Date on Master's Thesis/Doctoral Dissertation

5-2016

Document Type

Doctoral Dissertation

Degree Name

Ph. D.

Department

Chemical Engineering

Degree Program

Chemical Engineering, PhD

Committee Chair

Berson, R. Eric

Committee Co-Chair (if applicable)

Willing, Gerold

Committee Member

Willing, Gerold

Committee Member

Fu, Xiao-An

Committee Member

Lian, Yongsheng

Abstract

Mean age theory is a useful tool for analyzing mixing by providing spatial distributions of time based data for material inside a system using a steady-state CFD approach, but has been limited to single phase systems. Mean age theory was extended here to multiphase systems by defining the scalar tracer concentration independently for individual phases, which allows mean age to be solved at steady-state for each phase independently within a multiphase system. The theory was validated by comparing multiphase mean age (MMA) distributions extracted from spatial distributions determined computationally at two locations where RTDs were experimentally measured in a water-oil flow system. Mean residence times from MMA theory were within 1-3% of experimental values and variances were within 3-11%. MMA was then modified for applicability to closed systems and applied towards predicting just suspended speed in mixing tanks by evaluating MMA near the bottom surface through strategic zone selection. MMA equations were solved only in a thin section along the bottom of the vessel (~1% of the vessel height), allowing the mean age in proximity to the bottom to be computed. The technique was accurate within 1-3% of experimental values across a range of solid densities, solid fractions, and particle sizes while using multiple impeller types and vessel geometries. At high-solids conditions, biomass slurries exhibit non-Newtonian single phase behavior with a yield stress and require high power input for mixing. The goals was to determine the effect of scale and geometry on power number, P0, and estimate the power for mixing a biomass slurry in a million gallon hydrolysis reactor of conventional design. A lab-scale CFD model was validated against experimental data and then scaled up. A pitched-blade turbine and A310 hydrofoil were tested for various geometric arrangements. Flow was transitional; laminar and turbulence models resulted in equivalent P0 which increased with scale. The ratio of impeller diameter to tank diameter affected P0 for both impellers, but impeller clearance to tank diameter affected P0 only for the A310. At least 2 MW is required to operate at this scale.

Share

COinS