This study introduces a novel approach to Large Eddy Simulations (LES) of turbulent reacting flows, crucial for understanding combustion systems. By integrating machine learning and model reduction methods, the Sample-Partitioning Adaptive Chemistry (SPARC) algorithm automates LES with adaptive chemistry. Enhanced with Local Principal Component Analysis and Computational Singular Perturbation, the strategy minimizes user intervention, boosting efficiency. The method’s effectiveness is demonstrated in simulating the Adelaide Jet in Hot Coflow (AJHC) burner, showcasing reduced computational effort and robust performance in complex turbulent-reacting flows.

This study proposes a method for determining rate constants in the heterogeneous growth of pyrocarbon deposits, utilizing theory-based corrections of analogous gas-phase reactions. Applied to H-abstraction reactions, the methodology informs the revision of a pyrocarbon deposition model, integrated with a state-of-the-art gas-phase dynamics model for species evolution and growth of polycyclic aromatic hydrocarbons (PAHs) and soot. Validated against literature data, the comprehensive kinetic framework accurately predicts pyrocarbon formation under varied operating conditions, offering insights into relevant reaction pathways and the competition between carbon deposition and amorphous carbon formation.

This study investigates n-heptane soot formation in low-strain rate counterflow flames. Flame temperatures and soot fractions are measured under varied oxygen mole fractions. Good agreement is observed between measured and predicted maximum temperature and peak soot fraction. Discrepancies in soot volume profiles between 1D and 2D simulations are attributed to thermophoretic and radial effects. Higher oxygen leads to increased temperatures, fostering n-heptane pyrolysis and larger, more dehydrogenated soot particles at xO2=0.45 compared to xO2=0.3.