Yali Tang

Direct Numerical Simulation of Dense Gas-Fluidized Beds

Project aim

Dense gas-solid interaction is of great importance in many industrial processes, especially fluidized beds which are frequently applied in a variety of engineering, e.g., chemical, environmental, energy, petrochemical, etc. Fully resolved simulations can give accurate prediction of the interfacial exchanges, which significantly contribute to control and optimize the flow behavior.


An iterative Immersed Boundary Method (IBM) is adopted to carry out fully resolved simulations of gas-solid flows. However, the grid dependency of this method might result in a large error for the computed drag force when a coarse grid is used, especially for dense packing and high Reynolds numbers. In this work, we propose a methodolody for highly accurate results of the drag force, by making use of this grid dependency and introducing an effective diameter. It finally allows for IBM simulations at relatively low resolutions for large particulate systems.

Firstly, the validation of our code has been done with simulations for  Stokes flow around single sphere, and moderate Reylnods (Re) flows past regular arrays as well. Subsequently, the methodology we proposed was successfully applied for modeling the drag force in gaas-solid flows at Re=50 and 100. Those simulations reproduce the results from the exist drag correlations in literatures. Moreover, this application of the effective diameter calculated in the way from our methodology is verified for the simulations of large random arrays. Then, with this methodology, more Reynolds numbers have been investigated for the gas-solid flow simulations. Finnaly, we improve the drag correlation based on the one proposed by Ladd (1994) and extend its application to a wider range of Reynolds number (0<Re<1000) as well as a completed range of the packing fraction. Besides, we also obtain a function of the effective diameter as Reylnolds number and the resolution. With this function, the IBM simulations of dense gas-solid flows at relativelu large system would be feasible to be done at low resolution, which refers to relatively low computational cost.

The next step of our work is to study the effect of moving particles on the drag force, comparing to the corresponding system but with stationary particles. Further, the dynamics of a small gas-fluidized bed will be investigated by both experiment and IBM simulation.

Scientific publications

A methodology for highly accurate results of direct numerical simulations: drag force in  dense gas-solid flows at intermediate Reynolds number. Submitted to the International Journal of Multiphase Flows, July, 2013.