(437g) Investigating Sources of Discrepancy between CFD-DEM Simulations and Experimental Granular Flow Data

Accurate, experiment-informed modelling of interparticle and interfacial interactions is crucial for the development of industrially representative CFD-DEM systems. One example of such a system is the spout-fluidized bed (SFB) reactor, which comprises complex hydrodynamic interactions between solids and fluid. The SFB is widely applicable to many industries and applications, owing to its highly efficient heat and mass transfer and ability to process high-density feedstock. Accordingly, the reactor and chemical engineering community would benefit from empirically validated SFB models to help guide experimental design for improved process performance.

In the present work, feedstock particle velocities in a cold-flow, non-reactive SFB are tracked for an experiment and an analogous CFD-DEM model using particle image velocimetry on the spouting region and at the wall of the reactor. Particle motion in the spout demonstrates good agreement, as quantified by a high correlation coefficient (> 0.8), a high structural similarity index (> 0.75), and a low mean squared error (< 5 × 10^-2) of the spatial velocity distribution. However, particle motion at the wall within the bed indicated a more limited agreement, as the CFD-DEM model overpredicts particle motion along the wall by, at most, ½ an order of magnitude in mean velocity. This discrepancy between velocities at the wall—where hydrodynamic behavior of the slow-moving feedstock particles is strongly dictated by interparticle and interfacial interactions—suggests that some aspects of these interactions are being underpredicted, or even omitted from the model.

This talk will discuss these results, and some of the potential sources of discrepancy between the hydrodynamic behavior in the modelled SFB and in the experiment, including experimental observations of electrostatic interactions and aspects of the model parameterization. Overall, these investigations will improve our understanding of the existing limitations in these interparticle and interfacial effects which dictate hydrodynamic behavior in granular flow models.