO2 into deep reservoirs is an effective and innovative CUS technology that not only improves oil production and economics, but also provides better CO2 storage, slowing the growth of global carbon emissions [8–10]. Laboratory and field observations have confirmed that CO2 is a potential displacement agent for deep reservoirs [11–15].

    View all citing articles on Scopus
    View full text