(473b) Computational Study of the Bubbling-to-Slugging Transition in a Laboratory-Scale Fluidized Bed

Bubbling fluidized bed reactors are utilized in a wide range of chemical industries, including biomass conversion, petroleum refining, and pharmaceutical and commodity chemicals production. Establishing a comprehensive understanding of how the gas-solid hydrodynamics affects interphase mixing and associated chemical reactions is critical for optimizing process performance. In that regard, one of the most important remaining challenges is to be able to accurately detect and predict the transition from bubbling to slugging and utilize that understanding to control interphase mixing.

In this presentation we summarize observations of the bubbling-to-slugging transition in a laboratory-scale fluidized bed of Geldart Group B particles based on computational simulations utilizing the two-fluid, Eulerian-Eulerian MFIX model. Our goal is to resolve previously unrecognized details of the underlying physics. The specific details of interest in this case are the dynamic trends revealed by fluctuations in simulated pressure and void fraction measurements at different locations within the bed as the simulated gas flow is increased from near minimum fluidization to well above the point of maximum slugging. We quantified the observed trends in our simulations using standard statistical measurements for complex time series and estimated bubble statistics derived from the void fraction measurements using the MS3DATA algorithm recently developed by researchers at MIT. In contrast to several correlations in literature, our simulations indicate that the bubbling-to-slugging transition is a gradual process involving a series of bubble coalescence events that occur over a range of gas flows rather than at a single critical value. The spatial progression of these bubble coalescence processes can be seen in the statistical shifts in both pressure and bubble size time series measured at different locations in the bed. With the proper interpretation, we expect that these patterns can be exploited for real-time reactor diagnostics and controls. We propose that the next steps for research in this area should include: 1) experimental validation of the computational simulations reported here; and 2) computational and experimental studies of the impact of the dynamical transitions involved on solid-solid and gas-solid mixing.