Search Results

A three-dimensional numerical model has been developed to predict spray formation process of swirl or slit type injectors which are currently used in direct-injection gasoline engines. The Discrete Droplet Model (DDM) is totally enhanced: a new droplet deformation model is developed, which is theoretically introduced with a spheroidal shape assumption. The droplet drag model and droplet break-up model via Kelvin-Helmholtz and Rayleigh-Taylor instabilities are modified taken into account with the deformation. The break-up model parameters are dynamically changed according to a droplet Weber number. The model functions are developed using single droplet breakup measurement data. A liquid sheet injection and breakup models are incorporated into the DDM. A new parcel radius model is also introduced to get rid of the grid dependence of the droplet collision-coalescence model.

In direct-injection (DI) gasoline engines, spray characteristics greatly affect engine combustion. For the rapid development of new gasoline direct-injectors, it is necessary to predict the spray characteristics accurately by numerical analysis based on the injector nozzle geometry. In this study, two-phase flow inside slit nozzle injectors is calculated using the volume of fluid method in a three-dimensional CFD. The calculation results are directly applied to the boundary conditions of spray calculations, of which the submodels are recently developed to predict spray formation process in direct injection gasoline engines. The calculation results are compared with the experiments. Good agreements are obtained for typical spray characteristics such as spray shape, penetration and Sauter mean diameter at both low and high ambient pressures. Two slit nozzle injectors of which the slit thickness is different are compared.

A new spray/wall impingement model for gasoline engines has been developed. The model is based on experimental analysis of impinging spray droplets using a phase doppler particle analizer (PDPA). Three new equations were obtained in terms of droplet size, Weber number and the angle from a wall for droplets which were splashed after impinging or created by the impact of a droplet on the liquid film layer on the wall. The three-dimensional calculation results using the model agree very well with the experimental data. The model is also applied to the fuel mixture formation process in a lean-burn gasoline engine.

For the measurements of flow rate, pressure and/or temperature in an engine exhaust pipe, probes are often inserted into the exhaust pipe depending on the application. These measurement probes differ a lot in terms of their size and shape. The flow around the probes become further complicated due to the pulsation of engine exhaust flow. In this study, computational fluid dynamics (CFD) simulations were carried out and a zero-dimensional (0D) model was constructed to analyze the flow field around the probe and flow rate of a pulsating flow. The simulations and the measurements of the flow rate and pressure were performed on flows around a hexagonal prism inserted in a circular pipe which is intended to be a differential pressure flow meter. The velocity field was also measured using the particle image velocimetry (PIV) technique. The CFD simulation results were validated with the experiments for both steady and pulsating flows.

A novel total hydrocarbon (THC) emission concentration estimation model is proposed for reduction of engine development cost and simplification of exhaust measurements. The proposed method uses the absorbance spectra of a Fourier transform infrared (FTIR) spectrometer, which contains the information on a wide variety of hydrocarbons, as input. The model is based on machine learning algorithms including the least absolute shrinkage and selection operator (LASSO) regression and bagging techniques. To train the model, we created a dataset containing pairs of a spectrum of engine exhaust gas and the THC concentration. In addition, we incorporate absorbance spectra of individual hydrocarbon components and several inorganic components so that the model learns the contribution of each hydrocarbon to THC concentration and to ignore interferences of irrelevant gas components.