Pore-Network Modeling of Reactive Transport and the Use of Statistical Data Analysis to Classify Dissolution Regimes

Abstract:

Reactive transport in porous media and mineral dissolution are of extreme interest for many subsurface applications, where various dissolution patterns of the porous medium are observed and impact directly the success of injection/production processes. For example, formation of preferential pathways (i.e., wormholes) leads to poor utilization of the pore space for the storage of CO₂, surface dissolution alters the mechanical integrity of the rock near the well, and uniform dissolution might lead to subsidence due to weakening of mechanical properties of the reservoir. A pore-network modeling approach is proposed to simulate mineral dissolution regimes during single-phase reactive transport at the pore scale. Statistical tools for data analysis are presented as a quantitative criterion to classify the different dissolution patterns. Our approach solves a coupled transport and reaction pore network model that implements a kinetic model with heterogeneous chemical reaction describing the dissolution of calcite by acidic solutions. The reactive transport problem is described by the acid mass balance equation in pores and throats. The network geometry is updated based on the dissolution process happening at the mineral surface. The fluid flow field is also updated due to these new larger sizes of pores and throats. Three-dimensional random networks were used to study the reactive transport problem. Comparisons to experimental data demonstrate the applicability of pore-network modeling as a tool to explore various dissolution regimes observed at larger scales. The statistical approach proposed to evaluate the result data and classify the different patterns of dissolution was extensively and successfully tested in extensive simulation cases. Finally, behavior diagrams based on dimensionless Péclet and Damköhler numbers were constructed to summarize succinctly the different regimes and their boundaries.