A Numerical Study of the Fluid Dynamics of Bubbling Fluidized Beds

Doctoral thesis, 1997

The two-fluid model was evaluated as a tool for predicting local flow behaviour in bubbling fluidized beds. A review is given of the two-fluid model approach for simulation of the flow in bubbling fluidized beds. Several alternatives for drag functions, mixture viscosity models, particle-particle interaction force and turbulence models for the gas and particulate phases are discussed. A presentation of publications on the application of the two-fluid model to bubbling and circulating fluidization is also included. The main effort is focused on the validation of different closures by comparing computed versus experimentally obtained local statistical bubble parameters. These are the bubble frequency, mean bubble rise velocity, mean pierced bubble length and the mean bubble volume fraction. The simulations used for the validations require a rather fine degree of mesh refinement and a large number of time steps, which raises the need for a more efficient two-fluid solver. In view of this, a parallel multi-block solver was developed for possible use on distributed or shared memory parallel computers. The simulations were carried out for the operational pressures of 0.1, 0.8 and 1.6 MPa and an excess gas velocity of 0.6 m/s using similar bed geometry and bed material properties as were used for the previous experimental investigation. Four different closures of the gas and particulate phase stress terms were investigated. The simplest model, which does not take gas and particle turbulence into account, was used for the two-dimensional mesh refinement study, a three-dimensional simulation and for validation purposes. The three remaining models, used for validation purposes, contain particle turbulence, particle and gas turbulence and, finally, particle and gas turbulence taking into account the effects of the drift velocity. The results show that the drift velocity has no major influence on the fluid dynamic behaviour of the bubbling fluidized bed under consideration. The results obtained from the models including gas turbulence suggest that gas turbulence is of no importance at atmospheric conditions, whereas the effects are considerable at higher opertional pressures. The study also indicates that three-dimensional effects are of importance. Furthermore, a simple empirical dimensional bed expamsion model is presented for bubbling fluidized beds with and without internal heat exchanger tubes. This empirical model was derived from different experimental studies and is applicable for group B/D particles and a range of fluidization velocities, pressures and staggered horizonta tube bank geometries.