Date of Award

January 2019

Document Type


Degree Name

Master of Science (MS)


Chemical Engineering

First Advisor

Gautham Krishnamoorthy

Second Advisor

Wayne Seames


Oxy-fuel combustion processes are promising technologies for power generation that allow CO2 recovery and sequestration using essentially conventional equipment. Although oxy-fuel combustion has been well researched over the years, there are fundamental issues related to the associated ash deposition processes and their subsequent impact on radiative heat transfer that need to be understood before oxy-fuel combustion can be fully scaled up to a commercial scale. Ash formation and deposition is a complex physio-chemical process consisting of: vaporization, condensation, melting, fragmentation, nucleation and coagulation of the mineral matter and organically bound metals in the parent fuel that results in a distinct tri-modal distribution of the ash particles within the combustor. Recent experimental evidence from the University of Utah has shown that a correlation exists between the rate of deposition of the tightly bound “inner” ash deposit layer adjacent to the heat transfer surfaces (which is difficult to remove) and the concentrations of the submicron aerosols in the flue gas. However, predicting these time-dependent particle size distribution characteristics during combustion as the parent fuel transitions to ash using the commonly employed Lagrangian tracking based particle simulation methodologies in Computational Fluid Dynamics (CFD) frameworks is extremely challenging. Further, Lagrangian tracking methods assume isotropic turbulence near the walls (that can impact deposition characteristics) as well as optically thin radiative losses from particles and exhibit low parallelization efficiencies. On the other hand, Eulerian particle tracking methods, like the population balance framework, are more amenable to capturing the nucleation, fragmentation and coagulation characteristics of ash. In addition, near-wall turbulence characteristics and radiative heat losses can potentially be modeled more rigorously. The overall goal of the research reported in this thesis was to examine the oxy-fuel combustion simulation characteristics of two multiphase modeling frameworks (Lagrangian and Euler-Euler) with the goal of evaluating their potential to simulate ash formation and deposition processes. Consistent and identical phenomenological laws for the interphase interactions were utilized across all three frameworks by implementing user

defined (add-on) subroutines to model the diffusional and kinetic resistances associated with the heterogeneous char oxidation, non-grey effects of gas radiation and the variations in the radiative properties of the solid phase during combustion. Prediction accuracies of the different frameworks were assessed by comparisons against measurements of: (1) flame stand-off, (2) temperature and velocity at different axial locations, (3) radiative heat transfer and (4) ash deposition rates of a nonswirling and a swirling oxy-coal flames (burning an Utah Bituminous and Sub-Bituminous coal, respectively) carried out at the University of Utah.

Only the Euler-Euler model was able to capture the experimental observed trends for flame stand-off as a function of oxygen concentration in the primary burner. Non-swirling flame temperature and velocity predictions were in reasonable agreement with measurements across both multiphase modeling frameworks. Swirling flame temperature and velocity predictions were not in reasonable agreement with measurements. Radiation was the dominant mode of heat transfer with the radiative heat loss fraction (Radiative heat loss/Total chemical heat release) determined to be 0.6 for both non-swirling flames and 0.7-0.9 for both swirling flames. Radiation from the participating gases accounted for 75% of the radiative heat transfer. The incident radiative flux predictions were in good agreement with measured values from similar flames in this furnace. The predicted relative concentrations of the submicron aerosols in flue gas in two flames (Oxy27 and Oxy70) were in reasonable agreement with measurements for the Oxy27 flame but not the Oxy70 flame.