Search Results

In this study, a piezo disk was used to generate a cloud of n-decane fuel drops, which were mixed with air, then carried into a combustion chamber and ignited by a platinum wire. Microgravity data obtained at the Japan Microgravity Center (JAMIC) were compared to normal gravity data, all at 1Atm pressure and 20+/-1°C initial temperature. Under normal gravity the lean limit was found to be 7.6x106/mm3 (Φ = 1.0), and from this point the flame front speed steadily increased from 20cm/s up to a maximum flame front speed of 210cm/s at a fuel drop density of about 14x106/mm3 (Φ = 1.85). Microgravity data showed a much richer lean limit - about 14.5x106/mm3 (Φ = 1.9), and the flame front speed did not gradually rise to a peak value. Instead, the measurements indicated a peak value of about 250cm/s, with a steep increase followed by a gradual decrease at richer fuel air ratios. A cellular flame structure appeared, and the cell size decreased as the mixture density increased.

Flame propagation at elevated pressures for propane, butane and autogas (20% propane and 80% butane by mass) were investigated. Flame arrival time was measured using ionization probes installed along the wall of a cylindrical combustion chamber. Flame radius was also measured using a laser schlieren technique. Results showed that the flame front speed decreased with increasing initial pressure, and the initial pressure effect on maximum flame front speed was correlated by the relationship Sf = 175·pi-0.15 (for Φ=1.0). Characteristics of flame front speed between propane, butane and autogas were very similar, whereas at fuel-rich conditions flame front speed of butane and autogas were higher than that of propane. A thermodynamic model to predict flame radius and speed as a function of time was derived and tested using measured pressure-time curves.

The combustion mechanism of a fuel droplet cloud was studied by numerical simulation. We investigated how the flame front speed and combustion products changed depending on the equivalence ratio and initial temperature. Modeling was performed using the KIVA-III software package, a three dimensional analysis software used mainly for internal combustion engine applications. The computational domain was a horizontal 1x1x100 cell sector of a spherical combustion chamber and the fuel was n-decane. Results showed that when all the fuel droplets were assumed to have evaporated, the flame front speed increased from 28 cm/s to 152 cm/s as the equivalence ratio increased. The maximum flame front speed was reached at ϕ=1.1, beyond which it decreased (at richer overall equivalence ratios). With a constant equivalence ratio, the flame front speed decreased near the outside region, because the unburned gas was compressed by the expanding burned gas.

To date, the DME combustion mechanism has been investigated by in-cylinder gas sampling, numerical calculations and observation of combustion radicals. It has been possible to quantify the emission intensities of in-cylinder combustion using a monochromator, and to observe the emitting species as images by using band-pass filters. However, the complete band images were not observed since the broadband (thermal) intensity may be stronger than band spectra intensities. Emission intensities of DME combustion radicals from a pre-mixed burner flame have been measured using a spectroscope and photomultiplier. Results were compared to other fuels, such as n-butane and methane, then, in this study, to better understand the combustion characteristics of DME, emission intensities near CH bands of an actual DI diesel engine fueled with DME were measured, and band spectra emitted from the engine were defined. Near TDC, emission intensities did not vary with wavelength.

This work aimed to develop an LPG fueled direct injection SI engine, especially in order to improve the exhaust emission quality while maintaining high thermal efficiency comparable to a conventional engine. In-cylinder direct injection engines developed recently worldwide utilizes the stratified charge formation technique at low load, whereas at high load, a close-to-homogeneous charge is formed. Thus, compared to a conventional port injection engine, a significant improvement of fuel consumption and power can be achieved. To implement such a combustion strategy, the stratification of mixture charge is very important, and an understanding of its combustion process is also inevitably necessary. In this work, a numerical simulation was performed using a CFD code (KIVA-3), where the shape of a combustion chamber, swirl intensity, injection timing and duration, etc. were varied and their effects on the mixture formation and combustion process were investigated.

For better understanding of the in-cylinder combustion characteristics of DME, combustion radicals of a direct injection DME-Fueled compression ignition engine were observed using a spectroscopic method. In this initial report, the emission intensity of OH, CH, CHO, C2 and NO radicals was measured using a photomultiplier. These radicals could be measured with wavelength resolution (half-width) as about 3.3 nm. OH and CHO radicals appeared first, and then CH radical emission was detected. After that, the combustion radicals were observed using a high-speed image intensified video camera with band-pass filter. All of radicals were able to observe as images with half-width as 6 or about 10 nm. Rich DME leaked from nozzle was burning at the end of combustion. Therefore, the second light emission of C2 radical after the main combustion was observed.

This is a preliminary work for the development of a stratified combustion engine using liquefied petroleum gas(LPG) as an alternative fuel. The main objective of this research is to find out the optimizing engine parameters from the viewpoint of mixture formation with the aid of simulation, where the KIVA_ code was used. The combustion characteristics of LPG and gasoline are different because of their different physical properties. Therefore, the numerical simulation was performed for optimizing engine parameters by changing the piston and cylinder geometry, as well as injection conditions. Result showed that geometry of combustion chamber has a great influence on mixture stratification. Also, weaker swirl seems to be better for mixture formation in the vicinity of the spark plug.

Flame propagation characteristics, in a heavy-duty type LPG lean burn SI engine, were investigated by simulation methodology, using the global one step and the ten step chemical kinetic reaction mechanisms, respectively. The swirl ratio and equivalence ratio were varied to investigate their effects on flame front speed. The effect of increased swirl intensity on flame speed was very minor at ranges of equivalence ratio of this study. Flame front shape, however, was affected by swirl intensity. Circular flame front formed for a higher swirl ratio, which is in a qualitative accordance with that of measurements. Comparison between calculation and measurements of flame propagation characteristics shows a good agreement for both the global one step and the ten step chemical kinetic model. This work concludes that the reduced chemical kinetic reactions, consisting of ten steps, is useful for flame propagation study in an LPG SI engine.

Band spectrum images for CH, OH and CHO were taken in a heavy duty type LPG lean-burn SI engine, to investigate the combustion process as it pertains to the pollutant formation process in the post flame region. Full spectra and band spectrum flame images were observed with a bottom view single cylinder research engine and two high speed cameras. NOx emissions were also measured for excess air ratios ranging from 1.0 to 1.6. A thermodynamic model, including the detailed chemical kinetic mechanism for LPG and NOx formation reactions, was developed to predict the major reaction species in the post flame region, and NOx emissions during the combustion process. The model qualitatively described the flame images for each band spectrum and could predict the measured NOx emissions very well.