Application of a multi-zone model for the prediction of species concentrations in rapid compression machine experiments

Abstract

Accurate chemical kinetic models, which predict species evolution and heat release rates in chemically reactive systems, are essential for further advancements in fuel and combustion technology. An experimental facility that is widely used for evaluating the accuracy of kinetic models is a rapid compression machine (RCM), which creates a well-defined reaction environment by compressing a reactive mixture inside a chamber. Generally, RCM experiments are conducted in order to obtain ignition delay data. However, chemical speciation data provides greater insight into reaction pathways, and is therefore a more rigorous benchmark for validating kinetic models.In order for a chemical kinetic model to be evaluated using RCM data, the kinetic model must be coupled with a thermodynamic model that can predict the temporally varying conditions that evolve during an RCM experiment. The most common approach is to utilize a thermally and compositionally homogeneous 0-dimensional reactor model (HRM), which predicts conditions inside the hot core region of the main combustion chamber of an RCM, where a significant portion of the chemical reaction in an RCM takes place. This approach requires an effective volume profile, which is derived from the pressure profile of either a non-reactive experiment with similar transport properties as the condition of interest, or a separate multi-zone model (MZM), via the relationship between pressure and volume for an isentropic process. While HRMs have been shown to yield adequate ignition delay predictions, they cannot be used to predict average speciation data, since the conditions in the core region vary considerably from the average conditions of the total reaction chamber.This work introduces a modified MZM, which simulates chemical reaction throughout the entire temperature-varying main combustion chamber of an RCM, in addition to boundary work, conduction, and crevice flows as the traditional MZM approach. Simulating chemistry in the MZM allows for average speciation predictions, and eliminates the need for an HRM. The new approach is shown to yield similar average speciation data as CFD simulations (within 15% difference) for the combustion of primary reference fuels at various conditions. iii ACKNOWLEDGEMENTS

Advances in rapid compression machine studies of low- and intermediate-temperature autoignition phenomena

“…Stratified reactor models, like the 'onion-layer', or 'concentric-balloon' model, have been applied to simulate RCM conditions in just a few studies [165], but there is an extensive literature on their application to modeling stratified autoignition in HCCI engines [281,282]. Potential exists for their use in various RCM studies, including autoignition chemistry investigations to address and reduce uncertainties under NTC conditions and long ignition delay times where the measurements can be significantly affected by boundary layer gradients [267,283] [140].…”

Heat release analysis for rapid compression machines: Challenges and opportunities

“…A more refined approach utilizes the adiabatic core hypothesis where reactivity is neglected in the boundary layer gas [13], with heat and enthalpy flows derived empirically [14], or via reduced-order models [15,16]. An even more robust methodology considers the gas as a consisting of multiple, concentric zones in the reaction chamber [17] with potential to account for flow to the piston crevice volume. In the reacting zone, the gas could consist as a mixture of unburned charge and burned products, e.g., fuel/oxidizer and CO2/H2O, or could include important intermediates, e.g., CO. A detailed review of these, and assessments of the impact on uncertainties is outside the scope of this paper, and saved for future work.…”

A Multi-Zone Reaction-Based Diesel Combustion Model for Model-Based Control

“…The two-zone model improved the prediction accuracy of cylinder pressure, exhaust gas temperature, emission concentrations and SOC, compared to a single-zone model. More complicated multi-zone models were presented in [9][10][11][12][13] and shown to address the inhomogeneity fairly well. However, these models are too complicated for real-time simulations, and thus, a simplification must be applied.…”