Reactive Pore Network Modeling Dedicated to the Determination of the Petrophysical Property Changes while Injecting CO2

Abstract

A Pore Network Model is an efficient tool to account for phenomena occurring at the pore scale. Its explicit three-dimensional network of pores interconnected by throats enables to easily consider the topology and geometry effects on upscaled and homogenized petrophysical parameters. In particular, this modeling approach is appropriate to study the rock/fluid interactions. It can provide quantitative information both on the effective transport property modifications due to the reactions and on the structure evolution resulting from dissolution/precipitation mechanisms. The developed model is based on the resolution of the macroscopic reactive transport equation between the nodes of the network. By upscaling the results, we have then determined the effective transport properties at the core-scale. A sensitivity study on reactive and flow regimes has been conducted in the case of single-phase flow in the limit of long times. It has been observed that the mean reactive solute velocity and dispersion can vary up to one order of magnitude compared to the tracer values because of the concentration profile heterogeneity at the pore scale resulting from the surface reactions. As for the reactive apparent coefficient, when the kinetics is limited by the mass transfer, it can decrease by several orders of magnitude with regard to the one calculated by the usual perfect mixing assumption. That is why scale factors should be added to the classical macroscopic transport equation implemented in reservoir simulator to accurately predict the reactive flow impacts. For each study case we have also obtained the permeability variation versus the porosity evolution in a physical way which accounts for reactive transport conditions. It appears that the wall deformation pattern and its impact on petrophysical properties must be explained by considering both microscopic and macroscopic scales of the reactive transport, each one governed by a dimensionless number comparing reaction and transport characteristic times. This work contributes to improve the understanding of surface reactions impacts on reactive flow in one hand, and on permeability and porosity modifications in the other hand. Using the PNM approach, scale factors parameters and permeability versus porosity relations can be determined for various rock-types and reactive flow regimes. Once integrated as inputs in a reservoir simulator, these relations are a powerful and convenient means to enhance the modelling accuracy of the petrophysical properties evolution during a reactive fluid injection such as CO2-rich brines. Introduction The CO2 geological storage is considered as a solution to reduce the greenhouse gas emission due to human activities. However, the carbon dioxide is not inert. Its dissolution in water perturbs the thermodynamical and chemical equilibrium, which may cause rock dissolution or precipitation. An induced modification of the pore-size distribution impacts the petrophysical properties of the porous medium, in particular the permeability and the porosity. But, as the pore-size distribution and the wettability of the new rock/fluid interface evolves, other properties, such as capillary pressure or relative permeabilities, can also vary. In the aim of predicting the CO2 migration and the risks of leakage or well-injectivity damage, operators need numerical simulators able to account for the reactive transport specificities. On this base the site selection will be made and the optimal flow rate will be estimated. To improve the simulation reliability, the reactive flow understanding and modeling are key-points. To simulate the reactive transport at Darcy-scale, modifying the advection-dispersion equation by only adding a sink/source term, usually determined by a 0D geochemical module, is not sufficient. Firstly, such a chemical model, based on the "perfect mixing" concept, calculates the chemical equilibrium by using mean concentrations whereas dissolution and precipitation depend on the concentration at the rock/fluid interface. Secondly, the parameters of the macroscopic transport equation are modified when surface reactions occur: the velocity and the dispersion of a reactive solute are different from the values of an inert tracer, since the pore-scale concentration profile is not uniform any more. Thus, scale factors must be added to the macroscopic transport equation initially developed for non-reactive species (Shapiro and Brenner, 1988; Bekri et al., 1995).

Quantification and Prediction of the 3D Pore Network Evolution in Carbonate Reservoir Rocks

“…The pore throats are modeled by basic geometries, such as cylindrical, triangular and elliptical capillaries, whereas the pore-bodies are represented by spheres or cubes. Thus, in each element (pore body or throat), the reactive transport can easily be upscaled and the scale factors determined, according to the microscopic dimensionless numbers of Peclet and Peclet-Damköhler (Algive et al, 2009b). The used Reactive Pore Network Model can be divided into three main tasks.…”

“…PNM-R computations can then provide macroscopic numbers at the core scale (see Algive, 2009b) to help in improving reactive transport modeling during reservoir studies. As for the pore scale, macroscopic coefficients representing apparent reaction rate, velocity vector and dispersive tensor of the solute can be defined.…”

“…This approach provides a way of correctly describing the pore scale phenomena by using well-known continuum equations. The problem can then be upscaled to the core scale by calculating effective and global properties (Algive et al, 2009b). The Pore Network Model is a conceptual representation of a porous medium.…”

“…The local concentration profile and the pore structure variations must be taken into account when solving the governing equations. Reactive-PNMs (PNM-R) have been applied to predict porosity-permeability changes during and following the injection of CO 2 or its geological storage in aquifers (Li et al, 2006;Algive et al, 2009b;Bemer and Lombard, 2010;Izgec et al, 2010). In contrast, the application and validation of PNM-R in studies of the evolution of reservoir rock properties during early and late diagenetic processes, remain yet poorly explored, though of interest for reservoir modeling.…”

Impact of Diagenetic Alterations on the Petrophysical and Multiphase Flow Properties of Carbonate Rocks Using a Reactive Pore Network Modeling Approach

“…This modeling approach is flexible, and can account for various phenomena occurring at the pore scale. Algive et al in 2007 and2009 proposed the PNM approach to study the mineral dissolution and precipitation caused by CO 2 sequestration. They developed a PNM tool, dedicated to reactive flow modeling, which has been validated by experimental data.…”