(421u) Industrial-Scale Mixing of High Solids Biomass Slurries

Berson, R. E., University of Louisville
Thomas, J. M., University of Louisville
Miller, Q. S., University of Louisville

Designing reactors for viscous slurries and suspensions is complicated due to the non-Newtonian flow behavior and the high power requirements.  For biomass slurries, this is further complicated by significant changes in the slurry characteristics as the reaction proceeds due to changing insoluble solids concentration (ISS).  Pretreated corn stover (PCS) slurries were used in trials designed to understand how ISS affects mixing and power draw at large-scale, with the objective of designing a system capable of maintaining good mixing and suspension of solids while minimizing power draw.  A computational fluid dynamics (CFD) model representing a conventional baffled mixing tank with a pitched-blade impeller was created and used to predict solids distribution throughout the tank and torque requirements for low and high solids concentrations.  The low solids content was represented by a 5% slurry where the system exists as two phases.  The high solids content was represented by a 12.5% slurry, which results in little free water so the system behaves as a single phase.  Model validity was satisfactorily established by comparing CFD results to experimentally measured solids concentration strata and torque on the impeller shaft in a lab-scale reactor.  ISS stratification was determined experimentally by pipette sampling at varying depths. Power requirements were measured with a torque sensor aligned with the shaft.  The rheology of the PCS slurries, an important parameter input for the CFD model, was determined using a cup-and-vane rheometer and the data was fit to a Herschel-Bulkley model.  Interestingly, torque and power requirements were actually higher for the 5% slurry above a certain rpm (and shear) due to the shear thinning nature and low flow away from the impeller of the higher viscosity slurry.  The validated CFD model was applied to predict power per unit volume and solid suspensions at large-scale.