CO2 Injectivity in geological storages: an overview of program and results of the GeoCarbone-Injectivity Project

CO2 Injectivity in Geological Storages: an Overview of Program and Results of the GeoCarboneInjectivity Project — The objective of the GeoCarbone-Injectivity project was to develop a methodology to study the complex phenomena involved in the near wellbore region during CO2 injection. This paper presents an overview of the program and results of the project, and some further necessary developments. The proposed methodology is based on experiments and simulations at the core scale, in order to understand (physical modelling and definition of constitutive laws) and quantify (calibration of simulation tools) the mechanisms involved in injectivity variations: fluid/rock interactions, transport mechanisms, geomechanical effects. These mechanisms and the associated parameters have then to be integrated in the models at the wellbore scale. The methodology has been applied for the study of a potential injection of CO2 in the Dogger geological formation of the Paris Basin, in collaboration with the other ANR GeoCarbone projects. Oil & Gas Science and Technology – Rev. IFP, Vol. 65 (2010), No. 4, pp. 533-539 Copyright © 2010, Institut français du pétrole DOI: 10.2516/ogst/2010013 CO2 Storage in the Struggle against Climate Change Le stockage du CO2 au service de la lutte contre le changement climatique D o s s i e r ogst100034_Lombard 8/07/10 16:54 Page 533


INTRODUCTION
During the lifetime of a CO 2 geological storage operation, large volumes of CO 2 will have to be injected through a minimum number of wells, for economical reasons but also to minimise leakage risks. The design of a storage will thus require the control of the injectivity: which are the phenomena and the parameters governing the injectivity, how do they vary during the injection, and how to modify them in case of an injectivity loss?
The main goal of the "GeoCarbone-Injectivity" project was to develop a methodology to study these complex phenomena involved in the near wellbore region, and to apply it to the context of the Paris Basin which is studied for site screening and characterization within the framework of the companion project "GeoCarbone-PICOREF".
The proposed methodology is based on experiments and simulations at the core scale, in order to understand through physical modelling and determination of constitutive laws, and to quantify (calibration of simulation tools) the mechanisms involved in injectivity variations. Figure 1 presents the main mechanisms which can potentially impact the CO 2 injectivity within a CO 2 geological storage operation. Three main classes of phenomena can be distinguished and justify the project organization: -fluid/rock interactions (Phase 1 of the project), -transport mechanisms (Phase 2), -geomechanical effects (Phase 3).
These mechanisms and the associated parameters have then to be integrated in the models at the wellbore scale (Phase 4).
The first three steps of the methodology were applied on samples representative of the Dogger formation of Paris Basin site: Lavoux limestone and/or Charmottes carbonate plugs, and in the different laboratories of the project partners. In-house and commercial softwares were used for the interpretation of these experiments and for simulations of CO 2 injection at the wellbore scale.

Rock/Fluid Interactions
In the first phase of the project, the influence of geochemical reactions on CO 2 injectivity was studied through batch and percolation experiments. The interpretations were performed through numerical simulations using models taking into account both transport and geochemical phenomena (Coores™and ToughReact simulators).
Experiments in batch reactors have been conducted on plugs of Lavoux oolitic limestones in presence of CO 2 + brine, N 2 + brine and supercritical CO 2 , without thermal gradient at 80°C and at 150 bar bulk pressure. Weak dissolution and/or precipitation phenomena have been detected. Analysis of water chemistry after experiments shows that less than 1% of limestone is dissolved. On the contrary, important mass transfers have been observed in the case of experiments in grain columns (COTAGES device) with thermal gradient (35 to 100°C on 70 cm). Dissolution occurred in the coldest part of the reactor whereas clear sealing calcite precipitated in the hottest part (Fig. 2). This mechanism is accompanied by strong pressure variations with time marking dissolution of calcite during the first 10 days and precipitation after. This behaviour has been qualitatively reproduced by numerical simulations (Sterpenich et al., 2009;André et al., 2009).
Whereas batch experiments with supercritical CO 2 showed only a slight modification of the pore structure due to calcite dissolution, the dissolution/precipitation phenomenon was more strongly highlighted during brine/CO 2 displacement experiments. The results of the percolation experiments performed in Lavoux and Charmottes samples clearly show the influence of the flow rate (effect of the Peclet and Damkhöler parameters) on the dissolution patterns observed by NMR and CT-scanner: dissolution with wormhole formation for high flow rates and compact dissolution for low flow rates. These observations gathered in Table 1 are in good agreement with the Peclet/Damkhöler diagram of the literature (Fig. 3) established for acid flows (Bazin, 2001;Egermann et al., 2005;Lombard et al., 2007). In most of the cases, the evolution of pressure and effluent composition during the percolation experiments as well as final porosity and permeability, were reproduced qualitatively by simulation (Coores™). Mechanisms involved in injectivity evolution. Injectivity Index II depends on flow rate and pressures. K: absolute permeability, K r : relative permeability, μ: viscosity of the injected fluid.
Nevertheless, it was observed that particle displacements may occur and modify the pressure profiles which were not reproducible by simulation. At the present time, this impact is not taken into account by the simulators (Bazin et al., 1994;Bazin, 2001) and this still represents a research challenge.

Transport Properties
Without considering the pore structure alteration due to the acidified brine, transport parameters such as capillary pressure and relative permeability are directly linked to interfacial properties such as interfacial tension (IFT) and wettability.  Effect of the thermal gradient on the mass transfer (C a CO 3 ) for the COTAGES experiment. Left: calcite dissolution at the cold side, right: calcite precipitation at the hot side. Chalbaud et al. (2009) (in this volume), present an experimental study, performed in parallel to the Injectivity project, which allowed new brine/CO 2 IFT data necessary for CO 2 storage issues to be acquired, and propose a correlation with thermodynamical conditions. Observations at the pore and core scales have also shown the possible wetting behaviour of CO 2 and the authors propose a discussion on its influence on the storage capacity of the investigated formations. Within the Injectivity project, the evolution of transport properties (single phase and two phase flows) generated by the pore structure alteration of the rock near the well was studied through the characterization of both native samples representative of the Dogger formation (Lavoux limestone) and samples altered uniformly by an acid solution (Egermann et al., 2006). The acidity has implied a slight increase of permeability and porosity and a decrease of inertial coefficients. Both relative permeabilities and Klinkenberg effects remained unchanged. When no particle displacement takes place, the injectivity of the studied rock seems not to be altered by the acid solution (Radilla et al., 2010, in this volume).
In the first phase of the project it has been shown that the structural modification patterns and the porosity/permeability variations with the reactive and flow regimes can be interpreted with Peclet and Damkhöler numbers. In parallel to this Reactive percolations on Lavoux limestone samples -PeDa diagram.  Natural trends for carbonates drained elastic moduli (Bemer et al., 2009). work, Algive et al. (2009) used these dimensionless numbers and a Pore Network Modelling approach to study the impact of such a reactive flow on transport parameters. For different flow regimes (Peclet and Damkhöler numbers) they calculated the porosity/permeability evolution due to dissolution/precipitation phenomena, and the modified two phase flow parameters: capillary pressure and relative permeability.

Geomechanical Parameters
The consequences of the injection of an acid solution on geomechanical parameters of Lavoux samples were investigated. During percolation experiments of a CO 2 saturated brine, the same dissolution patterns were observed as in the first phase, with an axial deformation of the samples. Water+CO 2 percolation experiments have shown an increase of limestone dissolution with the increase of the fluid flow. Dissolution is accompanied by an increase of the size of the "wormholes". CO 2 -saturated water, chemically equilibrated with the limestones in room conditions, dissolves calcite until the solution becomes equilibrated at P, T conditions of the experiment. Mechanical behaviour is not affected by the percolation of dry CO 2 (gaseous or supercritical). Replacement of liquid water by CO 2 at the same pressure does not induce any deformation of the samples (Grgic et al., 2010).
The poroelastic properties as well as the rupture criteria (induced fracturing) have been measured on native and altered Lavoux and Comblanchien Charmottes carbonate samples. The homogeneous alteration procedure with an acid solution was the same as in Phase 2 (Egermann et al., 2006). Clear trends of chemically induced mechanical weakening have been outlined (Fig. 4). The experimental data have been compared to the data obtained by phenomenological and empirical models (Bemer et al., 2009, in the previous OGST volume). Nevertheless additional experimental data are needed to further investigate the effects of chemical alteration on carbonates geomechanical properties and to integrate them in reactive transport simulators.

Coupling Phenomena
The information on the coupling of Thermo-Hydro-Mechanical-Chemical (THMC) phenomena was partially integrated in the simulators used by the partners (ToughReact, Coores™) within Phase 4 of the project. The effects of several types of coupling (HC, TH, THC, THM) were also tested (ToughReact) in the context of various scenarios of CO 2 injection in a generic aquifer representative of Paris Basin. 1D and 2D radial models were used. The petrophysical properties of the Lavoux limestone were considered. The initial Structured zones and physical (evaporation, overpressure, cooling and relative saturation) and reactive chemical (pH, mineral dissolution/precipitation/transformation) phenomena within different zones around the CO 2 injection well (schematic concept based on the integrated results on experiments and numerical simulations) .
pressure and temperature of the formation were 180 bar and 75°C.
In terms of Pressure-Rate-Time relationship, the modifications of CO 2 injection flow rate (temperature T = 75°C) led to only slight variations of the (stable) final pressure (Tab. 2). For a CO 2 injection at a constant flow rate of Q = 5 kg/s (T = 50°C), the Thermo-Hydro-Mechanical coupling (THM) also does not show significant overpressures . Finally Figure 5 summarizes the main phenomena occurring in the different zones around the injection well and partially highlighted within this study. The phenomena of drying and dehydration of the near wellbore region were not studied within the Injectivity project. The observed displacement of fines is also not yet taken into account in the simulators and require further research work.

CONCLUSIONS AND PERSPECTIVES
This Injectivity project objective was to develop a methodology to study the complex phenomena involved in the near wellbore region during CO 2 injection. It allowed new experimental tools to be developed and first sets of geochemical, petrophysical and mechanical experimental data to be acquired before and after CO 2 injection in the Dogger Paris Basin formation: -the dissolution/precipitation phenomena linked to the acidity of the brine/CO 2 solution (not equilibrated with the rock) have been outlined and can be studied through dimensionless numbers. In presence of water, equilibrated with the rock and saturated with CO 2 , dissolution rates and pore rearrangements are weak, at the detection limit of the analytical techniques; -the measured interfacial and transport parameters are presented in Chalbaud et al. (2009) and in Radilla et al. (2010). When no particle displacement takes place, the injectivity of the studied formation is not altered by the acid solution. A Pore Network Modelling Approach should allow a better prediction of the evolution of pore structure and transport parameters during a CO 2 injection; -clear trends of chemically induced mechanical weakening have been outlined on altered Lavoux and Charmottes samples (Bemer et al., 2009); -within the simulation of various scenarios of CO 2 injection in a generic aquifer representative of Paris Basin, several types of coupling (HC, TH, THC, THM) were tested. For this aquifer, large overpressures were never obtained throughout the simulations. In order to improve the interpretation models and for a better implementation in the reactive transport simulators, the experimental data would need to be completed.
The influence of particle displacements which has been observed is not yet well taken into account and should be studied within further research works. Furthermore, drying and dehydration are two other important phenomena occurring near the injection well which were not considered in this project. They are studied within the "Near Well" ANR project (2008)(2009)(2010). The influence of impurities in CO 2 on geochemical and dissolution/precipitation mechanisms also has to be better understood. This is one of the goals of the "Annex Gas" ANR project (2007)(2008)(2009)(2010).

ACKNOWLEDGMENTS
The GeoCarbone-Injectivity project was funded by the CO 2 programme of the French National Agency for Research (ANR). It was supported by a consortium of companies (GDF Suez, Geostock, Total) and research institutions (IFP, BRGM, INPL and LFC). The authors thank Toreador for providing rock samples and all the researchers who contributed to this work.