Biomass as a clean and renewable source of energy has immense potential to aid in solving the energy crisis in the world. In order to accurately predict the fluidization behavior of biomass particles using the Eulerian–Eulerian approach and the kinetic theory for granular flows (KTGF), employing appropriate models that adapt to irregularly shaped particles and can precisely predict the interaction between particles is crucial. In this study, the effects of varying radial distribution functions (RDF), frictional viscosity models (FVM), angles of internal friction ($\varphi$), and stress blending functions (SBF) on the performance of two-fluid models (TFM) were investigated. Simulation predictions were compared and validated with the previous experiments in the literature on Geldart B biomass particles of walnut shells. When applying sphericity to account for size irregularities of biomass particles, the results of this study demonstrated that predictions of both the Ma–Ahmadi and the Carnahan–Starling RDFs along with the Schaeffer FVM agree with experimental data. More specifically, the bubbling behavior prediction slightly favored the use of the Ma–Ahmadi RDF for biomass particles. The results also revealed the importance of using FVM regardless of the initial void fraction. The use of the Schaeffer FVM became more important as time proceeded and particle bulk density decreased. With the change of $\varphi$ and the application of SBF, no significant differences in the time-averaged results were observed. However, when $\varphi$ ranges were between 30 and 40, the predictions of bubbling behavior became more greatly aligned with experimental results.

A new method was developed to predict the pressure drop and minimum fluidization velocity as a function of inlet gas velocity for multicomponent fluidized beds. The method attributes the changing bed composition during the fluidization process by using a definition for the mass fractions of the bed not yet fluidized. First, a literature survey presents single and multicomponent system correlations that estimate the minimum fluidization velocity. Then, development of the “MASS method” was described and compared to published experimental data and numerical simulations for pressure drop from single-, binary-, and ternary-component fluidized bed systems, where good agreement was observed. Minimum fluidization velocities predicted using correlations in the literature were compared with the MASS method and corresponding experimental data and numerical simulations. The predicted minimum fluidization velocity from the MASS method provided very good results with a relative error of ±4% compared to experimental data. The MASS method was shown to accurately predict fluidization of complex multicomponent systems of granular material and can be used to predict other systems.