Comparison of First-Order Upwind (DWF) vs. Superbee Scheme for TFM Bubble Simulations

Hi everyone,

I’m currently running Two-Fluid Model (TFM) simulations in MFIX and would like to ask for your insights regarding spatial discretization schemes.

Specifically, I’m curious whether using the first-order upwind scheme (DWF) versus the Superbee high-resolution scheme leads to noticeable differences in predicted bubble characteristics—such as size, shape, and dynamics—in gas–solid fluidized beds.

While it’s generally accepted that second-order (or higher-resolution) schemes like Superbee provide better accuracy by reducing numerical diffusion, I’ve observed that the first-order upwind scheme offers better stability, especially during long-duration simulations.

Has anyone compared these two schemes in similar contexts? If so, did you notice significant differences in bubble behavior? And how do you typically balance accuracy versus stability when choosing a discretization scheme for TFM?

Thanks in advance for your advice!

Best regards,

Hi Wuming,

We did have tried to study the effect of spatial discretization scheme on the prediction of bubble behavior in our past study ( A two fluid modeling study of bubble collapse due to bubble interaction in a fluidized bed - ScienceDirect ). We cited the work of The effect of numerical diffusion on simulation of isolated bubbles in a gas–solid fluidized bed - ScienceDirect, where you can find more detailed comparsion.

Hope this helps.

3 Likes