Prediction of NOx in premixed high-pressure lean methane flames with a MMC-partially stirred reactor

Abstract

The multiple mapping conditioning (MMC) method is used to extend the stochastic PDF approach to the flamelet regimes of premixed combustion. MMC permits for a desired degree of flamelet-like conditions while reflecting the fluctuating nature of turbulent flames and preserving the integrity and universality of the chosen mixing model. The model is implemented in the context of the partially stirred reactor (PaSR), which is therefore generalised to have a wider range of applicability. A stochastic formulation of original MMC is deployed, where mixing of particle scalar values is conditioned on a Markovian reference variable which emulates an implied particle position relative to a flame. The model interactions with the reference variable are controlled through the flamelet localness parameter, Λ, which is also related to the ratio of diffusive to convective time scales. The model is implemented in a Monte Carlo numerical scheme using detailed GRI3.0 chemical kinetics without adjustments of kinetic coefficients. Predictions of NOx emissions are validated against experimental data for a lean premixed high-pressure combustor in which reactions fall between the flamelet and distributed regimes. There is good agreement between model predictions and experimental data.