Numerical Simulation of Particle Collision and Agglomeration in Turbulent Channel Flows

The study described in this thesis concerns the simulation of dispersed and dense particle-laden turbulent channel flows. The research primarily investigates the role of gravity; in terms of its contribution to particle collision, agglomeration and re-distribution. Large eddy simulation is employed to predict the fluid-phase, with solutions coupled with a Lagrangian particle tracking routine to model the particle-phase. In order to establish the validity of the preferred numerical method, results generated from the single-phase and the dilute particle-phase predictions were compared with those based on DNS, with good agreements found. Results obtained for horizontal zero gravity channel flows, show effects of particle size, particle concentration and turbulence on colliding and agglomerating particles. All variables were shown to strongly impact on collision and agglomeration, with the number of events reaching maximum towards the channel walls due to increased particle concentrations and turbulence levels in these regions. The collision and agglomeration is, however, shown to enhance exponentially with the inclusion of gravity and accentuated on the lower wall of the channel. An extension of the investigation into vertical channels of upward and downward flow configurations, also demonstrated the significance of the gravity force on particle collision and agglomeration. The effect of the particles on the flow is small, owing to the low mass-loading. Agglomeration is found to be most favourable for flows of low turbulence; and unlike collisions, dominantly forms in the channel centre. The investigation presented is a novel contribution to literature that provides a fundamental improvement on the understanding of turbulent fluid-particle flows. Particularly, it extends the existing knowledge on cohesive particle behaviours in turbulent flows by examining the effect of gravity on such flows. The contribution finds relevance in many engineering and industrial flow processes and should aide the design of better flow processes