Mixing and segregation behavior of ternary biomass-sand-olivine fluidized bed: detailed comparison between experiments and modelling predictions

• Addition of olivine, as tar-reduction catalyst, used to control the mixing/segregation of biomass and sand. • Olivine particle size should be the smallest to counterbalance the density effect on segregation. • Olivine fractions above 0.25 could lead to mixed ternary mixtures upon complete fluidization. • The Particle Segregation Model can be succe extended to ternary mixtures. • Segregation direction/mixing of ternary mixtures effectively characterized in ternary plot. Multi-solid fluidized beds are commonly employed in industry. Conversion of solid biomass into syngas is often operated in fluidized inert material, where severe operational issues like tar condensation and aerosol formation, corrosion, fouling, are mitigated by integrating specific tar-reducing catalysts (e.g. olivine). However, the resulting ternary bed (biomass, inert and catalyst) is prone to inefficiency owing to solids’ segregation. The present study examines the fluidization and segregation behavior of mixtures of a biomass (crushed olive pits), sand and olivine of given size, shape and density, at different compositions, by an experimental characterization of the voidage, fluidization velocity intervals, vertical concentration profiles and mixing index. Interpretation of these profiles is carried out according to the Particle Segregation Model (PSM). Originally developed for segregation in two-component fluidized beds, the PSM is conceptually extended to three-component systems. Initially, the role of composition is investigated for a biomass-rich and a sand-rich mixture. Given the contrasting sizes and densities, the former system maintains its state of mixing at low olivine fractions (maximum deviation from nominal composition: Δx S = 0.06). In the sand-rich mixture, the size-segregating behavior leads to modest segregation of the biomass and olivine towards the bed surface (Δx B = 0.15 max deviation) and sand at the bottom (Δx S = 0.13 max deviation). In a third examined mixture with extended olivine fraction, a similar behavior to the sand-rich mixture is observed (Δx S = 0.27 max deviation). Both tendencies are well captured by the PSM, whose analytical formulation allows drawing triangular plots that effectively compare experimental points and equilibrium lines.