A two-dimensional, unsteady-state, kinetic-diffusion-vaporization-controlled numerical model for aluminum particle combustion is presented. The model solves the conservation equations, while accounting for species generation and destruction with a 15-reaction kinetic mechanism. Two of the major phenomena that differentiate aluminum combustion from hydrocarbon-droplet combustion, namely, condensation of the aluminum-oxide product and subsequent deposition of part of the condensed oxide onto the particle, are accounted for in detail with a submodel for each phenomenon. The effect of the oxide cap in the distortion of the species and temperature profiles around the particle is included into the model. The results obtained from the model, which include two-dimensional species and temperature profiles, are analyzed and compared with experimental data. The combustion process is found to approach a diffusion-controlled process for the oxidizers (O 2 , CO 2 , and H 2 O) and conditions treated. The flame-zone location and thickness are found to vary with the oxidizer. The result shows that the exponent of the particle-diameter dependence of the burning time is not a constant and changes from ≈1.2 for smaller-diameter particles to ≈1.9 for larger-diameter particles. Owing to deposition of the aluminum oxide onto the particle surface, the particle velocity oscillates. The effect of pressure is analyzed for a few oxidizers. Key words: aluminum particle, two-dimensional unsteady-state numerical model, kinetic mechanism, effect of the nature of the gaseous oxidizer, condensed oxide, oxide dissociation, oxide cap, burning time, species and temperature fields INTRODUCTIONAluminum has been added to propellants for many years as an extra energy source for the propellant. Thus, research on the combustion mechanism of burning aluminum has been an ongoing effort. A very significant effort was expended in the 1960s and 1970s shortly after the effects of aluminum were first conceived. In an early study, Glassman [1] and Brzustowski and Glassman [2] recognized that metal combustion would be analogous to hydrocarbon-droplet combustion, that the D 2 law ought to be applied, and that ignition and combustion ought to depend on the melting and boiling points of the metal and the oxide. Glassman speculated that ig-

No abstract

No abstract

No abstract

In this work the G-equation combustion model has been implemented into StarCD code and coupled with an ODE-kinetic solver for predicting knock in a DISI engine based on the Shell autoignition model. The combustion modelhas been validated for selected engine operating points. A mesh sensitivity analysis highlighting the influences of mesh topology and grid type (hexahedral and polyhedral) is presented. Tb account for sensitivities of octane number and EGR ratio on knock behaviour, the governing parameters of the Shell model have been identified and optimized with the help of DoE. The knock models presented here can be operated in both a passive and an active way. The validatien of both knock models with respect to temporal and local knock onset is done with fibre optical spark plug measurements and evaluation of statistical results of pressure trace analysis for a direct-iajection spark-ignition engine.Both passive and active knock model approaches are compared in regards of pros and cons as well as engine perfbrmance. Finally the necessity of thin turbulent fiame brushes realized with G-equation combustion model and their relevance for knock simulation is discussed.The thickness of the turbulent flame brush is in the orderCopyright @ 2012 by the Japan Society of Mechanical Engineers -579 -NII-Electionic