Abstract.Soot particles are nanoparticles consisting primarily of carbon that are formed during the combustion of fuel-rich mixtures. Due to environmental and health concerns, soot emissions from combustion systems are tightly regulated, and this regulation will only become stricter in the future. To enable the design of the next generation of low-emission combustion systems, predictive numerical simulations will be required. However, soot is a particularly difficult modeling problem due to the need for high-fidelity models for soot itself in addition to chemistry and turbulence. In this talk, an integrated modeling framework based on Large Eddy Simulation (LES) for soot evolution in turbulent reacting flows will be presented, and the model will be demonstrated and evaluated in an actual aircraft combustor.

In order to enable these high-fidelity simulations, three component models have been developed.  First, a detailed soot model has been developed within the framework of the Method of Moments.  Closure of the moment source terms is achieved with the Hybrid Method of Moments (HMOM), an accurate yet computationally efficient method.  Second, a new turbulent combustion model has been developed based on the Radiation Flamelet/Progress Variable (RFPV) model that can account for the removal of precursors from the gas-phase to form soot particles.  Third, a subfilter PDF model has been developed to account for the unresolved small-scale interactions between soot, turbulence, and chemistry.  The subfilter PDF approach is validated a priori against a DNS database of soot evolution in a turbulent nonpremixed flame.

The integrated modeling approach is validated against experimental measurements in two laboratory-scale turbulent nonpremixed flames: a natural gas piloted jet flame and an ethylene bluff body flame.  Differences in soot evolution due to the differences in the large-scale mixing in the two flames are highlighted and discussed.  The validated model is then applied to the simulation of a Pratt & Whitney aircraft combustor.   Two operating points are simulated to assess the ability of the integrated model to reproduce quantitative trends in soot emissions.  In absolute terms, the LES model overpredicts the combustor exit smoke number by about 25% compared to experimental measurements.  However, the model is able to capture the relative difference between the operating points very well.