Mechanical Theory for Particle Detachment in Porous Media

Abstract

The detachment of particles from surfaces is prevalent in numerous environmental, chemical, biological and industrial applications. Some examples include the resuspension of dust in indoor environments, the removal of contaminants from manufactured components, the delivery of pharmaceutical drugs and the transport of bacteria in underground aquifers. Specific to the petroleum industry, the detachment of clay particles induced by hydrocarbon production leads to the plugging of rock pores and hinders the rate of production. The phenomenon of particle detachment is therefore of wide scientific and practical interest. However, the current understanding is often limited to simplified cases of perfectly spherical particles, which limits their applicability for real-world applications. This thesis aims to further understand and characterise the detachment of nonspherical particles, accounting for parameters such as the particle size, shape, aspect ratio, orientation angle, distribution of parameters and favourable/unfavourable conditions. The study focuses on kaolinite particles as a specific application due to the ubiquity of the mineral in petroleum reservoirs. The numerical study comprises the modelling of the detaching hydrodynamic forces in creeping flow and the resisting Derjaguin-Landau-Verwey-Overbeek (DLVO) adhesion force. The former and latter are executed using computational fluid dynamics (CFD) and surface element integration (SEI), respectively. A laminar steady-state model based on the CFD package ANSYS/CFX was used throughout the study. Similarly, SEI is a theoretical method that scales the DLVO adhesion force between two parallel plates to that of cylinders and oblate spheroids considered in this study. The torque balance model was used as the main criterion for studying rolling detachment, incorporating both adhesion and hydrodynamic forces; Hertz contact theory was used to evaluate the torque lever arm. Sliding and lifting detachments were also considered, where necessary. The numerical studies are further complemented by direct visualisation experiments of engineered polystyrene microparticles (radii of 1.5 to 5 μm) in spherical and spheroidal shapes. Kaolinite particles were also studied to evaluate the effectiveness of the model in characterising natural particles. The particles were deposited in a rectangular microfluidic glass channel and subjected to fluid flows of different pH, salinities and flow rates to assess their detachment rates. The detachment rates of the sphere particles evaluated from the experiment were validated against the established theory for sphere particle detachment before subsequent visualisations of non-spherical particles were conducted. By comparing the experimental and the simulation results, a new fundamental understanding of the detachment of non-spherical particles was gained. It was shown for the first time that the detachment of spheroidal particles with decreasing aspect ratio is nonmonotonic: there is a particular aspect ratio where detachment rates reach a minimum, before increasing again. Rolling and sliding detachment are most likely to occur for oblate spheroids and cylinders, respectively. In addition, a two-stage detachment process was observed for both engineered latex particles and natural kaolinite under unfavourable conditions. The observation supports the theory of particle detachment from the primary and secondary minima. Separately, the experimentally observed gradual detachment behaviour of kaolinites was reproduced in our modelling, which accounted for the variations in particle parameters using probabilistic distributions. Lastly, spheroidal particles were observed to be deposited at an angle in the visualisation experiments. At low orientation angles, particles with the smallest aspect ratio (flattest) are the hardest to detach, but this trend is reversed at high orientation angles. Consequently, the investigation has advanced the existing understanding of the detachment of particles, which is more relevant to real-world applications.