Particulate drag models are commonly used to predict the behavior of fluidized beds and are examined experimentally and computationally here. Different drag models are used depending on the density of the bed. In dense beds the Ergun relation is commonly used, and for beds of intermediate density (void fractions above 0.8, although there is no sharply defined transition) other models are necessary. In each case, drag depends on the void fraction of the bed, particle diameter, and Reynolds number. In dilute flows (void fractions nearing unity) models for drag on single particles without reference to void fraction are used.

The most common difficulty in using drag models is in applying them to particles with distributions of sizes and shapes, as models typically specify a single particle diameter, although some involve an effective diameter and/or particle sphericity. Here, experiments are performed with particles with normal (e.g., Gaussian) size distributions and mixtures of particles with distinct size distributions. Particles are fluidized with air in a 10.2 cm inner diameter bed, through a porous distributor plate. The primary measurement is the differential pressure across the bed (used to calculated the drag force), and in addition visual observations and pressures measured at several locations within the bed are used to gauge void fractions.

Data are plotted alongside a number of commonly used drag models. Comparisons are made to point out where corrections to the models are needed to account for the size distribution. Computational simulations are also performed using the software Arena-flow, and any disagreements with the experimental data suggest corrections to the drag models used in software.