Analysis of the overlap of devolatilisation and char conversion during single coal particle burning using a detailed multi-step kinetic approach

The interaction and overlap between the devolatilisation and char conversion processes of a single coal particle are investigated using a detailed boundary layer resolved simulation (BLRS) approach, in which all relevant gradients are resolved and gas solid interactions are represented through a set of boundary conditions. The gas phase is modelled with a detailed homogeneous reaction mechanism comprising 76 species and 973 reactions. The solid phase is treated as a porous medium with a detailed multi-step kinetic approach for both devolatilisation and char conversion (CRECK-S-C). The devolatilisation mechanism includes 32 species and 35 reactions, while the char conversion mechanism involves 8 species and 14 reactions. In total, 45 simulations are performed under varying ambient conditions, i.e. oxygen concentration, gas temperature, and varying particle size. The overlap between devolatilisation and char conversion is quantified by a criterion where the respective conversion rates differ by less than one order of magnitude, indicating a strong coupling that should not be neglected. Moreover, the overlap is expressed as the fraction of volatile mass released when char conversion becomes significant, Y$_{OL}$, which can be directly applied in reduced-order models, e.g. ... mehrsub-models in the Euler-Lagrange context. The results reveal a consistent trend for smaller particle sizes (dp $\leq$ 100 µ m). Increasing ambient temperature, oxygen concentration and/or particle diameter leads to earlier start of overlap, corresponding to a lower value of Y$_{OL}$. For larger particle sizes (dp $\leq$ 100 µ m) a slightly different behaviour has been observed, particularly under a certain condition, where larger overlap regions, i.e. lower Y$_{OL}$ values, occur than would be expected from the general trend. Based on these findings, a linear regression model for Y$_{OL}$(X$_{O2}$, T, d$_p$) was derived to predict the activation point of char conversion sub-models under varying conditions that can be implemented in reduced-order simulations.