G Kumar, Reddy, J Joshi
2009 Seventh International Conference on CFD in the Minerals and Process Industries CSIRO   unpublished
Solid-liquid fluidized beds are used widely in industry for hydrometallurgical, catalytic cracking, ion exchange, adsorption, crystallisation, sedimentation, and particle classification. In each of these operations, size and density distribution of particles is present and influence bed expansion, particle segregation and overall motion of both liquid and particles. It is these properties that govern heat and mass transfer and reaction rates and determine bed volume and residence time
more » ... ence time requirements. This study investigates experimentally and computationally a number of phenomena underlying the behaviour of fixed and fluidised beds. Both instantaneous and time-average velocity measurements have bee performed in a refractive index matched bed to obtain local energy dissipation rates. A commercially available CFD code (FLUENT) was used to simulate bed composition, liquid flow rate and end effects on fluid-particle drag, bed expansion, segregation and intermixing. Finally, direct numerical simulation was developed to resolve detailed flow structures around individual and interacting particles. Each of these aspects of the research is more full described below: EXPERIMENTAL MEASUREMENTS PIV experiments were performed using an index-matched (RI=1.47) borosilicate glass turpentine/ tetrahydronaphthalene solution fluidised bed in the creeping, transition and turbulent regimes (0.320) D/d p ratio and in the creeping, transition and the turbulent flow regimes. Using the drag law of Joshi (1983) and Pandit and Joshi (1998), the computed index (n) was in excellent agreement with the published values of Richardson and Zaki (1954). Simulations were also carried out at low D/d p ratios (3, 5, and 10) to explore the influence of wall effects on the fluid-particle drag across fixed beds. Computationally, the particles were fixed artificially in a regular configuration but not in mutual contact. As expected, it was found that wall influence on the fluid-particle drag coefficient was reduced with increasing D/d p ratio, such that at D/d p =10 the deviation from the Ergun equation was only 13.2 percent. The effect of particle concentration on the drag coefficient for both fixed and expanded beds was also investigated. It was found that the drag coefficient, increased with increasing particle concentration. For instance, at a Re of 1000 the drag coefficient increased by 2.5 and 8 times as the particle concentration went from 0.217 to 0.365, to 0.577, respectively. For the fluidised beds, the simulations were able to capture the channelling effects through high voidage regions near the wall. The FLUENT analysis was extended to binary particle size systems to explore the behaviour of segregation and intermixing. Binary mixtures with ratio of terminal settling velocity range 1.2-3.2 and Reynolds number from 0.33 to 2080 were investigated. The computational model was in good agreement with experimental observations and predicted the layer inversion phenomena due to different size and density as well as the critical velocity at which the complete mixing of the two particle species occurred.