Simulation of Combustion and Mixture Formation for Gasoline Direct Injection Engine Application

The development and introduction of new engine technologies are primarily motivated by the need to comply with increasingly stringent emissions legislation and to reduce fuel consumption. One of the most important of these new engine technologies is direct gasoline injection, which is considered to be an important and cost-effective measure to meet both targets. Computational Fluid Dynamics (CFD) simulations and optical methods are important tools in the development of direct injection gasoline engines. The aim of the work described in this thesis was to develop models, methods, and a numerical platform for simulating the behavior of Direct Injection Spark Ignition (DISI) engines using a variety of fuels, including gasoline-ethanol blends. One of the most important goals of this work was to devise improvements to OpenFOAM (a free, open source CFD package) that would increase its utility as a tool for studying SIDI engines, as there is strong industrial demand for inexpensive software. The work described in this thesis addressed two important problems relevant to modeling combustion in a DISI engine. First, to facilitate the simulation of turbulent burning and pollutant formation, a chemical mechanism for the combustion of gasoline-ethanol blends was refined and thoroughly validated under various conditions (equivalence ratios Φ, initial temperatures Tu, and pressures p). The gasoline surrogate used in this project was composed of iso-octane, toluene, and n-heptane in volumetric ratios of 55%:35%:10%, respectively. The hydrogen to carbon ratio (the H/C ratio) of this blend is similar to that of gasoline, which is around 1.87, as is its equivalence ratio; this is particularly important when studying DISI engines. The integrated mechanism for gasoline-ethanol blends features 120 species participating in 677 reactions and is suitable for use in CFD engine modeling. The mechanism was tested against experimental data on ignition delay times and laminar flame speeds, obtained for various n-heptane/iso-octane/toluene/ethanol-air mixtures under various equivalence ratios, initial temperatures, and pressures. Second, the gasoline and ethanol hollow cone sprays released by an outward-opening pintle-type piezo-controlled injector commonly associated with GDI engines were studied numerically, since accurate simulation of fuel-air mixing and the flow field is critical for subsequent combustion modeling. A pintle injector model was implemented into OpenFOAM in order to simulate the spray discharged by an outward-opening piezo injector. The flow field calculated using the pintle injector model is more realistic than that predicted by the default unit injector model normally used in OpenFOAM. A number of modifications were made to the standard spray submodels in OpenFOAM, including the LISA, TAB and Reitz-KHRT breakup models and the O'Rourke and Trajectory collision models. For instance, three different modified Reitz-KHRT models were implemented into OpenFOAM; these modifications were found to have noticeable effects on the accuracy of the simulated liquid penetration and SMD. Extensive sensitivity studies were carried out on the hollow cone sprays, focusing on the effects of varying the initial and boundary conditions, spray model constants, and other parameters. Validation studies showed that several combinations of spray submodels yield acceptable results in liquid penetration and SMD, including the combination of the Rosin-Rammler + Reitz-Diwakar models and that of the uniform droplet size + Reitz-KHRT models; the latter combination offered the best performance under the studied conditions