Abstract
Friday, April 14 at 2:30 p.m.
Montgomery Knight Building 317
Stringent government regulations regarding the emissions of non-volatile particulate matter
(nVPM) have motivated efforts to better predict soot formation and growth in a practical gas 
turbine engine. Reliable numerical methodologies can be a viable tool here but predicting soot 
numerically is an inherently complex problem that is further aggravated by the presence of the 
multiscale and multi-physics nature of soot-turbulence-chemistry interactions in a turbulent 
environment. Moreover, the current understanding of soot formation and growth is crippled with 
major uncertainties at each stage of soot formation and growth starting from its nucleation, and 
surface growth, to aggregation, etc. The purpose of this research is to understand key factors 
influencing soot formation and growth in turbulent reacting flows in a systematic manner from 
canonical to complex flows. First, zero-dimensional perfectly stirred reactors are used to 
establish the role of gas phase kinetics and key model sensitivities (sizes, and concentrations of 
inception species as well as rates of coagulation and surface growth for soot particles) on global 
soot predictions as well as soot particle size distribution functions. To simulate complex three- 
dimensional turbulent reacting flows, a multi-scale and multi-physics Linear Eddy Mixing (LEM) 
model, that takes into account the reaction, diffusion, and turbulent stochastic stirring at their 
respective scales within the subgrid of Large Eddy Simulations (LES), is extended to account for 
soot physics. It will be used to conduct simulations of the sooting turbulent bluff body stabilized 
flames of gaseous ethylene fuels for verification against the experimental data. Furthermore, this 
verified LEMLES based modeling framework will be used to assess the impact of model sensitivities 
such as the stages of nucleation, surface growth, and fractal evolution on a more canonical 
non-premixed temporal mixing jet. Eventually, model simulations of soot formation and growth will 
be shown in a more complex liquid fueled Rich-Burn-Quick-Quench-Lean-Burn (RQL) gas turbine 
combustor while acknowledging key understandings of its sensitivities.

Committee
•  Prof. Suresh Menon – School of Aerospace Engineering (advisor)
•  Prof. Adam Steinberg – School of Aerospace Engineering
•  Prof. Joseph Oefelein – School of Aerospace Engineering
•  Prof. Wenting Sun   – School of Aerospace Engineering
•  Prof. Michael Mueller – School of Mechanical and Aerospace Engineering, Princeton 
University