Soot formation and evolution characteristics of premixed hydrocarbon flames

Abstract

Soot particles formed during the incomplete combustion of hydrocarbons not only reduce the efficiency of many combustion devices but also adversely affect the global climate, air quality and human health. The study of soot formation and evolution processes is of great importance to predict and control soot emissions. In the present study, the newly proposed and developed weighted fraction Monte Carlo (WFMC) method with reactive force fields (ReaxFF) molecular dynamics (MD) simulations are used to gain better insight into the soot formation and evolution processes. The new WFMC method is firstly proposed by introducing the additional fraction function based on the concept of weighted numerical particles to reduce the stochastic error. The WFMC method is validated by comparing its numerical simulation results with the corresponding analytical solutions, sectional method and other Monte Carlo (MC) method schemes (i.e., direct simulation Monte Carlo (DSMC) and multi-Monte Carlo (MMC) methods) in excellent agreement. The stochastic errors obtained by different MC method schemes are further studied, and the results show that the WFMC method can significantly reduce the stochastic error for higher-order moments and particle size distribution (PSD) over the larger particle size regime with a slightly higher computational cost. The further development of this new WFMC method is then combined with the Lagrangian particle tracking (LPT) and coupled with the detailed soot model to study soot formation and evolution in ethylene laminar premixed flames. The LPT-WFMC method is validated by both the experimental results and the numerical results of the DSMC and MMC methods. The evolution of soot number density, volume fraction and particle size distribution are investigated for different flame conditions which shows that LPT-WFMC method can extend the soot PSDs and reduce the statistical error for larger particle size regime. The effects of different parameters in the soot model on the soot PSDs are investigated by the parametric sensitivity analysis. In order to study the atomic aspect of soot nucleation, the physical dimerization of different polycyclic aromatic hydrocarbon (PAH) structures is studied by the MD simulations. The collisional association and dissociation processes of each PAH species are investigated for different temperatures, impact parameters and orientations. The dissociation of the PAH dimers is also statistically analyzed by using the Rice-Ramsperger-Kassel (RRK) theory of unimolecular dissociation to gain a deeper insight of the energy transformation, and the contributions of intermolecular and intramolecular degrees of freedom. Finally, the soot formation and evolution processes are studied by using ReaxFF MD simulations for different carbon dioxide additions. The transformation from PAH precursors to the final soot nanoparticle under three stages of soot formation and evolution processes including nucleation, surface growth and coagulation, and graphitization is observed. The chemical effects of different carbon dioxide additions on the soot properties and reaction pathways are also analyzed. In summary, the newly proposed and developed LPT-WFMC method in the present study has demonstrated high capability in predicting soot PSDs. Meanwhile, the ReaxFF MD simulations in the present study also provide a better understanding of the soot nucleation process.