High-energy short-duration spark igniters are common in combustion devices where the discharge often occurs in a nonfueled region. This requires convection between where the spark kernel is created and where combustion can begin. Here, a reduced-order model is developed to study an ignition kernel in conditions similar to a gas turbine combustor. The kernel is generated in air and entrains ambient fluid into the reactor to model such observed fluid dynamics in a parallel set of experiments. Continuous dilution with cold reactants and endothermic fuel decomposition reactions are shown to compete with heat release in the entraining kernel. Simulation results also reveal a pressure influence on the kernel’s initial energy density and entrainment that can exceed the direct dependence of the reaction rates. A Monte Carlo perturbation approach is directly input to simulations, resulting in ignition probabilities that are directly comparable to experimental results. A computationally efficient support vector machine is trained in multiple ways with simulation results and used to quickly classify ignition success. Including the statistical variation of input parameters, the simulated ignition probabilities match the experimental results, indicating the capability of this approach for predicting ignition probability in non-premixed combustors.
Modeling Ignition Probability for Stratified Flows
Journal of Propulsion and Power ; 33 , 5 ; 1294-1304
01.09.2017
Aufsatz (Zeitschrift)
Elektronische Ressource
Englisch
Modeling Ignition Probability for Stratified Flows
Online Contents | 2017
|Ignition Probability in a Stratified Turbulent Flow With a Sunken Fire Igniter
Online Contents | 2014
|Stratified Torch Ignition Engine: Raw Emissions
British Library Conference Proceedings | 2016
|Stratified Torch Ignition Engine: Combustion Analysis
British Library Conference Proceedings | 2016
|Stratified Torch Ignition Engine: Combustion Analysis
SAE Technical Papers | 2016
|