Book Description

The overall objective of this dissertation is the development of a modeling and simulation approach for turbulent two-phase chemically reacting flows. A new full velocity-scalar filtered mass density function (FMDF) formulation for large eddy simulation (LES) of a separated two-phase flow is developed. In this formulation several terms require modeling that include important conditionally averaged phase-coupling terms (PCT). To close the PCT a new derivation of the local instantaneous two-phase equations is presented and important identities are derived relating the PCT to surface averages. The formulation is then applied for two particle laden flow cases and solved using a full particle based Monte-Carlo numerical solution procedure. The first case is a temporally developing counter-current mixing layer dilutely seeded with evaporating water droplets. Validation studies reveal excellent agreement of the full particle method to previous hybrid FDF studies and direct numerical simulations for single-phase flows. One-way coupled simulations reveal that the overall dispersion is maximized with unity Stokes number droplets. Two-way coupled simulations reveal the advantages of two FDF approaches where the subgrid variation of droplet properties are explicitly taken into account. Comparisons of the fully-coupled FDF approach are compared to more approximate means of determining phase-coupling based on filtered properties and local and compounded global errors are assessed. The second case considered is the combustion aluminum particles. A new mechanistic model for the ignition and combustion of aluminum particulate is developed that accounts for unsteady heating, melting, heterogeneous surface reactions (HSR) and quasi-steady burning. Results of this model agree well with experimental data for overall burn rates and ignition times. Two-phase simulations of aluminum particulate seeded mixing layer reveal the variations in flame radius resulting in local extinguishment from SGS variations in gas oxidizer.