Experimental Investigation and Coarse-Grained SuperDEM-CFD Modeling of Sand–Plastic Mixing in a Fluidized Bed

Modeling the fluidization behavior of sand–plastic mixtures containing nonspherical particles is significant for designing, optimizing, and scaling up biomass/plastic fluidized bed reactors. However, the large particle number and nonspherical shape are challenging in DEM-CFD simulations. This study developed a coarse-grained SuperDEM-CFD method that lumps spherical particles or nonspherical particles into parcels to save computational costs. To validate the developed method, sand–plastic binary fluidization experiments with varying cylinder mass fractions of 0, 5.4, and 18.0% were carried out in a fluidized bed, including pressure drops, expansion height, particle height, and orientation distributions. The machine learning-aided image processing method was employed to analyze bubble properties. Compared with experimental results, the coarse-grained SuperDEM-CFD method can successfully predict pressure drops, fluidization behavior, and expanded bed height. The predicted particle height, orientation distributions, and bubble properties including size and circularity also agree well with experiments, demonstrating the ability and effectivity of the novel method to simulate sand–plastic binary fluidization containing nonspherical particles.