2015

Kallio, Sirpa

Computational fluid dynamic (CFD) modeling of industrial scale fluidized beds is a challenging task due to the mismatch between a large process size and fine flow structures. In the present work, methods are developed to overcome the problems in order to make it possible to use CFD as a cost-effective tool for development of processes based on the fluidized bed concept. Two approaches to tackle the problems related to fine flow structures are discussed: 1) transient simulation using a coarse computational mesh and subgrid-scale closure relations and 2) a steady-state simulation approach that applies time-averaged transport equations for mass and momentum. The biggest benefit of the transient coarse-mesh simulation approach is that the closure laws need to describe a much smaller fraction of the total momentum transfer than what is the case in steady-state modeling. The biggest drawback is that a long simulation is required to produce the average flow field. An additional complication is that the closure laws have mesh resolution as a parameter. Steady-state simulations produce the average flow field directly and thus significantly reduce the computation time. In this work, length scales of flow patterns in fluidized beds are analyzed from experiments. Averaged transport equations for mass and momentum are presented and the terms in the equations are analyzed. It is shown that drag force is one of the main terms to be modeled. A drag correction coefficient is defined and ways to determine it from transient CFD simulation data are presented. In the work, correlations for both the space-averaged and the timeaveraged drag forces are applied in riser simulations.