Computational Modeling and Analysis of Low Temperature Combustion Regimes for Advanced Engine Applications.

Abstract

To achieve cleaner and more efficient energy utilization, novel strategies in modern combustion devices operate using lean, premixed reactant mixtures at high pressures. Under these conditions, auto-ignition often becomes a dominant process for burning. Therefore, accurate prediction of auto-ignition characteristics is of paramount importance in successful implementation of these advanced combustion systems.

The first part of this dissertation focuses on auto-ignition characteristics at high-pressure, low-temperature conditions, relevant to modern gas turbine engines. In particular, strong (homogeneous) and weak (deflagration-dominant) ignition regimes in the presence of thermal inhomogeneities are computationally investigated. Predictive criteria based on Zel’dovich’s theory and passive scalar mixing, which can capture the ignition behavior a priori, are proposed and validated using extensive parametric tests of one-dimensional laminar systems of a lean syngas/air mixture. Subsequently, a non-dimensional scaling analysis is performed to derive regime criteria for turbulent reacting flows, leading to a turbulent ignition regime diagram. The regime diagram is then numerically validated against two-dimensional direct numerical simulations of syngas/air auto-ignition. A number of parametric test cases, by varying the turbulent Damköhler and Reynolds numbers, are considered. The auto-ignition phenomena are characterized by analyzing the corresponding heat release rates and resultant combustion modes. It is demonstrated that the observed ignition behaviors are consistent with the regime diagram predictions.

In the second part of the dissertation, applicability of a Reynolds-Averaged Navier Stokes based spray-interactive flamelet (SIF) combustion model to stratified LTC in direct-injection compression ignition (DICI) engines is assessed, which incorporates the interaction between spray evaporation, gas-phase combustion and turbulent mixing. A number of parametric cases are considered by way of varying the fuel start-of-injection (SOI) timing. The numerical results are validated against available experimental data for in-cylinder pressure trace and CO/NO emissions. It is shown that the SIF model performs well over a wide range of stratified conditions due to the incorporation of the effects of small-scale turbulent transport on combustion. Finally, the SIF model is employed to further investigate the impact of fuel injection parameters such as injection pressure and spray cone angle on the NO-CO trade-off of the DICI engine for the most delayed SOI timing.