An in-silico open-source approach for the investigation of multi-phase packed systems: rigid body simulations and computational fluid dynamics

In this work the problem of fluid flow and particle transport and deposition in porous media is investigated. These phenomena are of paramount importance in the oil, metallurgical and process industries, in aquifer dynamics, as well as in the case of chemical reactions carried out in fixed bed catalytic reactors. Due to the difficulty of obtaining the parameters of interest (such as mass transfer or dispersion coefficients) caused by many small-scale complexities arising both in fluid flow and in transport processes, an interesting alternative lies in the use of a fully in-silico simulation framework, which is the way explored in this dissertation. Different tools have been developed in the last decades and most of them rely on Discrete Element Method (DEM), for generating the porous media model, and Computational Fluid Dynamics (CFD), for simulating fluid flow and scalar transport. This workflow presents the main drawbacks of being computationally expensive, mostly since DEM codes are designed to describe with very high accuracy particle- fluid interactions, that very often are negligible during packing generation, and of making use of in-house or commercial codes, that are either difficult to access or costly. In this thesis we thus employed a novel open-source and easily accessible workflow based on Blender, a rigid-body simulation tool developed for computer graphics applications, and OpenFOAM, a very well-known CFD code. This approach presents the additional advantage of being computationally fast. The first step of this work was the validation of this workflow's robustness by comparison with experimental data for global porosity and porosity profiles. Then, pressure drops through the different realizations of porous media were calculated and compared with Ergun's law predictions, showing a very good agreement. A number of different grain shapes, namely spheres, cylindrical beads and trilobes, were considered for this study. Great emphasis was also placed on the numerical issues related to mesh generation and spatial discretization, which both play an important role in determining the final accuracy of the finite-volume scheme and are often overlooked. The next validation step was performed by analyzing fluid velocity distributions and hydrodynamic dispersion rates in a wide range of operating conditions (when compared with other works carried out by solving the Stokes equation as opposed to the full Navier-Stokes equations as is done in this work). Results showed that dispersion within the analyzed porous medium is adequately described by classical power laws obtained by analytic homogenization. Moreover, the validity of Fickian diffusion to treat dispersion in porous media is also assessed. Subsequently, particle deposition was investigated, with only Brownian motion and steric interception being accounted for as deposition mechanisms, while gravitational settling was neglected. Particle transport in the system is investigated via Eulerian steady-state simulations, where particle concentration is solved for, not following explicitly particles trajectories, but solving the corresponding advection-diffusion equation. An assumption was made in considering favorable collector-particle interactions, resulting in a perfect sink boundary condition for the collectors. The gathered simulation data are used to calculate the deposition efficiency due to Brownian motions and steric interception. The original Levich law for one simple circular collector is verified; subsequently the full porous media packings are considered. Results show that the interactions between the different collectors result in behaviors which are not in line with the theory developed by Happel and co-workers, highlighting a different dependency of the deposition efficiency on the dimensionless groups involved in the relevant correlations. Lastly, this dissertation deals with a more theoretical approach to the problem of obtaining an appropriate upscaled form of the equations governing scalar and particle transport at the pore-scale, employing the method of asymptotic homogenization via multiple-scale expansion. This method was successfully used in the past to obtain formally correct macroscale forms of fluid flow and particle transport equations, and was used in this work to approach the upscaling of the particle transport and deposition equations. Concluding, this work proves the success and viability of a fully in-silico open-source approach for the investigation of multi-phase packed systems via rigid-body simulations and computational fluid dynamic