Elsevier

Computers & Geosciences

Volume 98, January 2017, Pages 9-20
Computers & Geosciences

Research paper
Reactive transport in porous media for CO2 sequestration: Pore scale modeling using the lattice Boltzmann method

https://doi.org/10.1016/j.cageo.2016.09.008Get rights and content

Highlights

  • A geochemical numerical model for pore scale reactive transport is proposed.

  • 3D digital sandstone sample reconstruction using Micro-CT/QEMSCAN scans.

  • Multi-mineral surfaces with complex contacts and different reactivities.

  • Validated with laboratory CO2-water reaction of Hutton Sandstone core.

  • Predicted 10% porosity increase by calcite dissolution at equilibrium reaction status.

Abstract

Injection of CO2 subsurface may lead to chemical reactivity of rock where CO2 is dissolved in groundwater. This process can modify pore networks to increase or decrease porosity through mineral dissolution and precipitation. A lattice Boltzmann (LB) based computational model study on the pore scale reactive transport in three dimensional heterogeneous porous media (sandstone consisting of both reactive and non-reactive minerals) is described. This study examines how fluid transport in porous materials subject to reactive conditions is affected by unsteady state local reactions and unstable dissolution fronts. The reaction of a calcite cemented core sub-plug from the Hutton Sandstone of the Surat Basin, Australia, is used as a study case. In particular, the work studies the interaction of acidic fluid (an aqueous solution with an elevated concentration of carbonic acid) with reactive (e.g. calcite) and assumed non-reactive (e.g. quartz) mineral surfaces, mineral dissolution and mass transfer, and resultant porosity change. The proposed model is implemented in our custom LBM code and suitable for studies of multiple mineral reactions with disparate reaction rates. A model for carbonic acid reaction with calcite cemented sandstone in the CO2-water-rock system is verified through laboratory experimental data including micro-CT characterization before and after core reaction at reservoir conditions. The experimentally validated model shows: (1) the dissolution of calcite cement forms conductive channels at the pore scale, and enables the generation of pore throats and connectivity; (2) the model is able to simulate the reaction process until the reaction equilibrium status is achieved (around 1440 days); (3) calcite constituting a volume of around 9.6% of the whole core volume is dissolved and porosity is consequently increased from 1.1% to 10.7% on reaching equilibrium; (4) more than a third of the calcite (constituting 7.4% of the total core volume) is unaffected, which suggests that this calcite is not connected with open pores (at the resolution of the model) that the acidic fluid can access. The model enables exploration of the porosity change in systems as they react, which has applications for analysis of the induced permeability change at the macroscale.

Introduction

CO2 geosequestration is widely recognized as an effective means to mitigate greenhouse gas emission and climate change for long-term environment safety. Even where the geological subsurface is well defined, however, many uncertainties still exist regarding the injected CO2 plume behavior, due to the difficulties in evaluating fluid dynamics within reacting systems. Research and prediction of reactive transport in heterogeneous porous media is topical and motivated by urgent demands in both natural resources management and industrial exploration (Gaus et al., 2005, Johnson et al., 2004; Massarotto et al., 2010; Sahimi, 2011).

There has been much observation and analysis of macroscale phenomena in systems disturbed by CO2 injection for geosequestration, but the underlying detail of the small-scale processes that dominate behavior, including transport of dissolving and precipitating species, remains challenging. This is due to the fine-scale complexity and heterogeneity of porous media, and the complicated nature of acidic fluid processes at different conditions of temperature, pressure and chemistry. Changes in pore scale flow features lead to significant variation of macroscopic phenomenon, and the potential for creation of permeability and porosity around the wellbore area as CO2 injection proceeds. This requires understanding of the pore scale processes, and development of methods to exchange, aggregate and disaggregate information across different scales.

To date, studies on the mechanisms of pore scale geochemical transport have been conducted through laboratory experiments and mathematical models. At pore scale, the dissolution of solid-phase reactants in heterogeneous porous media first occurs as surface reactions, and change the pore scale geo-structures by opening additional conductive channels and generating new or different pore throats and connectivity. Thus, porosity enhancement in reservoir rocks has been observed (Egermann et al., 2005, Farquhar et al., 2013, Farquhar et al., 2015; Pearce et al., 2012). Conversely, previously dissolved minerals (e.g. ferroan carbonates) may also be precipitated in response to chemical changes such as pH buffering, leading to occlusion of the pore throats and associated porosity reduction (Baker et al., 1995; Farquhar et al., 2015; Steefel and Lasage, 1994; Strazisar et al., 2009; Wachi and Jones, 1991). Since it takes quite a long time (~years) for geochemical systems to reach the reaction equilibrium, this makes it almost impossible for laboratory experiments to observe the whole process in tractable research time frames. Effective modeling of such coupled processes at pore scale is hence essential for predicting porosity and permeability variation through the whole sequestration sequence (Gaus et al., 2008, Massarotto et al., 2010).

Amongst numerical methods that have been proposed and applied to simulating pore scale fluid flow in the heterogeneous and fractured porous media, the lattice Boltzmann method (LBM) provides some important advantages. Gao et al., 2014, Gao et al., 2015 studied the pore scale fluid dynamics modeling through complex tortuosity with multiple mineral surface contacts in porous media. Huang et al. proposed a regularized boundary condition for solving convection-diffusion equations (Huang et al., 2011) and presented a comprehensive review of all popular multiphase Lattice Boltzmann Methods developed thus far (Huang et al., 2015). Meakin and Tartakovsky (2009) have done research on the pore scale multiphase fluid flow and reactive transport in the fractured and porous media, and Steefel et al. (2013) have focused on the pore scale process associated with the subsurface CO2 injection and sequestration (Steefel et al., 2013). Kang et al., 2010a, Kang et al., 2010b developed a pore scale LBM model for simulating the injection of CO2-saturated brine into structured porous media, focusing on the fundamental physics occurring at the pore scale for reactions involved in geologic CO2 sequestration. Recently, a further study in pore scale reactive transport was also done by Yoon and Kang (Yoon et al., 2015). A drawback of these studies, however, is the lack of model validation against experiment.

The current paper considers the acidification process in carbonate or carbonate cemented reservoirs using a LBM model and incorporating experimental validation. We present a generalized pore scale geochemical reaction computational model, for studying complex geochemical reaction systems, including mechanisms of: mutually soluble acidic fluid transport in pore networks within bulk structures; acidic fluid interactions on reactive and less-reactive mineral surfaces; dissolution/precipitation and mass transfer of the reactive minerals; and the impact of temperature on the reactivity and potential chemical reactions. In the model, simulations of reactive transport have considerable complexity due to the geometric heterogeneity and minerals diversity, contributing different reactivities and distributions that occur in reality.

The Surat Basin in Queensland, Australia (Fig. 1), has been identified as a highly prospective site for CO2 storage, estimated to have a capacity of 2962 Mt (Hodgkinson et al., 2010). The Precipice Sandstone, a low salinity aquifer, is a target reservoir for CO2 injection (or brine reinjection from coal seam gas operations) with a mean reported porosity of 16±5% (Kellett et al., 2012). The Evergreen Formation is an aquitard, overlying the Precipice Sandstone, with variable porosity, which would act as an unconventional seal or cap-rock. The overlying Hutton Sandstone (Fig. 1) that may act as a secondary seal, has been suggested as a reservoir in its own right, or represents an area for sub-surface monitoring for potential CO2 leakage from the underlying Precipice Sandstone (Hodgkinson et al., 2010, Horner et al., 2015). Mean porosity in the Hutton Sandstone is 22±6% as reported (Kellett et al., 2012). Several geochemical studies have recently reacted core from all three formations from the GSQ Chinchilla 4 well (Fig. 1) with CO2 and water or CO2-SO2 and water at simulated reservoir conditions (Farquhar et al., 2015, Horner et al., 2015; Pearce et al., 2015). The experimental data used for model verification here are from a laboratory CO2-water-rock experiment using Hutton Sandstone core (Farquhar et al., 2015). The focus of this case is on the dissolution reactions that are likely to occur in the immediate wellbore environment (Bacci et al., 2011, Massarotto et al., 2010). This region is the most influential with respect to controlling the possible injection rates, since the fluid flux here is highest.

Section snippets

Methods

The subsurface system typically encompasses multiple minerals, some of which are reactive, and a complex network of flow paths. Thus, a model must encompass many aspects including fluid transport, the representation of tortuous physical geometries with heterogeneous chemical surfaces, the fluid-solid interactions and chemical reactions, and the coupling of these multiple processes. As introduced in our previous work (Gao et al., 2014, Gao et al., 2015), fluid transport model (including fluid

Materials for experimental validation

The model was compared with a laboratory experiment carried out previously and reported elsewhere, which studied the pore scale geometry and geochemistry changes of a fresh water wetted Hutton Sandstone core sample sub-plug when exposed to CO2 injection at reservoir conditions of 60 °C and 120 bar (Farquhar et al., 2013, Farquhar et al., 2015). The experimental batch-soak reactor has been described in detail in Pearce et al. (2015). The core sample sub-plug was reacted over a period of 384 h, and

Specialized carbonate acid reaction LBM model

In the CO2-water-rock application, the reaction LBM model is used to study the transport of bulk CO2-water mixture (modeled as a single-phase aqueous solution with an elevated concentration of carbonic acid), carbonic acid interactions on assumed non-reactive solid surfaces, and carbonic acid reactions on calcite surfaces and the resultant dissolution of reactive minerals. In the simulation, the temperature and pressure were set as 60 °C and 120 bar for the reaction modeling.

Generally, dry CO2 is

Discussion and implications

One of the main difficulties associated with geological reactive transport modeling is model validation, associated with the difficulties of sufficient sample characterization and experimental conditions. This also reflects the value of a good model, insofar as it provides a way to probe and gain detailed insight from different process strategies, without the laboratory burden. The current reactive model has been validated against a single, but well characterized experimental sample data set

Conclusions

A geochemical reaction LBM model for pore scale reactive transport in heterogeneous porous media has been developed and validated against some experimental data.

The following conclusions can be drawn:

  • (1)

    A general geochemical reaction model acting at the pore scale is presented that describes coupled mass transport and dissolution processes, set up to include heterogeneity in the porous media and its grain structures, fluid transport and reactive mineral dissolution, and linked with the associated

Acknowledgments

The authors wish to acknowledge financial assistance provided through Australian National Low Emissions Coal Research and Development (ANLEC R&D) Project 3-1110-0101. We also wish to acknowledge Australian Research Council (ARC) Project DP110103024, and National Natural Science Foundation of China (NSFC) Project 11232003. The authors are grateful to Dr. Alexandra Golab and FEI for access to the micro-CT data sets.

References (43)

  • J.K. Pearce et al.

    SO2 impurity impacts on experimental and simulated CO2-water-reservoir rock reactions at carbonstorage conditions

    Chem. Geol.

    (2015)
  • C.A. Peters

    Accessibilities of reactive minerals in consolidated sedimentary rock: An imaging study of three sandstones

    chem. Geol.

    (2009)
  • O.S. Pokrovsky et al.

    Dissolution kinetics of calcite, dolomite and magnesite at 25° C and 0 to 50 atm pCO2

    Chem. Geol.

    (2005)
  • O.S. Pokrovsky et al.

    Calcite, dolomite and magnesite dissolution kinetics in aqueous solutions at acid to circumneutral pH, 25 to 150° C and 1 to 55 atm pCO2: Newconstraints on CO2 sequestration in sedimentary basins.

    Chem. Geol.

    (2009)
  • O. Regnault et al.

    measurement of portlandite carbonation kinetics with supercritical CO2

    Chem. Geol.

    (2009)
  • B. Strazisar et al.

    Chemical reactions of wellbore cement under CO2 storage conditions: Effects of cement additives.

    Energy Procedia

    (2009)
  • S. Wachi et al.

    Mass transfer with chemical reaction and precipitation

    Chem. Eng. Sci.

    (1991)
  • T. Xu et al.

    Mineral sequestration of carbon dioxide in a sandstone–shale system.

    Chem. Geol.

    (2005)
  • S. Yin et al.

    Coupled THMC modeling of CO2 injection by finite element methods

    J. Pet. Sci. Eng.

    (2011)
  • J.C. Baker et al.

    Continental-scale magmatic carbon dioxide seepage recorded by dawsonite in the Bowen-Gunnedah-Sydney Basin system, eastern Australia

    J. Sediment. Res.

    (1995)
  • G.K.W. Dawson et al.

    ANLEC project 3-1110-0101: review of laboratory-scale geochemical and geomechanical experiments simulating geosequestration of CO2 in sandstone, and associated modelling studies

    Aust. Natl. Low Emiss. Coal Res. Dev.

    (2013)
  • Cited by (44)

    • Investigation on reactive flow through porous media by quadtree Lattice Boltzmann

      2022, Journal of Natural Gas Science and Engineering
      Citation Excerpt :

      More recently, Montemore et al., (2017) used LB models to investigate reactive transport in nano-porous catalysts and catalytic performances. Gao et al., (2017) proposed a geochemical numerical model for pore scale reactive transport using LB models and investigated reactive fluid transport in 3D porous media with different reactive mineral surfaces. On the other hand, there are some studies that used a combination of LBM with other numerical methods like Direct Numerical Simulation (DNS) or particle tracking (PT) for studying reactive or active transport phenomena in porous media.

    • Pore-scale modeling of complex transport phenomena in porous media

      2022, Progress in Energy and Combustion Science
    View all citing articles on Scopus
    View full text