Physical Insights into Turbulent Stratified Combustion using Direct Numerical Simulations

Abstract

Partial premixing of reactants occurs frequently in practical combustion devices despite most research e orts being directed towards the idealised premixed and non-premixed con gurations. Combustion with deliberate compositional variations within the ammability limits, known as strati ed combustion, has increasingly been exploited to provide greater control over the ignition and ame propagation behaviour. Strati ed ames can achieve stable operation with leaner overall mixtures, giving rise to e ciency improvements and reduced pollutant emissions. is research aims to advance the understanding and capability to predict combustion in strati ed mixtures using three-dimensional direct numerical simulations. e central focus of the investigation is the analysis and modelling of strati ed ame propaga tion and the cross-scalar dissipation rate between the reaction progress variable and the mixture fraction. Two direct numerical simulation con gurations, namely planar ames and slot jets, were studied to reveal the in uence of mixture strati cation on ame dynamics. e cross-scalar dissipation rate is a crucial and o en overlooked parameter in many turbulent strati ed combus tion models. Unlike the individual scalar dissipation rates, it can rarely be approximated using a simple algebraic model. A modelled transport equation is presented which may lead to a greater predictive performance when solved. To preserve the mixture strati cation in the unburned gas against the in uence of turbulent mixing, a scalar forcing mechanism is introduced and validated. is approach enables precise control over the compositional distribution of the reactant mixture, facilitating detailed parametric analyses that can further enhance the understanding of strati ed ames. e outcomes of this research support the development and adoption of strati ed combustion technology by enhancing the physical understanding of strati ed ame propagation, improving the predictive capability of industrial simulations, and extending the simulation methodology for ongoing academic research.