(189d) An Exercise of Sootlib Models in Unsteady Flames

Sootlib is a library that has been developed to allow for access to several soot chemistry and particle size distribution (PSD) models that can be integrated into existing combustion model codes. The soot formation processes included in Sootlib are nucleation, growth, oxidation, and coagulation models. Five PSD models are included: monodispersed, sectional, an assumed shape lognormal distribution, the method of moments closed using quadrature method of moments (QMOM), and the method of moments with interpolative closure (MOMIC). Models currently in development in Sootlib include: conditional QMOM (CQMOM) and direct QMOM (DQMOM). These models allow for multiple internal coordinates for the PSD, which allows for other soot formation processes, like aggregation, to be modeled.

This work uses a 1-D unsteady flame code that includes heat loss to compare the different soot formation and PSD models. The flame code models non-premixed diffusion flames and burner-stabilized flames. The diffusion flame is 1-D without an imposed velocity or flux. Flame strain is imposed by using fixed domain length with Dirichlet boundary conditions. For both flames, a finite volume approximation is used. Cantera is used to calculate the reaction rates, thermochemical properties, and transport properties. The sundials package with CVODE is used for the chemistry integration. The combination of chemistry and PSD models has a significant impact on simulation results when compared with experimental data. Several different combinations of models are simulated and compared with each other and with available experimental data from ISF data.

This work is also looking at using flamelet progress variable (FPV) methods parameterized by heat loss, progress variable, mean mixture fraction, mixture fraction variance, and soot and radiative effect. Data is generated using the 1-D flame code and accessed using a lookup table.