Experimental and numerical investigations of geomechanical controls on petrophysical changes of carbonates during fluid flow

Abstract

In flooding experiments, porosity and permeability of carbonate rocks is enhanced through the dissolution of the rock matrix, which further increase the permeability as well as the inter-connection of the pre-existing porosity. Authors often refer to this process as wormholing or channelling, which define preferential pathways for any fluid circulating through the rock’s matrix (Hoefner and Fogler, 1988; Fredd and Fogler, 1998; Golfier et al., 2002). A wormhole’s shape and size ranges from face dissolution at very low fluid flow rates (where the reactive fluid is rapidly consumed after the injection point), to uniform dissolution (where the acid is brought to far-ends within the rock matrix, which allows the creation of a large network of connected pores). Authors have studied the factors influencing the relationship between dissolution fronts, injection rate, rock nature, and acidity of the circulating fluid (Frick et al., 1994a; Bazin et al., 1995; Fredd et al., 1996; Fredd and Fogler, 1998; Golfier et al., 2002; Egermann et al., 2006; Luquot and Gouze, 2009; Menke et al., 2015; Ott and Oedai, 2015; Barri et al., 2016; Luquot et al., 2016; Teles et al., 2016; Zhang et al., 2016). The current status of knowledge present strong connections between the reaction rate and the diffusion rate (referred to as the Damköhler number – Da Bekri et al. (1995); Egermann et al. (2010)), as well as the study between the fluid velocity and the ability for a medium to diffuse into a solvent (referred to as the Péclet number – Pe Golfier et al. (2002); Menke et al. (2015)). The Da number measures the relative importance of the reaction rate constant versus advection over some length scale, while the Pe number gives the ratio of advective to dispersive flux for a given length scale (Zhang and Kang, 2004; Steefel and Lasaga, 1990). Large Da correspond to rapid chemical reaction in comparison to all other processes. On the other hand, smaller Da testify of very slow chemical reactions in comparison to all other processes taking place during fluid flow (Zhang and Kang, 2004). A low Péclet number suggests that transport is governed by diffusion and not by convection, and inversely at high Pe number (De Boever et al., 2012).

Along with these dynamically controlled numbers, studies have tried to unpick the relationship between rock-fluid interaction for a variety of injection fluid, as well as rock-stress interaction. These studies have been done through the analysis of key variables (resistivity, porosity, permeability, etc.), using acidic and non-acidic fluids (Hoefner and Fogler, 1988; Frick et al., 1994b; Bazin et al., 1995; Fredd et al., 1996; Fredd and Fogler, 1998; Golfier et al., 2002; Egermann et al., 2006; Luquot and Gouze, 2009; Menke et al., 2015; Ott and Oedai, 2015; Barri et al., 2016; Luquot et al., 2016; Teles et al., 2016; Zhang et al., 2016). The experimental rationales of these studies usually implies large changes in the variables representing the reservoirs conditions, such as the temperature, confining pressure, and the effective stress. A large gap has been found between the actual state of knowledge and the absolute impact of effective stress on reservoir rock alteration, at steady reservoir condition of pressure and temperature. In this study, we have created an experimental matrix where the variable representing the reservoir conditions are kept constant during an experimental flooding, while varied between experiments. By doing so, we can isolate and cross-compare the effect of each variable on the rock alteration. We have flooded a total of twelve 38 mm large diameter carbonate cores of different nature (Indiana limestones, Saturnia travertines, and pre-salt shrubs) under constant geo-reservoir condition of P-T: Pc = 50 MPa, and T= 60 ◦C. The effective stress and pore volume rate was varied between experiments while kept constant during each experimental flooding. We used porosity, permeability, Ca-Mg analysis, and μCT scanning as proxies for stress state related rock matrix alteration. While it is agreed that injection rate plays a major role in carbonate dissolution, through a higher dissolution rate corresponding to a high injection rate, and our work confirms this, we also demonstrate that for a constant given confining pressure, the effective stress can have a stimulant role in rock matrix alteration and wormhole development (Indiana limestone). Inversely, effective stress has a reverse role in less consolidated, more heterogeneous rocks (travertines). The pre-salt rock samples have shown interesting and mixed results, whose behaviour falls in between the Indiana limestone’s ones and the travertines’ ones: the chemical response behaved like an Indiana limestone while the physical response can be compared to a travertine.

We think that our results highlight the importance of the stress state in a reservoir, and while the confining pressure cannot be varied during injection or depletion of a reservoir, the pore pressure can be affected. The processes involved behind this are not yet clarified by the experimental work, but we believe that they are time and chemistry related, with further study by the authors indicating that our results are energy-dependent. Therefore, in a carbonate sample, the expected wormhole shape and spread can be predicted thanks to the reservoir conditions, the experimental conditions, and the rock’s petromorphology. Finally, our numerical work further demonstrates that the heterogeneities within the porosity arrangement and geometry drive the fluid flow and could represent the main driving variable for the creation of pore space and carbonate dissolution.