SIMULATION OF FLOW AND PARTICLE TRANSPORT AND DEPOSITION IN POROUS MEDIA WITH COMPUTATIONAL FLUID DYNAMICS

The simulation of transport and deposition of colloidal particles in porous media finds important applications in many engineering and environmental problems, such as particle filtration, catalytic processes carried out in filter beds, chromatographic separation and aquifer remediation. This study focuses in particular on remediation of contaminated groundwater via direct injection of nano-sized zerovalent iron particles, which have been shown to be able to efficiently degrade a large variety of contaminants. Application of this technology on full scale applications poses a number of challenges, the most important of which regards the mobility of the particles and their delivery to the contaminated site in the soil. Particles migration is usually quantitatively expressed by a single parameter: the deposition efficiency in the porous bed, whose theoretical reference lies in the classical colloid filtration theory, which moreover further subdivides the process of deposition in the three mechanisms by which particles can reach the solid grain: Brownian diffusion, steric interception, and gravitational sedimentation. This theory, however, has been developed only for very simple geometrical representations of the porous media and a narrow range of fluid conditions. The difficulties in investigating this kind of systems from the experimental point of view have prevented the development of accurate models able to account for the high degree of complexity which characterizes a porous medium, both in the grain arrangement and in their shape. The aim of this study is therefore to simulate the transport of the nanoparticles and their interaction with the porous media (at the microscopic scale), in order to improve the current understanding of these phenomena and obtain predictive models for the deposition efficiency of the colloids on the surface of the grains constituting the porous medium; moreover, eventually, to evaluate the effectiveness of the zerovalent iron technology. Several two and three dimensional microscale (the order of millimiters) representations of grain packings with different degrees of complexity were analyzed. First, two dimensional random arrangements of spheres were considered. Then, the analysis was extended to domains reconstructed from SEM images of a real porous medium. The work was then expanded in three dimensions, first considering simplified domains constituted by irregular packings of spheres, and finally geometries constituted by grains of realistic shapes. These last geometries were created using an algorithm simulating the grain sedimentation process in porous media (Settledyn). Flow field and particle transport was then investigated using finite volume CFD codes (Fluent and OpenFoam), solving the Navier-Stokes equations for the flow and using an Eulerian approach for the colloid transport, eventually obtaining, for each case, an estimate of the colloidal transport efficiency. After having validated the methodology used in this work by comparing our results with proved analytical results available for simplified cases, new predictive equations for each of the individual contributions of the three deposition mechanisms were derived, highlighting the differences from the theoretical model due to the wider range of operating conditions investigated and/or the different geometrical characteristics of the porous media