435284 Quantitative Assessment of Fine-Grid Kinetic-Theory-Based Predictions of Mean-Flow Properties in Unbounded Fluidization

Monday, November 9, 2015: 3:40 PM
254B (Salt Palace Convention Center)
William Fullmer, Chemical and Biological Engineering, University of Colorado, Boulder, CO and Christine M. Hrenya, Chemical and Biological Engineering, University of Colorado at Boulder, Boulder, CO

For several decades, it has been well known that kinetic-theory-based continuum models are qualitatively able to capture dynamic particle clustering in high-velocity, gas-solid flows.  The quantitative accuracy of such descriptions, however, remains unclear mostly due to the computational requirements involved in such simulations. In particular, these particle clusters - regions where the local solids concentration exceeds the mean - occur at the so-called meso-scale, a length scale intermediate to the particle size and the system size. Numerical grids on the order of 1 to 20 particle diameters are typically required to fully resolve these heterogeneous structures. Such an approach is very computationally expensive, precluding its use in most industrial and pilot-scale applications. Therefore strategies to model, rather than directly simulate, the full range of meso-scale structures currently remains a highly active area of research. At present, the approach taken most frequently uses fine-grid continuum simulations to constitute additional closure models for coarse-grid filtered models. However, assessing the quantitative accuracy of these underlying kinetic-theory-based continuum models, which have many simplifying assumptions of their own, has not received the same amount of attention. Recent high-resolution CFD-DEM simulations (Radl and Sundarsean, 2014) of an unbounded sedimenting (or fluidization) system is used as ideal data to validate mean flow properties predicted by a new two-fluid model of Garzo, Tenneti, Subramaniam, and Hrenya (2012). Overall the comparisons are quite good both qualitatively and quantitatively, which is a bit surprising given the low-gradient assumption inherent in kinetic-theory models and the high concentration gradients present at the interface between clustered and dilute regions.  Small differences at the highest concentrations are noted, and discussed in terms of the assumptions inherent in kinetic theory.

Extended Abstract: File Not Uploaded