The impact of particle polydispersity on fluidization phenomena

Abstract

Particles naturally found in polydisperse form, but systematic study on the impact of particle polydispersity in fluidization field is overshadowed by the more common particle properties (e.g., diameter, density). Recent development in computational fluid dynamics (CFD) codes and computing capabilities enable detailed study of inter-species interactions in a complex and chaotic particle-fluid flow, hence advancing mechanistic understanding of fluidization phenomena. The impact of particle polydispersity (i.e., monodisperse, binary, ternary, and continuous PSD mixtures) on column hydrodynamics (i.e., local particle and fluid velocities, mixing/segregation, inter-species interactions), fluidization phenomena (i.e., entrainment flux, TDH, cluster characteristics), and system performance (i.e., energy consumption, membrane fouling mitigation in Anaerobic Fluidized Bed Membrane Bio-Reactor/AnFMBR) are studied. Analysis of entrainment and TDH values of monodisperse and polydisperse mixtures (Chapters 2 – 4) showed (i) polydispersity effects are poorly incorporated into fluidization correlation, while the use of empirical constants are inadequate to represent such impact, as indicated by huge discrepancies among prediction of correlations intended for similar system; (ii) physical insights into the relation between particle polydispersity and inter-species interactions, i.e., cohesive agglomeration, cluster formation, and inter-species collisions, are guaranteed. Detailed investigation into cluster characteristics of polydisperse particles in CFB riser revealed (i) lower cluster break-up of monodisperse particle relative to polydisperse particles for complete/partial mixing conditions, because of the higher inter-species momentum transfer for multi-species mixtures; (ii) under segregated condition, polydisperse particles also hinders cluster formation (i.e., bidisperse < continuous PSD) because only smaller constituents form clusters at centre column, while larger particles concentrated near wall area destroys smaller constituent’s clusters as they travel radially. Subsequently, fluidization of mono-, bi-, and tri-disperse GAC to induce particle-membrane scouring in AnFMBR are evaluated (Chapters 5 – 7). CFD and experimental results found (i) GAC mean diameter increases particle-membrane scouring (or decreases membrane fouling) through higher superficial liquid velocity required to scour entire membrane height, which indicates the trade-off between fouling mitigation and energy usage; (ii) for fully segregated condition (i.e., high superficial liquid velocity or power input), number of particle constituents (i.e., tridisperse > bidisperse > monodisperse) and size differences among constituents (i.e., larger dp2/dp1 and σ/dsm for bidisperse and continuous PSD, accordingly) decrease particle-membrane momentum transfer, because small constituents at upper reactor obtained less energy due to high pressure drop at lower reactor consist of highly-concentrated larger GACs; (iii) the opposite findings of the (ii) conclusion found in partial mixing condition (i.e., low superficial liquid velocity), presumably due to the more dominant inter-species collisions.